View TS80C51RA2-LCBB_6751429.PDF datasheet online --- IC-ON-LINE

Datasheet File OCR Text:

features ? high-performance, low-power 32-bit atmel ? avr ? microcontroller ? compact single-cycle risc instruction set including dsp instructions ? read-modify-write instructions and atomic bit manipulation ? performance ? up to 64dmips running at 50mhz from flash (1 flash wait state) ? up to 36dmips running at 25mhz from flash (0 flash wait state) ? memory protection unit (mpu) ? secure access unit (sau) providing user-defined peripheral protection ? picopower ? technology for ultra-low power consumption ? multi-hierarchy bus system ? high-performance data transfers on separate buses for increased performance ? 12 peripheral dma channels improve speed for peripheral communication ? internal high -speed flash ? 256kbytes and 128kbytes versions ? single-cycle access up to 25mhz ? flashvault technology allows pre-programmed secure library support for end user applications ? prefetch buffer optimizing instru ction execution at maximum speed ? 100,000 write cycles, 15-year data retention capability ? flash security locks and us er-defined configuration area ? internal high-speed sram, si ngle-cycle access at full speed ?32kbytes ? interrupt controller (intc) ? autovectored low-latency interrupt service with programmable priority ? external interrupt controller (eic) ? peripheral event system for direct pe ripheral to periph eral communication ? system functions ? power and clock manager ? sleepwalking power saving control ? internal system rc oscillator (rcsys) ? 32 khz oscillator ? multipurpose oscillator, phase locked loop (pll), and digital frequency locked loop (dfll) ? windowed watchdog timer (wdt) ? asynchronous timer (ast) with real-time clock capability ? counter or calendar mode supported ? frequency meter (freqm) for accurate measuring of clock frequency ? six 16-bit timer/co unter (tc) channels ? external clock inputs, pwm, capture, and various counting capabilities ? pwm channels on all i/o pins (pwma) ? 8-bit pwm with a source clock up to 150mhz ? four universal synchronous/asynchro nous receiver/transmitters (usart) ? independent baudrate generator, support for spi ? support for hardware handshaking ? one master/slave serial peripheral inte rface (spi) with chip select signals ? up to 15 spi slaves can be addressed 32145c?06/2013 32-bit atmel avr microcontroller at32uc3l0256 at32uc3l0128
2 32145c?06/2013 at32uc3l0128/256 ? two master and two slave two-wire interfaces (twi), 400kbit/s i 2 c-compatible ? one 8-channel analog-to -digital converter (adc) with up to 12 bits resolution ? internal temperature sensor ? eight analog comparators (ac) with optional window detection ? capacitive touch (cat) module ? hardware-assisted atmel ? avr ? qtouch ? and atmel ? avr ? qmatrix touch acquisition ? supports qtouch and qmatrix capture from capacitive touch sensors ? qtouch library support ? capacitive touch buttons, sliders, and wheels ? qtouch and qmatrix acquisition ? on-chip non-intrusive debug system ? nexus class 2+, runtime control, non-intrusive data and program trace ? awire single-pin programming trace and debug interface muxed with reset pin ? nanotrace provides trace capabilities through jtag or awire interface ? 48-pin tqfp/qfn/tllga (36 gpio pins) ? five high-drive i/o pins ? single 1.62-3.6 v power supply
3 32145c?06/2013 at32uc3l0128/256 1. description the atmel ? avr ? at32uc3l0128/256 is a complete system-on-chip microcontroller based on the avr32 uc risc processor running at frequencies up to 50mhz. avr32 uc is a high-per- formance 32-bit risc microprocessor core, designed for cost-sensitive embedded applications, with particular emphasis on low power consumption, high code density, and high performance. the processor implements a memory protection unit (mpu) and a fast and flexible interrupt con- troller for supporting m odern and real-tim e operating systems. the se cure access unit (sau) is used together with the mpu to provide the required security and integrity. higher computation capability is achieved using a rich set of dsp instructions. the at32uc3l0128/256 embeds state-of-the-art picopower technology for ultra-low power con- sumption. combined power control techniques are used to bring active current consumption down to 174a/mhz, and leakage down to 220na while still retaining a bank of backup regis- ters. the device allows a wide range of trade-offs between functionality and power consumption, giving the user the ability to reach the lowest possible power consumptio n with the feature set required for the application. the peripheral direct memory access (dma) controller enables data transfers between periph- erals and memories without processor involvement. the peripheral dma controller drastically reduces processing overhead when transferring continuous and large data streams. the at32uc3l0128/256 incorporates on-chip flash and sram memories for secure and fast access. the flashvault technology allows secure libraries to be programmed into the device. the secure libraries can be executed while the cpu is in secure state, but not read by non- secure software in the device. the device can thus be shipped to end customers, who will be able to program their own code into the device to access the secure libraries, but without risk of compromising the proprietary secure code. the external interrupt controller (eic) allows pins to be configured as external interrupts. each external interrupt has its own interrupt request and can be individually masked. the peripheral event system allows peripherals to receive, react to, and send peripheral events without cpu intervention. asynchronous interrupts allow advanced peripheral operation in low power sleep modes. the power manager (pm) improves design flexibility and securi ty. the power ma nager supports sleepwalking functionality, by which a module can be selectively activated based on peripheral events, even in sleep modes where the module clock is stopped. power monitoring is supported by on-chip power-on reset (por), brown-out detector (bod), and supply monitor (sm). the device features several oscillators, such as phase locked loop (pll), digital frequency locked loop (dfll), oscillator 0 (osc0), and syst em rc oscillator (rcsys). either of these oscillators can be used as source for the system clock. the dfll is a programmable internal oscillator from 20 to 150mh z. it can be tuned to a high accura cy if an accurate reference clock is running, e.g. the 32khz crystal oscillator. the watchdog timer (wdt) will reset the device unless it is periodically serviced by the soft- ware. this allows the device to recover from a condition that has caused the system to be unstable. the asynchronous timer (ast) combined with th e 32khz crystal oscillator supports powerful real-time clock capabilities, with a maximum timeou t of up to 136 years. the ast can operate in counter mode or calendar mode.
4 32145c?06/2013 at32uc3l0128/256 the frequency meter (freqm) allows accurate measuring of a clock frequency by comparing it to a known reference clock. the device includes six identical 16-bit timer/counter (tc) channels. each channel can be inde- pendently programmed to perform frequency measurement, event counting, interval measurement, pulse generation, delay timing, and pulse width modulation. the pulse width modulation controller (pwma) provides 8-bit pwm channels which can be syn- chronized and controlled from a common timer. one pwm channel is available for each i/o pin on the device, enabling applications that require multiple pwm outputs, such as lcd backlight control. the pwm channels can operate independently , with duty cycles se t individually, or in interlinked mode, with multiple ch annels changed at the same time. the at32uc3l0128/256 also features many communication interfaces, like usart, spi, and twi, for communication intensive applications . the usart supports different communication modes, like spi mode and lin mode. a general purpose 8-channel adc is provided, as well as eight analog comparators (ac). the adc can operate in 10-bit mode at full speed or in enhanced mode at reduced speed, offering up to 12-bit resolution. the adc also provides an internal temperature sensor input channel. the analog comparators can be paired to detect when the sensing voltage is within or outside the defined reference window. the capacitive touch (cat) module senses touch on external capacitive touch sensors, using the qtouch technology. capacitive touch sensor s use no external mechanical components, unlike normal push buttons, and therefore demand less maintenance in the user application. the cat module allows up to 17 touch sensors, or up to 16 by 8 matrix sensors to be interfaced. all touch sensors can be configured to oper ate autonomously without software interaction, allowing wakeup from sleep modes when activated. atmel offers the qtouch library for embedding capacitive touch buttons, sliders, and wheels functionality into avr microcontrollers. the patented charge-transfer signal acquisition offers robust sensing and includes fully debounced reporting of touch keys as well as adjacent key suppression ? (aks ? ) technology for unambiguous detection of key events. the easy-to-use qtouch suite toolchain allows you to explore, develop, and debug your own touch applications. the at32uc3l0128/256 integrates a class 2+ nexus 2.0 on-chip debug (ocd) system, with non-intrusive real-time trace and full-speed read/write memory access, in addition to basic run- time control. the nanotrace interface enables trace feature for awire- or jtag-based debuggers. the single-pin awire interface allows all features available through the jtag inter- face to be accessed through the reset pin, a llowing the jtag pins to be used for gpio or peripherals.
5 32145c?06/2013 at32uc3l0128/256 2. overview 2.1 block diagram figure 2-1. block diagram system control interface interrupt controller asynchronous timer peripheral dma controller hsb-pb bridge b hsb-pb bridge a s mm m s s m external interrupt controller high speed bus matrix generalpurpose i/os general purpose i/os pa pb extint[5..1] nmi g c l k [ 4 . . 0 ] pa pb spi dma miso, mosi npcs[3..0] usart0 usart1 usart2 usart3 dma rxd txd clk rts, cts watchdog timer sck jtag interface mcko mdo[5..0] mseo[1..0] evti_n tdo tdi tms configuration registers bus 128/256 kb flash s flash controller evto_n avr32uc cpu nexus class 2+ ocd instr interface data interface memory interface local bus 32 kb sram memory protection unit local bus interface frequency meter pwm controller pwma[35..0] timer/counter 0 timer/counter 1 a[2..0] b [ 2 . . 0 ] twi master 0 twi master 1 dma twi slave 0 twi slave 1 dma 8-channel adc interface dma ad[8..0] advrefp power manager reset controller sleep controller clock controller xin32 xout32 osc32k rcsys x i n 0 x o u t 0 osc0 dfll tck awire reset_n capacitive touch module dma csb[16:0] smp csa[16:0] sync ac interface acrefn acan[3..0] acbn[3..0] acbp[3..0] acap[3..0] twck twd twalm twck twd twalm rc32k rc120m glue logic controller in[7..0] out[1:0] dataout sau s/m vdiven dis trigger adp[1..0] rc32out pll g c l k _ i n [ 1 . . 0 ] clk[2..0]
6 32145c?06/2013 at32uc3l0128/256 2.2 configuration summary table 2-1. configuration summary feature at32uc3l0256 at32uc3l0128 flash 256kb 128kb sram 32kb gpio 36 high-drive pins 5 external interrupts 6 twi 2 usart 4 peripheral dma channels 12 peripheral event system 1 spi 1 asynchronous timers 1 timer/counter channels 6 pwm channels 36 frequency meter 1 watchdog timer 1 power manager 1 secure access unit 1 glue logic controller 1 oscillators digital frequency locked loop 20-150 mhz (dfll) phase locked loop 40-240 mhz (pll) crystal oscillator 0.45-16 mhz (osc0) crystal oscillator 32 khz (osc32k) rc oscillator 120mhz (rc120m) rc oscillator 115 khz (rcsys) rc oscillator 32 khz (rc32k) adc 8-channel 12-bit temperature sensor 1 analog comparators 8 capacitive touch module 1 jtag 1 awire 1 max frequency 50 mhz packages tqfp48/qfn48/tllga48
7 32145c?06/2013 at32uc3l0128/256 3. package and pinout 3.1 package the device pins are multiplexed with pe ripheral functions as described in section 3.2.1 . figure 3-1. tqfp48/qfn48 pinout gnd 1 pa09 2 pa08 3 pa03 4 pb12 5 pb00 6 pb02 7 pb03 8 pa22 9 pa06 10 pa00 11 pa05 12 pa02 13 pa01 14 pa07 15 pb01 16 vddin 17 vddcore 18 gnd 19 pb05 20 pb04 21 reset_n 22 pb10 23 pa21 24 pa14 36 vddana 35 advrefp 34 gndana 33 pb08 32 pb07 31 pb06 30 pb09 29 pa04 28 pa11 27 pa13 26 pa20 25 pa15 37 pa16 38 pa17 39 pa19 40 pa18 41 vddio 42 gnd 43 pb11 44 gnd 45 pa10 46 pa12 47 vddio 48
8 32145c?06/2013 at32uc3l0128/256 figure 3-2. tllga48 pinout 3.2 peripheral multiplexing on i/o lines 3.2.1 multiplexed signals each gpio line can be assigned to one of the peripheral functions. the following table describes the peripheral signals multiplexed to the gpio lines. gnd 1 pa09 2 pa08 3 pa03 4 pb12 5 pb00 6 pb02 7 pb03 8 pa22 9 pa06 10 pa00 11 pa05 12 pa02 13 pa01 14 pa07 15 pb01 16 vddin 17 vddcore 18 gnd 19 pb05 20 pb04 21 reset_n 22 pb10 23 pa21 24 pa14 36 vddana 35 advrefp 34 gndana 33 pb08 32 pb07 31 pb06 30 pb09 29 pa04 28 pa11 27 pa13 26 pa20 25 pa15 37 pa16 38 pa17 39 pa19 40 pa18 41 vddio 42 gnd 43 pb11 44 gnd 45 pa10 46 pa12 47 vddio 48 table 3-1. gpio controller func tion multiplexing 48- pin pin g p i o supply pin type gpio function ab c d e f gh 11 pa00 0 vddio normal i/o u s a r t 0 txd u s a r t 1 rts s p i npcs[2] p w m a pwma[0] s c i f gclk[0] c a t csa[2] 14 pa01 1 vddio normal i/o u s a r t 0 rxd u s a r t 1 cts s p i npcs[3] u s a r t 1 clk p w m a pwma[1] a c i f b acap[0] t w i m s 0 twalm c a t csa[1]
9 32145c?06/2013 at32uc3l0128/256 13 pa02 2 vddio high- drive i/o u s a r t 0 rts a d c i f b trigger u s a r t 2 txd t c 0 a0 p w m a pwma[2] a c i f b acbp[0] u s a r t 0 clk c a t csa[3] 4 pa03 3 vddio normal i/o u s a r t 0 cts s p i npcs[1] u s a r t 2 txd t c 0 b0 p w m a pwma[3] a c i f b acbn[3] u s a r t 0 clk c a t csb[3] 28 pa04 4 vddio normal i/o s p i miso t w i m s 0 twck u s a r t 1 rxd t c 0 b1 p w m a pwma[4] a c i f b acbp[1] c a t csa[7] 12 pa05 5 vddio normal i/o (twi) s p i mosi t w i m s 1 twck u s a r t 1 txd t c 0 a1 p w m a pwma[5] a c i f b acbn[0] t w i m s 0 twd c a t csb[7] 10 pa06 6 vddio high- drive i/o, 5v tolerant s p i sck u s a r t 2 txd u s a r t 1 clk t c 0 b0 p w m a pwma[6] e ic extint[2] s c i f gclk[1] c a t csb[1] 15 pa07 7 vddio normal i/o (twi) s p i npcs[0] u s a r t 2 rxd t w i m s 1 twalm t w i m s 0 twck p w m a pwma[7] a c i f b acan[0] e i c nmi (extint[0]) c a t csb[2] 3 pa08 8 vddio high- drive i/o u s a r t 1 txd s p i npcs[2] t c 0 a2 a d c i f b adp[0] p w m a pwma[8] c a t csa[4] 2 pa09 9 vddio high- drive i/o u s a r t 1 rxd s p i npcs[3] t c 0 b2 a d c i f b adp[1] p w m a pwma[9] s c i f gclk[2] e i c extint[1] c a t csb[4] 46 pa10 10 vddio normal i/o t w i m s 0 twd t c 0 a0 p w m a pwma[10] a c i f b acap[1] s c i f gclk[2] c a t csa[5] 27 pa11 11 vddin normal i/o p w m a pwma[11] 47 pa12 12 vddio normal i/o u s a r t 2 clk t c 0 clk1 c at smp p w m a pwma[12] a c i f b acan[1] s c i f gclk[3] c a t csb[5] 26 pa13 13 vddin normal i/o g l o c out[0] g l o c in[7] t c 0 a0 s c i f gclk[2] p w m a pwma[13] c at smp e i c extint[2] c a t csa[0] 36 pa14 14 vddio normal i/o a d c i f b ad[0] t c 0 clk2 u s a r t 2 rts c at smp p w m a pwma[14] s c i f gclk[4] c a t csa[6] 37 pa15 15 vddio normal i/o a d c i f b ad[1] t c 0 clk1 g l o c in[6] p w m a pwma[15] c at sync e i c extint[3] c a t csb[6] 38 pa16 16 vddio normal i/o a d c i f b ad[2] t c 0 clk0 g l o c in[5] p w m a pwma[16] a c i f b acrefn e i c extint[4] c a t csa[8] 39 pa17 17 vddio normal i/o (twi) t c 0 a1 u s a r t 2 cts t w i m s 1 twd p w m a pwma[17] c at smp c at dis c a t csb[8] 41 pa18 18 vddio normal i/o a d c i f b ad[4] t c 0 b1 g l o c in[4] p w m a pwma[18] c at sync e i c extint[5] c a t csb[0] 40 pa19 19 vddio normal i/o a d c i f b ad[5] t c 0 a2 t w i m s 1 twalm p w m a pwma[19] s c i f gclk_in[0] c at sync c a t csa[10] 25 pa20 20 vddin normal i/o u s a r t 2 txd t c 0 a1 g l o c in[3] p w m a pwma[20] s c i f rc32out c a t csa[12] 24 pa21 21 vddin normal i/o (twi, 5v tolerant, smbus) u s a r t 2 rxd t w i m s 0 twd t c 0 b1 a d c i f b trigger p w m a pwma[21] p w m a pwmaod[21] s c i f gclk[0] c a t smp 9 pa22 22 vddio normal i/o u s a r t 0 cts u s a r t 2 clk t c 0 b2 c at smp p w m a pwma[22] a c i f b acbn[2] c a t csb[10] 6 pb00 32 vddio normal i/o u s a r t 3 txd a d c i f b adp[0] s p i npcs[0] t c 0 a1 p w m a pwma[23] a c i f b acap[2] t c 1 a0 c a t csa[9] 16 pb01 33 vddio high- drive i/o u s a r t 3 rxd a d c i f b adp[1] s p i sck t c 0 b1 p w m a pwma[24] t c 1 a1 c a t csb[9] 7 pb02 34 vddio normal i/o u s a r t 3 rts u s a r t 3 clk s p i miso t c 0 a2 p w m a pwma[25] a c i f b acan[2] s c i f gclk[1] c a t csb[11] table 3-1. gpio controller func tion multiplexing
10 32145c?06/2013 at32uc3l0128/256 see section 3.3 for a description of the various peripheral signals. refer to ?electrical characteristics? on page 791 for a description of the electrical properties of the pin types used. 3.2.2 twi, 5v tolerant, and smbus pins some normal i/o pins offer twi, 5v tolerance, and smbus features. these features are only available when either of the twi functions or the pwmaod function in the pwma are selected for these pins. refer to the ?twi pin characteristics(1)? on page 798 for a description of the electrical proper- ties of the twi, 5v tolerance, and smbus pins. 8 pb03 35 vddio normal i/o u s a r t 3 cts u s a r t 3 clk s p i mosi t c 0 b2 p w m a pwma[26] a c i f b acbp[2] t c 1 a2 c a t csa[11] 21 pb04 36 vddin normal i/o (twi, 5v tolerant, smbus) t c 1 a0 u s a r t 1 rts u s a r t 1 clk t w i m s 0 twalm p w m a pwma[27] p w m a pwmaod[27] t w i m s 1 twck c a t csa[14] 20 pb05 37 vddin normal i/o (twi, 5v tolerant, smbus) t c 1 b0 u s a r t 1 cts u s a r t 1 clk t w i m s 0 twck p w m a pwma[28] p w m a pwmaod[28] s c i f gclk[3] c a t csb[14] 30 pb06 38 vddio normal i/o t c 1 a1 u s a r t 3 txd a d c i f b ad[6] g l o c in[2] p w m a pwma[29] a c i f b acan[3] e i c nmi (extint[0]) c a t csb[13] 31 pb07 39 vddio normal i/o t c 1 b1 u s a r t 3 rxd a d c i f b ad[7] g l o c in[1] p w m a pwma[30] a c i f b acap[3] e i c extint[1] c a t csa[13] 32 pb08 40 vddio normal i/o t c 1 a2 u s a r t 3 rts a d c i f b ad[8] g l o c in[0] p w m a pwma[31] c at sync e i c extint[2] c a t csb[12] 29 pb09 41 vddio normal i/o t c 1 b2 u s a r t 3 cts u s a r t 3 clk p w m a pwma[32] a c i f b acbn[1] e i c extint[3] c a t csb[15] 23 pb10 42 vddin normal i/o t c 1 clk0 u s a r t 1 txd u s a r t 3 clk g l o c out[1] p w m a pwma[33] s c i f gclk_in[1] e i c extint[4] c a t csb[16] 44 pb11 43 vddio normal i/o t c 1 clk1 u s a r t 1 rxd a d c i f b trigger p w m a pwma[34] c at vdiven e i c extint[5] c a t csa[16] 5 pb12 44 vddio normal i/o t c 1 clk2 t w i m s 1 twalm c at sync p w m a pwma[35] a c i f b acbp[3] s c i f gclk[4] c a t csa[15] table 3-1. gpio controller func tion multiplexing
11 32145c?06/2013 at32uc3l0128/256 3.2.3 peripheral functions each gpio line can be assigned to one of several peripheral functions. the following table describes how the various peripheral functions are selected. the last listed function has priority in case multiple functions are enabled on the same pin. 3.2.4 jtag port connections if the jtag is enabled, the jtag will take control over a number of pins, irrespectively of the i/o controller configuration. 3.2.5 nexus ocd aux port connections if the ocd trace system is enabled, the trace system will take control over a number of pins, irre- spectively of the i/o controller configurat ion. two different ocd trace pin mappings are possible, depending on the configuration of the ocd axs register. for details, see the avr32 uc technical reference manual . table 3-2. peripheral functions function description gpio controller function multiplexing gpio and gpio peripheral selection a to h nexus ocd aux port connections ocd trace system awire dataout awire output in two-pin mode jtag port connections jtag debug port oscillators osc0, osc32 table 3-3. jtag pinout 48-pin pin name jtag pin 11 pa00 tck 14 pa01 tms 13 pa02 tdo 4pa03tdi table 3-4. nexus ocd aux po rt connections pin axs=1 axs=0 evti_n pa05 pb08 mdo[5] pa10 pb00 mdo[4] pa18 pb04 mdo[3] pa17 pb05 mdo[2] pa16 pb03 mdo[1] pa15 pb02 mdo[0] pa14 pb09
12 32145c?06/2013 at32uc3l0128/256 3.2.6 oscillator pinout the oscillators are not mapped to the normal gp io functions and their muxings are controlled by registers in the system control interface (scif). please refer to the scif chapter for more information about this. 3.2.7 other functions the functions listed in table 3-6 are not mapped to the normal gpio functions. the awire data pin will only be active after the awire is e nabled. the awire dataout pin will only be active after the awire is enabled an d the 2_pin_mode command has been sent. the wake_n pin is always enabled. please refer to section 6.1.4 on page 40 for constraints on the wake_n pin. evto_n pa04 pa04 mcko pa06 pb01 mseo[1] pa07 pb11 mseo[0] pa11 pb12 table 3-4. nexus ocd aux po rt connections pin axs=1 axs=0 table 3-5. oscillator pinout 48-pin pin name oscillator pin 3pa08xin0 46 pa10 xin32 26 pa13 xin32_2 2pa09xout0 47 pa12 xout32 25 pa20 xout32_2 table 3-6. other functions 48-pin pin function 27 pa11 wake_n 22 reset_n awire data 11 pa00 awire dataout
13 32145c?06/2013 at32uc3l0128/256 3.3 signal descriptions the following table gives details on signal names classified by peripheral. table 3-7. signal descriptions list signal name function type active level comments analog comparator interface - acifb acan3 - acan0 negative inputs for comparators "a" analog acap3 - acap0 positive inputs for comparators "a" analog acbn3 - acbn0 negative inputs for comparators "b" analog acbp3 - acbp0 positive inputs for comparators "b" analog acrefn common negative reference analog adc interface - adcifb ad8 - ad0 analog signal analog adp1 - adp0 drive pin for resistive touch screen output trigger external trigger input awire - aw data awire data i/o dataout awire data output for 2-pin mode i/o capacitive touch module - cat csa16 - csa0 capacitive sense a i/o csb16 - csb0 capacitive sense b i/o dis discharge current control analog smp smp signal output sync synchronize signal input vdiven voltage divider enable output external interrup t controller - eic nmi (extint0) non-maskable interrupt input extint5 - extint1 exte rnal interrupt input glue logic controller - gloc in7 - in0 inputs to lookup tables input out1 - out0 outputs from lookup tables output jtag module - jtag tck test clock input tdi test data in input tdo test data out output tms test mode select input
14 32145c?06/2013 at32uc3l0128/256 power manager - pm reset_n reset input low pulse width modulation controller - pwma pwma35 - pwma0 pwma channel waveforms output pwmaod35 - pwmaod0 pwma channel waveforms, open drain mode output not all channels support open drain mode system control interface - scif gclk4 - gclk0 generic clock output output gclk_in1 - gclk_in0 generic clock input input rc32out rc32k output at startup output xin0 crystal 0 input analog/ digital xin32 crystal 32 inpu t (primary location) analog/ digital xin32_2 crystal 32 input (secondary location) analog/ digital xout0 crystal 0 output analog xout32 crystal 32 output (primary location) analog xout32_2 crystal 32 output (secondary location) analog serial peripheral interface - spi miso master in slave out i/o mosi master out slave in i/o npcs3 - npcs0 spi peripheral chip select i/o low sck clock i/o timer/counter - tc0, tc1 a0 channel 0 line a i/o a1 channel 1 line a i/o a2 channel 2 line a i/o b0 channel 0 line b i/o b1 channel 1 line b i/o b2 channel 2 line b i/o clk0 channel 0 external clock input input clk1 channel 1 external clock input input clk2 channel 2 external clock input input two-wire interface - twims0, twims1 twalm smbus smbalert i/o low twck two-wire serial clock i/o twd two-wire serial data i/o table 3-7. signal descriptions list
15 32145c?06/2013 at32uc3l0128/256 note: 1. adcifb: ad3 does not exist. universal synchronous asynchronous receiver transmitter - usart0, usart1, usart2, usart3 clk clock i/o cts clear to send input low rts request to send output low rxd receive data input txd transmit data output table 3-7. signal descriptions list table 3-8. signal description list, continued signal name function type active level comments power vddcore core power supply / voltage regulator output power input/output 1.62v to 1.98v vddio i/o power supply power input 1.62v to 3.6v. vddio should always be equal to or lower than vddin. vddana analog power supply power input 1.62v to 1.98v advrefp analog reference voltage power input 1.62v to 1.98v vddin voltage regulator input power input 1.62v to 3.6v (1) gndana analog ground ground gnd ground ground auxiliary port - aux mcko trace data output clock output mdo5 - mdo0 trace data output output mseo1 - mseo0 trace frame control output evti_n event in input low evto_n event out output low general purpose i/o pin pa22 - pa00 parallel i/o controller i/o port 0 i/o pb12 - pb00 parallel i/o controller i/o port 1 i/o 1. see section 6.1 on page 36
16 32145c?06/2013 at32uc3l0128/256 3.4 i/o line considerations 3.4.1 jtag pins the jtag is enabled if tck is low while the reset_n pin is re leased. the tck, tms, and tdi pins have pull-up resistors when jtag is enabled. the tck pin always has pull-up enabled dur- ing reset. the tdo pin is an output, driven at vddio, and has no pull-up resistor. the jtag pins can be used as gpio pins and multiplex ed with peripherals when the jtag is disabled. please refer to section 3.2.4 on page 11 for the jtag port connections. 3.4.2 pa00 note that pa00 is multiplexed with tck. pa00 gpio function must only be used as output in the application. 3.4.3 reset_n pin the reset_n pin is a schmitt input and integrates a permanent pull-up resistor to vddin. as the product integrates a power-on reset detector, the reset_n pin can be left unconnected in case no reset from the system nee ds to be applied to the product. the reset_n pin is also used for the awire de bug protocol. when the pin is used for debug- ging, it must not be driven by external circuitry. 3.4.4 twi pins pa21/pb04/pb05 when these pins are used for twi, the pins are open-drain outputs with slew-rate limitation and inputs with spike filtering. when used as gpio pins or used for other peripherals, the pins have the same characteristics as other gpio pins. sele cted pins are also smbus compliant (refer to section 3.2.1 on page 8 ). as required by the smbus specification, these pins provide no leakage path to ground when the at32uc3l0128/256 is powered down. this allows other devices on the smbus to continue communicating even though the at32uc3l0128/256 is not powered. after reset a twi function is selected on these pins instead of the gpio. please refer to the gpio module configuration chapter for details. 3.4.5 twi pins pa05/pa07/pa17 when these pins are used for twi, the pins are open-drain outputs with slew-rate limitation and inputs with spike filtering. when used as gpio pins or used for other peripherals, the pins have the same characteristics as other gpio pins. after reset a twi function is selected on these pins instead of the gpio. please refer to the gpio module configuration chapter for details. 3.4.6 gpio pins all the i/o lines integrate a pull-up resistor ? programming of this pull-up resistor is performed independently for each i/o line through the gpio controllers. after reset, i/o lines default as inputs with pull-up resistors disabled, except pa00 which has the pull-up resistor enabled. pa20 selects scif-rc32out (gpio function f) as default enabled after reset. 3.4.7 high-drive pins the five pins pa02, pa06, pa08, pa09, and pb01 have high-drive output capabilities. refer to section 32. on page 791 for electrical characteristics.
17 32145c?06/2013 at32uc3l0128/256 3.4.8 rc32out pin 3.4.8.1 clock output at startup after power-up, the clock generated by the 32khz rc oscillator (rc32k) will be output on pa20, even when the device is still reset by the powe r-on reset circuitry. this clock can be used by the system to start other devices or to clock a switching regulator to rise the power supply volt- age up to an acceptable value. the clock will be available on pa20, but will be di sabled if one of the following conditions are true: ? pa20 is configured to use a gpio function other than f (scif-rc32out) ? pa20 is configured as a general purpose input/output (gpio) ? the bit frc32 in the power manager ppcr register is written to zero (refer to the power manager chapter) the maximum amplitude of the clock signal will be defined by vddin. once the rc32k output on pa20 is disabled it can never be enabled again. 3.4.8.2 xout32_2 function pa20 selects rc32out as default enabled after reset. this function is not automatically dis- abled when the user enables the xout32_2 function on pa20. this disturbs the oscillator and may result in the wrong frequency. to avoid this, rc32out must be disabled when xout32_2 is enabled. 3.4.9 adc input pins these pins are regular i/o pins powered from the vddio. however, when these pins are used for adc inputs, the voltage applied to the pin must not exceed 1.98v. internal circuitry ensures that the pin cannot be used as an analog input pin when the i/o drives to vdd. when the pins are not used for adc inputs, the pins may be driven to the full i/o voltage range.
18 32145c?06/2013 at32uc3l0128/256 4. processor and architecture rev: 2.1.2.0 this chapter gives an overview of the avr32uc cpu. avr32uc is an implementation of the avr32 architecture. a summary of the programming model, instruction set, and mpu is pre- sented. for further details, see the avr32 architecture manual and the avr32uc technical reference manual . 4.1 features ? 32-bit load/store avr32a risc architecture ? 15 general-purpose 32-bit registers ? 32-bit stack pointer, program counter and link register reside in register file ? fully orthogonal instruction set ? privileged and unprivileged modes enabling efficient and secure operating systems ? innovative instruction set together with variable instruction length ensu ring industry leading code density ? dsp extension with saturating arithmetic, and a wide variety of multiply instructions ? 3-stage pipeline allowing one instruction per clock cy cle for most instructions ? byte, halfword, word, and double word memory access ? multiple interrupt priority levels ? mpu allows for operating s ystems with memory protection ? secure state for supporting flashvault technology 4.2 avr32 architecture avr32 is a new, high-performance 32-bit risc mi croprocessor architectu re, designed for cost- sensitive embedded applications, with particul ar emphasis on low power consumption and high code density. in addition, the instruction set architecture has been tuned to allow a variety of microarchitectures, enabling the avr32 to be implemented as low-, mid-, or high-performance processors. avr32 extends the avr family into the world of 32- and 64-bit applications. through a quantitative approach, a large set of industry recognized benchmarks has been com- piled and analyzed to achieve the best code density in its class. in addition to lowering the memory requirements, a compact code size also contributes to the core?s low power characteris- tics. the processor supports byte and halfword data types without penalty in code size and performance. memory load and store operations are provided for byte, halfword, word, and double word data with automatic sign- or zero extension of halfw ord and byte data. the c-compiler is closely linked to the architecture and is able to expl oit code optimization features, both for size and speed. in order to reduce code size to a minimum, so me instructions have multiple addressing modes. as an example, instructions with immediates often have a compact format with a smaller imme- diate, and an extended format with a larger immediate. in this way, the compiler is able to use the format giving the smallest code size. another feature of the instruction set is that frequently used instructions, like add, have a com- pact format with two operands as well as an extended format with three operands. the larger format increases performance, allowing an addition and a data move in the same instruction in a
19 32145c?06/2013 at32uc3l0128/256 single cycle. load and store instructions have seve ral different formats in order to reduce code size and speed up execution. the register file is organized as sixteen 32-bi t registers and includes the program counter, the link register, and the stack pointer. in addition, register r12 is designed to hold return values from function calls and is used im plicitly by some instructions. 4.3 the avr32uc cpu the avr32uc cpu targets low- and mediu m-performance applications, and provides an advanced on-chip debug (ocd) system, no caches, and a memory protection unit (mpu). java acceleration hardware is not implemented. avr32uc provides three memory interfaces, one high speed bus master for instruction fetch, one high speed bus master for data access, an d one high speed bus slave interface allowing other bus masters to access data rams internal to the cpu. keeping data rams internal to the cpu allows fast access to the rams, reduces latency, and guarantees deterministic timing. also, power consumption is reduced by not needing a full high speed bus access for memory accesses. a dedicated data ram interface is prov ided for communicating with the internal data rams. a local bus interface is provided for connecting the cpu to device-specific high-speed systems, such as floating-point units and i/o controller port s. this local bus has to be enabled by writing a one to the locen bit in the cpucr system regi ster. the local bus is able to transfer data between the cpu and the local bus slave in a single clock cycle. the local bus has a dedicated memory range allocated to it, and data transfers are performed using regular load and store instructions. details on which devices that are mapped into the local bus space is given in the cpu local bus section in the memories chapter. figure 4-1 on page 20 displays the contents of avr32uc.
20 32145c?06/2013 at32uc3l0128/256 figure 4-1. overview of the avr32uc cpu 4.3.1 pipeline overview avr32uc has three pipeline stages, instruction fetch (if), instruction decode (id), and instruc- tion execute (ex). the ex stage is split into three parallel subsections, one arithmetic/logic (alu) section, one multiply (mul) sect ion, and one load/store (ls) section. instructions are issued and complete in order. certain operations require several clock cycles to complete, and in this case, the instruction resides in the id and ex stages for the required num- ber of clock cycles. since there is only three pipeline stages, no inte rnal data forwarding is required, and no data dependencies can arise in the pipeline. figure 4-2 on page 21 shows an overview of the avr32uc pipeline stages. avr32uc cpu pipeline instruction memory controller mpu high speed bus high speed bus ocd system ocd interface interrupt controller interface high speed bus slave high speed bus high speed bus master power/ reset control reset interface cpu local bus master cpu local bus data memory controller cpu ram high speed bus master
21 32145c?06/2013 at32uc3l0128/256 figure 4-2. the avr32uc pipeline 4.3.2 avr32a microarchitecture compliance avr32uc implements an avr32a microarchitecture. the avr32a microarchitecture is tar- geted at cost-sensitive, lower-end applications like smaller microcontrollers. this microarchitecture does not provide dedicated hard ware registers for shadowing of register file registers in interrupt contexts. additionally, it does not provide hardware registers for the return address registers and return status registers. instead, all this information is stored on the system stack. this saves chip area at the expense of slower interrupt handling. 4.3.2.1 interrupt handling upon interrupt initiation, registers r8-r12 are automatically pushed to the system stack. these registers are pushed regardless of the priority level of the pending interrupt. the return address and status register are also automatically pushed to stack. the interrupt handler can therefore use r8-r12 freely. upon interrupt completion, the old r8-r12 registers and status register are restored, and execution continues at the return address stored popped from stack. the stack is also used to store the status register and return address for exceptions and scall . executing the rete or rets instruction at the completion of an exception or system call will pop this status register and continue execution at the popped return address. 4.3.2.2 java support avr32uc does not provide java hardware acceleration. 4.3.2.3 memory protection the mpu allows the user to check all memory accesses for privilege violations. if an access is attempted to an illegal memory address, the access is aborted and an exception is taken. the mpu in avr32uc is specified in t he avr32uc technical reference manual. 4.3.2.4 unaligned reference handling avr32uc does not support unaligned accesses, except for doubleword accesses. avr32uc is able to perform word-aligned st.d and ld.d . any other unaligned memory access will cause an if id alu mul regfile write prefetch unit decode unit alu unit multiply unit load-store unit ls regfile read
22 32145c?06/2013 at32uc3l0128/256 address exception. doubleword -sized accesses with word-align ed pointers will automatically be performed as two word-sized accesses. the following table shows the instructions with support for unaligned addresses. all other instructions requir e aligned addresses. 4.3.2.5 unimplemented instructions the following instructions are unimplemented in avr32uc, and will cause an unimplemented instruction exception if executed: ? all simd instructions ? all coprocessor instructions if no coprocessors are present ? retj, incjosp, popjc, pushjc ? tlbr, tlbs, tlbw ? cache 4.3.2.6 cpu and architecture revision three major revisions of the avr32uc cpu currently exist. the device described in this datasheet uses cpu revision 3. the architecture revision field in the config0 system register identifies which architecture revision is implemented in a specific device. avr32uc cpu revision 3 is fully backward-compatibl e with revisions 1 and 2, ie. code compiled for revision 1 or 2 is binary-compatible with revision 3 cpus. table 4-1. instructions with una ligned reference support instruction supported alignment ld.d word st.d word
23 32145c?06/2013 at32uc3l0128/256 4.4 programming model 4.4.1 register file configuration the avr32uc register file is shown below. figure 4-3. the avr32uc register file 4.4.2 status register configuration the status register (sr) is split into two halfwords, one upper and one lower, see figure 4-4 and figure 4-5 . the lower word contains the c, z, n, v, and q condition code flags and the r, t, and l bits, while the upper halfword contains information about the mode and state the proces- sor executes in. refer to the avr32 architecture manual for details. figure 4-4. the status register high halfword application bit 0 supervisor bit 31 pc sr int0pc fintpc int1pc smpc r7 r5 r6 r4 r3 r1 r2 r0 bit 0 bit 31 pc sr r12 int0pc fintpc int1pc smpc r7 r5 r6 r4 r11 r9 r10 r8 r3 r1 r2 r0 int0 sp_app sp_sys r12 r11 r9 r10 r8 exception nmi int1 int2 int3 lr lr bit 0 bit 31 pc sr r12 int0pc fintpc int1pc smpc r7 r5 r6 r4 r11 r9 r10 r8 r3 r1 r2 r0 sp_sys lr bit 0 bit 31 pc sr r12 int0pc fintpc int1pc smpc r7 r5 r6 r4 r11 r9 r10 r8 r3 r1 r2 r0 sp_sys lr bit 0 bit 31 pc sr r12 int0pc fintpc int1pc smpc r7 r5 r6 r4 r11 r9 r10 r8 r3 r1 r2 r0 sp_sys lr bit 0 bit 31 pc sr r12 int0pc fintpc int1pc smpc r7 r5 r6 r4 r11 r9 r10 r8 r3 r1 r2 r0 sp_sys lr bit 0 bit 31 pc sr r12 int0pc fintpc int1pc smpc r7 r5 r6 r4 r11 r9 r10 r8 r3 r1 r2 r0 sp_sys lr bit 0 bit 31 pc sr r12 int0pc fintpc int1pc smpc r7 r5 r6 r4 r11 r9 r10 r8 r3 r1 r2 r0 sp_sys lr secure bit 0 bit 31 pc sr r12 int0pc fintpc int1pc smpc r7 r5 r6 r4 r11 r9 r10 r8 r3 r1 r2 r0 sp_sec lr ss_status ss_adrf ss_adrr ss_adr0 ss_adr1 ss_sp_sys ss_sp_app ss_rar ss_rsr bit 31 0 0 0 bit 16 interrupt level 0 mask interrupt level 1 mask interrupt level 3 mask interrupt level 2 mask 1 0 0 0 0 1 1 0 0 0 0 0 0 fe i0m gm m1 - d m0 em i2m dm - m2 lc 1 ss initial value bit name i1m mode bit 0 mode bit 1 - mode bit 2 reserved debug state - i3m reserved exception mask global interrupt mask debug state mask secure state
24 32145c?06/2013 at32uc3l0128/256 figure 4-5. the status register low halfword 4.4.3 processor states 4.4.3.1 normal risc state the avr32 processor supports several diff erent execution contexts as shown in table 4-2 . mode changes can be made under software control, or can be caused by external interrupts or exception processing. a mode can be interrupted by a higher priority mode, but never by one with lower priority. nested exceptions can be supported with a minimal software overhead. when running an operating system on the avr32, user processes will typically execute in the application mode. the programs executed in this mode are restricted from executing certain instructions. furthermore, most system registers together with the upper halfword of the status register cannot be accessed. protected memory areas are also not available. all other operating modes are privileged and are collectively called system modes. they have full access to all priv- ileged and unprivileged re sources. after a reset, the proc essor will be in su pervisor mode. 4.4.3.2 debug state the avr32 can be set in a debug state, which allows implementation of software monitor rou- tines that can read out and alter system information for use during application development. this implies that all system and application regist ers, including the status registers and program counters, are accessible in debug state. th e privileged instructions are also available. all interrupt levels are by default disabled when debug state is entered, but they can individually be switched on by the monitor routine by clearing the respective mask bit in the status register. bit 15 bit 0 reserved carry zero sign 0 0 0 0 0 0 0 0 0 0 0 0 0 0 - - - - t - bit name initial value 0 0 l q v n z c - overflow saturation - - - lock reserved scratch table 4-2. overview of execution modes, thei r priorities and privilege levels. priority mode securi ty description 1 non maskable interrupt privileged non maskable high priority interrupt mode 2 exception privileged execute exceptions 3 interrupt 3 privileged general purpose interrupt mode 4 interrupt 2 privileged general purpose interrupt mode 5 interrupt 1 privileged general purpose interrupt mode 6 interrupt 0 privileged general purpose interrupt mode n/a supervisor privileged runs supervisor calls n/a application unprivileged normal program execution mode
25 32145c?06/2013 at32uc3l0128/256 debug state can be entered as described in the avr32uc technical reference manual . debug state is exited by the retd instruction. 4.4.3.3 secure state the avr32 can be set in a secure state, that allows a part of the code to execute in a state with higher security levels. the rest of the code can not access resources reserved for this secure code. secure state is used to implemen t flashvault technology. refer to the avr32uc techni- cal reference manual for details. 4.4.4 system registers the system registers are placed outside of the virtual memory space, and are only accessible using the privileged mfsr and mtsr instructions. the table below lis ts the system registers speci- fied in the avr32 architecture, some of which are unused in avr32uc. the programmer is responsible for maintaining correct sequen cing of any instructions following a mtsr instruction. for detail on the system registers, refer to the avr32uc technical reference manual . table 4-3. system registers reg # address name function 0 0 sr status register 1 4 evba exception vector base address 2 8 acba application call base address 3 12 cpucr cpu control register 4 16 ecr exception cause register 5 20 rsr_sup unused in avr32uc 6 24 rsr_int0 unused in avr32uc 7 28 rsr_int1 unused in avr32uc 8 32 rsr_int2 unused in avr32uc 9 36 rsr_int3 unused in avr32uc 10 40 rsr_ex unused in avr32uc 11 44 rsr_nmi unused in avr32uc 12 48 rsr_dbg return status register for debug mode 13 52 rar_sup unused in avr32uc 14 56 rar_int0 unused in avr32uc 15 60 rar_int1 unused in avr32uc 16 64 rar_int2 unused in avr32uc 17 68 rar_int3 unused in avr32uc 18 72 rar_ex unused in avr32uc 19 76 rar_nmi unused in avr32uc 20 80 rar_dbg return address register for debug mode 21 84 jecr unused in avr32uc 22 88 josp unused in avr32uc 23 92 java_lv0 unused in avr32uc
26 32145c?06/2013 at32uc3l0128/256 24 96 java_lv1 unused in avr32uc 25 100 java_lv2 unused in avr32uc 26 104 java_lv3 unused in avr32uc 27 108 java_lv4 unused in avr32uc 28 112 java_lv5 unused in avr32uc 29 116 java_lv6 unused in avr32uc 30 120 java_lv7 unused in avr32uc 31 124 jtba unused in avr32uc 32 128 jbcr unused in avr32uc 33-63 132-252 reserved reserved for future use 64 256 config0 configuration register 0 65 260 config1 configuration register 1 66 264 count cycle counter register 67 268 compare compare register 68 272 tlbehi unused in avr32uc 69 276 tlbelo unused in avr32uc 70 280 ptbr unused in avr32uc 71 284 tlbear unused in avr32uc 72 288 mmucr unused in avr32uc 73 292 tlbarlo unused in avr32uc 74 296 tlbarhi unused in avr32uc 75 300 pccnt unused in avr32uc 76 304 pcnt0 unused in avr32uc 77 308 pcnt1 unused in avr32uc 78 312 pccr unused in avr32uc 79 316 bear bus error address register 80 320 mpuar0 mpu address register region 0 81 324 mpuar1 mpu address register region 1 82 328 mpuar2 mpu address register region 2 83 332 mpuar3 mpu address register region 3 84 336 mpuar4 mpu address register region 4 85 340 mpuar5 mpu address register region 5 86 344 mpuar6 mpu address register region 6 87 348 mpuar7 mpu address register region 7 88 352 mpupsr0 mpu privilege select register region 0 89 356 mpupsr1 mpu privilege select register region 1 table 4-3. system registers (continued) reg # address name function
27 32145c?06/2013 at32uc3l0128/256 4.5 exceptions and interrupts in the avr32 architecture, events are used as a common term for exceptions and interrupts. avr32uc incorporates a powerful event handling scheme. the different event sources, like ille- gal op-code and interrupt requests, have different priority levels, ensuring a well-defined behavior when multiple events are received simultaneously. additionally, pending events of a higher priority class may preempt handling of ongoing events of a lower priority class. when an event occurs, the execution of the instru ction stream is halted, and execution is passed to an event handler at an address specified in table 4-4 on page 31 . most of the handlers are placed sequentially in the code sp ace starting at the ad dress specified by evba, with four bytes between each handler. this gives ample space for a jump instruction to be placed there, jump- ing to the event routine itself. a few critical handlers have larg er spacing between them, allowing the entire event routine to be placed directly at the address specified by the evba-relative offset generated by hardware. all interrupt sources have autovectored interrupt service routine (isr) addresses. this allows the interrupt controller to directly specify the isr address as an address 90 360 mpupsr2 mpu privilege select register region 2 91 364 mpupsr3 mpu privilege select register region 3 92 368 mpupsr4 mpu privilege select register region 4 93 372 mpupsr5 mpu privilege select register region 5 94 376 mpupsr6 mpu privilege select register region 6 95 380 mpupsr7 mpu privilege select register region 7 96 384 mpucra unused in this version of avr32uc 97 388 mpucrb unused in this version of avr32uc 98 392 mpubra unused in this version of avr32uc 99 396 mpubrb unused in this version of avr32uc 100 400 mpuapra mpu access permission register a 101 404 mpuaprb mpu access permission register b 102 408 mpucr mpu control register 103 412 ss_status secure state status register 104 416 ss_adrf secure state address flash register 105 420 ss_adrr secure state address ram register 106 424 ss_adr0 secure state address 0 register 107 428 ss_adr1 secure state address 1 register 108 432 ss_sp_sys secure state stack pointer system register 109 436 ss_sp_app secure state stac k pointer application register 110 440 ss_rar secure state return address register 111 444 ss_rsr secure state return status register 112-191 448-764 reserved reserved for future use 192-255 768-1020 impl implementation defined table 4-3. system registers (continued) reg # address name function
28 32145c?06/2013 at32uc3l0128/256 relative to evba. the autovector offset has 14 address bits, giving an offset of maximum 16384 bytes. the target address of the event handle r is calculated as (evba | event_handler_offset), not (evba + event_handler_offse t), so evba and exception code segments must be set up appropriately. the same mechanisms are used to service all different types of events, including interrupt requests, yielding a uniform event handling scheme. an interrupt controller does the priority handling of the interrupts and provides the autovector off- set to the cpu. 4.5.1 system stack issues event handling in avr32uc uses the system stack pointed to by the system stack pointer, sp_sys, for pushing and popping r8 -r12, lr, status register, and return ad dress. since event code may be timing-critical, sp_sys should point to memory addresses in the iram section, since the timing of accesses to this memory section is both fast and deterministic. the user must also make sure that the system stack is large enough so that any event is able to push the required registers to stack. if the system stack is full, and an event occurs, the system will enter an undefined state. 4.5.2 exceptions and interrupt requests when an event other than scall or debug request is received by the core, the following actions are performed atomically: 1. the pending event will not be acce pted if it is masked. the i3 m, i2m, i1m, i0m, em, and gm bits in the status register are used to mask different events. not all events can be masked. a few critical events (nmi, unreco verable exception, tlb multiple hit, and bus error) can not be masked. when an event is accepted, hardware automatically sets the mask bits corresponding to all sources with equal or lower priority. this inhibits acceptance of other events of the same or lower priority, except for the critical events listed above. software may choose to clear some or all of these bits after saving the necessary state if other priority schemes are desired. it is the event source?s respons- ability to ensure that their events are left pending until accepted by the cpu. 2. when a request is accepted, the status register and program counter of the current context is stored to the system stack. if the event is an int0, int1, int2, or int3, reg- isters r8-r12 and lr are also automatically stored to stack. storing the status register ensures that the core is returned to the previous execution mode when the current event handling is completed. when exceptions occur, both the em and gm bits are set, and the application may manually enable nested exceptions if desired by clear- ing the appropriate bit. each exception handler has a dedicated handler address, and this address uniquely identifies the exception source. 3. the mode bits are set to reflect the priority of the accepted event, and the correct regis- ter file bank is selected. the address of the event handler, as shown in table 4-4 on page 31 , is loaded into the program counter. the execution of the event handler routine then continues from the effective address calculated. the rete instruction signals the end of the event. when encountered, the return status register and return address register are popped from the system stack and restored to the status reg- ister and program counter. if the rete instruction returns from int0, int1, int2, or int3, registers r8-r12 and lr are also popped from the system stack. the restored status register contains information allowing the core to resume operation in the previous execution mode. this concludes the event handling.
29 32145c?06/2013 at32uc3l0128/256 4.5.3 supervisor calls the avr32 instruction set provides a supervisor mode call instruction. the scall instruction is designed so that privileged routines can be called from any context. this facilitates sharing of code between different execution modes. the scall mechanism is designed so that a minimal execution cycle overhead is experienced when performing supervisor routine calls from time- critical event handlers. the scall instruction behaves differently depending on which mode it is called from. the behav- iour is detailed in the instruction se t reference. in order to allow the scall routine to return to the correct context, a return from supervisor call instruction, rets , is implemented. in the avr32uc cpu, scall and rets uses the system stack to store the return address and the status register. 4.5.4 debug requests the avr32 architecture defines a dedicated debug mode. when a debug request is received by the core, debug mode is entered. entry into debug mode can be masked by the dm bit in the status register. upon entry into debug mode, hardware sets the sr.d bit and jumps to the debug exception handler. by default, debug mode executes in the exception context, but with dedicated return address register and return status register. these dedicated registers remove the need for storing this data to the system stack, t hereby improving debuggability. the mode bits in the status register can freely be manipulated in debug mode, to observe registers in all contexts, while retaining full privileges. debug mode is exited by executing the retd instruction. this return s to the previous context. 4.5.5 entry points for events several different event handler entry points exist. in avr32uc, the reset address is 0x80000000. this places the reset address in the boot flash memory area. tlb miss exceptions and scall have a dedicated space relative to evba where their event han- dler can be placed. this speeds up execution by removing the need for a jump instruction placed at the program address jumped to by the event hardware. all other exceptions have a dedicated event routine entry point located relative to evba. the handler routine address identifies the exception source directly. avr32uc uses the itlb and dtlb protection exc eptions to signal a mp u protection violation. itlb and dtlb miss exceptions are used to signal that an access address did not map to any of the entries in the mpu. tlb multiple hit exception indicates that an access address did map to multiple tlb entries, signalling an error. all interrupt requests have entry points located at an offset relative to evba. this autovector off- set is specified by an interrupt controller. the programmer must make sure that none of the autovector offsets interfere with the placement of other code. the autovector offset has 14 address bits, giving an offset of maximum 16384 bytes. special considerations should be made when loading evba with a po inter. due to security con- siderations, the event handlers should be located in non-writeable flash memory, or optionally in a privileged memory protection region if an mpu is present. if several events occur on the same instruction, they are handled in a prioritized way. the priority ordering is presented in table 4-4 on page 31 . if events occur on several instructions at different locations in the pipeline, the events on the oldest instruction are always handled before any events on any younger instruction, even if the younger instruction has events of higher priority
30 32145c?06/2013 at32uc3l0128/256 than the oldest instruction. an instruction b is younger than an instruction a if it was sent down the pipeline later than a. the addresses and priority of simultaneous events are shown in table 4-4 on page 31 . some of the exceptions are unused in avr32uc since it has no mmu, coprocessor interface, or floating- point unit.
31 32145c?06/2013 at32uc3l0128/256 table 4-4. priority and handler addresses for events priority handler address name event source stored return address 1 0x80000000 reset external input undefined 2 provided by ocd system ocd stop cpu ocd system first non-compl eted instruction 3 evba+0x00 unrecoverable exception int ernal pc of offending instruction 4 evba+0x04 tlb multiple hit mpu pc of offending instruction 5 evba+0x08 bus error data fetch data bu s first non-completed instruction 6 evba+0x0c bus error instruction fetch dat a bus first non-completed instruction 7 evba+0x10 nmi external input first non-completed instruction 8 autovectored interrupt 3 request external input first non-completed instruction 9 autovectored interrupt 2 request external input first non-completed instruction 10 autovectored interrupt 1 request external input first non-completed instruction 11 autovectored interrupt 0 request external input first non-completed instruction 12 evba+0x14 instruction address cp u pc of offending instruction 13 evba+0x50 itlb miss mpu pc of offending instruction 14 evba+0x18 itlb protection mpu pc of offending instruction 15 evba+0x1c breakpoint ocd system firs t non-completed instruction 16 evba+0x20 illegal opcode instructio n pc of offending instruction 17 evba+0x24 unimplemented instruction instr uction pc of offending instruction 18 evba+0x28 privilege violation instruc tion pc of offending instruction 19 evba+0x2c floating-point unused 20 evba+0x30 coprocessor absent instruct ion pc of offending instruction 21 evba+0x100 supervisor call instru ction pc(supervisor call) +2 22 evba+0x34 data address (read) cp u pc of offending instruction 23 evba+0x38 data address (write) cpu pc of offending instruction 24 evba+0x60 dtlb miss (read) mpu pc of offending instruction 25 evba+0x70 dtlb miss (write) mpu pc of offending instruction 26 evba+0x3c dtlb protection (read) mpu pc of offending instruction 27 evba+0x40 dtlb protection (write) m pu pc of offending instruction 28 evba+0x44 dtlb modified unused
32 32145c?06/2013 at32uc3l0128/256 5. memories 5.1 embedded memories ? internal high-speed flash ? 256kbytes (at32uc3l0256) ? 128kbytes (at32uc3l0128) ? 0 wait state access at up to 25mhz in worst case conditions ? 1 wait state access at up to 50mhz in worst case conditions ? pipelined flash architecture, allowing burst r eads from sequen tial flash loca tions, hiding penalty of 1 wait state access ? pipelined flash architecture typically reduce s the cycle penalty of 1 wait state operation to only 8% compared to 0 wait state operation ? 100 000 write cycles, 15-year data retention capability ? sector lock capabilities, boot loader protection, security bit ? 32 fuses, erased during chip erase ? user page for data to be preserved during chip erase ? internal high-speed sram, sing le-cycle access at full speed ?32kbytes 5.2 physical memory map the system bus is implemented as a bus matrix . all system bus addresses are fixed, and they are never remapped in any way, not even during boot. note that avr32 uc cpu uses unseg- mented translation, as described in the avr32 ar chitecture manual. the 32-bit physical address space is mapped as follows: table 5-1. at32uc3l0128/256 physical memory map device start address size at32uc3l0256 at32uc3l0128 embedded sram 0x00000000 32kbytes 32kbytes embedded flash 0x80000000 256kbytes 128kbytes sau channels 0x90000000 256 bytes 256 bytes hsb-pb bridge b 0xfffe0000 64kbytes 64kbytes hsb-pb bridge a 0xffff0000 64kbytes 64kbytes table 5-2. flash memory parameters part number flash size ( flash_pw ) number of pages ( flash_p ) page size ( flash_w ) at32uc3l0256 256kbytes 512 512bytes at32uc3l0128 128kbytes 256 512bytes
33 32145c?06/2013 at32uc3l0128/256 5.3 peripheral address map table 5-3. peripheral address mapping address peripheral name 0xfffe0000 flashcdw flash controller - flashcdw 0xfffe0400 hmatrix hsb matrix - hmatrix 0xfffe0800 sau secure access unit - sau 0xffff0000 pdca peripheral dma controller - pdca 0xffff1000 intc interrupt controller - intc 0xffff1400 pm power manager - pm 0xffff1800 scif system control interface - scif 0xffff1c00 ast asynchronous timer - ast 0xffff2000 wdt watchdog timer - wdt 0xffff2400 eic external interrupt controller - eic 0xffff2800 freqm frequency meter - freqm 0xffff2c00 gpio general-purpose input/output controller - gpio 0xffff3000 usart0 universal synchronous asynchronous receiver transmitter - usart0 0xffff3400 usart1 universal synchronous asynchronous receiver transmitter - usart1 0xffff3800 usart2 universal synchronous asynchronous receiver transmitter - usart2 0xffff3c00 usart3 universal synchronous asynchronous receiver transmitter - usart3 0xffff4000 spi serial peripheral interface - spi 0xffff4400 twim0 two-wire master interface - twim0
34 32145c?06/2013 at32uc3l0128/256 5.4 cpu local bus mapping some of the registers in the gpio module are mapped onto the cpu local bus, in addition to being mapped on the peripheral bus. these registers can therefore be reached both by accesses on the peripheral bus, and by accesses on the local bus. mapping these registers on the local bus allows cycle-deterministic toggling of gpio pins since the cpu and gpio are the only modules connected to this bus. also, since the local bus runs at cpu speed, one write or read operation can be pe rformed per clock cycle to the local bus- mapped gpio registers. 0xffff4800 twim1 two-wire master interface - twim1 0xffff4c00 twis0 two-wire slave interface - twis0 0xffff5000 twis1 two-wire slave interface - twis1 0xffff5400 pwma pulse width modulation controller - pwma 0xffff5800 tc0 timer/counter - tc0 0xffff5c00 tc1 timer/counter - tc1 0xffff6000 adcifb adc interface - adcifb 0xffff6400 acifb analog comparator interface - acifb 0xffff6800 cat capacitive touch module - cat 0xffff6c00 gloc glue logic controller - gloc 0xffff7000 aw awire - aw table 5-3. peripheral address mapping
35 32145c?06/2013 at32uc3l0128/256 the following gpio registers are mapped on the local bus: table 5-4. local bus mapped gpio registers port register mode local bus address access 0 output driver enable register (oder) write 0x40000040 write-only set 0x40000044 write-only clear 0x40000048 write-only toggle 0x4000004c write-only output value register (ovr) write 0x40000050 write-only set 0x40000054 write-only clear 0x40000058 write-only toggle 0x4000005c write-only pin value register (pvr) - 0x40000060 read-only 1 output driver enable register (oder) write 0x40000140 write-only set 0x40000144 write-only clear 0x40000148 write-only toggle 0x4000014c write-only output value register (ovr) write 0x40000150 write-only set 0x40000154 write-only clear 0x40000158 write-only toggle 0x4000015c write-only pin value register (pvr) - 0x40000160 read-only
36 32145c?06/2013 at32uc3l0128/256 6. supply and startup considerations 6.1 supply considerations 6.1.1 power supplies the at32uc3l0128/256 has several types of power supply pins: ?vddio: powers i/o lines. voltage is 1.8 to 3.3v nominal. ?vddin: powers i/o lines and the internal regulator. voltage is 1.8 to 3.3v nominal. ?vddana: powers the adc. voltage is 1.8v nominal. ?vddcore: powers the core, memories, and peripherals. voltage is 1.8v nominal. the ground pins gnd are common to vddcore, vddio, and vddin. the ground pin for vddana is gndana. when vddcore is not connected to vddin, t he vddin voltage must be higher than 1.98v. refer to section 32. on page 791 for power consumption on the various supply pins. for decoupling recommendations for the different power supplies, please refer to the schematic checklist. refer to section 3.2 on page 8 for power supply connections for i/o pins. 6.1.2 voltage regulator the at32uc3l0128/256 embeds a voltage regulator that converts from 3.3v nominal to 1.8v with a load of up to 60ma. the regulator supplies the output voltage on vddcore. the regula- tor may only be used to drive internal circuitr y in the device. vddcore should be externally connected to the 1.8v domains. see section 6.1.3 for regulator connection figures. adequate output supply decoupling is mandat ory for vddcore to reduce ripple and avoid oscillations. the best way to achieve this is to use two capacitors in parallel between vddcore and gnd as close to the device as possible. please refer to section 32.8.1 on page 805 for decoupling capacitors values and regulator characteristics. figure 6-1. supply decoupling the voltage regulator can be turned off in the shutdown mode to power down the core logic and keep a small part of the system powered in order to reduce power consumption. to enter this mode the 3.3v supply mode, with 1.8v regulated i/o lines power supply configuration must be used. 3.3v 1.8v vddin vddcore 1.8v regulator c in1 c out1 c out2 c in2 in3 c
37 32145c?06/2013 at32uc3l0128/256 6.1.3 regulator connection the at32uc3l0128/256 supports three power supply configurations: ? 3.3v single supply mode ? shutdown mode is not available ? 1.8v single supply mode ? shutdown mode is not available ? 3.3v supply mode, with 1.8v regulated i/o lines ? shutdown mode is available 6.1.3.1 3.3v single supply mode in 3.3v single supply mode the internal regulator is connected to the 3.3v source (vddin pin) and its output feeds vddcore. figure 6-2 shows the power schematics to be used for 3.3v single supply mode. all i/o lines will be po wered by the same power (vddin=vddio). figure 6-2. 3.3v single supply mode vddio vddcore + - 1.98-3.6v vddana adc vddin gnd gndana cpu, peripherals, memories, scif, bod, rcsys, dfll, pll osc32k, rc32k, por33, sm33 i/o pins i/o pins osc32k_2, ast, wake, regulator control linear regulator
38 32145c?06/2013 at32uc3l0128/256 6.1.3.2 1.8 v single supply mode in 1.8v single supply mode the internal regul ator is not used, and vddio and vddcore are powered by a single 1.8 v supply as shown in figure 6-3 . all i/o lines will be powered by the same power (vddin = vddio = vddcore). figure 6-3. 1.8v single supply mode. vddio vddcore + - 1.62-1.98v vddana adc vddin gnd gndana cpu, peripherals, memories, scif, bod, rcsys, dfll, pll osc32k, rc32k, por33, sm33 i/o pins i/o pins osc32k_2, ast, wake, regulator control
39 32145c?06/2013 at32uc3l0128/256 6.1.3.3 3.3v supply mode with 1.8 v regulated i/o lines in this mode, the internal regulator is connecte d to the 3.3v source and its output is connected to both vddcore and vddio as shown in figure 6-4 . this configuration is required in order to use shutdown mode. figure 6-4. 3.3v supply mode with 1.8v regulated i/o lines in this mode, some i/o lines are powered by v ddin while other i/o lines are powered by vddio. refer to section 3.2.1 on page 8 for description of power supply for each i/o line. refer to the power manager chapter for a description of what parts of the system are powered in shutdown mode. important note: as the regulator has a maximum output current of 60 ma, this mode can only be used in applications where the maximum i/o current is known and compatible with the core and peripheral power consumption. typically, great care must be used to ensure that only a few i/o lines are toggling at the same time and drive very small loads. vddio vddcore + - 1.98-3.6v vddana adc vddin gnd gndana cpu, peripherals, memories, scif, bod, rcsys, dfll, pll osc32k, rc32k, por33, sm33 i/o pins i/o pins osc32k_2, ast, wake, regulator control linear regulator
40 32145c?06/2013 at32uc3l0128/256 6.1.4 power-up sequence 6.1.4.1 maximum rise rate to avoid risk of latch-up, the rise rate of the power supplies must not exceed the values described in table 32-3 on page 792 . recommended order for power supplies is also described in this chapter. 6.1.4.2 minimum rise rate the integrated power-on reset (por33) circuitry monitoring the vddin powering supply requires a minimum rise rate for the vddin power supply. see table 32-3 on page 792 for the minimum rise rate value. if the application can not ensure that the minimum rise rate condition for the vddin power sup- ply is met, one of the following configurations can be used: ? a logic ?0? value is applied during power-up on pin pa11 (wake_n) until vddin rises above 1.2v. ? a logic ?0? value is applied during power-up on pin reset_n until vddin rises above 1.2v. 6.2 startup considerations this chapter summarizes the boot sequence of the at32uc3l0128/256. the behavior after power-up is controlled by the power manager. for specific details, refer to the power manager chapter. 6.2.1 starting of clocks after power-up, the device will be held in a reset state by the power-on reset (por18 and por33) circuitry for a short time to allow t he power to stabilize thr oughout the device. after reset, the device will use the system rc oscillat or (rcsys) as clock source. please refer to table 32-17 on page 804 for the frequency fo r this oscillator. on system start-up, all high-speed clocks are disabled. all clocks to all modules are running. no clocks have a divided frequency; all parts of th e system receive a clock with the same frequency as the system rc oscillator. when powering up the device, there may be a delay before the voltage has stabilized, depend- ing on the rise time of the supply used. the cpu can start executing code as soon as the supply is above the por18 and por33 thresholds, and bef ore the supply is stable. before switching to a high-speed clock source, the user should use the bod to make sure the vddcore is above the minimum level (1.62v). 6.2.2 fetching of initial instructions after reset has been released, the avr32 uc cpu starts fetching instructions from the reset address, which is 0x80000000. this address points to the first address in the internal flash. the code read from the internal flash is free to configure the clock system and clock sources. please refer to the pm and scif chapters for more details.
41 32145c?06/2013 at32uc3l0128/256 7. peripheral dma controller (pdca) rev: 1.2.3.1 7.1 features ? multiple channels ? generates transfers between memories and peripherals such as usart and spi ? two address pointers/counters per ch annel allowing double buffering ? performance monitors to measure average and maximum transfer latency ? optional synchronizing of data transfers with extenal peripheral events ? ring buffer functionality 7.2 overview the peripheral dma controller (pdca) transfers data between on-chip peripheral modules such as usart, spi and memories (those memories ma y be on- and off-chip memories). using the pdca avoids cpu intervention for data transfers, improving the performance of the microcon- troller. the pdca can transfer data from memory to a peripheral or from a peripheral to memory. the pdca consists of multiple dma channels. each channel has: ? a peripheral select register ? a 32-bit memory pointer ? a 16-bit transfer counter ? a 32-bit memory pointer reload value ? a 16-bit transfer counter reload value the pdca communicates with the peripheral modules over a set of handshake interfaces. the peripheral signals the pdca when it is ready to receive or transmit data. the pdca acknowl- edges the request when the transmission has started. when a transmit buffer is empty or a receive buffer is full, an optional interrupt request can be generated.
42 32145c?06/2013 at32uc3l0128/256 7.3 block diagram figure 7-1. pdca block diagram 7.4 product dependencies in order to use this module, other parts of the system must be configured correctly, as described below. 7.4.1 power management if the cpu enters a sleep mode that disables the pdca clocks, the pdca will stop functioning and resume operation after the system wakes up from sleep mode. 7.4.2 clocks the pdca has two bus clocks connected: one high speed bus clock (clk_pdca_hsb) and one peripheral bus clock (clk_pdca_pb). thes e clocks are generated by the power man- ager. both clocks are enabled at reset, and can be disabled in the power manager. it is recommended to disable the pdca before disabling the clocks, to avoid freezing the pdca in an undefined state. 7.4.3 interrupts the pdca interrupt request lines are connected to the interrupt controller. using the pdca interrupts requires the interrupt controller to be programmed first. hsb to pb bridge peripheral dma controller (pdca) peripheral 0 high speed bus matrix handshake interfaces peripheral bus irq hsb hsb interrupt controller peripheral 1 peripheral 2 peripheral (n-1) ... memory hsb
43 32145c?06/2013 at32uc3l0128/256 7.4.4 peripheral events the pdca peripheral events are connected via the peripheral event system. refer to the peripheral event system chapter for details. 7.5 functional description 7.5.1 basic operation the pdca consists of multiple independent pdca channels, each capable of handling dma requests in parallel. each pdca channels contains a set of configuration registers which must be configured to start a dma transfer. in this section the steps necessary to configure one pdca channel is outlined. the peripheral to transfer data to or from must be configured correctly in the peripheral select register (psr). this is performed by writing the peripheral identity (pid) value for the corre- sponding peripheral to the pid field in the ps r register. the pid also encodes the transfer direction, i.e. memory to peripheral or peripheral to memory. see section 7.5.6 . the transfer size must be written to the transfer size field in the mode register (mr.size). the size must match the data size produced or consumed by the selected peripheral. see section 7.5.7 . the memory address to transfer to or from, depending on the psr, must be written to the mem- ory address register (mar). for each transfer the memory address is increased by either a one, two or four, depending on the size set in mr. see section 7.5.2 . the number of data items to transfer is written to the tcr register. if the pdca channel is enabled, a transfer will start imm ediately after writing a non-zero value to tc r or the reload ver- sion of tcr, tcrr. after each transfer the tcr value is decreased by one. both mar and tcr can be read while the pdca channel is active to monitor the dma progress. see section 7.5.3 . the channel must be enabled for a transfer to start. a channel is enable by writing a one to the en bit in the control register (cr). 7.5.2 memory pointer each channel has a 32-bit memory address regi ster (mar). this register holds the memory address for the next transfer to be performed. the register is automatically updated after each transfer. the address will be increa sed by either one, two or four depending on the size of the dma transfer (byte, halfword or word). the mar can be read at any time during transfer. 7.5.3 transfer counter each channel has a 16-bit transfer counter register (tcr). this register must be written with the number of transfers to be performed. the tcr register should contain the number of data items to be transferred independently of the transfer size. the tcr can be read at any time dur- ing transfer to see the number of remaining transfers. 7.5.4 reload registers both the mar and the tcr have a reload regist er, respectively memory address reload regis- ter (marr) and transfer counter reload register (tcrr). these registers provide the possibility for the pdca to work on two memory buffer s for each channel. when one buffer has completed, mar and tcr will be re loaded with the values in marr and tcrr. the reload logic is always enabled and will trigger if the tcr reaches zero while tcrr holds a non-zero value. after reload, the marr and tcrr registers are cleared.
44 32145c?06/2013 at32uc3l0128/256 if tcr is zero when writing to tcrr, the tcr and mar are automatically updated with the value written in tcrr and marr. 7.5.5 ring buffer when ring buffer mode is enabled the tcrr and marr regi sters will not be cleared when tcr and mar registers reload. th is allows the pdca to read or write to the same memory region over and over again until the transfer is actively stopped by the user. ring buffer mode is enabled by writing a one to the ring buffer bit in the mode register (mr.ring). 7.5.6 peripheral selection the peripheral select register (psr) decides which peripheral should be connected to the pdca channel. a peripheral is selected by writing the corresponding peripheral identity (pid) to the pid field in the psr register. writing the pid will both select the direction of the transfer (memory to peripheral or peripheral to memory ), which handshake interface to use, and the address of the peripheral holding register. refer to the peripheral identity (pid) table in the mod- ule configuration section for the peripheral pid values. 7.5.7 transfer size the transfer size can be set individually for each channel to be either byte, halfword or word (8- bit, 16-bit or 32-bit respectively). transfer size is set by writing the desired value to the transfer size field in the mode register (mr.size). when the pdca moves data between peripherals and memory, data is automatically sized and aligned. when memory is accessed, the size specified in mr.size and system alignment is used. when a peripheral register is accessed t he data to be transferred is converted to a word where bit n in the data corresponds to bit n in the peripheral register. if the transfer size is byte or halfword, bits greater than 8 and16 respectively are set to zero. refer to the module configuration section for information regarding what peripheral registers are used for the different peripherals and then to the peripheral specific chapter for information about the size option available for the different registers. 7.5.8 enabling and disabling each dma channel is enabled by writing a one to the transfer enable bit in the control register (cr.ten) and disabled by writing a one to the transfer disable bit (cr.tdis). the current sta- tus can be read from the status register (sr). while the pdca channel is enabled all dma re quest will be handled as long the tcr and tcrr is not zero. 7.5.9 interrupts interrupts can be enabled by writing a one to the corresponding bit in the interrupt enable regis- ter (ier) and disabled by writing a one to the corresponding bit in the interrupt disable register (idr). the interrupt mask register (imr) can be read to see whether an interrupt is enabled or not. the current status of an interrupt source can be read through the interrupt status register (isr). the pdca has three interrupt sources: ? reload counter zero - the tcrr register is zero. ? transfer finished - both the tcr and tcrr registers are zero. ? transfer error - an error has occurred in accessing memory.
45 32145c?06/2013 at32uc3l0128/256 7.5.10 priority if more than one pdca channel is requesting transfer at a given time, the pdca channels are prioritized by their channel number. channels wi th lower numbers have priority over channels with higher numbers, giving channel zero the highest priority. 7.5.11 error handling if the memory address register (mar) is set to point to an invalid location in memory, an error will occur when the pdca tries to perform a transfer. when an error occurs, the transfer error bit in the interrup t status register (isr.terr) will be se t and the dma channel that caused the error will be stopped. in order to restart the channel, the user must program the memory address register to a valid address and then write a one to the error clear bit in the control register (cr.eclr). if the transfer error interrupt is enabled , an interrupt request will be gener- ated when a transfer error occurs. 7.5.12 peripheral event trigger peripheral events can be used to trigger pdca channel transfers. peripheral event synchroniza- tions are enabled by writing a one to the event trigger bit in the mode register (mr.etrig). when set, all dma requests will be blocked until a peripheral event is re ceived. for each periph- eral event received, only one data item is transferred. if no dma requests are pending when a peripheral event is received, the pdca will st art a transfer as soon as a peripheral event is detected. if multiple events are received while the pdca channel is busy transferring data, an overflow condition will be signaled in the peripheral ev ent system. refer to the peripheral event system chapter for more information. 7.6 performance monitors up to two performance monitors allow the user to measure the activity and stall cycles for pdca transfers. to monitor a pdca channel, the corresponding channel number must be written to one of the mon0/1ch fields in the performance control register (pcontrol) and a one must be written to the corresponding ch0/1en bit in the same register. due to performance monitor hardware resource sharing, the two monitor channels should not be programmed to monitor the same pdca channel. this may result in undefined perfor- mance monitor behavior. 7.6.1 measuring mechanisms three different parameters can be measured by each channel: ? the number of data transfer cycles since last channel reset, both for read and write ? the number of stall cycles since last channel reset, both for read and write ? the maximum latency since last channel reset, both for read and write these measurements can be extracted by soft ware and used to generate indicators for bus latency, bus load, and maximum bus latency. each of the counters has a fixed width, and may therefore overflow. when an overflow is encountered in either the performance channel data read/write cycle registers (prdata0/1 and pwdata0/1) or the performance channel read/write stall cycles registers (prstall0/1 and pwstall0/1) of a channel, all registers in the channel are reset. this behavior is altered if the channel overflow freeze bit is one in the performance control register (pcon- trol.ch0/1ovf). if this bit is one, the channel registers are frozen when either data or stall reaches its maximum value. this simplifies one-shot readout of the counter values.
46 32145c?06/2013 at32uc3l0128/256 the registers can also be manually reset by writing a one to the channel reset bit in the pcon- trol register (pcontrol.ch0/1res). the performance channel read/write latency registers (prlat0/1 and pwlat0/1) are saturating when their maximum count value is reached. the prlat0/1 and pwlat0/1 registers can only be reset by writing a one to the cor- responding reset bit in pcontrol (pcontrol.ch0/1res). a counter is enabled by writing a one to the channel enable bit in the performance control reg- ister (pcontrol.ch0/1en).
47 32145c?06/2013 at32uc3l0128/256 7.7 user interface 7.7.1 memory map overview the channels are mapped as shown in table 7-1 . each channel has a set of configuration regis- ters, shown in table 7-2 , where n is the channel number. 7.7.2 channel memory map note: 1. the reset values are device specific. please refer to the module config uration section at the end of this chapter. table 7-1. pdca register memory map address range contents 0x000 - 0x03f dma channel 0 configuration registers 0x040 - 0x07f dma channel 1 configuration registers ... ... (0x000 - 0x03f)+m*0x040 dma channel m configuration registers 0x800-0x830 performance monitor registers 0x834 version register table 7-2. pdca channel configuration registers offset register register name access reset 0x000 + n*0x040 memory address register mar read/write 0x00000000 0x004 + n*0x040 peripheral select register psr read/write - (1) 0x008 + n*0x040 transfer counter register tcr read/write 0x00000000 0x00c + n*0x040 memory address reload register marr read/write 0x00000000 0x010 + n*0x040 transfer counter reload register tcrr read/write 0x00000000 0x014 + n*0x040 control register cr write-only 0x00000000 0x018 + n*0x040 mode register mr read/write 0x00000000 0x01c + n*0x040 status register sr read-only 0x00000000 0x020 + n*0x040 interrupt enable register ier write-only 0x00000000 0x024 + n*0x040 interrupt disable register idr write-only 0x00000000 0x028 + n*0x040 interrupt mask register imr read-only 0x00000000 0x02c + n*0x040 interrupt status register isr read-only 0x00000000
48 32145c?06/2013 at32uc3l0128/256 7.7.3 performance monitor memory map note: 1. the number of performance monitors is device specific. if the device has only one perfor- mance monitor, the channel1 r egisters are not available. please refer to the module configuration section at the end of this chapter for the number of performance monitors on this device. 7.7.4 version register memory map note: 1. the reset values are device specific. please refer to the module configuration section at the end of this chapter. table 7-3. pdca performance monitor registers (1) offset register register name access reset 0x800 performance control register pcontrol read/write 0x00000000 0x804 channel0 read data cycles prdata0 read-only 0x00000000 0x808 channel0 read stall cycles prstall0 read-only 0x00000000 0x80c channel0 read max latency prlat0 read-only 0x00000000 0x810 channel0 write data cycles pwdata0 read-only 0x00000000 0x814 channel0 write stall cycles pwstall0 read-only 0x00000000 0x818 channel0 write max latency pwlat0 read-only 0x00000000 0x81c channel1 read data cycles prdata1 read-only 0x00000000 0x820 channel1 read stall cycles prstall1 read-only 0x00000000 0x824 channel1 read max latency prlat1 read-only 0x00000000 0x828 channel1 write data cycles pwdata1 read-only 0x00000000 0x82c channel1 write stall cycles pwstall1 read-only 0x00000000 0x830 channel1 write max latency pwlat1 read-only 0x00000000 table 7-4. pdca version register memory map offset register register name access reset 0x834 version register version read-only - (1)
49 32145c?06/2013 at32uc3l0128/256 7.7.5 memory address register name: mar access type: read/write offset: 0x000 + n*0x040 reset value: 0x00000000 ? maddr: memory address address of memory buffer. maddr should be programmed to poin t to the start of the memory buffer when configuring the pdca. during transfer, maddr will point to the next memory location to be read/written. 31 30 29 28 27 26 25 24 maddr[31:24] 23 22 21 20 19 18 17 16 maddr[23:16] 15 14 13 12 11 10 9 8 maddr[15:8] 76543210 maddr[7:0]
50 32145c?06/2013 at32uc3l0128/256 7.7.6 peripheral select register name: psr access type: read/write offset: 0x004 + n*0x040 reset value: - ? pid: peripheral identifier the peripheral identifier selects which peripheral should be con nected to the dma channel. writ ing a pid will select both which handshake interface to use, the direction of the transfer and also the address of the receive/transfer holding register for the peripheral. see the module configuration section of pdca for details. the width of the pid field is device specific and dependent on the number of peripheral modules in the device. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 pid
51 32145c?06/2013 at32uc3l0128/256 7.7.7 transfer counter register name: tcr access type: read/write offset: 0x008 + n*0x040 reset value: 0x00000000 ? tcv: transfer counter value number of data items to be transferred by the pdca. tcv must be programmed with the total number of transfers to be made. during transfer, tcv contains the number of remaining transfers to be done. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 tcv[15:8] 76543210 tcv[7:0]
52 32145c?06/2013 at32uc3l0128/256 7.7.8 memory address reload register name: marr access type: read/write offset: 0x00c + n*0x040 reset value: 0x00000000 ? marv: memory address reload value reload value for the mar register. this value will be loaded in to mar when tcr reaches zero if the tcrr register has a non- zero value. 31 30 29 28 27 26 25 24 marv[31:24] 23 22 21 20 19 18 17 16 marv[23:16] 15 14 13 12 11 10 9 8 marv[15:8] 76543210 marv[7:0]
53 32145c?06/2013 at32uc3l0128/256 7.7.9 transfer counter reload register name: tcrr access type: read/write offset: 0x010 + n*0x040 reset value: 0x00000000 ? tcrv: transfer counter reload value reload value for the tcr register. when tcr reaches zero, it will be reloaded with tcrv if tcrv has a positive value. if tcrv is zero, no more transfers will be performed for the chann el. when tcr is reloaded, the tcrr register is cleared. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 tcrv[15:8] 76543210 tcrv[7:0]
54 32145c?06/2013 at32uc3l0128/256 7.7.10 control register name: cr access type: write-only offset: 0x014 + n*0x040 reset value: 0x00000000 ? eclr: transfer error clear writing a zero to this bit has no effect. writing a one to this bit will clear the transfer error bit in the status register (sr.terr). clearing the sr.terr bit will all ow the channel to transmit data. the memory address must first be set to point to a valid location. ? tdis: transfer disable writing a zero to this bit has no effect. writing a one to this bit will disable transfer for the dma channel. ? ten: transfer enable writing a zero to this bit has no effect. writing a one to this bit will enable transfer for the dma channel. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------eclr 76543210 ------tdisten
55 32145c?06/2013 at32uc3l0128/256 7.7.11 mode register name: mr access type: read/write offset: 0x018 + n*0x040 reset value: 0x00000000 ? ring: ring buffer 0:the ring buffer functionality is disabled. 1:the ring buffer functionality is enabled. when enabled, the reload registers, marr and tcrr will not be cleared after reload. ? etrig: event trigger 0:start transfer when the peripheral selected in periph eral select register (psr) requests a transfer. 1:start transfer only when or after a peripheral event is received. ? size: size of transfer 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 - - - - ring etrig size table 7-5. size of transfer size size of transfer 0 byte 1 halfword 2word 3 reserved
56 32145c?06/2013 at32uc3l0128/256 7.7.12 status register name: sr access type: read-only offset: 0x01c + n*0x040 reset value: 0x00000000 ? ten: transfer enabled this bit is cleared when the tdis bit in cr is written to one. this bit is set when the ten bit in cr is written to one. 0: transfer is disabled for the dma channel. 1: transfer is enabled for the dma channel. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 -------ten
57 32145c?06/2013 at32uc3l0128/256 7.7.13 interrupt enable register name: ier access type: write-only offset: 0x020 + n*0x040 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will set the corresponding bit in imr. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 - - - - - terr trc rcz
58 32145c?06/2013 at32uc3l0128/256 7.7.14 interrupt disable register name: idr access type: write-only offset: 0x024 + n*0x040 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in imr. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 - - - - - terr trc rcz
59 32145c?06/2013 at32uc3l0128/256 7.7.15 interrupt mask register name: imr access type: read-only offset: 0x028 + n*0x040 reset value: 0x00000000 0: the corresponding interrupt is disabled. 1: the corresponding interrupt is enabled. a bit in this register is cleared when the corresponding bit in idr is written to one. a bit in this register is set when the corresponding bit in ier is written to one. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 - - - - - terr trc rcz
60 32145c?06/2013 at32uc3l0128/256 7.7.16 interrupt status register name: isr access type: read-only offset: 0x02c + n*0x040 reset value: 0x00000000 ? terr: transfer error this bit is cleared when no transfer errors ha ve occurred since the last write to cr.eclr. this bit is set when one or more transfer errors has occurred since reset or the last write to cr.eclr. ? trc: transfer complete this bit is cleared when the tcr and/or the tcrr holds a non-zero value. this bit is set when both the tcr and the tcrr are zero. ? rcz: reload counter zero this bit is cleared when the tcrr holds a non-zero value. this bit is set when tcrr is zero. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 - - - - - terr trc rcz
61 32145c?06/2013 at32uc3l0128/256 7.7.17 performance control register name: pcontrol access type: read/write offset: 0x800 reset value: 0x00000000 ? mon1ch: performance monitor channel 1 ? mon0ch: performance monitor channel 0 the pdca channel number to monitor with counter n due to performance monitor hardware resour ce sharing, the two performance monitor channels should not be programmed to monitor the same pdca channel. this may result in undefined monitor behavior. ? ch1res: performance channel 1 counter reset writing a zero to this bit has no effect. writing a one to this bit will reset the counter in the channel specified in mon1ch. this bit always reads as zero. ? ch0res: performance channel 0 counter reset writing a zero to this bit has no effect. writing a one to this bit will reset the counter in the channel specified in mon0ch. this bit always reads as zero. ? ch1of: channel 1 overflow freeze 0: the performance channel registers are reset if data or stall overflows. 1: all performance channel registers are frozen just before data or stall overflows. ? ch1of: channel 0 overflow freeze 0: the performance channel registers are reset if data or stall overflows. 1: all performance channel registers are frozen just before data or stall overflows. ? ch1en: performance channel 1 enable 0: performance channel 1 is disabled. 1: performance channel 1 is enabled. ? ch0en: performance channel 0 enable 0: performance channel 0 is disabled. 1: performance channel 0 is enabled. 31 30 29 28 27 26 25 24 -- mon1ch 23 22 21 20 19 18 17 16 -- mon0ch 15 14 13 12 11 10 9 8 ------ch1resch0res 76543210 - - ch1of ch0of - - ch1en ch0en
62 32145c?06/2013 at32uc3l0128/256 7.7.18 performance channel 0 read data cycles name: prdata0 access type: read-only offset: 0x804 reset value: 0x00000000 ? data: data cycles counted since last reset clock cycles are counted us ing the clk_pdca_hsb clock 31 30 29 28 27 26 25 24 data[31:24] 23 22 21 20 19 18 17 16 data[23:16] 15 14 13 12 11 10 9 8 data[15:8] 76543210 data[7:0]
63 32145c?06/2013 at32uc3l0128/256 7.7.19 performance channel 0 read stall cycles name: prstall0 access type: read-only offset: 0x808 reset value: 0x00000000 ? stall: stall cycles counted since last reset clock cycles are counted us ing the clk_pdca_hsb clock 31 30 29 28 27 26 25 24 stall[31:24] 23 22 21 20 19 18 17 16 stall[23:16] 15 14 13 12 11 10 9 8 stall[15:8] 76543210 stall[7:0]
64 32145c?06/2013 at32uc3l0128/256 7.7.20 performance channel 0 read max latency name: prlat0 access type: read/write offset : 0x80c reset value: 0x00000000 ? lat: maximum transfer initiation cycles counted since last reset clock cycles are counted us ing the clk_pdca_hsb clock this counter is saturating. the register is reset only when pcontrol.ch0res is written to one. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 lat[15:8] 76543210 lat[7:0]
65 32145c?06/2013 at32uc3l0128/256 7.7.21 performance channel 0 write data cycles name: pwdata0 access type: read-only offset: 0x810 reset value: 0x00000000 ? data: data cycles counted since last reset clock cycles are counted us ing the clk_pdca_hsb clock 31 30 29 28 27 26 25 24 data[31:24] 23 22 21 20 19 18 17 16 data[23:16] 15 14 13 12 11 10 9 8 data[15:8] 76543210 data[7:0]
66 32145c?06/2013 at32uc3l0128/256 7.7.22 performance channel 0 write stall cycles name: pwstall0 access type: read-only offset: 0x814 reset value: 0x00000000 ? stall: stall cycles counted since last reset clock cycles are counted us ing the clk_pdca_hsb clock 31 30 29 28 27 26 25 24 stall[31:24] 23 22 21 20 19 18 17 16 stall[23:16] 15 14 13 12 11 10 9 8 stall[15:8] 76543210 stall[7:0]
67 32145c?06/2013 at32uc3l0128/256 7.7.23 performance channel 0 write max latency name: pwlat0 access type: read/write offset : 0x818 reset value: 0x00000000 ? lat: maximum transfer initiation cycles counted since last reset clock cycles are counted us ing the clk_pdca_hsb clock this counter is saturating. the register is reset only when pcontrol.ch0res is written to one. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 lat[15:8] 76543210 lat[7:0]
68 32145c?06/2013 at32uc3l0128/256 7.7.24 performance channel 1 read data cycles name: prdata1 access type: read-only offset: 0x81c reset value: 0x00000000 ? data: data cycles counted since last reset clock cycles are counted us ing the clk_pdca_hsb clock 31 30 29 28 27 26 25 24 data[31:24] 23 22 21 20 19 18 17 16 data[23:16] 15 14 13 12 11 10 9 8 data[15:8] 76543210 data[7:0]
69 32145c?06/2013 at32uc3l0128/256 7.7.25 performance channel 1 read stall cycles name: prstall1 access type: read-only offset: 0x820 reset value: 0x00000000 ? stall: stall cycles counted since last reset clock cycles are counted us ing the clk_pdca_hsb clock 31 30 29 28 27 26 25 24 stall[31:24] 23 22 21 20 19 18 17 16 stall[23:16] 15 14 13 12 11 10 9 8 stall[15:8] 76543210 stall[7:0]
70 32145c?06/2013 at32uc3l0128/256 7.7.26 performance channel 1 read max latency name: prlat1 access type: read/write offset : 0x824 reset value: 0x00000000 ? lat: maximum transfer initiation cycles counted since last reset clock cycles are counted us ing the clk_pdca_hsb clock this counter is saturating. the register is reset only when pcontrol.ch1res is written to one. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 lat[15:8] 76543210 lat[7:0]
71 32145c?06/2013 at32uc3l0128/256 7.7.27 performance channel 1 write data cycles name: pwdata1 access type: read-only offset: 0x828 reset value: 0x00000000 ? data: data cycles counted since last reset clock cycles are counted us ing the clk_pdca_hsb clock 31 30 29 28 27 26 25 24 data[31:24] 23 22 21 20 19 18 17 16 data[23:16] 15 14 13 12 11 10 9 8 data[15:8] 76543210 data[7:0]
72 32145c?06/2013 at32uc3l0128/256 7.7.28 performance channel 1 write stall cycles name: pwstall1 access type: read-only offset: 0x82c reset value: 0x00000000 ? stall: stall cycles counted since last reset clock cycles are counted us ing the clk_pdca_hsb clock 31 30 29 28 27 26 25 24 stall[31:24] 23 22 21 20 19 18 17 16 stall[23:16] 15 14 13 12 11 10 9 8 stall[15:8] 76543210 stall[7:0]
73 32145c?06/2013 at32uc3l0128/256 7.7.29 performance channel 1 write max latency name: pwlat1 access type: read/write offset : 0x830 reset value: 0x00000000 ? lat: maximum transfer initiation cycles counted since last reset clock cycles are counted us ing the clk_pdca_hsb clock this counter is saturating. the register is reset only when pcontrol.ch1res is written to one. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 lat[15:8] 76543210 lat[7:0]
74 32145c?06/2013 at32uc3l0128/256 7.7.30 pdca version register name: version access type: read-only offset: 0x834 reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
75 32145c?06/2013 at32uc3l0128/256 7.8 module configuration the specific configuration for each pdca instance is listed in the following tables.the module bus clocks listed here are connected to the syst em bus clocks. please refer to the power man- ager chapter for details. the pdca and the peripheral modules communicate through a set of handshake signals. the following table defines the valid settings for the peripheral identifier (pid) in the pdca periph- eral select register (psr). the direction is specified as observed from the memory, so rx means transfers from peripheral to memory, and tx means from memory to peripheral. table 7-6. pdca configuration feature pdca number of channels 12 number of performance monitors 1 table 7-7. pdca clocks clock name description clk_pdca_hsb clock for the pdca hsb interface clk_pdca_pb clock for the pdca pb interface table 7-8. register reset values register reset value psr ch 0 0 psr ch 1 1 psr ch 2 2 psr ch 3 3 psr ch 4 4 psr ch 5 5 psr ch 6 6 psr ch 7 7 psr ch 8 8 psr ch 9 9 psr ch 10 10 psr ch 11 11 version 123 table 7-9. peripheral identity values pid direction peripheral in stance peripheral register 0 rx usart0 rhr 1 rx usart1 rhr 2 rx usart2 rhr
76 32145c?06/2013 at32uc3l0128/256 3 rx usart3 rhr 4 rx spi rdr 5rxtwim0 rhr 6rxtwim1 rhr 7rxtwis0 rhr 8rxtwis1 rhr 9 rx adcifb lcdr 10 rx aw rhr 11 rx cat acount 12 tx usart0 thr 13 tx usart1 thr 14 tx usart2 thr 15 tx usart3 thr 16 tx spi tdr 17 tx twim0 thr 18 tx twim1 thr 19 tx twis0 thr 20 tx twis1 thr 21 tx aw thr 22 tx cat mblen table 7-9. peripheral identity values pid direction peripheral in stance peripheral register
77 32145c?06/2013 at32uc3l0128/256 8. flash controller (flashcdw) rev: 1.2.0.0 8.1 features ? controls on-chip flash memory ? supports 0 and 1 wait state bus access ? buffers reducing penalty of wait state in sequential code or loops ? allows interleaved burst reads for systems with one wait state, outputti ng one 32-bit word per clock cycle for sequential reads ? secure state for supporting flashvault technology ? 32-bit hsb interface for reads from flash and writes to page buffer ? 32-bit pb interface for issuing commands to and configuratio n of the controller ? flash memory is divided into 16 regions can be individually prot ected or unprotected ? additional protection of the boot loader pages ? supports reads and writes of general-purpose non volatile memory (nvm) bits ? supports reads and writes of additional nvm pages ? supports device protection through a security bit ? dedicated command for chip-erase, first erasing all on-chip volatile me mories before erasing flash and clearing security bit 8.2 overview the flash controller (flashcdw) interfaces the on-chip flash memory with the 32-bit internal hsb bus. the controller manages the reading, writing, erasing, locking, and unlocking sequences. 8.3 product dependencies in order to use this module, other parts of the system must be configured correctly, as described below. 8.3.1 power management if the cpu enters a sleep mode that disabl es clocks used by the flashcdw, the flashcdw will stop functioning an d resume operation a fter the system wakes up from sleep mode. 8.3.2 clocks the flashcdw has two bus clocks co nnected: one high speed bus clock (clk_flashcdw_hsb) and one peripheral bu s clock (clk_flashcdw_pb). these clocks are generated by the power manager. both clocks are enabled at reset, and can be disabled by writing to the power manager. the user has to ensure that clk_flashcdw_hsb is not turned off before reading the flash or writing the pagebuffer and that clk_flashcdw_pb is not turned off before accessing the flashcdw confi guration and control regi sters. failing to do so may deadlock the bus. 8.3.3 interrupts the flashcdw interrupt r equest lines are connect ed to the interrupt controller. using the flashcdw interrupts requires the interrupt co ntroller to be pr ogrammed first.
78 32145c?06/2013 at32uc3l0128/256 8.3.4 debug operation when an external debugger forces the cpu into debug mode, the flashcdw continues nor- mal operation. if the flashcdw is configured in a way that requires it to be periodically serviced by the cpu through interrupts or similar, improper operation or data loss may result during debugging. 8.4 functional description 8.4.1 bus interfaces the flashcdw has two bus interfaces, one high speed bus (hsb) interface for reads from the flash memory and writes to the page buffer, and one peripheral bus (pb) interface for issuing commands and reading status from the controller. 8.4.2 memory organization the flash memory is divided into a set of pages. a page is the basic unit addressed when pro- gramming the flash. a page consists of several words. the pages are grouped into 16 regions of equal size. each of these regions can be locked by a dedicated fuse bit, protecting it from acci- dental modification. ? p pages ( flash_p ) ? w bytes in each page and in the page buffer ( flash_w ) ? pw bytes in total ( flash_pw ) ? f general-purpose fuse bits ( flash_f ), used as region lock bits and for other device-specific purposes ? 1 security fuse bit ? 1 user page 8.4.3 user page the user page is an additional page, outside the regular flash array, that can be used to store various data, such as calibration data and serial numbers. this page is not erased by regular chip erase. the user page can only be written and erased by a special set of commands. read accesses to the user page are performed just as any other read accesses to the flash. the address map of the user page is given in figure 8-1 on page 80 . 8.4.4 read operations the on-chip flash memory is typically used for storing instructions to be executed by the cpu. the cpu will address instructions using the hsb bus, and the flashcdw will access the flash memory and return the addressed 32-bit word. in systems where the hsb clock period is slower than the access time of the flash memory, the flashcdw can operate in 0 wait state mode, and output one 32-bit word on the bus per clock cycle. if the clock frequency allows, the user should use 0 wait state mode, because this gives the highest performance as no stall cycles are encountered. the flashcdw can also operate in systems where the hsb bus clock period is faster than the access speed of the flash memory. wait state support and a read granularity of 64 bits ensure efficiency in such systems. performance for systems with high clock freque ncy is increased since the internal read word width of the flash memory is 64 bits. when a 32-bit word is to be addressed, the word itself and
79 32145c?06/2013 at32uc3l0128/256 also the other word in the same 64-bit location is read. the first word is output on the bus, and the other word is put into an internal buffer. if a read to a sequential address is to be performed in the next cycle, the buffered word is output on the bus, while the next 64-bit location is read from the flash memory. thus, latency in 1 wait state mode is hidden for sequential fetches. the programmer can select the wait states required by writing to the fws field in the flash con- trol register (fcr). it is the responsibility of the prog rammer to select a nu mber of wait states compatible with the clock frequency and timing characteristics of the flash memory. in 0ws mode, no wait states are encountered on any flash read operations. in 1 ws mode, one stall cycle is encountered on the first access in a single or burst transfer. in 1 ws mode, if the first access in a burst access is to an address that is not 64-bit aligned, an additional stall cycle is also encountered when reading the second word in the burst. all subsequent words in the burst are accessed without any stall cycles. the flash controller provides two sets of buffers that can be enabled in order to speed up instruction fetching. these buffers can be enabled by writing a one to the fcr.seqbuf and fcr.brbuf bits. the seqbuf bit enables buffering hardware optimizing sequential instruction fetches. the brbuf bit enables buffering hardwar e optimizing tight inner loops. these buffers are never used when the flash is in 0 wait st ate mode. usually, both these buffers should be enabled when operating in 1 wait state mode. some users requiring absolute cycle determinism may want to keep the buffers disabled. the flash controller address space is displayed in figure 8-1 . the memory space between address pw and the user page is reserved, and reading addresses in this space returns an undefined result. the user page is permanently mapped to an offset of 0x00800000 from the start address of the flash memory. table 8-1. user page addresses memory type start address, byte sized size main array 0 pw bytes user 0x00800000 w bytes
80 32145c?06/2013 at32uc3l0128/256 figure 8-1. memory map for the flash memories 8.4.5 high speed read mode the flash provides a high speed read mode, offering slightly higher flash read speed at the cost of higher power consumption. two dedicated commands, high speed read mode enable (hsen) and high speed read mode disable (hsdis) control the speed mode. the high speed mode (hsmode) bit in the flash status register (fsr) shows which mode the flash is in. after reset, the high speed mode is disabled, and must be manually enabled if the user wants to. refer to the electrical characteristics chapter at the end of this datasheet for details on the max- imum clock frequencies in normal and high speed read mode. 0 pw reserved flash data array reserved user page flash with user page 0x0080 0000 all addresses are byte addresses flash base address offset from base address
81 32145c?06/2013 at32uc3l0128/256 figure 8-2. high speed mode 8.4.6 quick page read a dedicated command, quick page read (qpr), is provided to read all words in an addressed page. all bits in all words in this page are and?e d together, returning a 1-bi t result. this result is placed in the quick page read result (qprr) bi t in flash status register (fsr). the qpr command is useful to check that a page is in an erased state. the qpr instruction is much faster than performing the erased-page check using a regular software subroutine. 8.4.7 quick user page read a dedicated command, quick user page read (qp rup), is provided to read all words in the user page. all bits in all word s in this page are and?ed together, returning a 1-bit result. this result is placed in the quick page read result (qprr) bit in flash status register (fsr). the qprup command is useful to check that a page is in an erased state. the qprup instruction is much faster than performing the erased-page check using a regular software subroutine. 8.4.8 page buffer operations the flash memory has a write and erase granularity of one page; data is written and erased in chunks of one page. when programming a page, the user must first write the new data into the page buffer. the contents of the entire page buffer is copied into the desired page in flash memory when the user issues the write page command, refer to section 8.5.1 on page 83 . in order to program data into flash page y, write the desired data to locations y0 to y31 in the regular flash memory map. writ ing to an address a in the fl ash memory map will not update the flash memory, but will inst ead update location a% 32 in the page buffer. th e pagen field in the flash command (fcmd) register will at the same time be up dated with the value a/32. frequency frequency limit for 0 wait state operation n o r m a l h i g h speed mode 1 wait state 0 wait state
82 32145c?06/2013 at32uc3l0128/256 figure 8-3. mapping from page buffer to flash internally, the flash memory stores data in 64-bit doublewords. therefore, the native data size of the page buffer is also a 64-bit doubleword. all locations shown in figure 8-3 are therefore dou- bleword locations. since the hsb bus only has a 32-bit data width, two 32-bit hsb transfers must be performed to write a 64 -bit doubleword into the page buffer. the flashcdw has logic to combine two 32-bit hsb transfers into a 64-bit data before writing this 64-bit data into the page buffer. this logic requires the word with the low address to be written to the hsb bus before the word with the high address. to exemplify, to write a 64-bit value to doubleword x0 residing in page x, first write a 32-bit word to the byte address pointing to address x0, thereafter write a word to the byte address pointing to address (x0+4). the page buffer is word-addressable and should only be written with aligned word transfers, never with byte or halfword transfers. the page buffer can not be read. the page buffer is also used for writes to the user page. page buffer write operations are performed with 4 wait states. any accesses attempted to the flashcdw on the hsb bus du ring these cycles will be automatically stalled. writing to the page buffer can only change page buffer bits from one to zero, i.e. writing 0xaaaaaaaa to a page buffer loca tion that has the value 0x 00000000 will not change the page buffer value. the only way to change a bit from zero to one is to erase the entire page buffer with the clear page buffer command. z3 z2 z1 z0 z7 z6 z5 z4 z11 z10 z9 z8 z15 z14 z13 z12 z19 z18 z17 z16 z23 z22 z21 z20 z27 z26 z25 z24 z31 z30 z29 z28 y3 y2 y1 y0 y7 y6 y5 y4 y11 y10 y9 y8 y15 y14 y13 y12 y19 y18 y17 y16 y23 y22 y21 y20 y27 y26 y25 y24 y31 y30 y29 y28 x3 x2 x1 x0 x7 x6 x5 x4 x11 x10 x9 x8 x15 x14 x13 x12 x19 x18 x17 x16 x23 x22 x21 x20 x27 x26 x25 x24 x31 x30 x29 x28 3 2 1 0 7 6 5 4 11 10 9 8 15 14 13 12 19 18 17 16 23 22 21 20 27 26 25 24 31 30 29 28 page x page y page z page buffer 64-bit data flash all locations are doubleword locations
83 32145c?06/2013 at32uc3l0128/256 the page buffer is not automatically reset after a page write. the programmer should do this manually by issuing the clear page buffer flas h command. this can be done after a page write, or before the page buffer is loaded with data to be stored to the flash page. 8.5 flash commands the flashcdw offers a command set to manage programming of the flash memory, locking and unlocking of regions, and full flash erasing. see section 8.8.2 for a complete list of commands. to run a command, the cmd field in the flash command register (fcmd) has to be written with the command number. as soon as the fcmd register is written, the frdy bit in the flash status register (fsr) is automatically cleared. once the current command is complete, the fsr.frdy bit is automatically set. if an interrupt has been enabled by writing a one to fcr.frdy, the interrupt request line of the flash controller is activated. all flash commands except for quick page read (qpr) and quick us er page read (qprup ) will generate an inter- rupt request upon completion if fcr.frdy is one. any hsb bus transfers attempting to read flas h memory when the flashcdw is busy execut- ing a flash command will be stalled, and allo wed to continue when the flash command is complete. after a command has been written to fcmd, the programming algorithm should wait until the command has been executed before attempting to read instructions or data from the flash or writing to the page buffer, as the flash will be busy. the waiting can be pe rformed either by poll- ing the flash status register (fsr) or by waiting for the flash ready interrupt. the command written to fcmd is initiated on the first cloc k cycle where the hsb bus interface in flashcdw is idle. the user must make sure that the access pattern to the flashcdw hsb interface contains an idle cycle so that the command is allowed to start. make sure that no bus masters such as dma controllers are performing endless burst transfers from the flash. also, make sure that the cpu does not perform endless burst transfe rs from flash. this is done by letting the cpu enter sleep mode after writing to fcmd, or by polling fsr for command completion. this polling will result in an access pattern with idle hsb cycles. all the commands are protected by the same keyword, which has to be written in the eight high- est bits of the fcmd register. writing fcmd wi th data that does not contain the correct key and/or with an invalid command has no effect on the flash memory; however, the proge bit is set in the flash status register (fsr). this bit is automatically cleared by a read access to the fsr register. writing a command to fcmd while another command is being executed has no effect on the flash memory; however, the proge bit is set in the flash status register (fsr). this bit is automatically cleared by a read access to the fsr register. if the current command writes or erases a page in a locked region, or a page protected by the bootprot fuses, the command has no effect on the flash memory; however, the locke bit is set in the fsr register. this bit is automatically cleared by a read access to the fsr register. 8.5.1 write/erase page operation flash technology requires that an erase must be done before programming. the entire flash can be erased by an erase all command. alternativ ely, pages can be individually erased by the erase page command. the user page can be written and erased usin g the mechanisms descr ibed in this chapter.
84 32145c?06/2013 at32uc3l0128/256 after programming, the page can be locked to pr event miscellaneous write or erase sequences. locking is performed on a per-region basis, so locking a region locks all pages inside the region. additional protection is provided for the lowermost address space of the flash. this address space is allocated for the boot loader, and is protected both by the lock bit(s) corresponding to this address space, and the bootprot[2:0] fuses. data to be written is stored in an internal buffer called the page buffer. the page buffer contains w words. the page buffer wraps around within the internal memory area address space and appears to be repeated by the number of pages in it. writing of 8-bit and 16-bit data to the page buffer is not allowed and may lead to unpredictable data corruption. data must be written to the page buffer before the programming command is written to the flash command register (fcmd). the sequence is as follows: ? reset the page buffer with the clear page buffer command. ? fill the page buffer with the desired content s as described in section 8.4.8 on page 81 . ? programming starts as soon as the programming key and the programming command are written to the flash command register. the pagen field in the flash command register (fcmd) must contain the address of the page to write. pagen is automatically updated when writing to the page buffer, but can also be written to directly. the frdy bit in the flash status register (fsr) is automatically cleared when the page write operation starts. ? when programming is completed, the frdy bit in the flash status register (fsr) is set. if an interrupt was enabled by writing fcr.frdy to one, an interrupt request is generated. two errors can be detected in the fsr register after a programming sequence: ? programming error: a bad keyword and/or an invalid command have been written in the fcmd register. ? lock error: can have two different causes: ? the page to be programmed belongs to a locked region. a command must be executed to unlock the corresponding region before programming can start. ? a bus master without secure status attempted to program a page requiring secure privileges. 8.5.2 erase all operation the entire memory is erased if the erase all command (ea) is written to the flash command register (fcmd). erase all erases all bits in the flash array. the user page is not erased. all flash memory locations, the general-purpose fuse bi ts, and the security bit are erased (reset to 0xff) after an erase all. the ea command also ensures that all volatile me mories, such as register file and rams, are erased before the security bit is erased. erase all operation is allowed only if no regions are locked, and the bootprot fuses are con- figured with a bootprot region size of 0. thus, if at least one region is locked, the bit locke in fsr is set and the command is cancelled. if the locke bit in fcr is one, an interrupt request is set generated. when the command is complete, the frdy bit in th e flash status register (fsr) is set. if an interrupt has been enabled by writing fcr.frdy to one, an interrupt request is generated. two errors can be detected in the fsr register after issuing the command:
85 32145c?06/2013 at32uc3l0128/256 ? programming error: a bad keyword and/or an invalid command have been written in the fcmd register. ? lock error: at least one lock region is protected, or bootprot is different from 0. the erase command has been aborted and no page has been erased. a ?unlock region containing given page? (up) command must be executed to unlock any locked regions. 8.5.3 region lock bits the flash memory has p pages, and these pages are grouped into 16 lock regions, each region containing p /16 pages. each region has a dedicated lock bit preventing writing and erasing pages in the region. after production, the device may have some regions locked. these locked regions are reserved for a boot or default applic ation. locked regions can be unlocked to be erased and then programmed with another application or other data. to lock or unlock a region, the commands lock region containing page (lp) and unlock region containing page (up) are provided. writing one of these commands, together with the number of the page whose region should be locked/unlocked, performs the desired operation. one error can be detected in the fsr register after issuing the command: ? programming error: a bad keyword and/or an invalid command have been written in the fcmd register. the lock bits are implemented using the lowest 16 general-purpose fuse bits. this means that lock bits can also be set/cleared using the commands for writing/erasing general-purpose fuse bits, see section 8.6 . the general-purpose bit being in an erased (1) state means that the region is unlocked. the lowermost pages in the flash can additiona lly be protected by the bootprot fuses, see section 8.6 . 8.6 general-purpose fuse bits the flash memory has a number of general-purpose fuse bits that the application programmer can use freely. the fuse bits can be writ ten and erased using dedicated commands, and read
86 32145c?06/2013 at32uc3l0128/256 through a dedicated peripheral bus address. some of the general-purpose fuse bits are reserved for special purposes, and shou ld not be used for other functions: the bootprot fuses protects the following address space for the boot loader: table 8-2. general-purpose fuses with special functions general- purpose fuse number name usage 15:0 lock region lock bits. 16 epfl external privileged fetch lock. used to prevent the cpu from fetching instructions from external memories when in privileged mode. this bit can only be changed when the security bit is cleared. the address range corresponding to external memories is device-specific, and not known to the flash controller. this fuse bit is simply routed out of the cpu or bus system, the flash controller does not treat this fuse in any special way, except that it can not be altered when the security bit is set. if the security bit is set, only an external jtag or awire chip erase can clear epfl. no internal commands can alter epfl if the security bit is set. when the fuse is erased (i.e. "1"), the cpu can execute instructions fetched from external memories. when the fuse is programmed (i.e. "0"), instructions can not be executed from external memories. this fuse has no effect in de vices with no external memory interface (ebi). 19:17 bootprot used to select one of eight different bootloader sizes. pages included in the bootloader area can not be erased or programmed except by a jtag or awire chip erase. bootprot can only be changed when the security bit is cleared. if the security bit is set, only an external jtag or awire chip erase can clear bootprot, and thereby allow the pages protected by bootprot to be programmed. no internal commands can alter bootprot or the pages protected by bootprot if the security bit is set. 21:20 secure used to configure secure state and secure state debug capabilities. de coded into sse and ssde signals as shown in table 8-5 . refer to the avr32 architecture manual and the avr32uc technical reference manual for more details on sse and ssde. 22 uprot if programmed (i.e. ?0?), the jtag user protection feature is enabled. if this fuse is programmed some hsb addresses will be accessible by jtag access even if the flash security fuse is programmed. refer to the jtag documentation for more information on this functionality. this bit can only be changed when the security bit is cleared.
87 32145c?06/2013 at32uc3l0128/256 the secure fuses have the following functionality: to erase or write a general-purpose fuse bit, the commands write general-purpose fuse bit (wgpb) and erase general-purpose fuse bit (e gpb) are provided. writing one of these com- mands, together with the number of the fuse to write/erase, performs the desired operation. an entire general-purpose fuse byte can be wr itten at a time by using the program gp fuse byte (pgpfb) instruction. a pgpfb to gp fuse byte 2 is not allowed if the flash is locked by the security bit. the pfb command is issued with a parameter in the pagen field: ? pagen[2:0] - byte to write ? pagen[10:3] - fuse value to write all general-purpose fuses can be erased by the erase all general-purpose fuses (eagp) com- mand. an eagp command is not allowed if the flash is locked by the security bit. two errors can be detected in the fsr register after issuing these commands: ? programming error: a bad keyword and/or an invalid command have been written in the fcmd register. ? lock error: ? a write or erase of the bootprot or epfl or uprot fuse bits was attempted while the flash is locked by the security bit. ? a write or erase of the secure fuse bi ts was attempted when secure mode was enabled. the lock bits are implemented using the lowest 16 general-purpose fuse bits. this means that the 16 lowest general-purpose fuse bits can also be written/erased using the commands for locking/unlocking regions, see section 8.5.3 . table 8-3. boot loader area specified by bootprot bootprot pages protected by bootprot size of protected memory 7none 0 60-1 1kbyte 50-3 2kbyte 40-7 4kbyte 3 0-15 8kbyte 20-31 16kbyte 10-63 32kbyte 0 0-127 64kbyte table 8-5. secure state configuration secure functionality sse ssde 00 secure state disabled 0 0 01 secure enabled, secure state debug enabled 1 1 10 secure enabled, secure state debug disabled 1 0 11 secure state disabled 0 0
88 32145c?06/2013 at32uc3l0128/256 8.7 security bit the security bit allows the entire device to be locked from external jtag, awire, or other debug access for code security. the security bit can be written by a dedicated command, set security bit (ssb). once set, the only way to clear the security bit is through the jtag or awire chip erase command. once the security bit is set, the following flash controller comm ands will be unavailable and return a lock error if attempted: ? write general-purpose fuse bit (w gpb) to bootprot or epfl fuses ? erase general-purpose fuse bit (e gpb) to bootprot or epfl fuses ? program general-purpose fuse byte (pgpfb) of fuse byte 2 ? erase all general-purpose fuses (eagpf) one error can be detected in the fsr register after issuing the command: ? programming error: a bad keyword and/or an invalid command have been written in the fcmd register.
89 32145c?06/2013 at32uc3l0128/256 8.8 user interface note: 1. the value of the lock bits depend on their programmed state. all other bits in fsr are 0. 2. all bits in fgprhi/lo are dependent on the programmed state of the fuses they map to. any bits in these registers not mapped to a fuse read as 0. 3. the reset values for these registers are device specific. plea se refer to the module configuration section at the end of this chapter. table 8-6. flashcdw register memory map offset register register name access reset 0x00 flash control register fcr read/write 0x00000000 0x04 flash command register fcmd read/write 0x00000000 0x08 flash status register fsr read-only - (1) 0x0c flash parameter register fpr read-only - (3) 0x10 flash version register fvr read-only - (3) 0x14 flash general purpose fuse register hi fgpfrhi read-only - (2) 0x18 flash general purpose fuse register lo fgpfrlo read-only - (2)
90 32145c?06/2013 at32uc3l0128/256 8.8.1 flash control register name: fcr access type: read/write offset :0x00 reset value: 0x00000000 ? brbuf: branch target instruction buffer enable 0: the branch target instru ction buffer is disabled. 1: the branch target instruction buffer is enabled. ? seqbuf: sequential instru ction fetch buffer enable 0: the sequential instruction fetch buffer is disabled. 1: the sequential instruction fetch buffer is enabled. ? fws: flash wait state 0: the flash is read with 0 wait states. 1: the flash is read with 1 wait state. ? proge: programming error interrupt enable 0: programming error does not generate an interrupt request. 1: programming error generates an interrupt request. ? locke: lock error interrupt enable 0: lock error does not generate an interrupt request. 1: lock error generates an interrupt request. ? frdy: flash ready interrupt enable 0: flash ready does not generate an interrupt request. 1: flash ready generates an interrupt request. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -----brbufseqbuf- 76543210 - fws - - proge locke - frdy
91 32145c?06/2013 at32uc3l0128/256 8.8.2 flash command register name: fcmd access type: read/write offset :0x04 reset value: 0x00000000 the fcmd can not be written if the flash is in the process of performing a flash command. doing so will cause the fcr write to be ignored , and the proge bit in fsr to be set. ? key: write protection key this field should be written with the value 0xa5 to enable the command defined by the bits of t he register. if the field is wri tten with a different value, the write is not performed and no action is started. this field always reads as 0. ?pagen: page number the pagen field is used to address a page or fuse bit for certai n operations. in order to simplify programming, the pagen field is automatically updated every time the page buffer is written to. for every page buffer write, the pagen field is updated with the page number of the address being written to. hardware automatically masks writes to the pagen field so that only bits representing valid page numbers can be writt en, all other bits in pagen are always 0. as an example, in a flash with 1024 pages (page 0 - page 1023), bits 15:10 will always be 0. 31 30 29 28 27 26 25 24 key 23 22 21 20 19 18 17 16 pagen [15:8] 15 14 13 12 11 10 9 8 pagen [7:0] 76543210 -- cmd table 8-7. semantic of pagen field in different commands command pagen description no operation not used write page the number of the page to write clear page buffer not used lock region containing given page page number whose region should be locked unlock region containing given page page number whose region should be unlocked erase all not used write general-purpose fuse bit gpfuse # erase general-purpose fuse bit gpfuse # set security bit not used
92 32145c?06/2013 at32uc3l0128/256 ? cmd: command this field defines the flash command. issuing any unused command will cause the programming error bit in fsr to be set, and the corresponding interrupt to be requested if the proge bit in fcr is one. program gp fuse byte writed ata[7:0], byteaddress[2:0] erase all gp fuses not used quick page read page number write user page not used erase user page not used quick page read user page not used high speed mode enable not used high speed mode disable not used table 8-8. set of commands command value mnemonic no operation 0 nop write page 1 wp erase page 2 ep clear page buffer 3 cpb lock region containing given page 4 lp unlock region containing given page 5 up erase all 6 ea write general-purpose fuse bit 7 wgpb erase general-purpose fuse bit 8 egpb set security bit 9 ssb program gp fuse byte 10 pgpfb erase all gpfuses 11 eagpf quick page read 12 qpr write user page 13 wup erase user page 14 eup quick page read user page 15 qprup high speed mode enable 16 hsen high speed mode disable 17 hsdis reserved 16-31 table 8-7. semantic of pagen field in different commands command pagen description
93 32145c?06/2013 at32uc3l0128/256 8.8.3 flash status register name: fsr access type: read-only offset :0x08 reset value: 0x00000000 ? lockx: lock region x lock status 0: the corresponding lock region is not locked. 1: the corresponding lock region is locked. ? hsmode: high-speed mode 0: high-speed mode disabled. 1: high-speed mode enabled. ? qprr: quick page read result 0: the result is zero, i.e. the page is not erased. 1: the result is one, i.e. the page is erased. ? security: security bit status 0: the security bit is inactive. 1: the security bit is active. ? proge: programming error status automatically cleared when fsr is read. 0: no invalid commands and no bad keywords were written in the flash command register fcmd. 1: an invalid command and/or a bad keyword was/we re written in the flash command register fcmd. ? locke: lock error status automatically cleared when fsr is read. 0: no programming of at least one locked lock region has happened since the last read of fsr. 1: programming of at least one locked lock region has happened since the last read of fsr. ? frdy: flash ready status 0: the flash controller is busy and the application must wait before running a new command. 1: the flash controller is ready to run a new command. 31 30 29 28 27 26 25 24 lock15 lock14 lock13 lock12 lock11 lock10 lock9 lock8 23 22 21 20 19 18 17 16 lock7 lock6 lock5 lock4 lock3 lock2 lock1 lock0 15 14 13 12 11 10 9 8 -------- 76543210 - hsmode qprr security proge locke - frdy
94 32145c?06/2013 at32uc3l0128/256 8.8.4 flash parameter register name: fpr access type: read-only offset :0x0c reset value: - ? psz: page size the size of each flash page. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 ----- psz 76543210 ---- fsz table 8-9. flash page size psz page size 032 byte 164 byte 2128 byte 3256 byte 4512 byte 5 1024 byte 6 2048 byte 7 4096 byte
95 32145c?06/2013 at32uc3l0128/256 ? fsz: flash size the size of the flash. not all device families will provide all flash sizes indicated in the table. table 8-10. flash size fszflash sizefszflash size 0 4 kbyte 8 192 kbyte 1 8 kbyte 9 256 kbyte 2 16 kbyte 10 384 kbyte 3 32 kbyte 11 512 kbyte 4 48 kbyte 12 768 kbyte 5 64 kbyte 13 1024 kbyte 6 96 kbyte 14 2048 kbyte 7 128 kbyte 15 reserved
96 32145c?06/2013 at32uc3l0128/256 8.8.5 flash version register name: fvr access type: read-only offset :0x10 reset value: 0x00000000 ? variant: variant number reserved. no functi onality associated. ? version: version number version number of the module . no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 ---- version[11:8] 76543210 version[7:0]
97 32145c?06/2013 at32uc3l0128/256 8.8.6 flash general purpose fuse register high name: fgpfrhi access type: read-only offset :0x14 reset value: - this register is only used in syst ems with more t han 32 gp fuses. ? gpfxx: general purpose fuse xx 0: the fuse has a written/programmed state. 1: the fuse has an erased state. 31 30 29 28 27 26 25 24 gpf63 gpf62 gpf61 gpf60 gpf59 gpf58 gpf57 gpf56 23 22 21 20 19 18 17 16 gpf55 gpf54 gpf53 gpf52 gpf51 gpf50 gpf49 gpf48 15 14 13 12 11 10 9 8 gpf47 gpf46 gpf45 gpf44 gpf43 gpf42 gpf41 gpf40 76543210 gpf39 gpf38 gpf37 gpf36 gpf35 gpf34 gpf33 gpf32
98 32145c?06/2013 at32uc3l0128/256 8.8.7 flash general purpose fuse register low name: fgpfrlo access type: read-only offset :0x18 reset value: - ? gpfxx: general purpose fuse xx 0: the fuse has a written/programmed state. 1: the fuse has an erased state. 31 30 29 28 27 26 25 24 gpf31 gpf30 gpf29 gpf28 gpf27 gpf26 gpf25 gpf24 23 22 21 20 19 18 17 16 gpf23 gpf22 gpf21 gpf20 gpf19 gpf18 gpf17 gpf16 15 14 13 12 11 10 9 8 gpf15 gpf14 gpf13 gpf12 gpf11 gpf10 gpf09 gpf08 76543210 gpf07 gpf06 gpf05 gpf04 gpf03 gpf02 gpf01 gpf00
99 32145c?06/2013 at32uc3l0128/256 8.9 fuse settings the flash contains 32 general purpose fuses. these 32 fuses can be found in the flash general purpose fuse register low (fgpfrlo). the flash general purpose fuse register high (fgpfrhi) is not used. in addition to the general purpose fuses, parts of the flash user page can have a defined meaning outs ide of the flash controller and will also be described in this section. note that when writing to the user page the values do not get loaded by the other modules on the device until a chip reset occurs. the general purpose fuses are erased by a jtag or awire chip erase.
100 32145c?06/2013 at32uc3l0128/256 8.9.1 flash general purpose fuse register low (fgpfrlo) ? boden: brown out detector enable ? bodhyst: brown out detector hysteresis 0: the brown out detector hysteresis is disabled 1: the brown out detector hysteresis is enabled ? bodlevel: brown out detector trigger level this controls the voltage trigger leve l for the brown out detector. refer to ?electrical characteristics? on page 791 . ? uprot, secure, bootprot, epfl, lock these are flash controller fuses and are described in the flashcdw section. 8.9.1.1 default fuse value the devices are shipped with the fgpfrlo register value:0xe07fffff: ? boden fuses set to 11. bod is disabled. ? bodhyst fuse set to 1. the bod hysteresis is enabled. ? bodlevel fuses set to 000000. this is the mini mum voltage trigger level for bod. this level is lower than the por level, so when bod is e nabled, it will never trigger with this default value. ? uprot fuse set to 1. ? secure fuse set to 11. ? bootprot fuses set to 111. the bootloader protection is disabled. ? epfl fuse set to 1. external privileged fetch is not locked. ? lock fuses set to 1111111111111111. no region locked. after the jtag or awire chip erase command, the fgpfr register value is 0xffffffff. 31 30 29 28 27 26 25 24 boden bodhyst bodlevel[5:1] 23 22 21 20 19 18 17 16 bodlevel[0] uprot secure bootprot epfl 15 14 13 12 11 10 9 8 lock[15:8] 7 6543210 lock[7:0] boden description 00 bod disabled 01 bod enabled, bod reset enabled 10 bod enabled, bod reset disabled 11 bod disabled
101 32145c?06/2013 at32uc3l0128/256 8.9.2 first word of the user page (address 0x80800000) ? wdtauto: watchdog timer auto enable at startup 0: the wdt is automatical ly enabled at startup. 1: the wdt is not automatic ally enabled at startup. please refer to the wdt chapter for detail about tim eout settings when the wdt is automatically enabled. 8.9.2.1 default user page first word value the devices are shipped with the user page erased (all bits 1): ? wdtauto set to 1, wdt disabled. 31 30 29 28 27 26 25 24 - ------- 23 22 21 20 19 18 17 16 - ------- 15 14 13 12 11 10 9 8 - ------- 7 6543210 - ------wdtauto
102 32145c?06/2013 at32uc3l0128/256 8.9.3 second word of the user page (address 0x80800004) ? ssadrr: secure state end address for the ram ? ssadrf: secure state end address for the flash 8.9.3.1 default user page second word value the devices are shipped with the user page erased (all bits 1). 8.10 serial number each device has a unique 120 bits serial number readable from address 0x8080020c to 0x8080021a. 8.11 module configuration the specific configuration for each flashcdw in stance is listed in th e following tables. the module bus clocks listed here are c onnected to the system bus cloc ks. please refer to the power manager chapter for details. 31 30 29 28 27 26 25 24 ssadrr[15:8] 23 22 21 20 19 18 17 16 ssadrr[7:0] 15 14 13 12 11 10 9 8 ssadrf[15:8] 7 6543210 ssadrf[7:0] table 8-11. module configuration feature at32uc3l0256 at32uc3l0128 flash size 256kbytes 128kbytes number of pages 512 256 page size 512bytes 512bytes table 8-12. module clock name module name clock name description flashcdw clk_flashcdw_hsb clock for the flashcdw hsb interface clk_flashcdw_pb clock for the flashcdw pb interface
103 32145c?06/2013 at32uc3l0128/256 table 8-13. register reset values register at32uc3l0256 at32uc3l0128 fvr 0x00000120 0x00000120 fpr 0x00000409 0x00000407
104 32145c?06/2013 at32uc3l0128/256 9. secure access unit (sau) rev: 1.1.1.3 9.1 features ? remaps registers in memory regions protected by the mpu to regions not protected by the mpu ? programmable physical address for each channel ? two modes of operation: locked and open ? in locked mode, access to a channel must be preceded by an unlock action ? an unlocked channel remains op en only for a specific amount of time, if no access is performed during this time, the channel is relocked ? only one channel can be open at a time, op ening a channel while another one is open locks the first one ? access to a locked channel is denied, a bus error and optionally an interrupt is returned ? if a channel is relocked due to an unlock timeout, an interrupt can optionally be generated ? in open mode, all channels are permanently unlocked 9.2 overview in many systems, erroneous access to peripherals can lead to catastrophic failure. an example of such a peripheral is the pulse width modula tor (pwm) used to control electric motors. the pwm outputs a pulse train that controls the moto r. if the control registers of the pwm module are inadvertently updated with wrong values, the motor can start operating out of control, possi- bly causing damage to the application and the surrounding environment. however, sometimes the pwm control registers must be updated with new values, for example when modifying the pulse train to accelerate the motor. a mechanism must be used to protect the pwm control reg- isters from inadvertent access caused by for example: ? errors in the software code ? transient errors in the cpu caused by for example electrical noise altering the execution path of the program to improve the security in a computer system, the avr32uc implements a memory protection unit (mpu). the mpu can be set up to limit the accesses that can be performed to specific memory addresses. the mpu divides the memory space into regions, and assigns a set of access restrictions on each region. access restrictions can for example be read/write if the cpu is in supervisor mode, and read-on ly if the cpu is in application mode. the regions can be of dif- ferent size, but each region is usually quite larg e, e.g. protecting 1 kilobyte of address space or more. furthermore, access to each region is often controlled by the execution state of the cpu, i.e. supervisor or application mode. such a simple control mechanism is often too inflexible (too coarse-grained chunks) and with too much overhead (often requir ing system calls to access pro- tected memory locations) for simple or real-time systems such as embedded microcontrollers. usually, the secure access unit (sau) is used together with the mpu to provide the required security and integrity. the mpu is set up to protect regions of memory, while the sau is set up to provide a secure channel into specific me mory locations that are protected by the mpu. these specific locations can be thought of as fine-grained overrides of the general coarse- grained protection provided by the mpu.
105 32145c?06/2013 at32uc3l0128/256 9.3 block diagram figure 9-1 presents the sau integrated in an example system with a cpu, some memories, some peripherals, and a bus system. the sau is connected to both the peripheral bus (pb) and the high speed bus (hsb). configuration of the sau is done via the pb, while memory accesses are done via the hsb. the sau receives an access on its hsb slave interface, remaps it, checks that the channel is unlocked, and if so, initiates a transfer on its hsb master interface to the remapped address. the thin arrows in figure 9-1 exemplifies control flow when using the sau. the cpu wants to read the rx buffer in the usart. the mpu has been configured to protect all registers in the usart from user mode access, while the sau has been configured to remap the rx buffer into a memory space that is not protected by t he mpu. this unprotected memory space is mapped into the sau hsb slave space. when the cpu reads the appropriate address in the sau, the sau will perform an access to the desired rx buffer register in the usart, and therea fter return the read results to the cpu. th e return data flow will follow the op posite direction of the control flow arrows in figure 9-1 . figure 9-1. sau block diagram sau channel bus master mpu cpu bus slave usart pwm bus slave bus master bus slave flash bus slave ram bus bridge sau configuration interrupt request high speed bus sau peripheral bus
106 32145c?06/2013 at32uc3l0128/256 9.4 product dependencies in order to use this module, other parts of the system must be configured correctly, as described below. 9.4.1 power management if the cpu enters a sleep mode that disables clocks used by the sau, the sau will stop func- tioning and resume operation after the system wakes up from sleep mode. 9.4.2 clocks the sau has two bus clocks connected: o ne high speed bus clock (clk_sau_hsb) and one peripheral bus clock (clk_sau_pb). these clocks are generated by the power manager. both clocks are enabled at reset, and can be disabled by writing to the power manager. the user has to ensure that clk_sau_hsb is not turned off before accessing the sau. likewise, the user must ensure that no bus acce ss is pending in the sau before disabling clk_sau_hsb. failing to do so may deadlock the high speed bus. 9.4.3 interrupt the sau interrupt request line is connected to th e interrupt controller. using the sau interrupt requires the interrupt controller to be programmed first. 9.4.4 debug operation when an external debugger forces the cpu into debug mode, the sau continues normal opera- tion. if the sau is configured in a way that r equires it to be periodically serviced by the cpu through interrupts or similar, improper operation or data loss may result during debugging. 9.5 functional description 9.5.1 enabling the sau the sau is enabled by wr iting a one to the enable (en) bit in the control regist er (cr). this will set the sau enabled (en) bit in the status register (sr). 9.5.2 configuring the sau channels the sau has a set of channels, mapped in the hsb memory space. these channels can be configured by a remap target register (rtr), located at the same memory address. when the sau is in normal mode, the sau channel is addressed, and when the sau is in setup mode, the rtr can be addressed. before the sau can be used, the channels must be configured and enabled. to configure a channel, the corresponding rtr must be programmed with the remap target address. to do this, make sure the sau is in setup mode by writing a one to the setup mode enable (sen) bit in cr. this makes sure that a write to the rtr address accesses the rtr, not the sau chan- nel. thereafter, the rtr is written with the address to remap to, typically the address of a specific pb register. when all channels have been configured, return to normal mode by writing a one to the setup mode disable (sdis) in cr. the channels can now be enabled by writing ones to the corresponding bits in the channel enable registers (cerh/l). the sau is only able to remap addresses above 0xfffc0000.
107 32145c?06/2013 at32uc3l0128/256 9.5.2.1 protecting sau configuration registers in order to prevent the sau configuration registers to be changed by malicious or runaway code, they should be protected by the mpu as soon as they have been configured. maximum security is provided in the system if pr ogram memory does not contain any code to unprotect the config- uration registers in the mpu. this guarantees t hat runaway code can not accidentally unprotect and thereafter change the sau configuration registers. 9.5.3 lock mechanism the sau can be configured to use two different access mechanisms: open and locked. in open mode, sau channels can be accessed fr eely after they have been configured and enabled. in order to prevent accidental accesses to remapped addresses, it is possible to config- ure the sau in locked mode. writing a one to the open mode bit in the config register (config.open) will enable open mode. writing a zero to config.open will enable locked mode. when using locked mode, the lock mechanism must be configured by writing a user defined key value to the unlock key (ukey) field in the configuration register (config). the number of clk_sau_hsb cycles the channel remains unlock ed must be written to the unlock number of clock cycles (ucyc) field in config. access control to the sau channels is enabled by means of the unlock register (ur), which resides in the same address space as the sa u channels. before a channel can be accessed, the unlock register must be written with th correct key and channel number (single write access). access to the channel is then permitted for the next config.ucyc clock cycles, or until a suc- cessful access to the unlocked channel has been made. only one channel can be unlocked at a time. if any other channel is unlocked at the time of writ- ing ur, this channel will be automatically locked before the cha nnel addressed by the ur write is unlocked. an attempted access to a lo cked channel will be aborted, an d the channel access unsuccessful bit (sr.cau) will be set. any pending errors bits in sr must be cleared before it is possible to access ur. the following sr bits are defined as error bits: exp, cau, urread, urkey, ures, mberror, rtradr. if any of these bits are set while writing to ur, the write is aborted and the unlock register error status (ures) bit in sr is set. 9.5.4 normal operation the following sequence must be used in order to access a sau channel in normal operation (cr.sen=0): 1. if not in open mode, write the unlock key to ur.key and the channel number to ur.channel. 2. perform the read or write operation to the sau channel. if not in open mode, this must be done within config.ucyc clock cycles of unlocking the channel. the sau will use its hsb master interface to remap the access to the target address pointed to by the corresponding rtr. 3. to confirm that the access was successful, wait for the idle transfer status bit (sr.idle) to indicate the operation is completed. then check sr for possible error con- ditions. the sau can be configured to generate interrupt requests or a bus error exception if the access failed.
108 32145c?06/2013 at32uc3l0128/256 9.5.4.1 operation example figure 9-2 shows a typical memory map, consisting of some memories, some simple peripher- als, and a sau with multiple channels and an unlock register (ur). imagine that the mpu has been set up to disallow all accesses from the cpu to the grey modules. thus the cpu has no way of accessing for example the transmit holding register in the uart, present on address x on the bus. note that the sau rtrs are not protected by the mpu, thus the rtrs can be accessed. if for example rtr0 is configured to point to addres s x, an access to rtr0 will be remapped by the sau to address x according to the algorithm presented above. by program- ming the sau rtrs, specific addresses in modules that have generally been protected by the mpu can be performed. figure 9-2. example memory map for a system with sau 9.5.5 interrupts the sau can generate an interrupt request to signal different events. all events that can gener- ate an interrupt requ est have dedicated bits in the status register (sr). an interrupt request will be generated if the corresponding bit in the interrupt mask register (imr) is set. bits in imr are set by writing a one to the corresponding bit in the interrupt enable register (ier), and cleared by writing a one to the corresponding bit in the interrupt disable register (idr). the interrupt request remains active until the corresponding bit in sr is cleared by writing a one to the corre- sponding bit in the interrupt clear register (icr). the following sr bits are used for si gnalling the result of sau accesses: ? rtr address error (rtradr) is set if an ille gal address is written to the rtrs. only addresses in the range 0xfffc0000-0xffffffff are allowed. ? master interface bus error (mberror) is set if any of the conditions listed in section 9.5.7 occurred. transmit holding baudrate control receive holding channel 1 rtr0 rtr1 address x address z uart sau config sau channel ur rtr62 ...
109 32145c?06/2013 at32uc3l0128/256 ? unlock register error status (ures) is set if an attempt was made to unlock a channel by writing to the unlock regist er while one or more error bits in sr were set (see section 9.5.6 ). the unlock operation was aborted. ? unlock register key error (urkey) is set if the unlock register was attempted written with an invalid key. ? unlock register read (urread) is set if the unlock register was attempted read. ? channel access unsuccessful (cau) is set if the channel access was unsuccessful. ? channel access successful (cas) is set if the channel access was successful. ? channel unlock expired (exp) is set if the channel lock expired, with no channel being accessed after the channel was unlocked. 9.5.6 error bits if error bits are set when atte mpting to unlock a channel, sr.ures will be set. the following sr bits are considered error bits: ? exp ?cau ? urread ? urkey ?ures ? mberror ?rtradr 9.5.7 bus error responses by writing a one to the bus er ror response enable bit (cr.be rren), serious access errors will be configured to return a bus error to the cpu. this will ca use the cpu to execute its bus error data fetch exception routine. the conditions that can generate a bus error response are: ? reading the unlock register ? trying to access a locked channel ? the sau hsb master receiving a bus error response from its addressed slave 9.5.8 disabling the sau to disable the sau, the user must first ensure that no sau bus operations are pending. this can be done by checking that the sr.idle bit is set. the sau may then be disabled by writing a one to the disable (dis) bit in cr.
110 32145c?06/2013 at32uc3l0128/256 9.6 user interface the following addresses are used by sau channel configuration r egisters. all offsets are relative to the sau?s pb base address. note: 1. the reset values are device specific. please refer to the module configuration section at the end of this chapter. the following addresses are used by sau channel registers. a ll offsets are relative to the sau?s hsb base address. the number of channels implemented is device specific, refer to the module configuration section at the end of this chapter. table 9-1. sau configuration re gister memory map offset register register name access reset 0x00 control register cr write-only 0x00000000 0x04 configuration register config write-only 0x00000000 0x08 channel enable register high cerh read/write 0x00000000 0x0c channel enable register low cerl read/write 0x00000000 0x10 status register sr read-only 0x00000400 0x14 interrupt enable register ier write-only 0x00000000 0x18 interrupt disable register idr write-only 0x00000000 0x1c interrupt mask register imr read-only 0x00000000 0x20 interrupt clear register icr write-only 0x00000000 0x24 parameter register parameter read-only - (1) 0x28 version register version read-only - (1) table 9-2. sau channel register memory map offset register register name access reset 0x00 remap target register 0 rtr0 read/write n/a 0x04 remap target register 1 rtr1 read/write n/a 0x08 remap target register 2 rtr2 read/write n/a ... ... ... ... ... 0x04*n remap target register n rtrn read/write n/a 0xfc unlock register ur write-only n/a
111 32145c?06/2013 at32uc3l0128/256 9.6.1 control register name: cr access type: write-only offset: 0x00 reset value: 0x00000000 ? berrdis: bus error response disable writing a zero to this bit has no effect. writing a one to this bit disables bus error response from the sau. ? berren: bus error response enable writing a zero to this bit has no effect. writing a one to this bit enables bus error response from the sau. ? sdis: setup mode disable writing a zero to this bit has no effect. writing a one to this bit exits setup mode. ? sen: setup mode enable writing a zero to this bit has no effect. writing a one to this bit enters setup mode. ? dis: sau disable writing a zero to this bit has no effect. writing a one to this bit disables the sau. ? en: sau enable writing a zero to this bit has no effect. writing a one to this bit enables the sau. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 - - berrdis berren sdis sen dis en
112 32145c?06/2013 at32uc3l0128/256 9.6.2 configuration register name: config access type: write-only offset: 0x04 reset value: 0x00000000 ? open: open mode enable writing a zero to this bit disables open mode. writing a one to this bit enables open mode. ? ucyc: unlock number of clock cycles once a channel has been unlocked, it remains unlocked for this amount of clk_sau_hsb clock cycles or until one access to a channel has been made. ? ukey: unlock key the value in this field must be wri tten to ur.key to unlock a channel. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------open 15 14 13 12 11 10 9 8 ucyc 76543210 ukey
113 32145c?06/2013 at32uc3l0128/256 9.6.3 channel enable register high name: cerh access type: read/write offset: 0x08 reset value: 0x00000000 ? cerh[n]: channel enable register high 0: channel (n+32) is not enabled. 1: channel (n+32) is enabled. 31 30 29 28 27 26 25 24 - cerh[30:24] 23 22 21 20 19 18 17 16 cerh[23:16] 15 14 13 12 11 10 9 8 cerh[15:8] 76543210 cerh[7:0]
114 32145c?06/2013 at32uc3l0128/256 9.6.4 channel enable register low name: cerl access type: read/write offset: 0x0c reset value: 0x00000000 ? cerl[n]: channel enable register low 0: channel n is not enabled. 1: channel n is enabled. 31 30 29 28 27 26 25 24 cerl[31:24] 23 22 21 20 19 18 17 16 cerl[23:16] 15 14 13 12 11 10 9 8 cerl[15:8] 76543210 cerl[7:0]
115 32145c?06/2013 at32uc3l0128/256 9.6.5 status register name: sr access type: read-only offset: 0x10 reset value: 0x00000400 ?idle this bit is cleared when a read or write operation to the sau channel is started. this bit is set when the operation is completed and no sau bus operations are pending. ? sen: sau setup mode enable this bit is cleared when the sau exits setup mode. this bit is set when th e sau enters setup mode. ? en: sau enabled this bit is cleared when the sau is disabled. this bit is set when the sau is enabled. ? rtradr: rtr address error this bit is cleared when the corresponding bit in icr is written to one. this bit is set if, in the configuration phase, an rtr was wr itten with an illegal address, i.e. the upper 16 bits in the addre ss were different from 0xfffc, 0xfffd, 0xfffe or 0xffff. ? mberror: master interface bus error this bit is cleared when the corresponding bit in icr is written to one. this bit is set if a channel access generated a transfer on the master interface that received a bus error response from the addressed slave. ? ures: unlock register error status this bit is cleared when the corresponding bit in icr is written to one. this bit is set if an attempt was made to unlock a channel by writing to the unlock register while one or more error bits were set in sr. the unlock operation was aborted. ? urkey: unlock register key error this bit is cleared when the corresponding bit in icr is written to one. this bit is set if the unlock register was attempted written with an invalid key. ? urread: unlock register read this bit is cleared when the corresponding bit in icr is written to one. this bit is set if the unlock register was read. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -----idlesenen 76543210 rtradr mberror ures urkey urread cau cas exp
116 32145c?06/2013 at32uc3l0128/256 ? cau: channel access unsuccessful this bit is cleared when the corresponding bit in icr is written to one. this bit is set if channel access was unsuccessful, i.e. an access was attempted to a locked or disabled channel. ? cas: channel access successful this bit is cleared when the corresponding bit in icr is written to one. this bit is set if channel access successful, i.e. one access was made after the channel was unlocked. ? exp: channel unlock expired this bit is cleared when the corresponding bit in icr is written to one. this bit is set if channel unlock has expired, i.e. no access being made after the channel was unlocked.
117 32145c?06/2013 at32uc3l0128/256 9.6.6 interrupt enable register name: ier access type: write-only offset: 0x14 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will set the corresponding bit in imr. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 rtradr mberror ures urkey urread cau cas exp
118 32145c?06/2013 at32uc3l0128/256 9.6.7 interrupt disable register name: idr access type: write-only offset: 0x18 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in imr. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 rtradr mberror ures urkey urread cau cas exp
119 32145c?06/2013 at32uc3l0128/256 9.6.8 interrupt mask register name: imr access type: read-only offset: 0x1c reset value: 0x00000000 0: the corresponding interrupt is disabled. 1: the corresponding interrupt is enabled. a bit in this register is cleared when the corresponding bit in idr is written to one. a bit in this register is set when the corresponding bit in ier is written to one. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 rtradr mberror ures urkey urread cau cas exp
120 32145c?06/2013 at32uc3l0128/256 9.6.9 interrupt clear register name: icr access type: write-only offset: 0x20 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in sr and any corresponding interrupt request. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 rtradr mberror ures urkey urread cau cas exp
121 32145c?06/2013 at32uc3l0128/256 9.6.10 parameter register name: parameter access type: read-only offset: 0x24 reset value: - ? channels: number of channels implemented. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 - - channels
122 32145c?06/2013 at32uc3l0128/256 9.6.11 version register name: version access type: write-only offset: 0x28 reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
123 32145c?06/2013 at32uc3l0128/256 9.6.12 remap target register n name: rtrn access type: read/write offset: n*4 reset value: 0x00000000 ? rtr: remap target address for channel n rtr[31:16] must have one of the following values, any other value will result in undefined behavior: 0xfffc 0xfffd 0xfffe 0xffff rtr[1:0] must be written to 00, any other value will result in undefined behavior. 31 30 29 28 27 26 25 24 rtr[31:24] 23 22 21 20 19 18 17 16 rtr[23:16] 15 14 13 12 11 10 9 8 rtr[15:8] 76543210 rtr[7:0]
124 32145c?06/2013 at32uc3l0128/256 9.6.13 unlock register name: ur access type : write-only offset: 0xfc reset value: 0x00000000 ?key: unlock key the correct key must be written in order to unlock a channel. the key value written must correspond to the key value defined in config.ukey. ? channel: channel number number of the channel to unlock. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 key 76543210 - - channel
125 32145c?06/2013 at32uc3l0128/256 9.7 module configuration the specific configuratio n for each sau instance is listed in the following tables.the module bus clocks listed here are connecte d to the system bus clocks. pleas e refer to the power manager chapter for details. table 9-3. sau configuration feature sau sau channels 16 table 9-4. sau clock name module name clock name description sau clk_sau_hsb clock for the sau hsb interface sau clk_sau_pb clock for the sau pb interface table 9-5. register reset values register reset value version 0x00000111 parameter 0x00000010
126 32145c?06/2013 at32uc3l0128/256 10. hsb bus matrix (hmatrixb) rev: 1.3.0.3 10.1 features ? user interface on peripheral bus ? configurable number of masters (up to 16) ? configurable number of slaves (up to 16) ? one decoder for each master ? programmable arbitration for each slave ? round-robin ? fixed priority ? programmable default master for each slave ? no default master ? last accessed default master ? fixed default master ? one cycle latency for the first access of a burst ? zero cycle latency for default master ? one special function register for each slave (not dedicated) 10.2 overview the bus matrix implements a multi-layer bus structure, that enables parallel access paths between multiple high speed bus (hsb) masters and slaves in a system, thus increasing the overall bandwidth. the bus matrix interconnects up to 16 hsb masters to up to 16 hsb slaves. the normal latency to connect a master to a slave is one cycle except for the default master of the accessed slave which is connected directly (zero cycle latency). the bus matrix provides 16 special function registers (sfr) that allow the bus matrix to support application specific features. 10.3 product dependencies in order to configure this module by accessing the user registers, other parts of the system must be configured correctly, as described below. 10.3.1 clocks the clock for the hmatrix bus interface (clk_hmatrix) is generated by the power manager. this clock is enabled at reset, and can be disabled in the power manager. 10.4 functional description 10.4.1 special bus granting mechanism the bus matrix provides some speculative bus granting techniques in order to anticipate access requests from some masters. this mechanism reduc es latency at first access of a burst or single transfer. this bus granting mechanism sets a different default master for every slave. at the end of the current access, if no other re quest is pending, the slave remains connected to its associated default master. a slave can be as sociated with three kinds of default masters: no default master, last access master, and fixed default master.
127 32145c?06/2013 at32uc3l0128/256 to change from one kind of default master to another, the bus matrix user interface provides the slave configuration registers, one for each slave, that set a default master for each slave. the slave configuration register contains two fields: defmstr_type and fixed_defmstr. the 2-bit defmstr_type field selects the default mast er type (no default, last access master, fixed default master), whereas the 4-bit fixed_defmstr field selects a fixed default master pro- vided that defmstr_type is set to fixed default master. please refer to the bus matrix user interface description. 10.4.1.1 no default master at the end of the current access, if no other request is pending, the slave is disconnected from all masters. no default ma ster suits low-power mode. 10.4.1.2 last access master at the end of the current access, if no other re quest is pending, the slave remains connected to the last master that performed an access request. 10.4.1.3 fixed default master at the end of the current access, if no other r equest is pending, the slave connects to its fixed default master. unlike last access master, the fixed master does not change unless the user modifies it by a software action (field fixed_defmstr of the related scfg). 10.4.2 arbitration the bus matrix provides an arbitration mechanism that reduces latency when conflict cases occur, i.e. when two or more masters try to access the same slave at the same time. one arbiter per hsb slave is provided, thus ar bitrating each slave differently. the bus matrix provides the user with the possibility of choosing between 2 arbitration types for each slave: 1. round-robin arbitration (default) 2. fixed priority arbitration this is selected by the arbt field in the slave configuration registers (scfg). each algorithm may be complemented by selecting a default master configuration for each slave. when a re-arbitration must be done, specific conditions apply. this is described in ?arbitration rules? . 10.4.2.1 arbitration rules each arbiter has the ability to arbitrate between two or more different master requests. in order to avoid burst breaking and also to provide the maximum throughput for slave interfaces, arbitra- tion may only take place during the following cycles: 1. idle cycles: when a slave is not connected to any master or is connected to a master which is not currently accessing it. 2. single cycles: when a slave is currently doing a single access. 3. end of burst cycles: when the current cycle is the last cycle of a burst transfer. for defined length burst, predicted end of burst matches the size of the transfer but is man- aged differently for undefined length burst. this is described below. 4. slot cycle limit: when the slot cycle counte r has reached the limit value indicating that the current master access is too long and must be broken. this is described below.
128 32145c?06/2013 at32uc3l0128/256 ? undefined length burst arbitration in order to avoid long slave handling during unde fined length bursts (incr), the bus matrix pro- vides specific logic in order to re-arbitrate before the end of the incr transfer. a predicted end of burst is used as a defined length burst transfer and can be selected among the following five possibilities: 1. infinite: no predicted end of burst is gen erated and therefore i ncr burst transfer will never be broken. 2. one beat bursts: predicted end of burst is generated at each single transfer inside the incp transfer. 3. four beat bursts: predicted end of burst is generated at the end of each four beat boundary inside incr transfer. 4. eight beat bursts: predicted end of burst is generated at the end of each eight beat boundary inside incr transfer. 5. sixteen beat bursts: predicted end of burst is generated at the end of each sixteen beat boundary inside incr transfer. this selection can be done through the ulbt field in the master configuration registers (mcfg). ? slot cycle limit arbitration the bus matrix contains specific logic to break long accesses, such as very long bursts on a very slow slave (e.g., an external low speed memory). at the beginning of the burst access, a counter is loaded with the value previously written in the slot_cycle field of the related slave configuration register (scfg) and decreased at each clock cycle. when the counter reaches zero, the arbiter has the ability to re-arbitrate at the end of the current byte, halfword, or word transfer. 10.4.2.2 round-robin arbitration this algorithm allows the bus matrix arbiters to dispatch the requests from different masters to the same slave in a round-robin manner. if two or more master requests arise at the same time, the master with the lowest number is first serviced, then the others are serviced in a round-robin manner. there are three round-robin algorithms implemented: 1. round-robin arbitration without default master 2. round-robin arbitration with last default master 3. round-robin arbitration with fixed default master ? round-robin arbitration without default master this is the main algorithm used by bus matrix arbiters. it allows the bus matrix to dispatch requests from different masters to the same slave in a pure round-robin manner. at the end of the current access, if no other request is pending, the slave is disconnected from all masters. this configuration incurs one latency cycle for the first access of a burst. arbitration without default master can be used for masters that perform significant bursts. ? round-robin arbitration with last default master this is a biased round-robin algorithm used by bus matrix arbiters. it allows the bus matrix to remove the one latency cycle for the last master that accessed the slave. at the end of the cur-
129 32145c?06/2013 at32uc3l0128/256 rent transfer, if no other master request is pending, the slave remains connected to the last master that performed the acce ss. other non privileged masters still get one latency cycle if they want to access the same slave. this technique can be used for masters that mainly perform sin- gle accesses. ? round-robin arbitration with fixed default master this is another biased round-robin algorithm. it a llows the bus matrix arbiters to remove the one latency cycle for the fixed default master per slav e. at the end of the current access, the slave remains connected to its fixed default master. every request attempted by this fixed default mas- ter will not cause any latency whereas other non privileged masters w ill still get one latency cycle. this technique can be used for masters that mainly perform single accesses. 10.4.2.3 fixed priority arbitration this algorithm allows the bus matrix arbiters to dispatch the requests from different masters to the same slave by using the fixed priority defined by the user. if two or more master requests are active at the same time, the master with the highest priority number is serviced first. if two or more master requests with the same priority are active at the same time, the master with the highest number is serviced first. for each slave, the priority of each master may be defined through the priority registers for slaves (pras and prbs). 10.4.3 slave and master assignation the index number assigned to bus matrix slaves and masters are described in the module con- figuration section at the end of this chapter.
130 32145c?06/2013 at32uc3l0128/256 10.5 user interface table 10-1. hmatrix register memory map offset register name access reset value 0x0000 master configuration register 0 mcfg0 read/write 0x00000002 0x0004 master configuration register 1 mcfg1 read/write 0x00000002 0x0008 master configuration register 2 mcfg2 read/write 0x00000002 0x000c master configuration register 3 mcfg3 read/write 0x00000002 0x0010 master configuration register 4 mcfg4 read/write 0x00000002 0x0014 master configuration register 5 mcfg5 read/write 0x00000002 0x0018 master configuration register 6 mcfg6 read/write 0x00000002 0x001c master configuration register 7 mcfg7 read/write 0x00000002 0x0020 master configuration register 8 mcfg8 read/write 0x00000002 0x0024 master configuration register 9 mcfg9 read/write 0x00000002 0x0028 master configuration register 10 mcfg10 read/write 0x00000002 0x002c master configuration regi ster 11 mcfg11 read/write 0x00000002 0x0030 master configuration register 12 mcfg12 read/write 0x00000002 0x0034 master configuration register 13 mcfg13 read/write 0x00000002 0x0038 master configuration register 14 mcfg14 read/write 0x00000002 0x003c master configuration regi ster 15 mcfg15 read/write 0x00000002 0x0040 slave configuration register 0 scfg0 read/write 0x00000010 0x0044 slave configuration register 1 scfg1 read/write 0x00000010 0x0048 slave configuration register 2 scfg2 read/write 0x00000010 0x004c slave configuration register 3 scfg3 read/write 0x00000010 0x0050 slave configuration register 4 scfg4 read/write 0x00000010 0x0054 slave configuration register 5 scfg5 read/write 0x00000010 0x0058 slave configuration register 6 scfg6 read/write 0x00000010 0x005c slave configuration register 7 scfg7 read/write 0x00000010 0x0060 slave configuration register 8 scfg8 read/write 0x00000010 0x0064 slave configuration register 9 scfg9 read/write 0x00000010 0x0068 slave configuration register 10 scfg10 read/write 0x00000010 0x006c slave configuration register 11 scfg11 read/write 0x00000010 0x0070 slave configuration register 12 scfg12 read/write 0x00000010 0x0074 slave configuration register 13 scfg13 read/write 0x00000010 0x0078 slave configuration register 14 scfg14 read/write 0x00000010 0x007c slave configuration register 15 scfg15 read/write 0x00000010 0x0080 priority register a for slave 0 pras0 read/write 0x00000000 0x0084 priority register b for slave 0 prbs0 read/write 0x00000000 0x0088 priority register a for slave 1 pras1 read/write 0x00000000
131 32145c?06/2013 at32uc3l0128/256 0x008c priority register b for slave 1 prbs1 read/write 0x00000000 0x0090 priority register a for slave 2 pras2 read/write 0x00000000 0x0094 priority register b for slave 2 prbs2 read/write 0x00000000 0x0098 priority register a for slave 3 pras3 read/write 0x00000000 0x009c priority register b for slave 3 prbs3 read/write 0x00000000 0x00a0 priority register a for slave 4 pras4 read/write 0x00000000 0x00a4 priority register b for slave 4 prbs4 read/write 0x00000000 0x00a8 priority register a for slave 5 pras5 read/write 0x00000000 0x00ac priority register b for slave 5 prbs5 read/write 0x00000000 0x00b0 priority register a for slave 6 pras6 read/write 0x00000000 0x00b4 priority register b for slave 6 prbs6 read/write 0x00000000 0x00b8 priority register a for slave 7 pras7 read/write 0x00000000 0x00bc priority register b for slave 7 prbs7 read/write 0x00000000 0x00c0 priority register a for slave 8 pras8 read/write 0x00000000 0x00c4 priority register b for slave 8 prbs8 read/write 0x00000000 0x00c8 priority register a for slave 9 pras9 read/write 0x00000000 0x00cc priority register b for slave 9 prbs9 read/write 0x00000000 0x00d0 priority register a for slave 10 pras10 read/write 0x00000000 0x00d4 priority register b for slave 10 prbs10 read/write 0x00000000 0x00d8 priority register a for slave 11 pras11 read/write 0x00000000 0x00dc priority register b for slave 11 prbs11 read/write 0x00000000 0x00e0 priority register a for slave 12 pras12 read/write 0x00000000 0x00e4 priority register b for slave 12 prbs12 read/write 0x00000000 0x00e8 priority register a for slave 13 pras13 read/write 0x00000000 0x00ec priority register b for slave 13 prbs13 read/write 0x00000000 0x00f0 priority register a for slave 14 pras14 read/write 0x00000000 0x00f4 priority register b for slave 14 prbs14 read/write 0x00000000 0x00f8 priority register a for slave 15 pras15 read/write 0x00000000 0x00fc priority register b for slave 15 prbs15 read/write 0x00000000 0x0110 special function register 0 sfr0 read/write ? 0x0114 special function register 1 sfr1 read/write ? 0x0118 special function register 2 sfr2 read/write ? 0x011c special function register 3 sfr3 read/write ? 0x0120 special function register 4 sfr4 read/write ? 0x0124 special function register 5 sfr5 read/write ? 0x0128 special function register 6 sfr6 read/write ? table 10-1. hmatrix register memory map (continued) offset register name access reset value
132 32145c?06/2013 at32uc3l0128/256 0x012c special function register 7 sfr7 read/write ? 0x0130 special function register 8 sfr8 read/write ? 0x0134 special function register 9 sfr9 read/write ? 0x0138 special function register 10 sfr10 read/write ? 0x013c special function register 11 sfr11 read/write ? 0x0140 special function register 12 sfr12 read/write ? 0x0144 special function register 13 sfr13 read/write ? 0x0148 special function register 14 sfr14 read/write ? 0x014c special function register 15 sfr15 read/write ? table 10-1. hmatrix register memory map (continued) offset register name access reset value
133 32145c?06/2013 at32uc3l0128/256 10.5.1 master configuration registers name: mcfg0...mcfg15 access type: read/write offset: 0x00 - 0x3c reset value: 0x00000002 ? ulbt: undefined length burst type 31 30 29 28 27 26 25 24 ???????? 23 22 21 20 19 18 17 16 ???????? 15 14 13 12 11 10 9 8 ???????? 76543210 ????? ulbt table 10-2. undefined length burst type ulbt undefined length burst type description 000 inifinite length burst no predicted end of burst is generated and theref ore incr bursts coming from this master cannot be broken. 001 single-access the undefined length burst is treated as a succession of single accesses, allowing re- arbitration at each beat of the incr burst. 010 4 beat burst the undefined length bu rst is split into a four-beat burst, allowing re-arbitration at each four-beat burst end. 011 8 beat burst the undefined length burst is split into an eight-beat burst, allowing re-arbitration at each eight-beat burst end. 100 16 beat burst the undefined length burst is split in to a sixteen-beat burst, allowing re-arbitration at each sixteen-beat burst end.
134 32145c?06/2013 at32uc3l0128/256 10.5.2 slave configuration registers name: scfg0...scfg15 access type: read/write offset: 0x40 - 0x7c reset value: 0x00000010 ? arbt: arbitration type 0: round-robin arbitration 1: fixed priority arbitration ? fixed_defmstr: fixed default master this is the number of the default master for this slave. only used if defmstr_type is 2. sp ecifying the number of a master which is not connected to the selected slave is equivalent to sett ing defmstr_type to 0. ? defmstr_type: default master type 0: no default master at the end of the current slave access, if no other master r equest is pending, the slave is disconnected from all masters. this results in a one cycle latenc y for the first access of a burst transfer or for a single access. 1: last default master at the end of the current slave access, if no other master req uest is pending, the slave stays connected to the last master hav ing accessed it. this results in not having one cycle latency when the last master tries to access the slave again. 2: fixed default master at the end of the current slave access, if no other master request is pending, the slav e connects to the fixed master the numbe r that has been written in the fixed_defmstr field. this results in not having one cycle latency when the fixed master tries to access the slave again. ? slot_cycle: maximum number of allowed cycles for a burst when the slot_cycle limit is reached for a burst, it may be broken by another master trying to access this slave. this limit has been placed to avoid locking a very slow slave when very long bursts are used. this limit must not be very small. unreas onably small values break every burst and the bus matrix arbitrates without performing any data transfer. 16 cycles is a reasonable value for slot_cycle. 31 30 29 28 27 26 25 24 ???????arbt 23 22 21 20 19 18 17 16 ? ? fixed_defmstr defmstr_type 15 14 13 12 11 10 9 8 ???????? 76543210 slot_cycle
135 32145c?06/2013 at32uc3l0128/256 10.5.3 bus matrix priority registers a for slaves register name: pras0...pras15 access type: read/write offset: - reset value: 0x00000000 ? mxpr: master x priority fixed priority of master x for accessing the selected slave. the higher the number, the higher the priority. 31 30 29 28 27 26 25 24 - - m7pr - - m6pr 23 22 21 20 19 18 17 16 - - m5pr - - m4pr 15 14 13 12 11 10 9 8 - - m3pr - - m2pr 76543210 - - m1pr - - m0pr
136 32145c?06/2013 at32uc3l0128/256 10.5.4 priority registers b for slaves name: prbs0...prbs15 access type: read/write offset: - reset value: 0x00000000 ? mxpr: master x priority fixed priority of master x for accessing the selected slave. the higher the number, the higher the priority. 31 30 29 28 27 26 25 24 - - m15pr - - m14pr 23 22 21 20 19 18 17 16 - - m13pr - - m12pr 15 14 13 12 11 10 9 8 - - m11pr - - m10pr 76543210 - - m9pr - - m8pr
137 32145c?06/2013 at32uc3l0128/256 10.5.5 special function registers name: sfr0...sfr15 access type: read/write offset: 0x110 - 0x14c reset value: - ? sfr: special function register fields those registers are not a hmatrix spec ific register. the field of those will be defined where they are used. 31 30 29 28 27 26 25 24 sfr 23 22 21 20 19 18 17 16 sfr 15 14 13 12 11 10 9 8 sfr 76543210 sfr
138 32145c?06/2013 at32uc3l0128/256 10.6 module configuration the specific configuration for each hmatrix instance is listed in the following tables.the mod- ule bus clocks listed here are connected to the system bus clocks. plea se refer to the power manager chapter for details. 10.6.1 bus matrix connections the bus matrix has the several masters and slaves. each master has its own bus and its own decoder, thus allowing a different memory mapping per master. the master number in the table below can be used to index the hmatrix cont rol registers. for example, hmatrix mcfg0 register is associated with the cpu data master interface. each slave has its own arbiter, thus allowing a different arbitration per slave. the slave number in the table below can be used to index the hmatrix control registers. for example, scfg3 is associated with the internal sram slave interface. accesses to unused areas returns an error result to the master requesting such an access. table 10-3. hmatrix clocks clock name description clk_hmatrix clock for the hmatrix bus interface table 10-4. high speed bus masters master 0 cpu data master 1 cpu instruction master 2 cpu sab master 3 sau master 4 pdca table 10-5. high speed bus slaves slave 0 internal flash slave 1 hsb-pb bridge a slave 2 hsb-pb bridge b slave 3 internal sram slave 4 sau
139 32145c?06/2013 at32uc3l0128/256 figure 10-1. hmatrix master / slave connections cpu data 0 cpu instruction 1 cpu sab 2 sau 3 internal flash 0 hsb-pb bridge a 1 hsb-pb bridge b 2 internal sram 3 hmatrix slaves hmatrix masters sau 4 pdca 4
140 32145c?06/2013 at32uc3l0128/256 11. interrupt controller (intc) rev: 1.0.2.5 11.1 features ? autovectored low latency interrupt service with programmable priority ? 4 priority levels for regular, maskable interrupts ? one non-maskable interrupt ? up to 64 groups of interrupts with up to 32 interrupt requests in each group 11.2 overview the intc collects interrupt requests from the peripherals, prioritizes them, and delivers an inter- rupt request and an autovector to the cpu. the avr32 architecture supports 4 priority levels for regular, maskable interrupts, and a non-maskable interrupt (nmi). the intc supports up to 64 groups of interrupts. each group can have up to 32 interrupt request lines, these lines are connected to the peripherals. each group has an interrupt priority register (ipr) and an interrupt request register (irr). the iprs are used to assign a priority level and an autovector to each group, and the irrs are used to identify the active interrupt request within each group. if a group has only one interrupt request line, an active interrupt group uniquely identifies the active interrupt request line, and the corresponding irr is not needed. the intc also provides one interrupt cause register (icr) per priority level. these registers identify the group that has a pending interrupt of the corresponding priority level. if several groups have a pending interrupt of the same level, the group with the lowest number takes priority. 11.3 block diagram figure 11-1 gives an overview of the intc. the grey boxes represent registers that can be accessed via the user interface. the interrupt requests from the peripherals (ireqn) and the nmi are input on the left side of the figure. signals to and from the cpu are on the right side of the figure.
141 32145c?06/2013 at32uc3l0128/256 figure 11-1. intc block diagram 11.4 product dependencies in order to use this module, other parts of the system must be configured correctly, as described below. 11.4.1 power management if the cpu enters a sleep mode that disables clk_sync, the intc will stop functioning and resume operation after the system wakes up from sleep mode. 11.4.2 clocks the clock for the intc bus interface (clk_intc) is generated by the power manager. this clock is enabled at reset, and can be disabled in the power manager. the intc sampling logic runs on a clock which is stopped in any of the sleep modes where the system rc oscillator is not running. this cl ock is referred to as clk_sync. this clock is enabled at reset, and only turned off in sleep modes wher e the system rc oscillator is stopped. 11.4.3 debug operation when an external debugger forces the cpu into debug mode, the intc continues normal operation. 11.5 functional description all of the incoming interrupt requests (ireqs) are sampled into the corresponding interrupt request register (irr). the irrs must be accessed to identify which ireq within a group that is active. if several ireqs within the same group are active, the interrupt service routine must prioritize between them. all of the input lines in each group are logically ored together to form the grpreqn lines, indicating if there is a pending interrupt in the corresponding group. the request masking hardware maps each of the grpreq lines to a priority level from int0 to int3 by associating each grou p with the interrupt level (int level) field in the corresponding request masking or ireq0 ireq1 ireq2 ireq31 grpreq0 masks sreg masks i[3-0]m gm intlevel autovector prioritizer cpu interrupt controller or grpreqn nmireq or ireq32 ireq33 ireq34 ireq63 grpreq1 irr registers ipr registers icr registers int_level, offset int_level, offset int_level, offset ipr0 ipr1 iprn irr0 irr1 irrn valreq0 valreq1 valreqn . . . . . . . . .
142 32145c?06/2013 at32uc3l0128/256 interrupt priority register (ipr). the grpreq inputs are then masked by the mask bits from the cpu status register. any interrupt group that has a pending interrupt of a priority level that is not masked by the cpu status register, gets its corresponding valreq line asserted. masking of the interrupt requests is done based on five interrupt mask bits of the cpu status register, namely interrupt level 3 mask (i3m) to interrupt level 0 mask (i0m), and global inter- rupt mask (gm). an interrupt request is masked if either the gm or the corresponding interrupt level mask bit is set. the prioritizer hardware uses th e valreq lines and the intlevel fi eld in the iprs to select the pending interrupt of the highest priority. if an nm i interrupt request is pending, it automatically gets the highest priority of any pending interrupt. if several interrupt groups of the highest pend- ing interrupt level have pending interrupts, the interrupt group with the lowest number is selected. the intlevel and handler autovector offset (autovector) of the selected interrupt are transmitted to the cpu for interrupt handling an d context switching. the cpu does not need to know which interrupt is requesting handling, but only the level and the offset of the handler address. the irr registers contain the interrupt request lines of the groups and can be read via user interface registers for checking which interrupts of the group are actually active. the delay through the intc from the peripheral interrupt request is set until the interrupt request to the cpu is set is three cycles of clk_sync. 11.5.1 non-maskable interrupts a nmi request has priority over all other interrupt requests. nmi has a dedicated exception vec- tor address defined by the avr32 architec ture, so autovector is undefined when intlevel indicates that an nmi is pending. 11.5.2 cpu response when the cpu receives an interr upt request it checks if any othe r exceptions are pending. if no exceptions of higher priority are pending, interr upt handling is initiated. when initiating interrupt handling, the corresponding interrupt mask bit is se t automatically for this and lower levels in sta- tus register. e.g, if an interrupt of level 3 is approved for handling, the interrupt mask bits i3m, i2m, i1m, and i0m are set in status register. if an interrupt of level 1 is approved, the masking bits i1m and i0m are set in status register. th e handler address is calculated by logical or of the autovector to the cpu system register exception vector base address (evba). the cpu will then jump to the calculated address and start executing th e interrupt handler. setting the interrupt mask bits prevents the interrupts from the same and lower levels to be passed through the interrupt controller. setting of the same level mask bit prevents also multiple requests of the same interrupt to happen. it is the responsibility of the ha ndler software to clear the interrupt request that caused the inter- rupt before returning from the interrupt handler. if the conditions that caused the interrupt are not cleared, the interrupt request remains active. 11.5.3 clearing an interrupt request clearing of the interrupt request is done by writing to registers in the corresponding peripheral module, which then clears the corresponding nmireq/ireq signal. the recommended way of clearing an interrupt request is a store operation to the controlling peripheral register, followed by a dummy load operat ion from the same register. this causes a
143 32145c?06/2013 at32uc3l0128/256 pipeline stall, which prevents the interrupt from accidentally re-triggering in case the handler is exited and the interrupt mask is cleared before the interrupt request is cleared.
144 32145c?06/2013 at32uc3l0128/256 11.6 user interface table 11-1. intc register memory map offset register register name access reset 0x000 interrupt priority register 0 ipr0 read/write 0x00000000 0x004 interrupt priority register 1 ipr1 read/write 0x00000000 ... ... ... ... ... 0x0fc interrupt priority register 63 ipr63 read/write 0x00000000 0x100 interrupt request register 0 irr0 read-only n/a 0x104 interrupt request register 1 irr1 read-only n/a ... ... ... ... ... 0x1fc interrupt request regi ster 63 irr63 read-only n/a 0x200 interrupt cause register 3 icr3 read-only n/a 0x204 interrupt cause register 2 icr2 read-only n/a 0x208 interrupt cause register 1 icr1 read-only n/a 0x20c interrupt cause register 0 icr0 read-only n/a
145 32145c?06/2013 at32uc3l0128/256 11.6.1 interrupt priority registers name : ipr0...ipr63 access type: read/write offset: 0x000 - 0x0fc reset value: 0x00000000 ? intlevel: interrupt level indicates the evba-relative offs et of the interrup t handler of the co rresponding group: 00: int0: lowest priority 01: int1 10: int2 11: int3: highest priority ? autovector: autovector address handler offset is used to give the address of the interrupt handle r. the least significant bit should be written to zero to giv e halfword alignment. 31 30 29 28 27 26 25 24 intlevel ------ 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 - - autovector[13:8] 76543210 autovector[7:0]
146 32145c?06/2013 at32uc3l0128/256 11.6.2 interrupt request registers name : irr0...irr63 access type: read-only offset: 0x0ff - 0x1fc reset value: n/a ? irr: interrupt request line this bit is cleared when no interrupt request is pending on this input request line. this bit is set when an interrupt request is pending on this input request line. the are 64 irrs, one for each group. each irr has 32 bits, one for each possible interrupt request, for a total of 2048 possibl e input lines. the irrs are read by the software interrupt handler in order to determine which interrupt request is pending. the irrs are sampled continuously, and are read-only. 31 30 29 28 27 26 25 24 irr[32*x+31] irr[32*x+30] irr[32*x+ 29] irr[32*x+28] irr[32* x+27] irr[32*x+26] irr[32*x+25] irr[32*x+24] 23 22 21 20 19 18 17 16 irr[32*x+23] irr[32*x+22] irr[32*x+ 21] irr[32*x+20] irr[32* x+19] irr[32*x+18] irr[32*x+17] irr[32*x+16] 15 14 13 12 11 10 9 8 irr[32*x+15] irr[32*x+14] irr[32*x+ 13] irr[32*x+12] irr[32*x+ 11] irr[32*x+10] irr[32*x+9] irr[32*x+8] 76543210 irr[32*x+7] irr[32*x+6] irr[32*x+5] irr[32*x+4] irr[32*x+3] irr[32*x+2] irr[32*x+1] irr[32*x+0]
147 32145c?06/2013 at32uc3l0128/256 11.6.3 interrupt cause registers name : icr0...icr3 access type: read-only offset: 0x200 - 0x20c reset value: n/a ? cause: interrupt group causing interrupt of priority n icrn identifies the group with the highest pr iority that has a pending interrupt of le vel n. this value is only defined when at least one interrupt of level n is pending. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 -- cause
148 32145c?06/2013 at32uc3l0128/256 11.7 module configuration the specific configuration for each intc instance is listed in the following tables.the module bus clocks listed here are connected to the syst em bus clocks. please refer to the power man- ager chapter for details. 11.7.1 interrupt request signal map the interrupt controller supports up to 64 groups of interrupt requests. each group can have up to 32 interrupt request signals. all interrupt signals in the same group share the same autovector address and priority level. the table below shows how the interrupt re quest signals are connected to the intc. table 11-2. intc clock name module name clock name description intc clk_intc clock for the intc bus interface table 11-3. interrupt request signal map group line module signal 0 0 avr32uc3 cpu sysreg compare 1 0 avr32uc3 cpu ocd dcemu_dirty 1 avr32uc3 cpu ocd dccpu_read 2 0 flash controller flashcdw 3 0 secure access unit sau 4 0 peripheral dma controller pdca 0 1 peripheral dma controller pdca 1 2 peripheral dma controller pdca 2 3 peripheral dma controller pdca 3 5 0 peripheral dma controller pdca 4 1 peripheral dma controller pdca 5 2 peripheral dma controller pdca 6 3 peripheral dma controller pdca 7 6 0 peripheral dma controller pdca 8 1 peripheral dma controller pdca 9 2 peripheral dma controller pdca 10 3 peripheral dma controller pdca 11 7 0 power manager pm 8 0 system control interface scif 9 0 asynchronous timer ast alarm
149 32145c?06/2013 at32uc3l0128/256 10 0 asynchronous timer ast per 1 asynchronous timer ast ovf 2 asynchronous timer ast ready 3 asynchronous timer ast clkready 11 0 external interrupt controller eic 1 1 external interrupt controller eic 2 2 external interrupt controller eic 3 3 external interrupt controller eic 4 12 0 external interrupt controller eic 5 13 0 frequency meter freqm 14 0 general-purpose input/ou tput controller gpio 0 1 general-purpose input/ou tput controller gpio 1 2 general-purpose input/ou tput controller gpio 2 3 general-purpose input/ou tput controller gpio 3 4 general-purpose input/ou tput controller gpio 4 5 general-purpose input/ou tput controller gpio 5 6 general-purpose input/ou tput controller gpio 6 7 general-purpose input/ou tput controller gpio 7 15 0 universal synchronous asynchronous receiver transmitter usart0 16 0 universal synchronous asynchronous receiver transmitter usart1 17 0 universal synchronous asynchronous receiver transmitter usart2 18 0 universal synchronous asynchronous receiver transmitter usart3 19 0 serial peripheral interface spi 20 0 two-wire master interface twim0 21 0 two-wire master interface twim1 22 0 two-wire slave interface twis0 23 0 two-wire slave interface twis1 24 0 pulse width modulation controller pwma 25 0 timer/counter tc00 1 timer/counter tc01 2 timer/counter tc02 26 0 timer/counter tc10 1 timer/counter tc11 2 timer/counter tc12 table 11-3. interrupt request signal map
150 32145c?06/2013 at32uc3l0128/256 27 0 adc interface adcifb 28 0 analog comparator interface acifb 29 0 capacitive touch module cat 30 0 awire aw table 11-3. interrupt request signal map
151 32145c?06/2013 at32uc3l0128/256 12. power manager (pm) rev: 4.2.0.4 12.1 features ? generates clocks and resets for digital logic ? on-the-fly frequency change of cpu, hsb and pbx clocks ? sleep modes allow simple disabling of logic clocks and clock sources ? module-level clock gating through maskable peripheral clocks ? wake-up from internal or external interrupts ? automatic identificat ion of reset sources ? supports advanced shutdown sleep mode 12.2 overview the power manager (pm) provides synchronous clocks used to clock the main digital logic in the device, namely the cpu, and the modules and peripherals connected to the high speed bus (hsb) and the peripheral buses (pbx). the pm contains advanced power-saving features, allowing the user to optimize the power con- sumption for an application. the synchronous clocks are divided into a number of clock domains, one for the cpu and hsb, and one fo r each pbx. the clocks can run at different speeds, allowing the user to save power by runni ng peripherals relatively slow, whilst maintain- ing high cpu performance. the clocks can be independently cha nged on-the-fly, without halting any peripherals. the user may adjust cpu and memory speeds according to the dynamic appli- cation load, without disturbing or re-configuring active peripherals. each module has a separate clock, enabling the user to save power by switching off clocks to inactive modules. clocks and oscilla tors can be automatically switch ed off during idle periods by the cpu sleep instruction. the system will retu rn to normal operatio n when interrupts occur. to achieve minimal power usage, a special sleep mode, called shutdown is available, where power on all internal logic (cpu, peripherals) and most of the i/o lines is removed, reducing cur- rent leakage. only a sm all amount of logic, including the 32khz crystal oscillator (osc32k) and the ast remain powered. the power manager also contains a reset contro ller, which collects all possible reset sources, generates hard and soft resets, and allows the reset source to be identified by software.
152 32145c?06/2013 at32uc3l0128/256 12.3 block diagram figure 12-1. pm block diagram 12.4 i/o lines description 12.5 product dependencies 12.5.1 interrupt the pm interrupt line is connected to one of the interrupt controllers internal sources. using the pm interrupt requires the interrupt controller to be configured first. 12.5.2 clock implementation in at32uc3l0128/256, the hsb shares source clock with the cpu. write attempts to the hsb clock select register (hsbsel) will be ignored, and it will alwa ys read the sa me as the cpu clock select register (cpusel). the pm bus interface clock (clk_pm) is generated by the power manager. this clock is enabled at reset, and can be disabled in the power manager. if disabled it can only be re- enabled by a reset. 12.5.3 power considerations the shutdown mode is only available for the ?3.3v supply mode, with 1.8v regulated i/o lines? power configuration. table 12-1. i/o lines description name description type active level reset_n reset input low sleep controller synchronous clock generator reset controller main clock sources sleep instruction power-on reset detector(s) resets synchronous clocks cpu, hsb, pbx interrupts external reset pin reset sources
153 32145c?06/2013 at32uc3l0128/256 12.6 functional description 12.6.1 synchronous clocks the system rc oscillator (rcsys) and a select ion of other clock sources can provide the source for the main clock, which is the origin for the synchronous cpu/hsb and pbx module clocks. for details about the other main clock sources, please refer to the main clock control (mcctrl) register descr iption. the synchronous clocks can ru n of the main clock and all the 8- bit prescaler settings as long as f cpu ? f pbx, . the synchronous clock source can be changed on- the fly, according to variations in application load. the clock domains can be shut down in sleep mode, as described in section 12.6.3 . the module clocks in every synchronous clock domain can be individually masked to minimize power consumption in inactive modules. figure 12-2. synchronous clock generation 12.6.1.1 selecting the main clock source the common main clock can be co nnected to rcsys or a selection of other clock sources. for details about the other main clock sources, please refer to the mcctrl register description. by default, the main clock will be connected to rc sys. the user can connect the main clock to another source by writing to the main clock select (mcctrl.mcsel) field. the user must first assure that the source is enabled and ready in order to avoid a deadlock. care should also be taken so that the new synchronous clock frequencies do not exceed the maximum frequency for each clock domain. 12.6.1.2 selecting synchronous clock division ratio the main clock feeds an 8-bit prescaler, which can be used to generate the synchronous clocks. by default, the synchronous clocks run on the undivided main clock. the user can select a pres- caler division for the cpu clock by writing a on e to the cpu division bit in the cpu clock select register (cpusel.cpudiv), and a value to th e cpu clock select field (cpusel.cpusel), resulting in a cpu clock frequency: f cpu = f main / 2 (cpusel+1) mask prescaler main clock sources mcsel 0 1 cpusel cpudiv main clock sleep controller cpumask cpu clocks hsb clocks pbx clocks sleep instruction
154 32145c?06/2013 at32uc3l0128/256 similarly, the pbx clocks can be divided by writ ing their respective clock select (pbxsel) regis- ters to get the divided pbx frequency: f pbx = f main / 2 (pbsel+1) the pbx clock frequency can not exceed the cpu clock frequency. the user must select a pbx- sel.pbsel value greate r than or equal to the cpu sel.cpusel value, so that f cpu ? f pbx . if the user selects division factors that will result in f cpu < f pbx , the power manager will automatically change the pbxsel.pbsel/pbdiv values to ensure correct operation (f cpu ? f pbx ). the hsb clock will always be forced to the same division as the cpu clock. to ensure correct operation, the frequencies must never exceed the specified maximum fre- quency for each clock domain. for modules connected to the hsb bus, the pb cl ock frequency must be the same as the cpu clock frequency. 12.6.1.3 clock ready flag there is a slight delay from cpusel and pbx sel being written to the new clock setting taking effect. during this interval, the clock ready bi t in the status register (sr.ckrdy) will read as zero. when the clock settings change is complet ed, the bit will read as one. the clock select registers (cpusel, pbxsel) must not be writte n to while sr.ckrdy is zero, or the system may become unstable or hang. the clock ready bit in the interrupt status regi ster (isr.ckrdy) is set on a sr.ckrdy zero- to-one transition. if the clock ready bit in the interrupt mask register (imr.ckrdy) is set, an interrupt request is generated. imr.ckrdy is set by writing a one to the corresponding bit in the interrupt enable register (ier.ckrdy). 12.6.2 peripheral clock masking by default, the clocks for all modules are enab led, regardless of which modules are actually being used. it is possible to disable the clock for a module in the cpu, hsb, or pbx clock domain by writing a zero to the corresponding bit in the corresponding clock mask (cpu- mask/hsbmask/pbxmask) register. when a module is not clocked, it will cease operation, and its registers cannot be read nor written. the module can be re-enabled later by writing a one to the corresponding mask bit. a module may be connected to several clock domains, in which case it will have several mask bits. the maskable module clocks table in the clock mask regis- ter description contains a list of implemented maskable clocks. 12.6.2.1 cautionary note note that clocks should only be switched off if it is certain that the module will not be used. switching off the clock for the flash controller will cause a problem if the cpu needs to read from the flash. switching off the clock to the power manager, which contains the mask registers, or the corresponding pbx bridge, will make it impossible to write to the mask registers again. in this case, they can only be re-enabled by a system reset. 12.6.3 sleep modes in normal operation, all clock domains are active, allowing software execution and peripheral operation. when the cpu is idle, it is possible to switch it and other (optional) clock domains off to save power. this is done by the sleep instruction, which takes the sleep mode index number from table 12-2 on page 155 as argument.
155 32145c?06/2013 at32uc3l0128/256 12.6.3.1 entering and exiting sleep modes the sleep instruction will halt the cpu and all modules belonging to t he stopped clock domains. the modules will be halted regardless of th e bit settings in the mask registers. clock sources can also be switched off to save power. some of these have a relatively long start-up time, and are only switched off when very low power consumption is required. the cpu and affected modules are restarted when the sleep mode is exited. this occurs when an interrupt triggers. note that even if an interrupt is enabled in sleep mode, it may not trigger if the source module is not clocked. 12.6.3.2 supported sleep modes the following sleep modes are supported. these are detailed in table 12-2 on page 155 . ? idle: the cpu is stopped, the rest of the device is operational. ? frozen: the cpu and hsb modules are stopped, peripherals are operational. ? standby: all synchronous clocks are stopped, and the clock sources are running, allowing for a quick wake-up to normal mode. ? stop: as standby, but oscilla tors, and other clock sources ar e also stopped. 32khz oscillator osc32k (2) , rcsys, ast, and wdt will remain operational. ? deepstop: all synchronous clocks and clock sources are stopped. bandgap voltage reference and bod are turned off. osc32k (2) and rcsys remain operational. ? static: all clock sources, including rcsys ar e stopped. band gap voltage reference and bod are turned off. osc32k (2) remains operational. ? shutdown: all clock sources, including rcsys are stopped. bandgap voltage reference, bod detector, and voltage regulator are turned off. osc32k (2) remains operational. this mode can only be used in the ? 3.3v supply mode, with 1.8v regulated i/o lines ? configuration (described in power considerations chapter). refer to section 12.6.4 for more details. notes: 1. the sleep mode index is used as argument for the sleep instruction. 2. osc32k will only remain operational if pre-enabled. 3. clock sources other than those specifically listed in the table. 4. systimer is the clock for the cpu count and compare registers. the internal voltage regulator is also adjusted according to the sleep mode in order to reduce its power consumption. table 12-2. sleep modes index (1) sleep mode cpu hsb pbx, gclk clock sources (3) , systimer (4) osc32k (2) rcsys bod & bandgap voltag e regulator 0 idle stop run run run run run on normal mode 1 frozen stop stop run run run run on normal mode 2 standby stop stop stop run run run on normal mode 3 stop stop stop stop stop run run on low power mode 4 deepstop stop stop stop stop run run off low power mode 5 static stop stop stop stop run stop off low power mode 6 shutdown stop stop stop stop run stop off off
156 32145c?06/2013 at32uc3l0128/256 12.6.3.3 waking from sleep modes there are two types of wake-up sources from sleep mode, synchronous and asynchronous. synchronous wake-up sources are all non-masked interrupts. asynchronous wake-up sources are ast, wdt, external interrupts from eic, external reset, external wake pin (wake_n), and all asynchronous wake-ups enabled in the asynchronous wake up enable (awen) register. the valid wake-up sources for each sleep mode are detailed in table 12-3 on page 156 . in shutdown the only wake-up sources are external reset, external wake-up pin or ast. see section 12.6.4.3 on page 158 . notes: 1. the sleep mode index is used as argument for the sleep instruction. 2. only pb modules operational, as hsb module clocks are stopped. 3. wdt only available if clocked from pre-enabled osc32k. 12.6.3.4 sleepwalking in all sleep modes where the pbx clocks are stopped, except for shutdown mode, the device can partially wake up if a pbx module asynchronous ly discovers that it needs its clock. only the requested clocks and clock sources needed will be st arted, all other clocks will remain masked to zero. e.g. if the main clock source is osc0 , only osc0 will be started even if other clock sources were enabled in normal mode. generic clocks can also be started in a similar way. the state where only requested clocks are running is referred to as sleepwalking. the time spent to start the requested clock is mostly limited by the startup time of the given clock source. this allows pbx modules to handle incoming requests, while still keeping the power con- sumption at a minimum. when the device is sleepwalking any asynchronous wake-up can wake the device up at any time without stopping the requested pbx clock. all requests to start clocks can be masked by writing to the peripheral power control register (ppcr), all requests are enabled at reset. during sleepwalking the interrupt controller clock will be running. if an interrupt is pending when entering sleepwalki ng, it will wake the whole device up. 12.6.3.5 precautions when entering sleep mode modules communicating with external circuits should normally be disabled before entering a sleep mode that will stop the m odule operation. this will prevent errati c behavior caused by entering or exiting sleep modes. please refer to the relevant module documentation for recom- mended actions. table 12-3. wake-up sources index (1) sleep mode wake-up sources 0 idle synchronous, asynchronous 1 frozen synchronous (2) , asynchronous 2 standby asynchronous 3 stop asynchronous 4 deepstop asynchronous 5 static asynchronous (3) 6 shutdown external reset, external wake-up pin
157 32145c?06/2013 at32uc3l0128/256 communication between the synchronous clock domains is disturbed when entering and exiting sleep modes. bus transactions over clock domains affected by the sleep mode are therefore not recommended. the system may han g if the bus clocks are stopped during a bus transaction. the cpu is automatically stopped in a safe state to ensure that all cpu bus operations are com- plete when the sleep mode goes into effect. thus, when entering idle mode, no further action is necessary. when entering a sleep mode (except idle mode), all hsb masters must be stopped before entering the sleep mode. in order to let potential pbx write operations complete, the user should let the cpu perform a pbx register read operation before issuing the sleep instruction. this will stall the cpu until pending pbx operations have completed. the shutdown sleep mode requires extra care. please refer to section 12.6.4 . 12.6.4 shutdown sleep mode 12.6.4.1 description the shutdown sleep mode is available only when the device is used in the ? 3.3v supply mode, with 1.8v regulated i/o lines ? configuration (refer to the power considerations chapter). in this configuration, the voltage regulator supplies both vddcore and vddio power supplies. when the device enters shutdown mode, the regulator is turned off and only the following logic is kept powered by vddin: ? osc32k using alternate pinout pa13/pa20 ? ast core logic (internal counter and alarm detection logic) ? backup registers ? i/o lines pa11, pa13, pa20, pa21, pb04, pb05, and pb10 ? reset_n line the table below lists i/o line functionality that remains operational during shutdown sleep mode. if no special function is used the i/o line will ke ep its setting when en tering the sleep mode 12.6.4.2 entering shutdown sleep mode before entering the shutdown sleep mode, a few actions are required: ? all modules should normally be disabled before entering shutdown sleep mode (see section 12.6.3.5 ) table 12-4. i/o lines usage during shutdown mode pin possible usage during shutdown sleep mode pa11 wake_n signal (active low wake-up) pa13 xin32_2 (osc32k using alternate pinout) pa20 xout32_2 (osc32k using alternate pinout) pa 2 1 pb04 pb05 pb10 reset_n reset pin
158 32145c?06/2013 at32uc3l0128/256 ? the por33 must be masked to avoid spurious resets when the power is back. this must also be done when por33 is disabled, as por33 will be enabled automatically when the device wakes up from shutdown mode. disable the por33 by writing a one to the por33mask bit in the scif.vregcr register. due to internal synchronisation, this bit must be read as a one before the sleep instruction is executed by the cpu. refer to the system control interface (scif) chapter for more details. ? the 32khz rc oscillator (rc32k) must be running and stable . this is done by writing a one to the en bit in the scif.rc32kcr register. due to internal synchronisation, this bit must be read as a one to ensure that the oscillator is stable before the sleep instruction is executed by the cpu. as soon as the shutdown sleep mode is entered, all cpu and peripherals are reset to ensure a consistent state. por33 and rc32k are automatically disabled to save extra power. 12.6.4.3 leaving shutdown sleep mode exiting shutdown sleep mode can be done by the events described in table 12-5 . when a wake-up event occurs, th e regulator is turned on and th e device will wait for vddcore to be valid before starting. th e sleep reset bit in the reset cause register (rcause.sleep) is then set, allowing software running on the device to distinguish between the first power-up and a wake-up from shutdown mode. 12.6.4.4 special consideration re garding waking up from shutdown sleep mode usin g the wake_n pin by default, the wake_n pin will only wake the device up if it is pulled low after entering shut- down mode. if the wake_n is pulled low before the shutdown mode is entered, it will not wake the device from the sh utdown sleep mode. in order to wa ke the device by pulling wake_n low before entering shutdown mode, the user has to write a one to the bit corresponding to the waken wake-up source in the aw en register. in this scenar io, the cpu execution will proceed with the next instruction, and the rcau se register content will not be altered. 12.6.5 divided pb clocks the clock generator in the powe r manager provides di vided pbx clocks for use by peripherals that require a prescaled pbx clock. this is described in the documentation for the relevant mod- ules. the divided clocks are directly maskable, and are stopped in sleep modes where the pbx clocks are stopped. table 12-5. events that can wake up the device from shutdown mode source how pa11 (wake_n) pulling-down pa11 will wake up the device reset_n pulling-down reset_n pin will wake up the device the device is kept under re set until reset_n is tied high again ast osc32k must be set-up to use alternate pinout (xin32_2 and xout32_2) refer to the scif chapter ast must be configured to use the clock from osc32k ast must be configured to allow alarm, periodic, or overflow wake-up
159 32145c?06/2013 at32uc3l0128/256 12.6.6 reset controller the reset controller collects the various reset sources in the system and generates hard and soft resets for the digital logic. the device contains a power-on reset (por) detector, which keeps the system reset until power is stable. this eliminates the need for exte rnal reset circuitry to guarantee stable opera- tion when powering up the device. it is also possible to reset the device by pulling the reset_n pin low. this pin has an internal pull-up, and does not need to be driven externally during normal operation. table 12-6 on page 159 lists these and other reset sources supported by the reset controller. figure 12-3. reset controller block diagram in addition to the listed reset types, the jtag & awire can keep parts of the device statically reset. see jtag and awire documentation for details. table 12-6. reset description reset source description power-on reset supply voltage below the power-on reset detector threshold voltage v pot external reset reset_n pin asserted brown-out reset vddcore supply voltage belo w the brown-out detector threshold voltage jtag reset controller reset_n power-on reset detector(s) ocd watchdog reset rcause cpu, hsb, pbx ocd, ast, wdt, clock generator brown-out detector awire sm33 detector
160 32145c?06/2013 at32uc3l0128/256 depending on the reset source, when a reset occurs, some parts of the device are not always reset. only the power-on reset (por) will force a whole device reset. refer to the table in the module configuration section at the end of this chapter for further details. the latest reset cause can be read in the rcause register, and can be read during the applications boot sequence in order to determine proper action. 12.6.6.1 power-on reset detector the power-on reset 1.8v (por18) detector monitors the vddcore supply pin and generates a power-on reset (por) when the device is powered on. the por is active until the vddcore voltage is above the power-on threshold level (v pot ). the por will be re-generated if the voltage drops below the power-on threshold le vel. see electrical characteristics for para- metric details. the power-on reset 3.3v (por33) detector monitors the internal regulator supply pin and gen- erates a power-on reset (por) when the device is powered on. the por is active until the internal regulator supply voltage is above the regulator power-on threshold level (v pot ). the por will be re-generated if the voltage drops below the regulator power-on threshold level. see electrical characteristics for parametric details. 12.6.6.2 external reset the external reset detector monitors the reset_n pin state. by default, a low level on this pin will generate a reset. 12.6.7 clock failure detector this mechanism automatically s witches the main clock source to the safe rcsys clock when the main clock source fails. th is may happen when an external crystal is selected as a source for the main clock and the crystal is not mounted on the board. the main clock is compared with rcsys, and if no rising edge of the main clock is detecte d during one rcsys period, the clock is considered to have failed. the detector is enabled by writing a one to th e clock failure detection enable bit in the clock failure detector control register (cfdctrl.c fden). as soon as the detector is enabled, the clock failure detector will monitor the divided main clock. note that the detector does not monitor the main clock if rcsys is the source of the main clock, or if the main clock is temporarily not available (startup-time after a wake-up, switching timing etc.), or in sleep mode where the main clock is driven by the rcsys (s top and deepstop mode). when a clock failure is detected, the main clock automatically switches to the rc sys clock and the clock failure detected (cfd) interrupt is generated if enabled. the mcctrl register is also changed by hardware to indicate that the main cloc k comes from rcsys. 12.6.8 interrupts the pm has a number of interrupt sources: ? ae - access error, sm33 reset internal regulator supply voltage below the sm33 threshold voltage. this generates a power-on reset. watchdog timer see watchdog timer documentation ocd see on-chip debug documentation reset source description
161 32145c?06/2013 at32uc3l0128/256 ? a lock protected register is written to without first being unlocked. ? ckrdy - clock ready: ? new clock select settings in the cpusel /pbxsel registers have taken effect. (a zero-to-one transition on sr.ckrdy is detected). ? cfd - clock failure detected: ? the system detects that the main clock is not running. the interrupt status register contains one bit fo r each interrupt source. a bit in this register is set on a zero-to-one transition of the correspond ing bit in the status re gister (sr), and cleared by writing a one to the corresponding bit in the interrupt clear register (icr). the interrupt sources will generate an interrupt re quest if the corresponding bit in the interrupt mask register is set. the interrupt sources are ored together to form one interrupt request. the power man- ager will generate an interrupt request if at least one of the bits in the interrupt mask register (imr) is set. bits in imr are set by writing a one to the corresponding bit in the interrupt enable register (ier), and cleared by writing a one to the corresponding bit in the interrupt disable register (idr). the interrupt request remains active until the corresponding bit in the interrupt status register (isr) is cleared by writing a one to the corresponding bit in the interrupt clear register (icr). because all the interrupt sour ces are ored together, the interrupt request from the power manager will remain active un til all the bits in isr are cleared.
162 32145c?06/2013 at32uc3l0128/256 12.7 user interface note: 1. the reset value is device specific. please refer to th e module configuration section at the end of this chapter. 2. latest reset source. 3. latest wake source. table 12-7. pm register memory map offset register register name access reset 0x000 main clock control mcctrl read/write 0x00000000 0x004 cpu clock select cpusel read/write 0x00000000 0x008 hsb clock select hsbsel read-only 0x00000000 0x00c pba clock select pbasel read/write 0x00000000 0x010 pbb clock select pbbsel read/write 0x00000000 0x014 - 0x01c reserved 0x020 cpu mask cpumask read/write 0x00010001 0x024 hsb mask hsbmask read/write 0x0000007f 0x028 pba mask pbamask read/write 0x0fffffff 0x02c pbb mask pbbmask read/write 0x0000000f 0x030- 0x03c reserved 0x040 pba divided mask pbadivmask read/write 0x0000007f 0x044 - 0x050 reserved 0x054 clock failure detector control cfdctrl read/write 0x00000000 0x058 unlock register unlock write-only 0x00000000 0x05c - 0x0bc reserved 0x0c0 interrupt enable register ier write-only 0x00000000 0x0c4 interrupt disable register idr write-only 0x00000000 0x0c8 interrupt mask register imr read-only 0x00000000 0x0cc interrupt status r egister isr read-only 0x00000000 0x0d0 interrupt clear register icr write-only 0x00000000 0x0d4 status register sr read-only 0x00000020 0x0d8 - 0x15c reserved 0x160 peripheral power control register ppcr read/write 0x000001fa 0x164 - 0x17c reserved 0x180 reset cause register rcause read-only - (2) 0x184 wake cause register wcause read-only - (3) 0x188 asynchronous wake up enable register awen read/write 0x00000000 0x18c - 0x3f4 reserved 0x3f8 configuration register config read-only 0x00000043 0x3fc version register version read-only - (1)
163 32145c?06/2013 at32uc3l0128/256 12.7.1 main clock control name: mcctrl access type: read/write offset: 0x000 reset value: 0x00000000 ? mcsel: main clock select note: 1. if the 120mhz rc oscillator is selected as main clock sour ce, it must be divided by at least 4 before being used as cloc k source for the cpu. this division is selected by writing to the cpusel and cpudiv bits in the cpusel register, before switching to rc120m as main clock source. note that this register is protected by a lock. to write to this register the unlock register has to be written first. please refer to the unlock register description for details. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 ----- mcsel table 12-8. main clocks in at32uc3l0128/256. mcsel[2:0] main clock source 0 system rc oscillator (rcsys) 1 oscillator0 (osc0) 2dfll 3 120mhz rc oscillator (rc120m) (1)
164 32145c?06/2013 at32uc3l0128/256 12.7.2 cpu clock select name: cpusel access type: read/write offset: 0x004 reset value: 0x00000000 ? cpudiv, cpusel: cpu di vision and clock select cpudiv = 0: cpu clock equals main clock. cpudiv = 1: cpu clock equals main clock divided by 2 (cpusel+1) . note that if cpudiv is written to 0, cpusel should also be written to 0 to ensure correct operation. also note that writing this register clears poscsr.ckrdy. the register must not be re-written until ckrdy is set. note that this register is protected by a lock. to write to this register the unlock register has to be written first. please refer to the unlock register description for details. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 cpudiv - - - - cpusel
165 32145c?06/2013 at32uc3l0128/256 12.7.3 hsb clock select name: hsbsel access type: read offset: 0x008 reset value: 0x00000000 this register is read-only and its content is always equal to cpusel. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 hsbdiv - - - - hsbsel
166 32145c?06/2013 at32uc3l0128/256 12.7.4 pbx clock select name: pbxsel access type: read/write offset: 0x00c-0x010 reset value: 0x00000000 ? pbdiv, pbsel: pbx division and clock select pbdiv = 0: pbx clock equals main clock. pbdiv = 1: pbx clock equals main clock divided by 2 (pbsel+1) . note that if pbdiv is written to 0, pbsel should also be written to 0 to ensure correct operation. also note that writing this register clears sr.ckrdy. the register must not be re-written until sr.ckrdy goes high. note that this register is protected by a lock. to write to this register the unlock register has to be written first. please refer to the unlock register description for details. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 pbdiv - - - - pbsel
167 32145c?06/2013 at32uc3l0128/256 12.7.5 clock mask name: cpumask/hsbmask/pbamask/pbbmask access type: read/write offset: 0x020-0x02c reset value: - ? mask: clock mask if bit n is cleared, the clock for module n is stopped. if bit n is set, the clock for module n is enabled according to the cur rent power mode. the number of implemented bits in each mask register, as well as which module clock is controlled by each bit, is shown in table 12-9 . 31 30 29 28 27 26 25 24 mask[31:24] 23 22 21 20 19 18 17 16 mask[23:16] 15 14 13 12 11 10 9 8 mask[15:8] 76543210 mask[7:0] table 12-9. maskable module clocks in at32uc3l0128/256. bit cpumask hsbmask pbamask pbbmask 0 ocd pdca pdca flashcdw 1 - flashcdw intc hmatrix 2- sau pm sau 3 - pbb bridge scif - 4 - pba bridge ast - 5 - peripheral event system wdt - 6- - eic - 7- - freqm - 8- - gpio - 9 - - usart0 - 10 - - usart1 - 11 - - usart2 -
168 32145c?06/2013 at32uc3l0128/256 note that this register is protected by a lock. to write to this register the unlock register has to be written first. please refer to the unlock register description for details. 12 - - usart3 - 13 - - spi - 14 - - twim0 - 15 - - twim1 - 16 systimer - twis0 - 17 - - twis1 - 18 - - pwma - 19 - - tc0 - 20 - - tc1 - 21 - - adcifb - 22 - - acifb - 23 - - cat - 24 - - gloc - 25 - - aw - 31:26---- table 12-9. maskable module clocks in at32uc3l0128/256. bit cpumask hsbmask pbamask pbbmask
169 32145c?06/2013 at32uc3l0128/256 12.7.6 pba divided mask name: pbadivmask access type: read/write offset: 0x040 reset value: 0x0000007f ? mask: clock mask if bit n is written to zero, the clock divided by 2 (n+1) is stopped. if bit n is written to one, the clock divided by 2 (n+1) is enabled according to the current power mode. table 12-10 shows what clocks are affected by the different mask bits. note that this register is protected by a lock. to write to this register the unlock register has to be written first. please refer to the unlock register description for details. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 - ------ 15 14 13 12 11 10 9 8 -------- 76543210 - mask[6:0] table 12-10. divided clock mask bit usart0 usart1 usart2 usart3 tc0 tc1 0----timer_clock2timer_clock2 1---- - - 2 clk_usart/ div clk_usart/ div clk_usart/ div clk_usart/ div timer_clock3 timer_clock3 3---- - - 4----timer_clock4timer_clock4 5---- - - 6----timer_clock5timer_clock5
170 32145c?06/2013 at32uc3l0128/256 12.7.7 clock failure detector control register name: cfdctrl access type: read/write offset: 0x054 reset value: 0x00000000 ? sfv: store final value 0: the register is read/write 1: the register is read-only, to protect against further accidental writes. ? cfden: clock failure detection enable 0: clock failure detector is disabled 1: clock failure detector is enabled note that this register is protected by a lock. to write to this register the unlock register has to be written first. please refer to the unlock register description for details. 31 30 29 28 27 26 25 24 sfv------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 -------cfden
171 32145c?06/2013 at32uc3l0128/256 12.7.8 unlock register name: unlock access type: write-only offset: 0x058 reset value: 0x00000000 to unlock a write protected register, first write to the unlock register with the address of the register to unlock in the addr field and 0xaa in the key field. then, in the next pb access write to the register specified in the addr field. ?key: unlock key write this bit field to 0xaa to enable unlock. ? addr: unlock address write the address of the register to unlock to this field. 31 30 29 28 27 26 25 24 key 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 ------ addr[9:8] 76543210 addr[7:0]
172 32145c?06/2013 at32uc3l0128/256 12.7.9 interrupt enable register name: ier access type: write-only offset: 0x0c0 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will set the corresponding bit in imr. 31 30 29 28 27 26 25 24 ae------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 - - ckrdy - - - - cfd
173 32145c?06/2013 at32uc3l0128/256 12.7.10 interrupt disable register name: idr access type: write-only offset: 0x0c4 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in imr. 31 30 29 28 27 26 25 24 ae------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 - - ckrdy - - - - cfd
174 32145c?06/2013 at32uc3l0128/256 12.7.11 interrupt mask register name: imr access type: read-only offset: 0x0c8 reset value: 0x00000000 0: the corresponding interrupt is disabled. 1: the corresponding interrupt is enabled. this bit is cleared when the corresponding bit in idr is written to one. this bit is set when the corresponding bit in ier is written to one. 31 30 29 28 27 26 25 24 ae------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 - - ckrdy - - - - cfd
175 32145c?06/2013 at32uc3l0128/256 12.7.12 interrupt status register name: isr access type: read-only offset: 0x0cc reset value: 0x00000000 0: the corresponding interrupt is cleared. 1: the corresponding interrupt is pending. this bit is cleared when the corresponding bit in icr is written to one. this bit is set on a zero-to-one transition of the corresponding bit in the status register (sr). 31 30 29 28 27 26 25 24 ae------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 - - ckrdy - - - - cfd
176 32145c?06/2013 at32uc3l0128/256 12.7.13 interrupt clear register name: icr access type: write-only offset: 0x0d0 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in isr. 31 30 29 28 27 26 25 24 ae------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 - - ckrdy - - - - cfd
177 32145c?06/2013 at32uc3l0128/256 12.7.14 status register name: sr access type: read-only offset: 0x0d4 reset value: 0x00000020 ? ae: access error 0: no access error has occurred. 1: a write to lock protected register without unlocking it has occurred. ? ckrdy: clock ready 0: one of the cpusel/pbxsel registers has been writt en, and the new clock setting is not yet effective. 1: the synchronous clocks have frequencies as indicated in the cpusel/pbxsel registers. ? cfd: clock failure detected 0: main clock is running correctly. 1: failure on main clock detected. main clock is now running on rcsys. 31 30 29 28 27 26 25 24 ae------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 - - ckrdy - - - - cfd
178 32145c?06/2013 at32uc3l0128/256 12.7.15 peripheral power control register name : ppcr access type: read/write offset: 0x004 reset value: 0x000001fa ? rsttm: reset test mode 0: external reset not in test mode 1: external reset in test mode ? frc32: force rc32 out 0: rc32 signal is not forced as output 1: rc32 signal is forced as output ? rstpun: reset pull-up, active low 0: pull-up for external reset on 1: pull-up for external reset off 31 30 29 28 27 26 25 24 ppc[31:24] 23 22 21 20 19 18 17 16 ppc[23:16] 15 14 13 12 11 10 9 8 ppc[15:8] 76543210 ppc[7:0] table 12-11. peripheral power control bit name 0rstpun 1frc32 2rsttm 3 catrcmask 4 acifbcrcmask 5 adcifbrcmask 6 astrcmask 7 twis0rcmask 8 twis1rcmask 31:9 -
179 32145c?06/2013 at32uc3l0128/256 ? catrcmask: cat request clock mask 0: cat request clock is disabled 1: cat request clock is enabled ? acifbrcmask: acifb request clock mask 0: acifb request clock is disabled 1: acifb request clock is enabled ? adcifbrcmask: adcifb request clock mask 0: adcifb request clock is disabled 1: adcifb request clock is enabled ? astrcmask: ast request clock mask 0: ast request clock is disabled 1: ast request clock is enabled ? twis0rcmask: twis0 request clock mask 0: twis0 request clock is disabled 1: twis0 request clock is enabled ? twis1rcmask: twis1 request clock mask 0: twis1 request clock is disabled 1: twis1 request clock is enabled note that this register is protected by a lock. to write to this register the unlock register has to be written first. please refer to the unlock register description for details.
180 32145c?06/2013 at32uc3l0128/256 12.7.16 reset cause register name: rcause access type: read-only offset: 0x180 reset value: latest reset source ? awire: awire reset this bit is set when the last reset was caused by the awire. ? jtag: jtag reset this bit is set when the last reset was caused by the jtag. ? ocdrst: ocd reset this bit is set when the last reset was due to the res bit in the ocd development control register having been written to one. ? sleep: sleep reset this bit is set when the last reset was due to the device waking up from the shutdown sleep mode. ? wdt: watchdog reset this bit is set when the last reset was due to a watchdog time-out. ? ext: external reset pin this bit is set when the last reset was due to the reset_n pin being pulled low. ? bod: brown-out reset this bit is set when the last reset was due to the core supply voltage being lower than the brown-out threshold level. ? por: power-on reset this bit is set when the last reset was due to the core supp ly voltage vddcore being lower than the power-on threshold level (the reset is generated by the por18 detector), or the inter nal regulator supply voltage being lower than the regulator power-o n threshold level (generated by the por33 detector), or the internal regulator supply voltage being lower than the minimum required input voltage (generated by the 3.3v supply monitor sm33). 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 ---awire- jtagocdrst 76543210 - sleep - - wdt ext bod por
181 32145c?06/2013 at32uc3l0128/256 12.7.17 wake cause register name: wcause access type: read-only offset: 0x184 reset value: latest wake source a bit in this register is set on wake up caused by the peripheral referred to in table 12-12 on page 181 . 31 30 29 28 27 26 25 24 wcause[31:24] 23 22 21 20 19 18 17 16 wcause[23:16] 15 14 13 12 11 10 9 8 wcause[15:8] 76543210 wcause[7:0] table 12-12. wake cause bit wake cause 0cat 1acifb 2 adcifb 3twi slave 0 4twi slave 1 5 wake_n 6 adcifb pen detect 15:7 - 16 eic 17 ast 31:18 -
182 32145c?06/2013 at32uc3l0128/256 12.7.18 asynchronous wake up enable register name :awen access type: read/write offset: 0x188 reset value: 0x00000000 each bit in this register corresponds to an asynchronous wake-up source, according to table 12-13 on page 182 . 0: the corresponding wake-up source is disabled. 1: the corresponding wake-up source is enabled 31 30 29 28 27 26 25 24 awen[31:24] 23 22 21 20 19 18 17 16 awen[23:16] 15 14 13 12 11 10 9 8 awen[15:8] 76543210 awen[7:0] table 12-13. asynchronous wake-up sources bit asynchronous wake-up source 0cat 1acifb 2 adcifb 3twis0 4twis1 5 waken 6 adcifbpd 31:7 -
183 32145c?06/2013 at32uc3l0128/256 12.7.19 configuration register name: config access type: read-only offset: 0x3f8 reset value: - this register shows the configuration of the pm. ? hsbpevc:hsb pevc clock implemented 0: hsbpevc not implemented. 1: hsbpevc implemented. ? pbd: pbd implemented 0: pbd not implemented. 1: pbd implemented. ? pbc: pbc implemented 0: pbc not implemented. 1: pbc implemented. ? pbb: pbb implemented 0: pbb not implemented. 1: pbb implemented. ? pba: pba implemented 0: pba not implemented. 1: pba implemented. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 hsbpevc - - - pbd pbc pbb pba
184 32145c?06/2013 at32uc3l0128/256 12.7.20 version register name: version access type: read-only offset: 0x3fc reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
185 32145c?06/2013 at32uc3l0128/256 12.8 module configuration the specific configuration for ea ch pm instance is listed in t he following tables. the module bus clocks listed here are connected to the system bus clocks. please refer to the ?synchronous clocks?, ?peripheral clock masking? and ?s leep modes? sections for details. table 12-14. power manager clocks clock name description clk_pm clock for the pm bus interface table 12-15. register reset values register reset value version 0x00000420 table 12-16. effect of the different reset events power-on reset external reset watchdog reset bod reset sm33 reset cpu error reset ocd reset jtag reset cpu/hsb/pbx (excluding power manager) yyyyyyyy 32khz oscillator y n n n n n n n rc oscillator calibration register y n n n n n n n other oscillator control registers y y y y y y y y ast registers, except interrupt registers ynnnnnnn watchdog control register y y n y y y y y voltage calibration register y n n n n n n n sm33 control register y y y y y y y y bod control register y y y n y y y y clock control registers y y y y y y y y ocd system a nd ocd registers y y n y y y n y
186 32145c?06/2013 at32uc3l0128/256 13. system control interface (scif) rev: 1.1.0.0 13.1 features ? supports crystal oscill ator 0.45-16mhz (osc0) ? supports digital frequency locked loop 20-150mhz (dfll) ? supports phase locked loop 80-240mhz (pll) ? supports 32khz ultra-low-power oscillator (osc32k) ? supports 32khz rc oscillator (rc32k) ? integrated low-powe r rc oscillator (rcsys) ? generic clocks (gclk) with wide frequency range provided ? generic clock prescaler ? controls bandgap ? controls brown-out detectors (bod) and supply monitors ? controls voltage regulator (vreg) behavior and calibration ? controls temperature sensor ? controls supply monito r 33 (sm33) operating modes and calibration ? controls 120mhz integrated rc oscillator (rc120m) ? four 32-bit general-purpose backup registers 13.2 overview the system control interface (scif) controls the oscillators, generic clocks, bods, bandgap, vreg, temperature sensor, and backup registers. 13.3 i/o lines description 13.4 product dependencies in order to use this module, other parts of the system must be configured correctly, as described below. table 13-1. i/o lines description pin name pin description type rc32out rc32 output at startup output xin0 crystal 0 input analog/digital xin32 crystal 32 input (primary location) analog/digital xin32_2 crystal 32 input (sec ondary location) analog/digital xout0 crystal 0 output analog xout32 crystal 32 output (primary location) analog xout32_2 crystal 32 output (secondary location) analog gclk9-gclk0 generic clock output output gclk_in2-gclk_in0 generic clock input input
187 32145c?06/2013 at32uc3l0128/256 13.4.1 i/o lines the scif provides a number of generic clock outputs, which can be connected to output pins, multiplexed with gpio lines. the programmer must first program the gpio controller to assign these pins to their peripheral function. if the i/o pins of the scif are not used by the application, they can be used for ot her purposes by the gpio controller. oscillator pins are also multiplexed with gpio. when osc illators are used, the related pins are controlled directly by the scif, over- riding gpio settings. rc32out will be output after reset, and the gpio controller can assi gn this pin to other periph- eral function after start-up. 13.4.2 power management the bods and all the oscillators , except the 32khz oscillator (o sc32k) are turned off in some sleep modes and turned automatically on when the device wakes up. the voltage regulator is set in low power mode in some sleep modes and automatically set back in normal mode when the device wakes up. please refer to the power manager chapter for details. the bod control registers will not be reset by the po wer manager on a bod reset. 13.4.3 clocks the scif controls all o scillators in the device. the oscillators can be used as source for the cpu and peripherals. selection of sour ce is done in the power manage r. the oscillators can also be used as source for generic clocks. 13.4.4 interrupts the scif interrupt request line is connected to the interrupt controller. using the scif interrupt requires the interrupt controller to be programmed first. 13.4.5 debug operation the scif does not interact with debug operations. 13.5 functional description 13.5.1 oscillator (osc) operation rev: 1.1.1.0 the main oscillator (oscn) is de signed to be used wit h an external 0.450 to 16mhz crystal and two biasing capacitors, as shown in the electrical characteristics chapter, or with an external clock connected to t he xin. the oscillator can be used as source for the main clock in the device, as described in the powe r manager chapte r. the oscillator can be used as source for the generic clocks, as described in the generic clocks section. the oscillator is disabled by de fault after reset. wh en the oscillator is disabled, the xin and xout pins can be used as gener al purpose i/os. when the os cillator is enabl ed, the xin and xout pins are controlled directly by the scif , overriding gpio settings . when the oscillator is configured to use an external clock, the clock must be applied to the xin pin while the xout pin can be used as general purpose i/o. the oscillator is enabled by writin g a one to the oscillato r enable bit in the oscillator control reg- ister (oscctrln.oscen). operation mode (external clock or crystal) is selected by writing to the oscillator mode bit in oscctrln (oscctrln. mode). the oscillator is automatically dis-
188 32145c?06/2013 at32uc3l0128/256 abled in certain sleep modes to reduce power consumption, as described in the power manager chapter. after a hard reset, or wh en waking up from a sleep mode w here the oscillators were disabled, the oscillator will need a certain amount of time to stabilize on the correct frequen cy. this start- up time can be set in the oscctrln register. the scif masks the oscilla tor outputs during the st art-up time, to ensure that no unstable clocks propagate to the digital logic. the oscn ready bit in the power and clock status register (pclksr.oscnrdy) is set when the oscillator is stable and ready to be used as clock source. an interrupt can be generated on a zero-to-one transition on oscnrdy if the oscnrdy bit in the interrupt mask register (imr.oscnrdy) is set. this bit is set by writing a one to the corresponding bit in the interrupt enable register (ier.oscnrdy). 13.5.2 32khz oscillator (osc32k) operation rev: 1.1.0.1 the 32khz oscillator operates as described for the oscillator above. the 32khz oscillator can be used as source clock for the asynchronous timer (ast) and the watchdog timer (wdt). the 32khz oscillator can also be used as source for th e generic clocks. the oscillator is disabled by default after reset. when the os cillator is disa bled, the xin32 and xout32 pins can be used as general-purpose i/os. when the oscillator is enabled, the xin32 and xout32 pins are controlled directly by the scif, overriding gpio settings. when the oscil- lator is configured to use an external clock, the clock must be applied to the xin32 pin while the xout32 pin can be used as general-purpose i/o. the oscillator is enabled writin g a one to the osc32 enable bit in the 32khz oscillator control register (oscctrl32osc32en). th e oscillator is disabled by wr iting a zero to the osc32en bit, while keeping the other bits unchanged. writ ing to osc32en while also writing to other bits may result in unpredictable behavior. operation mode (external clock or crystal) is selected by writing to the oscillator mode bit in oscctrl32 (oscctrl32 .mode). the oscillator is an ultra-low-power design and rema ins enabled in all sleep modes. the start-up time of the 32khz oscillator is selected by writing to the oscillator start-up time field in the oscctrl32 regist er (oscctrl32.startup). the scif masks the oscillator out- put during the start-up time, to ensure that no uns table clock cycles propagate to the digital logic. the osc32 ready bit in the power and clock status register (pclksr.osc32rdy) is set when the oscillator is stable and ready to be used as clock so urce. an interrupt can be gener- ated on a zero-to-one transition on pclksr.osc 32rdy if the osc32rdy bit in the interrupt mask register (imr.osc32rdy) is set. this bit is set by writing a one to the corresponding bit in the interrupt enable register (ier.osc32rdy). .as a crystal oscillator usually r equires a very long start-up ti me (up to 1 second), the 32khz oscillator will keep running across re sets, except a power-on reset (por). the 32khz oscillator also has a 1khz output. this is enabled by writing a one to the enable 1khz output bit in oscctrl32 register (osc ctrl32.en1k). if the 32khz output clock is not needed when 1k is enabled, this can be disabled by writing a zero to the enable 32khz output bit in the oscctrl32 register (oscctrl32.en32k). oscctrl32 .en32k is set after a por. the 32khz oscillator has two possible sets of pins. to select between them write to the pin select bit in the oscctrl32 register (oscctrl32 .pinsel). if the 32khz oscillator is to be
189 32145c?06/2013 at32uc3l0128/256 used in shutdown mode, pinsel must be written to one, and xin32_2 and xout32_2 must be used. 13.5.3 pll operation rev: 1.1.0.0 the device contains one phase locked loop (pll), which is controlled by the phase locked loop interface (pllif). the pll is disabled by default, but can be enabled to provide high fre- quency source clocks for synchronous or gener ic clocks. the pll ca n use different clock sources as reference clock, please refer to th e ?pll clock sources? table in the scif module configuration section for details. the pll output is divided by a multiplication factor, and the pll compares the phase of the re sulting clock to the reference cl ock. the pll will adjust its out- put frequency until the two compar ed clocks phases are equal, thus locking the output frequency to a multiple of the reference clock frequency. when the pll is switched on, or when changing the clock source or multiplication factor for the pll, the pll is unlocked and the output frequency is undefined. the pll clock for the digital logic is automatically masked when the pll is unlocked, to prevent the connected digital logic from receiving a too high frequency and thus become unstable. the pll can be configured by writing the pll control register (plln). to prevent unexpected writes due to software bugs, write access to th e plln register is protected by a locking mecha- nism, for details please refer to the unlock register description. figure 13-1. pll with control logic and filters 13.5.3.1 enabling the pll before the pll is enabled it must be set up correctly. the pll oscillator select field ( pllosc) selects a source for the reference clock. the pll multiply factor (pllmul) and pll division phase detector output divider source clocks pllosc pllopt[0] pllmul lock bit mask pll clock input divider plldiv 1/2 pllopt[1] 0 1 vco f vco f pll lock counter f ref
190 32145c?06/2013 at32uc3l0128/256 factor (plldiv) fields must be written with the multiplication and division factors, respectively. the pllmul must always be greater than 1, creating the pll frequency: f vco = (pllmul+1)/plldiv ? f ref , if plldiv >0 f vco = 2?(pllmul+1) ? f ref , if plldiv = 0 the pll options (pllopt) field should be configured to proper values according to the pll operating frequency. the pllopt field can also be configured to divide the output frequency of the pll by 2 and wide-bandwidth mode, which allows faster startup time and out-of-lock time. it is not possible to change any of the pll configuration bits when the pll is enabled, any write to plln while the pll is enabled will be discarded. after setting up the pll, the pll is enabled by writing a one to the pll enable (pllen) bit in the plln register. 13.5.3.2 disabling the pll the pll is disabled by writing a zero to the p ll enable (pllen) bit in the plln register. after disabling the pll, the pll configuration fields becomes writable. 13.5.3.3 pll lock the lock signal for each pll is available as a plllockn flag in the pclksr register. if the lock for some reason is lost, the plllocklostn flag in pclksr register will be set. an interrupt can be generated on a 0 to 1 transition of these bits. 13.5.4 digital frequency locked loop (dfll) operation rev: 2.1.0.1 the dfll is controlled by the digital frequency locked loop interface (dfllif). the dfll is disabled by default, but can be enabled to prov ide a high-frequency source clock for synchro- nous and generic clocks. features: ? internal oscillator with no external components ? 20-150mhz frequency in closed loop mode ? can operate standalone as a high-frequency programmable oscillator in open loop mode ? can operate as an accurate frequency multiplier against a known frequency in closed loop mode ? optional spread-spectrum clock generation ? very high-frequency multiplication supported - can generate all frequencies from a 32khz clock the dfll can operate in both open loop mode and closed loop mode. in closed loop mode a low frequency clock with high accuracy can be used as reference clock to get high accuracy on the output clock (clk_dfll). to prevent unexpected writes due to software bugs , write access to the configuration registers is protected by a locking mechanism. for details please refer to the unlock register description.
191 32145c?06/2013 at32uc3l0128/256 figure 13-2. dfllif block diagram 13.5.4.1 enabling the dfll the dfll is enabled by writing a one to the enable bit (en) in the dflln configuration register (dfllnconf). no other bits or fields in dfllnconf must be changed simultaneously, or before the dfll is enabled. 13.5.4.2 internal synchronization due to multiple clock domains in the dfllif, values in the dfllif configuration registers need to be synchronized to other clock domains. the status of this synchronization can be read from the power and clocks status register (pclksr). before writing to a dfllif configuration reg- ister, the user must check that the dflln synchronization ready bit (dfllnrdy) in pclksr is set. when this bit is set, the dfll can be configured, and clk_dfll is ready to be used. any write to a dfllif configuration register while dfllnrdy is cleared will be ignored. before reading the value in any of the dfll configuration registers a one must be written to the synchronization bit (sync) in the dflln sync hronization register (dfllnsync). the dfll configuration registers are ready to be read when pclksr.dfllnrdy is set. 13.5.4.3 disabling the dfll the dfll is disabled by writing a zero to dfl lnconf.en. no other bits or fields in dflln- conf must be changed simultaneously. after disabling the dfll, pclksr.dfllnrdy will no t be set. it is not required to wait for pclksr.dfllnrdy to be set before re-enabling the dfll. 13.5.4.4 open loop operation after enabling the dfll, open loop mode is selected by writing a zero to the mode selection bit (mode) in dfllnconf. when operating in open loop mode the output frequency of the dfll will be determined by the values wr itten to the coarse calibration value field (coarse) and the fine calibration value field (fine) in the df llnconf register. when writing to coarse and dfll coarse fine 8 9 c l k _ d f l l imul fmul 32 clk_dfllif_ref frequency tuner dflllockc dflllocklostc dflllockf dflllocklostf dflllocka dflllocklosta cstep fstep 8+9 clk_dfllif_dither
192 32145c?06/2013 at32uc3l0128/256 fine, be aware that the output frequency must not exceed the maximum frequency of the device after the division in the clock generator. it is possible to change the value of coarse and fine, and thereby the output frequency of th e dfll, while the dfll is enabled and in use. the dfll clock is ready to be used when pclksr.dfllnrdy is cleared after enabling the dfll. the frequency range in open loop mode is 20-150mhz, but maximum frequency can be higher, and the minimum frequency can be lower. the best way to start the dfll at a specific frequency in open loop mode is to first configure it for closed loop mode, see section 13.5.4.5 . when a lock is achieved, read back the coarse and fine values and switch to open loop mode using these values. an alternative approach is to use the frequency meter (freqm) to monitor the dfll frequency and adjust the coarse and fine values based on measurement results form the freqm. please refer to the freqm chapter for more information on how to use it. note that the output frequency of the dfll will drift when in open loop mo de due to temper ature and voltage changes. please refer to the electrical characteristics chapter for details. 13.5.4.5 closed loop operation the dfll must be correctly configured before closed loop operation can be enabled. after enabling the dfll, enable and select a reference clock (clk_dfllif_ref). clk_dfllif_ref is a generic clock, please refer to generic clocks section for details. then set the maximum step size allowed in finding the coarse and fine values by setting the coarse maximum step field (cstep) and fine maximum step field (fstep) in the dflln max- imum step register (dfllnstep) . a small step size will ensure low overshoot on the output frequency, but can typically result in longer lock times. a high value might give a big overshoot, but can typically give faster locking. dfllnstep.cstep and dfllnstep.fstep must be lower than 50% of the maximum value of dfllnconf.coarse and dfllnconf.fine respectively. then select the multiplication factor in the integer multiply factor field (imul) and the fractional multiply field (fmu l) in the dflln mult iplier register (dfllnmul). care must be taken when choosing imul and fmul so the output frequency does not exceed the maximum frequency of the device. start the closed loop mode by writing a one to dfllnconf.mode bit. the frequency of clk_dfll (f dfll ) is given by: where f ref is the frequency of clk_dfllif_ref. coarse and fine in dfllnconf are read- only in closed loop mode, and are controlled by the dfllif to meet user specified frequency. the values in coarse when the closed loop mode is enabled is used by the frequency tuner as a starting point for coarse. se tting coarse to a value close to the final value will reduce the time needed to get a lock on coarse. frequency locking the locking of the frequency in closed loop mode is divided into three stages. in the coarse stage the control logic quickly finds the co rrect value for dfllnconf.coarse and thereby sets the output frequency to a value close to the correct frequency. the dflln locked on coarse value bit (dfllnlockc) in pclksr will be set when this is done. in the fine stage the control logic tunes the value in dfllnconf.fine so the out put frequency will be very close to the desired frequency. dflln locked on fine value bit (dfllnlockf) in pclksr will be set when this is done. in the accurate stage the dfll frequency tuning mechanism uses dither- ing on the fine bits to obtain an accurate average output frequency. dflln locked on accurate value bit (dfllnlocka) in pclksr will be se t when this is done. the accurate stage will f dfll imul fmul 2 16 ---------------- - + ?? ?? f ref =
193 32145c?06/2013 at32uc3l0128/256 only be executed if the dithering enable bit (dither) in dfllnconf has been written to a one. if dither is written to a zero dfllnlocka will never occur. if dithering is enabled, the fre- quency of the dithering is decided by a generic clock (clk_dfllif_dither). this clock has to be set up correctly before enabling dithering. pleas e refer to the generic clocks section for details. figure 13-3. dfll closed loop state diagram when dithering is enabled the accuracy of the average output frequency of the dfll will be higher. however, the actual frequency will be al ternating between two frequencies. if a fixed fre- quency is required, the dithering should not be enabled. figure 13-4. dfll locking in closed loop clk_dfll is ready to be used when the df lln synchronization ready bit (dfllnrdy) in pclksr is set after enabling the dfll. however, the accuracy of the output frequency depends on which locks are set. for lock times, please refer to the electrical characteristics chapter. measure f dflln calculate new coarse value dfllnlockc 0 calculate new fine value dfllnlockf 0 1 dfllnlocka 1 calculate new dithering dutycycle 0 compen- sate for drift 1 dither 1 compen- sate for drift 0 initial frequency target frequency dfllnlockc dfllnlockf dfllnlocka
194 32145c?06/2013 at32uc3l0128/256 drift compensation the frequency tuner will automatically compensate for drift in the f dfll without losing either of the locks. if the fine value overflows or underflow s, which should normally not happen, but could occur due to large drift in temp erature and voltage, all locks wi ll be lost, and the coarse and fine values will be recalibrated as described earl ier. if any lock is lost the corresponding bit in pclksr will be set, dflln lock lost on coarse value bit (dfllnlocklostc) for lock lost on coarse value, dflln lock lost on fine value bit (dfllnlocklostf) for lock lost on fine value and dflln lock lost on accurate value bit (dfllnlocklosta) for lock lost on accu- rate value. the corresponding lock status bit will be cleared when the lock lost bit is set, and vice versa. reference clock stop detection if clk_dfllif_ref stops or is running at a very slow frequency, the dflln reference clock stopped bit (dfllnrcs) in pclksr will be set. note that the detection of the clock stop will take a long time. the dfllif operate as if it was in open loop mode if it detects that the refer- ence clock has stopped. this means that the coarse and fine va lues will be kept constant while pclksr.dfllnrcs is set. closed loop mo de operation will automati cally resume if the clk_dfllif_ref is restarted, and compensate for any drift during the time clk_dfllif_ref was stopped. no locks will be lost. frequency error measurement the ratio between clk_dfllif_ref and clk_dfll is measured automatically by the dfllif. the difference between this ratio and dfllnmul is stored in the multiplication ratio difference field (ratiodiff) in the dflln ratio register (dfllnratio). the relative error on clk_dfll compared to the target frequency can be calculated as follows: where is the number of reference clock cycl es the dfllif is using for calculating the ratio. 13.5.4.6 dealing with delay in the dfll the time from selecting a new frequency until this frequency is output by the dfll, can be up to several micro seconds. if the difference between the desired output frequency (clk_dfll) and the frequency of clk_dfllif_ref is small this can lead to an in stability in the dfllif locking mechanism, which can prevent the dfllif from ac hieving locks. to avoid this, a chill cycle where the clk_dfll frequency is not measured can be enabled. the ch ill cycle is enabled by writing a one to the chill cycle enable (ccen) bit in the dfllnconf register. enabling chill cycles might double the lock time, another solution to the same problem can be to use less strict lock requirements. this is called quick lock (ql), which is enabled by writing a one to the quick lock enable (qlen) bit in the dfllnconf register. the ql might lead to bigger spread in the outputted frequency than chill cycles, but the average output frequency is the same. if the target frequency is below 40mhz, one of these methods should always be used. 13.5.4.7 spread spectrum generator (ssg) when the dfll is used as the main clock source for the device, the emi radiated from the device will be synchronous to f dfll . to provide better electromagnetic compatibility (emc) the error ratiodiff f ref ? 2 numref f dfll ? ------------------------------------------------ - = 2 numref
195 32145c?06/2013 at32uc3l0128/256 dfllif can provide a clock with the energy spread in the frequency domain. this is done by adding or subtracting values from the fine value. ssg is enabled by writing a one to the enable bit (en) in the dflln spread spectrum generator control register ( dfllnssg). a generic clock sets the rate at which the ssg changes the frequency of the dfll clock to gen- erate a spread spectrum (clk_dfllif_dither). this is the same clock used by the dithering mechanism. the frequency of this clock should be higher than f ref to ensure that the dfllif can lock. please refer to the ge neric clocks section for details. optionally, the clock ticks can be qualified by a pseudo random binary sequence (prbs) if the prbs bit in dfllnssg is one. this reduces the modulation effect of clk_dfllif_dither fre- quency onto f dfll . the amplitude of the frequency variation can be selected by setting the ssg amplitude field (amplitude) in dfllnssg. if amplitude is ze ro the ssg will toggle on the lsb of the fine value. if amplitude is one the ssg will ad d the sequence {1,- 1, 0} to fine. the step size of the ssg is selected by writing to the ssg step size field (stepsize) in dfllnssg. stepsize equal to zero or one will result in a step size equal to one. if the step size is set to n, the output value from the ssg will be incremented/decremented by n on every tick of the source clock. the spread spectrum generator is available in both open and closed loop mode. when spread spectrum is enabled in closed l oop mode, and the amplitude is high, an over- flow/underflow in fine is more likely to occur. figure 13-5. spread spectrum generator block diagram. 13.5.4.8 wake from sleep modes the dfllif may optionally reset it s lock bits when waking from a sleep mode which disables the dfll. this is configured by the lose lock af ter wake (llaw) bit in dfllnconf register. if dfllnconf.llaw is written to zero the dfll will be re-enabled and start running with the same configuration as before going to sleep even if the reference clock is not available. the locks will not be lost. when the refe rence clock has restar ted, the fine tracki ng will quickly com- pensate for any frequency drift during sleep. if a one is written to dfllnconf.llaw before going to a sleep mode where the dfll is turned off, the dfllif will lose all its locks when wak- ing up, and needs to regain these through the full lock sequence. 13.5.4.9 accuracy there are mainly three factors that decide the accuracy of the f dfll . these can be tuned to obtain maximum accuracy wh en fine lock is achieved. pseudorandom binary sequence spread spectrum generator fine 9 to dfll clk_dfllif_dither amplitude, stepsize prbs 1 0
196 32145c?06/2013 at32uc3l0128/256 ? fine resolution: the frequency step between two fine values. this is relatively smaller for high output frequencies. ? resolution of the measurement: if the resolution of the measured f dfll is low, i.e. the ratio between clk_dfll frequency and clk_dfllif_ref is small, then the dfllif might lock at a frequency that is lower than the targeted frequency. it is recommended to use a reference clock frequency of 32 khz or lower to avoid this issue for low target frequencies. ? the accuracy of the reference clock. 13.5.4.10 interrupts a interrupt can be generated on a zero-to-one transaction on dfllnlockc, dfllnlockf, dfllnlocka, dfllnlocklostc, dfllnlock lostf, dfllnlocklosta, dfllnrdy or dfllnrcs. 13.5.5 brown-out detection (bod) rev: 1.2.0.0 the brown-out detector monitors the vddcore supply pin and compares the supply voltage to the brown-out detection level. the bod is disabled by default, and is enabled by writing to the bod control field in the bod control register (bod.ctrl). this field can also be updated by flash fuses. the bod is powered by vddi o and will not be powered du ring shutdown sleep mode. to prevent unexpected writes to the bod register due to software bugs, write access to this reg- ister is protected by a locking mechanism. for details please refer to the unlock register description. to prevent further modifications by software, the content of the bod register can be set as read- only by writing a one to the store final value bit (bod.sfv). when this bit is one, software can not change the bod register content. this bit is cleared after flash calibration and after a reset except after a bod reset. the brown-out detection level is selected by writing to the bod level field in bod (bod.level). please refer to the electrical characteristics chapter for parametric details. if the bod is enabled (bod.ctrl is one or two) and the supply voltage goes below the detec- tion level, the brown-out detection bit in the power and clocks status register (pclksr.boddet) is set. this bit is cleared when the supply voltage goes above the detection level. an interrupt re quest will be generated on a zero-to-one tr ansition on pclksr.boddet if the brown-out detection bit in the interrupt mask register (imr.boddet) is set. this bit is set by writing a one to the corresponding bit in the interrupt enable register (ier.boddet). if bod.ctrl is one, a bod reset will be generated when the supply voltage goes below the detection level. if bo d.ctrl is two, the device will not be reset. writing a one to the bod hysteresis bit in bo d (bod.hyst) will add a hysteresis on the bod detection level. note that the bod must be disabled before ch anging bod.level, to avoid spurious reset or interrupt. after enabling the bo d, the bod output will be mask ed during one half of a rcsys clock cycle and two main clocks cycles to avoid false results. when the jtag or awire is enabled, the bod reset and interrupt are masked.
197 32145c?06/2013 at32uc3l0128/256 the ctrl, hyst, and level fields in the bod control register are loaded factory defined cal- ibration values from flash fuses after a reset. if the flash calibration done bit in the bod control register (bod.fcd) is zero, the flash calibra tion will be redone a fter any reset, and the bod.fcd bit will be set before program execution starts in the cpu. if bod.fcd is one, the flash calibration is redone after any reset except for a bod reset. the bod.fcd bit is cleared after a reset, except for a bod reset. bod.fcd is set when these fields have been updated after a flash calibration. it is possible to overri de the values in the bod.ctrl, bod.hyst, and bod.level fields after reset by writing to the bod control regi ster. please refer to the fuse settings chapter for more details about bod fuses and how to program the fuses. figure 13-6. bod block diagram 13.5.6 bandgap rev: 1.2.0.0 the flash memory, the bod, and the temperature sensor need a stable voltage reference to operate. this reference voltage is provided by an internal bandgap voltage reference. this refer- ence is automatically turned on at start-up and turned off during some sleep modes to save power. the bandgap reference is powered by the internal regulator supply voltage and will not be powered during shutdown sleep mode. please refer to the power manager chapter for details. vddcore por18 bod scif power manager(pm) intc reset bod detected enable b o d h y s t bod level reset i n t e r r u p t
198 32145c?06/2013 at32uc3l0128/256 13.5.7 system rc oscillator (rcsys) rev: 1.1.1.0 the system rc oscillator has a startup time of th ree cycles, and is alwa ys available except in some sleep modes. please refer to the power manager chapter for details. the system rc oscil- lator operates at a nominal frequency of 115khz, and is calibrated using the calibration value field (calib) in the rc oscillat or calibration register (rccr). after a power-on reset (por), the rccr.calib field is loaded with a factory defined value stored in the flash fuses. please refer to the fuse setting chapt er for more details about rccr fuses and how to program the fuses. if the flash calibration done (fcd) bit in the rccr is zero at any reset, the flash calibration will be redone and the rccr.fcd bit will be set before program executi on starts in the cpu. if the rccr.fcd is one, the flash calibration w ill only be redone afte r a power-on reset. to prevent unexpected writes to rccr due to software bugs, write access to this register is pro- tected by a locking mechanism. for details please refer to the unlock register description. although it is not reco mmended to override default factory settings, it is still possible to override the default values by writing to rccr.calib. 13.5.8 voltage regulator (vreg) rev: 1.1.0.0 the embedded voltage regulator can be used to provide the vddcore voltage from the inter- nal regulator supply vo ltage. it is controlled by the vo ltage regulator calibration register (vregcr). the voltage regulator is enabled by def ault at start-up but can be disabled by soft- ware if an external voltage is provided on th e vddcore pin. the vregcr also contains bits to control the por18 detector and the por33 detector. 13.5.8.1 register protection to prevent unexpected writes to vregcr due to software bugs, write access to this register is protected by a locking mechanism. for details please refer to the unlock register description. to prevent further modifications by software, the content of the vregcr register can be set as read-only by writing a one to the store final value bit (vregcr.sfv). once this bit is set, soft- ware can not change the vregcr content until a power-on reset (por) is applied. 13.5.8.2 controlling volt age regulator output the voltage regulator is always enabled at start-up, i.e. after a por or when waking up from shutdown mode. it can be disa bled by software by writing a zero to the enable bit (vregcr.en). this bit is set after a por. because of internal synchronization, the voltage reg- ulator is not immediately enabled or disabled. the actual state of the voltage regulator can be read from the on bit (vregcr.on). the voltage regulator output level is controlled by the select vdd field (selvdd) in vregcr. the default value of this field corresponds to a regulator output voltage of 1.8v. other values of this field are not defined, and it is not recommended to change the value of this field. the voltage regulator ok bit (vregcr.vregok) bit indicates when the voltage regulator out- put has reached the voltage threshold level.
199 32145c?06/2013 at32uc3l0128/256 13.5.8.3 factory calibration after a power-on reset (por) the vregcr.calib field is loaded with a factory defined calibra- tion value. this value is chosen so that the normal output voltage of the regulator after a power- up is 1.8v. although it is not reco mmended to override default factory settings, it is still possible to override these default values by writing to vregcr.calib. if the flash calibration done bit in vregcr (vregcr.fcd) is zero, the flash calibration will be redone after any reset, a nd the vregcr.fcd bit will be set be fore program exec ution starts in the cpu. if vregcr.fcd is one, the flas h calibration will only be redone after a por. 13.5.8.4 por33 control vregcr includes control bits for the power-on reset 3.3v (por33) detector that monitors the internal regulator supply voltage. the por33 dete ctor is enabled by default but can be disabled by software to reduce power consumption. the 3.3v supply monitor (sm33) can then be used to monitor the regulator power supply. the por33 detector is disabled by writing a zero to the por33 enable bit (vregcr.por33en). because of in ternal synchronisation, the por33 detector is not immedi- ately enabled or disabled. the actual state of the por33 detector can be read from the por33 status bit (vregcr.por33status). the 32khz rc oscillator (rc32k) must be enabl ed before disabling th e por33 detector. once the por33 detector has been di sabled, the rc32k oscillato r can be disabled again. to avoid spurious resets, it is mandatory to mask the power-on reset when enabling or dis- abling the por33 detector. the power-on reset generated by the por33 detector can be ignored by writing a one to the por33 mask bit (vregcr.por33mask). because of internal synchronization, the masking is not immediately effective, so software should wait for the vregcr.por33mask to read as a one before assuming the maski ng is effective. the output of the por33 detector is zero if the internal regulator supply voltage is below the por33 power-on threshold level, and one if the internal regulator supply voltage is above the por33 power-on threshold level. this output (before masking) can be read from the por33 value bit (vregcr.por33value). 13.5.8.5 por18 control vregcr includes control bits for the power-on reset 1.8v (por18) detector that monitors the vddcore voltage. the por18 detector is enabled by default but can be disabled by software to reduce power consumption. the por18 detector is disabled by writing a zero to the por18 enable bit (vregcr.por18en). because of in ternal synchronization, the por18 detector is not immedi- ately enabled or disabled. the actual state of the por18 detector can be read from the por18 status bit (vregcr.por18status). please note that the por18 detector cannot be disabled while the jtag or awire debug inter- face is used. writing a zero to vr egcr.por18en bit will have no effect. to avoid spurious resets, it is mandatory to mask the power-on reset when enabling or dis- abling the por18 detector. the power-on reset generated by the por18 detector can be ignored by writing a one to the por18 mask bit (vregcr.por18mask). because of internal
200 32145c?06/2013 at32uc3l0128/256 synchronisation, the masking is not immediately effective, so software should wait for the vregcr.por18mask to read as one befo re assuming the ma sking is effective. the output of the por18 detector is zero if the vddcore voltage is below the por18 power- on threshold level, and one if the vddcore vo ltage is above the por18 power-on threshold level. the output of the por18 detector (before masking) can be read from the por18 value bit (vregcr.por18value). 13.5.9 3.3 v supply monitor (sm33) rev: 1.1.0.0 the 3.3v supply monitor is a specific voltage detector for the internal regulator supply voltage. it will indicate if the internal re gulator supply voltage is above the minimum requi red input voltage threshold. the user can choose to generate either a power-on reset (por) and an interrupt request, or only an interrupt request, when the internal regulator supply voltage drops below this threshold. please refer to the electrical characteristics chapter for parametric details. 13.5.9.1 register protection to prevent unexpected writes to sm33 register due to software bugs, write access to this regis- ter is protected by a locking mechanism. for details please refer to the unlock register description. to prevent further modifications by software, the content of the register can be set as read-only by writing a one to the store final value bit (sm33.sfv). when this bit is one, software can not change the sm33 register content until the device is reset. 13.5.9.2 operating modes the sm33 is disabled by default and is enabled by writing to the supply monitor control field in the sm33 control register (sm33.ctrl). the current state of the sm33 can be read from the supply monitor on indicator bit in sm33 (sm33. onsm). enabling the sm33 will disable the por33 detector. the sm33 can operate in continuous mode or in sampling mode. in sampling mode, the sm33 is periodically enabled for a short period of time, just enough to make a a measurement, and then disabled for a longer time to reduce power consumption. by default, the sm33 operates in sampling mode during deepstop and static mode and in con- tinuous mode for other sleep modes. sampling mode can also be forced during sleep modes other than deepstop and static, and during normal operation, by writing a one to the force sampling mode bit in the sm33 register (sm33.fs). the user can select the sampling frequency by writing to the sampling frequency field in sm33 (sm33.sampfreq). the sampling mode uses the 32khz rc oscillator (rc32k) as clock source. the 32khz rc oscillator is automatically enabled when the sm33 operates in sampling mode. 13.5.9.3 interrupt and reset generation if the sm33 is enabled (sm33.ctrl is one or two) and the regulator supply voltage drops below the sm33 threshold, the sm33det bit in the power and clocks status register (pclksr.sm33det) is set. this bit is cleared when the supply voltage goes above the thresh- old. an interrupt request is generated on a zer-to-one transition of pclksr.sm33det if the
201 32145c?06/2013 at32uc3l0128/256 supply monitor 3.3v detection bit in the interrupt mask register (imr.sm33det) is set. this bit is set by writing a one to the corresponding bit in the interrupt enable register (ier.sm33det). if sm33.ctrl is one, a por will be generated when th e voltage drops belo w the threshold. if sm33.ctrl is two, the device will not be reset. 13.5.9.4 factory calibration after a reset the sm33.calib field is loaded with a factory defined value. this value is chosen so that the nominal threshold value is 1.75v. the flash calibration is redone after any reset, and the flash calibration done bit in sm33 (sm33.fcd) is set before program execution starts in the cpu. although it is not reco mmended to override default factory settings, it is still possible to override the default value by writing to sm33.calib 13.5.10 temperature sensor rev: 1.0.0.0 the temperature sensor is connected to an ad c channel, please refer to the adc chapter for details. it is enabled by writing one to the enab le bit (en) in the temperature sensor configura- tion register (tsens). the temperature sensor can not be calibrated. please refer to the electrical char acteristics chapter for more details. 13.5.11 120mhz rc oscillator (rc120m) rev: 1.1.0.0 the 120mhz rc oscillator can be used as source for the main clock in the device, as described in the power manager chapter. the oscillator can also be used as source for the generic clocks, as described in generic clock section. the rc120m must be enabled before it is used as a source clock. to enable the clock, the user mu st write a one to the enable bit in the 120mhz rc oscillator control register (rc120mcr.en), and read back the rc120mcr register until the en bit reads one. the clock is disabled by writ ing a zero to rc120mcr.en. the en bit must be read back as zero before the rc120m is re-enabled. if not, undefined behavior may occur. the oscillator is automatically di sabled in certain sleep modes to reduce power consumption, as described in the power manager chapter. 13.5.12 backup registers (br) rev: 1.0.0.1 four 32-bit backup registers are available to store values when the device is in shutdown mode. these registers will keep their content even when the vd dcore supply and the internal regulator supply voltage supplies are removed. the backup registers can be accessed by read- ing from and writing to the br0, br1, br2, and br3 registers. after writing to one of the backup registers the user must wait until the backup register interface ready bit in tne power and clocks status regist er (pclksr.brifardy) is set before writing to another backup register. writes to the backup register while pc lksr.brifardy is zero will be discarded. an interrupt can be generated on a zero-to-one transition on pclksr.brifardy if
202 32145c?06/2013 at32uc3l0128/256 the brifardy bit in the interrupt mask register (imr.brifardy) is set. this bit is set by writ- ing a one to the corresponding bit in the interrupt enable register (ier.brifardy). after powering up the device the backup register interface valid bit in pclksr (pclksr.bri- favalid) is cleared, indicating that the content of the backup registers ha s not been written and contains the reset value. after writing to one of the backup registers the pclksr.brifavalid bit is set. during writes to the backup registers (when brif ardy is zero) brifavalid will be zero. if a reset occurs when brif ardy is zero, brifavalid will be cleared after the reset, indi- cating that the content of the backup registers is not valid. if brifardy is one when a reset occurs, brifavalid will be one and the content is the sa me as before the reset. the user must ensure that brifavalid and brifardy are both set before reading the backup register values. 13.5.13 32khz rc oscillator (rc32k) rev: 1.1.0.0 the rc32k can be used as source for the generic clocks, as described in the generic clocks section. the 32khz rc oscillator (rc32k) is forced on after reset, and output on pa20. the clock is available on the pad until the ppcr.frc32 bit in the power manager has been cleared or a dif- ferent peripheral function has been chosen on pa20 (pa20 will st art with peripheral function f by default). note that the forcin g will only enable the clock outpu t. to be able to use the rc32k normally the oscillator must be enabled as described below. the oscillator is enabled by writ ing a one to the enable bit in the 32khz rc oscillator configura- tion register (rc32kcr.en) and disabled by writ ing a zero to rc32kcr.en. the oscillator is also automatically enabled when the sampling mode is requested for the sm33. in this case, writing a zero to rc32kcr.en will not disable the rc32k until the sampling mode is no longer requested. 13.5.14 generic clock prescalers rev: 1.0.0.0 the generic clocks can be sourced by two special prescalers to increase the generic clock fre- quency precision. these prescalers are named the high resolution prescaler (hrp) and the fractional prescaler (fp). 13.5.14.1 high resolution prescaler the hrp is a 24-bit counter that can generate a very accurate clock waveform. the clock obtained has 50% duty cycle.
203 32145c?06/2013 at32uc3l0128/256 figure 13-7. high resolution prescaler generation the hrp is enabled by writing a one to the high resolution prescaler enable (hrpen) bit in the high resolution prescaler co ntrol register (hrpcr). the user can select a clock source for the hrp by writing to the clock selection (cksel) field of the hrpcr register. the user must configure the high resolution prescaler clock (hrpclk) frequency by writing to the high resolution count (hrcount) field of th e high resolution counter (hrpcr) register. this results in the output frequency: f hrpclk = f src / (2*(hrcount+1)) the cksel field can not be changed dynamica lly but the hrcount field can be changed on- the-fly. 13.5.14.2 fractional prescaler the fp generates a clock whose average frequenc y is more precise than the hrp. however, this clock frequency is subject to jitter around th e target clock frequency. this jitter influence can be decreased by dividing this clock with the gclk divider. moreover the duty cycle of this clock is not precisely 50%. figure 13-8. fractional prescaler generation the fp is enabled by writing a one to the fpen bit in the fractional prescaler control register (fpcr). the user can select a clock source for the fp by writing to the cksel field of the fpcr register. divider cksel hrpclk hrcount mask hrpen divider cksel fpclk fpdiv mask fpen fpmul
204 32145c?06/2013 at32uc3l0128/256 the user must configure the fp frequency by writing to the fpmul and fpdiv fields of the fpmul and fpdiv registers. fpmul and fpdiv must not be equal to zero and fpdiv must be greater or equal to fpmul. this results in the output frequency: f fpclk = f src * fpmul/ (2*fpdiv) the cksel field can not be ch anged dynamically but the fpmul and fpdiv fields can be changed on-the-fly. ? jitter description as described in figure 13-9 , the clkfp half period lengths are integer multiples of the source clock period but are not always equals. however the difference between the low level half period length and the high level half period length is at the most one source clock period. this induces when fpdiv is not an integer mult iple of fpmul a jitter on the fpclk. the more the fpclk frequency is low, the more the jitter incidence is reduced. figure 13-9. fractional prescaler jitter examples 13.5.15 generic clocks rev: 1.1.0.0 timers, communication modules, and other modules connected to external circuitry may require specific clock frequencies to op erate correctly. the scif define s a number of generic clocks that can provide a wide range of accurate clock frequencies. each generic clock runs from either clock source listed in the ?generic clock sources? table in the scif module configuration section. the selected source can optionally be divided by any even integer up to 512. each clock can be i ndependently enabled and disabled, and is also automatically disabled along with peripheral clocks by the sleep controller in the power manager. src clock fpclk fmul= 5 fdiv=5 fmul=3 fdiv=10 fmul=7 fdiv=9
205 32145c?06/2013 at32uc3l0128/256 figure 13-10. generic clock generation 13.5.15.1 enabling a generic clock a generic clock is enabled by writing a one to the clock enable bit (cen) in the generic clock control register (gcctrl) . each generic clock can individually select a clock source by writing to the oscillator select field (oscsel). the sour ce clock can optionally be divided by writing a one to the divide enable bit (diven) and the divisio n factor field (div), resulting in the output frequency: where f src is the frequency of the selected source clock, and f gclk is the output frequency of the generic clock. 13.5.15.2 disabling a generic clock a generic clock is disabled by writing a zero to cen or entering a sleep mode that disables the pb clocks. in either case, the generic clock will be switched off on the first falling edge after the disabling event, to ensure that no glitches occu r. after cen has been written to zero, the bit will still read as one until the next falling edge occurs, and the clock is actually switched off. when writing a zero to cen the other bits in gcctrl should not be changed until cen reads as zero, to avoid glitches on the gene ric clock. the generic clocks will be automatically re-enabled when waking from sleep. 13.5.15.3 changing clock frequency when changing the generic clock frequency by changing oscsel or div, the clock should be disabled before being re-enabled with the new cl ock source or division setting. this prevents glitches during the transition. 13.5.15.4 generic clock allocation the generic clocks are allocated to different functions as shown in the ?generic clock allocation? table in the scif module configuration section. 13.5.16 interrupts the scif has the following interrupt sources: ? ae - access error: ? a protected scif register was accessed without first being correctly unlocked. divider oscsel generic clock div 0 1 diven mask cen sleep controller f src f gclk generic clock sources f gclk f src 2 div 1 + ?? ---------------------------- =
206 32145c?06/2013 at32uc3l0128/256 ? plllock - pll lock ? a 0 to 1 transition on the pclksr.plllock bit is detected. ? plllocklost - pll lock lost ? a to 1 transition on the pclksr.plllocklost bit is detected. ? brifardy - backup register interface ready. ? a 0 to 1 transition on the pclksr.brifardy bit is detected. ? dfll0rcs - dfll reference clock stopped: ? a 0 to 1 transition on the pclksr.dfllrcs bit is detected. ? dfll0rdy - dfll ready: ? a 0 to 1 transition on the pclksr.dfllrdy bit is detected. ? dfll0locklosta - dfll lock lost on accurate value: ? a 0 to 1 transition on the pclksr.dflllocklosta bit is detected. ? dfll0locklostf - dfll lock lost on fine value: ? a 0 to 1 transition on the pclksr.dflllocklostf bit is detected. ? dfll0locklostc - dfll lock lost on coarse value: ? a 0 to 1 transition on the pclksr.dflllocklostc bit is detected. ? dfll0locka - dfll locked on accurate value: ? a 0 to 1 transition on the pclksr.dflllocka bit is detected. ? dfll0lockf - dfll locked on fine value: ? a 0 to 1 transition on the pclksr.dflllockf bit is detected. ? dfll0lockc - dfll locked on coarse value: ? a 0 to 1 transition on the pclksr.dflllockc bit is detected. ? boddet - brown out detection: ? a 0 to 1 transition on the pclksr.boddet bit is detected. ? sm33det - supply monitor 3.3v detector: ? a 0 to 1 transition on the pclksr.sm33det bit is detected. ? vregok - voltage regulator ok: ? a 0 to 1 transition on the pclksr.vregok bit is detected. ? osc0rdy - oscillator ready: ? a 0 to 1 transition on the pclksr.osc0rdy bit is detected. ? osc32rdy - 32khz oscillator ready: ? a 0 to 1 transition on the pclksr.osc32rdy bit is detected. the interrupt sources will generate an interrupt request if the corr esponding bit in the interrupt mask register is set. the interrupt sources are ored together to form one interrupt request. the scif will generate an interrupt request if at least one of the bits in the interrupt mask register (imr) is set. bits in imr are set by writing a one to the corresponding bit in the interrupt enable register (ier), and cleared by writing a one to the corresponding bit in the interrupt disable register (idr). the interrupt request remains active until the corresponding bit in the interrupt status register (isr) is cleared by writing a one to the corresponding bit in the interrupt clear register (icr). because all the interrupt sour ces are ored together, the interrupt request from the scif will remain acti ve until all the bits in isr are cleared.
207 32145c?06/2013 at32uc3l0128/256 13.6 user interface table 13-2. scif register memory map offset register register name access reset 0x0000 interrupt enable register ier write-only 0x00000000 0x0004 interrupt disable register idr write-only 0x00000000 0x0008 interrupt mask register imr read-only 0x00000000 0x000c interrupt status register isr read-only 0x00000000 0x0010 interrupt clear register icr write-only 0x00000000 0x0014 power and clocks status register pclksr read-only 0x00000000 0x0018 unlock register unlock write-only 0x00000000 0x001c oscillator 0 control register oscctrl0 read/write 0x00000000 0x0020 oscillator 32 control register oscctrl32 read/write 0x00000004 0x0024 dfll config register dfll0conf read/write 0x00000000 0x0028 dfll multiplier regist er dfll0mul write-only 0x00000000 0x002c dfll step register dfll0step write-only 0x00000000 0x0030 dfll spread spectrum generator control register dfll0ssg write-only 0x00000000 0x0034 dfll ratio register dfll0ratio read-only 0x00000000 0x0038 dfll synchronization register dfll0sync write-only 0x00000000 0x003c bod level register bod read/write - (2) 0x0044 voltage regulator calibration register vregcr read/write - (2) 0x0048 system rc oscillator calibration register rccr read/write - (2) 0x004c supply monitor 33 calibration register sm33 read/write - (2) 0x0050 temperature sensor calibrati on register tsens read/write 0x00000000 0x0058 120mhz rc oscillator contro l register rc120mcr read/write 0x00000000 0x005c-0x0068 backup registers br read/write 0x00000000 0x006c 32khz rc oscillator control register rc32kcr read/write 0x00000000 0x0070 generic clock control0 gcctrl0 read/write 0x00000000 0x0074 generic clock control1 gcctrl1 read/write 0x00000000 0x0078 generic clock control2 gcctrl2 read/write 0x00000000 0x007c generic clock control3 gcctrl3 read/write 0x00000000 0x0080 generic clock control4 gcctrl4 read/write 0x00000000 0x0084 generic clock control5 gcctrl5 read/write 0x00000000 0x0088 generic clock control6 gcctrl6 read/write 0x00000000 0x008c generic clock control7 gcctrl7 read/write 0x00000000 0x0090 generic clock control8 gcctrl8 read/write 0x00000000 0x0094 generic clock control9 gcctrl9 read/write 0x00000000
208 32145c?06/2013 at32uc3l0128/256 note: 1. the reset value is device specific. please refer to th e module configuration section at the end of this chapter. 2. the reset value of this register depends on factory calibration. 0x0098 pll0 control register pll0 read/write 0x00000000 0x009c high resolution prescaler control register hrpcr read/write 0x00000000 0x00a0 fractional prescaler control register fpcr read/write 0x00000000 0x00a4 fractional prescaler multiplier register fpmul read/write 0x00000000 0x00a8 fractional prescaler divide r register fpdiv read/write 0x00000000 0x03bc commonly used modules version register cmversion read-only - (1) 0x03c0 generic clock prescaler version register gclkprescversion read-only - (1) 0x03c4 pll version register pllversion read-only - (1) 0x03c8 oscillator0 version register osc0version read-only - (1) 0x03cc 32 khz oscillator version register osc32version read-only - (1) 0x03d0 dfll version register dfllifversion read-only - (1) 0x03d4 bod version register bodifaversion read-only - (1) 0x03d8 voltage regulator version register vregifbversion read-only - (1) 0x03dc system rc oscillator version register rcoscifaversion read-only - (1) 0x03e0 3.3v supply monitor version register sm33ifaversion read-only - (1) 0x03e4 temperature sensor version register tsensifaversion read-only - (1) 0x03ec 120mhz rc oscillator version register rc120mifaversion read-only - (1) 0x03f0 backup register interface version register brifaversion read-only - (1) 0x03f4 32khz rc oscillator version register rc32kifaversion read-only - (1) 0x03f8 generic clock version register gclkversion read-only - (1) 0x03fc scif version register version read-only - (1) table 13-2. scif register memory map offset register register name access reset
209 32145c?06/2013 at32uc3l0128/256 13.6.1 interrupt enable register name: ier access type: write-only offset: 0x0000 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will set the corresponding bit in imr. 31 30 29 28 27 26 25 24 ae------- 23 22 21 20 19 18 17 16 ----- plllocklo st0 plllock0 brifardy 15 14 13 12 11 10 9 8 dfll0rcs dfll0rdy dfll0lock losta dfll0lock lostf dfll0lock lostc dfll0lock a dfll0lock f dfll0lock c 76543210 boddet sm33det vregok - - - osc0rdy osc32rdy
210 32145c?06/2013 at32uc3l0128/256 13.6.2 interrupt disable register name: idr access type: write-only offset: 0x0004 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in imr. 31 30 29 28 27 26 25 24 ae------- 23 22 21 20 19 18 17 16 ----- plllocklo st0 plllock0 brifardy 15 14 13 12 11 10 9 8 dfll0rcs dfll0rdy dfll0lock losta dfll0lock lostf dfll0lock lostc dfll0lock a dfll0lock f dfll0lock c 76543210 boddet sm33det vregok - - - osc0rdy osc32rdy
211 32145c?06/2013 at32uc3l0128/256 13.6.3 interrupt mask register name: imr access type: read-only offset: 0x0008 reset value: 0x00000000 0: the corresponding interrupt is disabled. 1: the corresponding interrupt is enabled. a bit in this register is cleared when the corresponding bit in idr is written to one. a bit in this register is set when the corresponding bit in ier is written to one. 31 30 29 28 27 26 25 24 ae------- 23 22 21 20 19 18 17 16 ----- plllocklo st0 plllock0 brifardy 15 14 13 12 11 10 9 8 dfll0rcs dfll0rdy dfll0lock losta dfll0lock lostf dfll0lock lostc dfll0lock a dfll0lock f dfll0lock c 76543210 boddet sm33det vregok - - - osc0rdy osc32rdy
212 32145c?06/2013 at32uc3l0128/256 13.6.4 interrupt status register name: isr access type: read-only offset: 0x000c reset value: 0x00000000 0: the corresponding interrupt is cleared. 1: the corresponding interrupt is pending. a bit in this register is cleared when the corresponding bit in icr is written to one. a bit in this register is set when the corresponding interrupt occurs. 31 30 29 28 27 26 25 24 ae------- 23 22 21 20 19 18 17 16 ----- plllocklo st0 plllock0 brifardy 15 14 13 12 11 10 9 8 dfll0rcs dfll0rdy dfll0lock losta dfll0lock lostf dfll0lock lostc dfll0lock a dfll0lock f dfll0lock c 76543210 boddet sm33det vregok - - - osc0rdy osc32rdy
213 32145c?06/2013 at32uc3l0128/256 13.6.5 interrupt clear register name: icr access type: write-only offset: 0x0010 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in isr. 31 30 29 28 27 26 25 24 ae------- 23 22 21 20 19 18 17 16 ----- plllocklo st0 plllock0 brifardy 15 14 13 12 11 10 9 8 dfll0rcs dfll0rdy dfll0lock losta dfll0lock lostf dfll0lock lostc dfll0lock a dfll0lock f dfll0lock c 76543210 boddet sm33det vregok - - - osc0rdy osc32rdy
214 32145c?06/2013 at32uc3l0128/256 13.6.6 power and clocks status register name: pclksr access type: read-only offset: 0x0014 reset value: 0x00000000 ? brifavalid: backup register interface valid 0: the values in the backup registers are not valid. 1: the values in the backup registers are valid. ? plll0locklost: pll0 lock lost value 0: pll0 has not lost it?s lock or has never been enabled. 1: pll0 has lost it?s lock, either by disabling the pll0 or due to faulty operation. ? pll0lock: pll0 locked on accurate value 0: pll0 is unlocked on accurate value. 1: pll0 is locked on accurate value, and is ready to be selected as clock source with an accurate output clock. ? brifardy: backup register interface ready 0: the backup register interface is busy updating th e backup registers. writes to brn will be discarded. 1: the backup register interface is ready to accept new writes to the backup registers. ? dfll0rcs: dfll0 reference clock stopped 0: the dfll reference clock is running, or has never been enabled. 1: the dfll reference clock has stopped or is too slow. ? dfll0rdy: dfll0 synchronization ready 0: read or write to dfll registers is invalid 1: read or write to dfll registers is valid ? dfll0locklosta: dfll0 lock lost on accurate value 0: dfll has not lost its accurate lock or has never been enabled. 1: dfll has lost its accurate lock, either by disabling the dfll or due to faulty operation. ? dfll0locklostf: dfll0 lock lost on fine value 0: dfll has not lost its fine lock or has never been enabled. 1: dfll has lost its fine lock, either by disabling the dfll or due to faulty operation. 31 30 29 28 27 26 25 24 -brifavalid------ 23 22 21 20 19 18 17 16 ----- plllocklo st0 plllock0 brifardy 15 14 13 12 11 10 9 8 dfll0rcs dfll0rdy dfll0lock losta dfll0lock lostf dfll0lock lostc dfll0lock a dfll0lock f dfll0lock c 76543210 boddet sm33det vregok - - - osc0rdy osc32rdy
215 32145c?06/2013 at32uc3l0128/256 ? dfll0locklostc: dfll0 lock lost on coarse value 0: dfll has not lost its coarse lock or has never been enabled. 1: dfll has lost its coarse lock, either by disabling the dfll or due to faulty operation. ? dfll0locka: dfll0 locked on accurate value 0: dfll is unlocked on accurate value. 1: dfll is locked on accurate value, and is ready to be selected as clock source with an accurate output clock. ? dfll0lockf: dfll0 locked on fine value 0: dfll is unlocked on fine value. 1: dfll is locked on fine value, and is ready to be selected as clock source with a high accuracy on the output clock. ? dfll0lockc: dfll0 lock ed on coarse value 0: dfll is unlocked on coarse value. 1: dfll is locked on coarse value, and is ready to be selected as clock source with medium accuracy on the output clock. ? boddet: brown-out detection 0: no bod event. 1: bod has detected that the supply voltage is below the bod reference value. ? sm33det: supply monitor 3.3v detector 0: sm33 not enabled or the supply voltage is above the sm33 threshold. 1: sm33 enabled and the supply voltage is below the sm33 threshold. ? vregok: voltage regulator ok 0: voltage regulator not enabled or not ready. 1: voltage regulator has reached its output threshold value after being enabled. ? osc0rdy: osc0 ready 0: oscillator not enabled or not ready. 1: oscillator is stable and ready to be used as clock source. ? osc32rdy: 32 khz oscillator ready 0: osc32k not enabled or not ready. 1: osc32k is stable and ready to be used as clock source.
216 32145c?06/2013 at32uc3l0128/256 13.6.7 unlock register name: unlock access type: write-only offset: 0x0018 reset value: 0x00000000 to unlock a write protected register, first write to the unlock r egister with the address of the re gister to unlock in the addr field and 0xaa in the key field. then, in the next pb ac cess write to the register specified in the addr field. the lock is by default off. to turn on the lock, first write 0x aa to the key field and unlock address offset to the addr field in the unlock register, followed by writing 0x5a5a5a5a to the unlock register. to turn off the lock, first write 0xaa to the key field and unlock address offset to the addr field in th e unlock register, followed by writing 0xa5aa5a55 to the unlock register. ? key: unlock key write this bit field to 0xaa to enable unlock. ? addr: unlock address write the address offset of the register to unlock to this field. 31 30 29 28 27 26 25 24 key 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 ------ addr[9:8] 76543210 addr[7:0]
217 32145c?06/2013 at32uc3l0128/256 13.6.8 oscillator control register name: oscctrln access type: read/write reset value: 0x00000000 ? oscen: oscillator enable 0: the oscillator is disabled. 1: the oscillator is enabled. ? startup: oscillator start-up time select start-up time for the oscillator. please refer to the ?o scillator startup time? table in the scif module configuration section for details. ? agc: automatic gain control for test purposes. ? gain: gain selects the gain for the oscillator. please refer to the ?oscilla tor gain settings? table in the scif module configuration sect ion for details. ? mode: oscillator mode 0: external clock connected on xin. xout can be used as general-purpose i/o (no crystal). 1: crystal is connected to xin/xout. note that this register is protected by a lock. to write to this register the unlock register has to be written first. please refer to the unlock register description for details. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------oscen 15 14 13 12 11 10 9 8 ---- startup[3:0] 76543210 ----agc gain[1:0] mode
218 32145c?06/2013 at32uc3l0128/256 13.6.9 32khz oscillator control register name: oscctrl32 access type: read/write reset value: 0x00000004 note: this register is only reset by power-on reset ? reserved this bit must always be written to zero. ? startup: oscillator start-up time select start-up time for 32 khz oscillator 31 30 29 28 27 26 25 24 reserved ------- 23 22 21 20 19 18 17 16 ----- startup[2:0] 15 14 13 12 11 10 9 8 ----- mode[2:0] 76543210 ----en1ken32kpinselosc32en table 13-3. start-up time fo r 32 khz oscillator startup number of rcsys clock cycle approximative equivalent time (rcosc = 115 khz) 00 0 1128 1.1ms 2 8192 72.3 ms 3 16384 143 ms 4 65536 570 ms 5 131072 1.1 s 6 262144 2.3 s 7 524288 4.6 s
219 32145c?06/2013 at32uc3l0128/256 ? mode: oscillator mode ? en1k: 1 khz output enable 0: the 1 khz output is disabled. 1: the 1 khz output is enabled. ? en32k: 32 khz output enable 0: the 32 khz output is disabled. 1: the 32 khz output is enabled. ? pinsel: pins select 0: default pins used. 1: alternate pins: xin32_2 pin is used instead of xin32 pin, xout32_2 pin is used instead of xout32. ? osc32en: 32 khz oscillator enable 0: the 32 khz oscillator is disabled 1: the 32 khz oscillator is enabled table 13-4. operation mode for 32 khz oscillator mode description 0 external clock connected to xin32, xout32 c an be used as general-purpose i/o (no crystal) 1 crystal mode. crystal is connected to xin32/xout32. 2reserved 3reserved 4 crystal and high current mode. cryst al is connected to xin32/xout32. 5reserved 6reserved 7reserved
220 32145c?06/2013 at32uc3l0128/256 13.6.10 dflln configuration register name: dfllnconf access type: read/write reset value: 0x00000000 ? coarse: coarse calibration value set the value of the coarse calibration register. if in closed loop mode, this field is read-only. ? fine: fine calibration value set the value of the fine calibration register. if in closed loop mode, this field is read-only. ? qlen: quick lock enable 0: quick lock is disabled. 1: quick lock is enabled. ? ccen: chill cycle enable 0: chill cycle is disabled. 1: chill cycle is enabled. ? llaw: lose lock after wake 0: locks will not be lost after waking up from sleep modes. 1: locks will be lost after waking up from sle ep modes where the dfll clock has been stopped. ? dither: enable dithering 0: the fine lsb input to the vco is constant. 1: the fine lsb input to the vco is dithered to achieve sub-lsb approximation to the correct multiplication ratio. ? mode: mode selection 0: the dfll is in open loop operation. 1: the dfll is in closed loop operation. ? en: enable 0: the dfll is disabled. 1: the dfll is enabled. note that this register is protected by a lock. to write to this register the unlock register has to be written first. please refer to the unlock register description for details. 31 30 29 28 27 26 25 24 coarse[7:0] 23 22 21 20 19 18 17 16 -------fine[8] 15 14 13 12 11 10 9 8 fine[7:0] 76543210 - qlen ccen - llaw dither mode en
221 32145c?06/2013 at32uc3l0128/256 13.6.11 dflln multiplier register name: dfllnmul access type: read/write reset value: 0x00000000 ? imul: integer multiply factor this field, together with fmul , determines the ratio between f dfll and f ref the dfll. imul is the integer part, while the fmul is the fractional part. in open loop mode, writing to this register has no effect. ? fmul: fractional multiply factor this field, together with imul, determines the ratio between f dfll and f ref the dfll. imul is the integer part, while the fmul is the fractional part. in open loop mode, writing to this register has no effect. note that this register is protected by a lock. to write to this register the unlock register has to be written first. please refer to the unlock register description for details. 31 30 29 28 27 26 25 24 imul[15:8] 23 22 21 20 19 18 17 16 imul[7:0] 15 14 13 12 11 10 9 8 fmul[15:8] 76543210 fmul[7:0]
222 32145c?06/2013 at32uc3l0128/256 13.6.12 dflln maximum step register name: dfllnstep access type: read/write reset value: 0x00000000 ? fstep: fine maximum step this indicates the maximum step size during fine adjustment in closed-loop mode. when adjusting to a new frequency, the expected overshoot of that freq uency depends on this step size. ? cstep: coarse maximum step this indicates the maximum step size during coarse adjustment in closed-loop mode. when adjusting to a new frequency, the expected overshoot of that freq uency depends on this step size. note that this register is protected by a lock. to write to this register the unlock register has to be written first. please refer to the unlock register description for details. 31 30 29 28 27 26 25 24 -------fstep[8] 23 22 21 20 19 18 17 16 fstep[7:0] 15 14 13 12 11 10 9 8 -------- 76543210 cstep[7:0]
223 32145c?06/2013 at32uc3l0128/256 13.6.13 dflln spread spectrum generator control register name: dfllnssg access type: read/write reset value: 0x00000000 ? stepsize: ssg step size sets the step size of the spread spectrum. ? amplitude: ssg amplitude sets the amplitude of the spread spectrum. ? prbs: pseudo random bit sequence 0: each spread spectrum frequency is applied at constant intervals 1: each spread spectrum frequency is applied at pseudo-random intervals ? en: enable 0: ssg is disabled. 1: ssg is enabled. note that this register is protected by a lock. to write to this register the unlock register has to be written first. please refer to the unlock register description for details. 31 30 29 28 27 26 25 24 - ------- 23 22 21 20 19 18 17 16 - - - stepsize[4:0] 15 14 13 12 11 10 9 8 - - - amplitude[4:0] 76543210 ------prbsen
224 32145c?06/2013 at32uc3l0128/256 13.6.14 dflln ratio register name: dfllnratio access type: read-only reset value: 0x00000000 ? ratiodiff: multiplica tion ratio difference in closed-loop mode, this field indicates the error in th e ratio between the vco frequency and the target frequency. ? numref: numerical reference the number of reference clock cycles used to measure the vco frequency equals 2^numref. 31 30 29 28 27 26 25 24 ratiodiff[15:8] 23 22 21 20 19 18 17 16 ratiodiff[7:0] 15 14 13 12 11 10 9 8 - ------- 76543210 - -- numref[4:0]
225 32145c?06/2013 at32uc3l0128/256 13.6.15 dflln synchronization register name: dfllnsync access type: write-only reset value: 0x00000000 ? sync: synchronization to be able to read the current value of dfllnconf or dfllnratio in closed-loop mode, this bit should be written to one. the updated value is available in dfllnconf and dfllnratio when pclksr.dfllnrdy is set. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 -------sync
226 32145c?06/2013 at32uc3l0128/256 13.6.16 bod control register name: bod access type: read/write reset value: - ? sfv: store final value 0: the register is read/write 1: the register is read-only, to protect against further accidental writes. this bit is cleared after any reset except for a bod reset, and during flash calibration. ? fcd: fuse calibration done 0: the flash calibration will be redone after any reset. 1: the flash calibration will be redone after any reset except for a bod reset. this bit is cleared after any reset, except for a bod reset. this bit is set when the ctrl, hyst and level fields have been updated by the flash fuses after a reset. ? ctrl: bod control ? hyst: bod hysteresis 0: no hysteresis. 1: hysteresis on. ? level: bod level this field sets the triggering threshold of the bod. se e electrical characteristics for actual voltage levels. note that any change to the level field of the bod register should be done with the bod deactivated to avoid spurious reset or interrupt. 31 30 29 28 27 26 25 24 sfv------- 23 22 21 20 19 18 17 16 -------fcd 15 14 13 12 11 10 9 8 ------ ctrl 76543210 - hyst level table 13-5. operation mode for bod ctrl description 0 bod is disabled. 1 bod is enabled and can reset the device. an interrupt request will be generated, if enabled in the imr register. 2 bod is enabled but cannot reset the device. an interrupt request will be generated, if enabled in the imr register. 3 reserved.
227 32145c?06/2013 at32uc3l0128/256 note that this register is protected by a lock. to write to this register the unlock register has to be written first. please refer to the unlock register description for details.
228 32145c?06/2013 at32uc3l0128/256 13.6.17 voltage regulator calibration register name: vregcr access type: read/write reset value: - ? sfv: store final value 0: the register is read/write. 1: the register is read-only, to protect against further accidental writes. this bit is cleared by a power-on reset. ? intpd: internal pull-down this bit is used for test purposes only. 0: the voltage regulator output is not pulled to ground. 1: the voltage regulator output has a pull-down to ground. ? por18value: power-on reset 1.8v output value 0: vddcore voltage is below the por18 power-on threshold level. 1: vddcore voltage is above the por18 power-on threshold level. this bit is read-only. writing to this bit has no effect. ? por33value: power-on reset 3.3v output value 0: internal regulator supply voltage is below the por33 power-on threshold level. 1: internal regulator supply voltage is above the por33 power-on threshold level. this bit is read-only. writing to this bit has no effect. ? por18mask: power-on r eset 1.8v output mask 0: power-on reset is not masked. 1: power-on reset is masked. ? por18status: power-on reset 1.8v status 0: power-on reset is disabled. 1: power-on reset is enabled. this bit is read-only. writing to this bit has no effect. ? por18en: power-on reset 1.8v enable writing a zero to this bit disables the por18 detector. writing a one to this bit enables the por18 detector. ? por33mask: power-on r eset 3.3v output mask 0: power-on reset 3.3v is not masked. 31 30 29 28 27 26 25 24 sfv intpd - - - dbg- por18value por33value 23 22 21 20 19 18 17 16 por18mask por18stat us por18en por33mask por33stat us por33en deepdis fcd 15 14 13 12 11 10 9 8 ---- calib 76543210 on vregok en - - selvdd
229 32145c?06/2013 at32uc3l0128/256 1: power-on reset 3.3v is masked. ? por33status: power-on reset 3.3v status 0: power-on reset is disabled. 1: power-on reset is enabled. this bit is read-only. writing to this bit has no effect. ? por33en: power-on reset 3.3v enable 0: writing a zero to this bit disables the por33 detector. 1: writing a one to this bi t enables the por33 detector. ? deepdis: disable regulator deep mode 0: regulator will enter deep mode in low-power sleep modes for lower power consumption. 1: regulator will stay in full-power mode in all sleep modes for shorter start-up time. ? fcd: flash calibration done 0: the flash calibration will be redone after any reset. 1: the flash calibration will only be redone after a power-on reset. this bit is cleared a fter a power-on reset. this bit is set when the calib field has been updated by flash calibration after a reset. ? calib: calibration value calibration value for voltage regulator. this is calibrated during production and should not be changed. ? on: voltage regulator on status 0: the voltage regulator is currently disabled. 1: the voltage regulator is currently enabled. this bit is read-only. writing to this bit has no effect. ? vregok: voltage regulator ok status 0: the voltage regulator is disabled or has not yet reached a stable output voltage. 1: the voltage regulator has reached the output voltage threshold level after being enabled. this bit is read-only. writing to this bit has no effect. ? en: enable 0: the voltage regulator is disabled. 1: the voltage regulator is enabled. note : this bit is set after a power-on reset (por). ? selvdd: select vdd output voltage of the voltage regulator. the default value of this bit corresp onds to an output voltage of 1.8v. note that this register is protected by a lock. to write to this register the unlock register has to be written first. please refer to the unlock register description for details.
230 32145c?06/2013 at32uc3l0128/256 13.6.18 system rc oscillator calibration register name: rccr access type: read/write reset value: - ? fcd: flash calibration done 0: the flash calibration will be redone after any reset. 1: the flash calibration will only be redone after a power-on reset. this bit is cleared after a por. this bit is set when the calib field has been updated by the flash fuses after a reset. ? calib: calibration value calibration value for the system rc oscillator (rcsys). note that this register is protected by a lock. to write to this register the unlock register has to be written first. please refer to the unlock register description for details. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------fcd 15 14 13 12 11 10 9 8 ------ calib[9:8] 76543210 calib[7:0]
231 32145c?06/2013 at32uc3l0128/256 13.6.19 supply monitor 33 calibration register name: sm33 access type: read/write reset value: - ? sampfreq: sampling frequency selects the sampling mode frequency of the 3.3v supply monitor. in sampling mode, the sm33 performs a measurement every 2 (sampfreq+5) cycles of the internal 32khz rc oscillator. ? onsm: supply monitor on indicator 0: the supply monitor is disabled. 1: the supply monitor is enabled. this bit is read-only. writing to this bit has no effect. ? sfv: store final value 0: the register is read/write 1: the register is read-only, to protect against further accidental writes. this bit is cleared after a reset. ? fcd: flash calibration done this bit is cleared after a reset. this bit is set when calib field has been updated after a reset. ? calib: calibration value calibration value for the sm33. ? fs: force sampling mode 0: sampling mode is enabled in deepstop and static mode only. 1: sampling mode is always enabled. ? ctrl: supply monitor control 31 30 29 28 27 26 25 24 ---- sampfreq 23 22 21 20 19 18 17 16 -----onsmsfvfcd 15 14 13 12 11 10 9 8 ---- calib 76543210 fs - - - ctrl
232 32145c?06/2013 at32uc3l0128/256 selects the operating mode for the sm33. note that this register is protected by a lock. to write to this register the unlock register has to be written first. please refer to the unlock register description for details. table 13-6. operation mode for sm33 ctrl description 0 sm33 is disabled. 1 sm33 is enabled and can reset the device. an interrupt request will be generated if the corresponding interrupt is enabled in the imr register. 2 sm33 is enabled and cannot reset the device. an interrupt request will be generated if the corresponding interrupt is enabled in the imr register. 3 sm33 is disabled 4-7 reserved
233 32145c?06/2013 at32uc3l0128/256 13.6.20 temperature sensor configuration register name: tsens access type: read/write reset value: 0x00000000 ? en: temperature sensor enable 0: the temperature sensor is disabled. 1: the temperature sensor is enabled. note that this register is protected by a lock. to write to this register the unlock register has to be written first. please refer to the unlock register description for details. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 --- --en
234 32145c?06/2013 at32uc3l0128/256 13.6.21 120mhz rc oscillator configuration register name: rc120mcr access type: read/write reset value: 0x00000000 ? en: rc120m enable 0: the 120 mhz rc oscillator is disabled. 1: the 120 mhz rc oscillator is enabled. note that this register is protected by a lock. to write to this register the unlock register has to be written first. please refer to the unlock register description for details. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 -------en
235 32145c?06/2013 at32uc3l0128/256 13.6.22 backup register n name: brn access type: read/write reset value: 0x00000000 this is a set of general-purpose read/write registers. data stored in these registers is retained when the device is in shut- down. before writing to these registers the user must ensure that pclksr.brifardy is not set. note that this registers are protected by a lock. to write to these registers the unlock register has to be written first. please refer to the unlock register description for details. 31 30 29 28 27 26 25 24 data[31:24] 23 22 21 20 19 18 17 16 data[23:16] 15 14 13 12 11 10 9 8 data[15:8] 76543210 data[7:0]
236 32145c?06/2013 at32uc3l0128/256 13.6.23 32khz rc oscillator configuration register name: rc32kcr access type: read/write reset value: 0x00000000 ? en: rc32k enable 0: the 32 khz rc oscillator is disabled. 1: the 32 khz rc oscillator is enabled. note that this register is protected by a lock. to write to this register the unlock register has to be written first. please refer to the unlock register description for details. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 -------en
237 32145c?06/2013 at32uc3l0128/256 13.6.24 generic clock control name: gcctrl access type: read/write reset value: 0x00000000 there is one gcctrl register per generic clock in the design. ? div: division factor the number of div bits for each generic clock is as shown in the ?generic clock number of div bi ts? table in the scif module configuration section. ? oscsel: oscillator select selects the source clock for the generic clock. please refer to the ?generic clock sources? table in the scif module configuration section. ? diven: divide enable 0: the generic clock equals the undivided source clock. 1: the generic clock equals the source clock divided by 2*(div+1). ? cen: clock enable 0: the generic clock is disabled. 1: the generic clock is enabled. 31 30 29 28 27 26 25 24 div[15:8] 23 22 21 20 19 18 17 16 div[7:0] 15 14 13 12 11 10 9 8 - - - oscsel[4:0] 76543210 ------divencen
238 32145c?06/2013 at32uc3l0128/256 13.6.25 pll control register name: plln access type: read/write reset value: 0x00000000 ? pllcount: pll count specifies the number of rcsys clock cycles before isr.plllockn will be set after pl ln has been written, or after plln has been automatically re-enabled after exiting a sleep mode. ? pllmul: pll multiply factor ? plldiv: pll division factor these fields determine the ratio of the pll ou tput frequency to the source oscillator frequency: f vco = (pllmul+1)/plldiv ? f ref if plldiv >0 f vco = 2?(pllmul+1) ? f ref if plldiv = 0 note that the pllmul field should always be greater than 1 or the behavior of the pll will be undefined. ? pllopt: pll option pllopt[0]: selects the vco frequency range (f vco ). 0: 80mhz 239 32145c?06/2013 at32uc3l0128/256 ? pllen: pll enable 0: pll is disabled. 1: pll is enabled. note that it is not possible to change an y of the pll configuration bits when the pl l is enabled, any write to plln while the p ll is enabled will be discarded. note that this register is protected by a lock. to write to this register the unlock register has to be written first. please refer to the unlock register description for details.
240 32145c?06/2013 at32uc3l0128/256 13.6.26 high resolution prescaler control register name: hrpcr access type: read/write reset value: 0x00000000 ? hrcount: high resolution counter specify the input clock period to count to generate the output clock edge. hrcount can be written to dynamically in order to tune the hrpclk frequency on-the-go. ? cksel: clock in put selection this field selects the clock input for the prescaler. see the ?hrp clock sources? table in the scif module configuration sectio n for details. it must not be changed if the hrpen is one. ? hrpen: high resolution prescaler enable 0: the high resolution prescaler is disabled. 1: the high resolution prescaler is enabled. 31 30 29 28 27 26 25 24 hrcount[23:16] 23 22 21 20 19 18 17 16 hrcount[15:8] 15 14 13 12 11 10 9 8 hrcount[7:0] 76543210 - - - - cksel hrpen
241 32145c?06/2013 at32uc3l0128/256 13.6.27 fractional prescaler control register name: fpcr access type: read/write reset value: 0x00000000 ? cksel: clock in put selection this field selects the clock input for the prescaler. see the ?fp clock sources? table in the scif module configuration section for details. it must not be changed if the fpen is one. ? fpen: high resolution prescaler enable 0: the fractional prescaler is disabled. 1: the fractional prescaler is enabled. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 - - - - cksel fpen
242 32145c?06/2013 at32uc3l0128/256 13.6.28 fractional prescaler mul register name: fpmul access type: read/write reset value: 0x00000000 ? fpmul: fractional prescal er multiplication factor this field selects the multiplication factor for the prescaler. notice that fpmul is always smaller than fpdiv. fpmul can be written to dynamically in order to tune the fpclk frequency on-the-go. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 fpmul[15:8] 76543210 fpmul[7:0]
243 32145c?06/2013 at32uc3l0128/256 13.6.29 fractional prescaler div register name: fpdiv access type: read/write reset value: 0x00000000 ? fpdiv: fractional prescaler division factor this field selects the divisi on factor for the prescaler. notice that fpmul must be smaller than fpdiv. fpdiv can be written to dynamically in order t o tune the fpclk frequency on- the-go. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 fpdiv[15:8] 76543210 fpdiv[7:0]
244 32145c?06/2013 at32uc3l0128/256 13.6.30 commonly used modules version register name: cmversion access type: read-only offset: 0x03bc reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
245 32145c?06/2013 at32uc3l0128/256 13.6.31 gclk prescaler version register name: gclkprescversion access type: read-only offset: 0x03c0 reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
246 32145c?06/2013 at32uc3l0128/256 13.6.32 pll version register name: pllversion access type: read-only offset: 0x03c4 reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
247 32145c?06/2013 at32uc3l0128/256 13.6.33 oscillator 0 version register name: osc0version access type: read-only offset: 0x03c8 reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
248 32145c?06/2013 at32uc3l0128/256 13.6.34 32khz oscillator version register name: osc32version access type: read-only offset: 0x03cc reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
249 32145c?06/2013 at32uc3l0128/256 13.6.35 digital frequency locked loop version register name: dfllif version access type: read-only offset: 0x03d0 reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
250 32145c?06/2013 at32uc3l0128/256 13.6.36 brown-out detector version register name: bodifaversion access type: read-only offset: 0x03d4 reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
251 32145c?06/2013 at32uc3l0128/256 13.6.37 voltage regulator version register name: vregifbversion access type: read-only offset: 0x03d8 reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
252 32145c?06/2013 at32uc3l0128/256 13.6.38 rc oscillator version register name: rcoscifaversion access type: read-only offset: 0x03dc reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
253 32145c?06/2013 at32uc3l0128/256 13.6.39 3.3v supply monitor version register name: sm33ifaversion access type: read-only offset: 0x03e0 reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
254 32145c?06/2013 at32uc3l0128/256 13.6.40 temperature sensor version register name: tsensifaversion access type: read-only offset: 0x03e4 reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
255 32145c?06/2013 at32uc3l0128/256 13.6.41 120mhz rc oscillator version register name: rc120mifaversion access type: read-only offset: 0x03ec reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
256 32145c?06/2013 at32uc3l0128/256 13.6.42 backup register interface version register name: brifaversion access type: read-only offset: 0x03f0 reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
257 32145c?06/2013 at32uc3l0128/256 13.6.43 32khz rc oscillator version register name: rc32kifaversion access type: read-only offset: 0x03f4 reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
258 32145c?06/2013 at32uc3l0128/256 13.6.44 generic clock version register name: gclkversion access type: read-only offset: 0x03f8 reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
259 32145c?06/2013 at32uc3l0128/256 13.6.45 scif version register name: version access type: read-only offset: 0x03fc reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 - - - - version[11:0] 76543210 version[7:0]
260 32145c?06/2013 at32uc3l0128/256 13.7 module configuration the specific configuration for each scif instance is listed in the following tables.the module bus clocks listed here are connecte d to the system bus clocks. pleas e refer to the power manager chapter for details. table 13-7. module clock name module name clock name description scif clk_scif clock for the scif bus interface table 13-8. oscillator startup times startup number of system rc oscillator clock cycle approximative equivalent time (rcsys = 115 khz) 00 0 1 64 557 us 2128 1.1ms 32048 18ms 44096 36ms 58192 71ms 6 16384 143 ms 7 32768 285 ms 8 4 35 us 9 8 70 us 10 16 139 us 11 32 278 us 12 256 2.2 ms 13 512 4.5 ms 14 1024 8.9 ms 15 32768 285 ms table 13-9. oscillator gain settings gain[1:0] function 0 oscillator is used with gain g0 (xin from 0.45 mhz to 12.0 mhz) 1 oscillator is used with gain g1 (xin from 12.0 mhz to 16.0 mhz) 2 oscillator is used with gain g2 (xin equals 16.0 mhz. used for e.g. increasing s/n ratio, better drive strength for high esr crystals) 3 oscillator is used with gain g3 (xin equals 16.0 mhz. used for e.g. increasing s/n ratio, better drive strength for high esr crystals)
261 32145c?06/2013 at32uc3l0128/256 in at32uc3l0128/256, there are 10 generic cloc ks. these are allocated to different functions as shown in table 13-10 . note that only gclk4-0 are routed out. note: 1. gclk is not routed out. table 13-10. generic clock allocation clock number function 0 dfllif main reference and gclk0 pin (clk_dfllif_ref) 1 dfllif dithering and ssg reference and gclk1 pin (clk_dfllif_dither) 2 ast and gclk2 pin 3 pwma and gclk3 pin 4 cat, acifb and gclk4 pin 5 (1) gloc 6 (1) 7 (1) 8 (1) pll0 source clock 9 (1) master generic clock. can be used as source for other generic clocks. table 13-11. generic clock sources oscsel clock/oscillator description 0 rcsys system rc oscillator clock 1 osc32k output clock from osc32k 2 dfll0 output clock from dfll0 3 osc0 output clock from oscillator0 4 rc120m output from 120mhz rcosc 5 clk_cpu the clock the cpu runs on 6 clk_hsb high speed bus clock 7 clk_pba peripheral bus a clock 8 clk_pbb peripheral bus b clock 9 rc32k output from 32khz rcosc 10 reserved 11 clk_1k 1khz output clock from osc32k 12 pll0 output clock from pll0 13-14 reserved 15 gclk_in0 gclk_in0 pin, digital clock input
262 32145c?06/2013 at32uc3l0128/256 . 16-17 reserved 18 gclk9 generic clock 9. can not be used as an input to itself 19-31 reserved table 13-12. pll clock sources pllosc clock/oscillator description 0 osc0 output clock from oscillator0 1 gclk8 generic clock 8 2-3 reserved table 13-13. generic clock number of div bits generic clock number of div bits 08 18 28 38 48 58 68 78 88 916 table 13-14. register reset values register reset value cmversion 0x00000100 pllversion 0x00000110 osc0version 0x00000111 osc32version 0x00000110 dfllifversion 0x00000210 bodifaversion 0x00000120 vregifbversion 0x00000110 table 13-11. generic clock sources oscsel clock/oscillator description
263 32145c?06/2013 at32uc3l0128/256 rcoscifaversion 0x00000111 sm33ifaversion 0x00000110 tsenseifaversion 0x00000100 rc120mifaversion 0x00000110 brifaversion 0x00000100 rc32kifaversion 0x00000110 gclkversion 0x00000110 version 0x00000110 table 13-14. register reset values register reset value
264 32145c?06/2013 at32uc3l0128/256 14. asynchronous timer (ast) rev: 3.1.0.1 14.1 features ? 32-bit counter with 32-bit prescaler ? clocked source ? system rc oscillator (rcsys) ? 32khz crystal oscillator (osc32k) ?pb clock ? generic clock (gclk) ? 1khz clock from 32khz oscillator ? operation and wakeup during shutdown ? optional calendar mode supported ? digital prescaler tuning for increased accuracy ? periodic interrup t(s) and peripheral event(s) supported ? alarm interrupt(s) and peripheral event(s) supported ? optional clear on alarm 14.2 overview the asynchronous timer (ast) enables periodi c interrupts and periodic peripheral events, as well as interrupts and peripheral events at a specifi ed time in the future. the ast consists of a 32-bit prescaler which feeds a 32-bit up-counter. the prescaler can be clocked from five differ- ent clock sources, including the low-power 32khz oscillator, which allows the ast to be used as a real-time timer with a maximum timeout of more than 100 years. also, the pb clock or a generic clock can be used for high-speed operation, allowing the ast to be used as a general timer. the ast can generate periodic interrupts and peripheral events from output from the prescaler, as well as alarm interrupts and peripheral events, which can trigger at any counter value. addi- tionally, the timer can trigger an overflow interrupt and peripheral event, and be reset on the occurrence of any alarm. this allows periodic interrupts and peripheral events at very long and accurate intervals. to keep track of time during shutdown the ast can run while the rest of the core is powered off. this will reduce the power consumption when the system is idle. the ast can also wake up the system from shutdown using either the alarm wakeup, periodic wakeup. or overflow wakeup mechanisms. the ast has been designed to meet the system tick and real time clock requirements of most embedded operating systems.
265 32145c?06/2013 at32uc3l0128/256 14.3 block diagram figure 14-1. asynchronous timer block diagram 14.4 product dependencies in order to use this module, other parts of the system must be configured correctly, as described below. 14.4.1 power management when the ast is enabled, it will re main clocked as long as its sele cted clock source is running. it can also wake the cpu from the currently active sleep mode. refer to the power manager chap- ter for details on the different sleep modes. 14.4.2 clocks the clock for the ast bus interface (clk_ast) is generated by the power manager. this clock is turned on by default, and can be enabled and disabled in the power manager. a number of clocks can be selected as source for the internal presca ler clock clk_ast_prsc. the prescaler, counter, an d interrupt will function as long as this selected clock source is active. the selected clock must be enabled in the system control interface (scif). the following clock sources are available: ? system rc oscillator (rcsys). th is oscillator is always enabl ed, except in some sleep modes. please refer to the electrical characteri stics chapter for the characteristic frequency of this oscillator. ? 32khz crystal oscillator (osc32k). this oscillator must be enabled before use. ? peripheral bus clock (pb clock). this is the cl ock of the peripheral bus the ast is connected to. 32-bit prescaler 32-bit counter alarm interrupts counter value clk_ast_cnt ovf control register en cssel psel periodic interrupts alarm register interrupt status and control irqs periodic interval register events wake control wake wake enable register digital tuner register rcsys osc32 pb clock gclk others clk_ast clk_ast_prsc clk_ast clk_ast
266 32145c?06/2013 at32uc3l0128/256 ? generic clock (gclk). one of the generic clocks is connected to the ast. this clock must be enabled before use, and remains enabled in sleep modes when the pb clock is active. ? 1khz clock from the 32khz oscillator (clk_1k). th is clock is only available in crystal mode, and must be enabled before use. in shutdown mode only the 32kh z oscillator and the 1khz clock are available, using certain pins. please refer to the power manager chapter for details. 14.4.3 interrupts the ast interrupt request lines are connected to the interrupt controller. using the ast inter- rupts requires the interrupt controller to be programmed first. 14.4.4 peripheral events the ast peripheral events are connected via the peripheral event system. refer to the periph- eral event system chapter for details. 14.4.5 debug operation the ast prescaler and counter is frozen during debug operation, unless the run in debug bit in the development control register is set and the bit corresponding to the ast is set in the peripheral debug register (pdbg). please refer to the on-chip debug chapter in the avr32uc technical reference manual, and the ocd module configuration section, for details. if the ast is configured in a way that requires it to be periodically serviced by the cpu through interrupts or similar, improper operation or data loss may result during debugging. 14.5 functional description 14.5.1 initialization before enabling the ast, the internal ast clock clk_ast_prsc must be enabled, following the procedure specified in section 14.5.1.1 . the clock source select field in the clock register (clock.cssel) selects the source for this clock. the clock en able bit in the clock register (clock.cen) enables the clk_ast_prsc. when clk_ast_prsc is enabled, the ast can be enabled by writing a one to the enable bit in the control register (cr.en). 14.5.1.1 enabling and disabling the ast clock the clock source selection field (clock.cssel ) and the clock enable bit (clock.cen) can- not be changed simultaneously. special procedures must be followed for enabling and disabling the clk_ast_prsc and for changing the source for this clock. to enable clk_ast_prsc: ? write the selected value to clock.cssel ? wait until sr.clkbusy reads as zero ? write a one to clock.cen, without changi ng clock.cssel ? wait until sr.clkbusy reads as zero to disable the clock: ? write a zero to clock.cen to disable the clock, without changing clock.cssel ? wait until sr.clkbusy reads as zero
267 32145c?06/2013 at32uc3l0128/256 14.5.1.2 changing the source clock the clk_ast_prsc must be disabled before swit ching to another source clock. the clock busy bit in the status register (sr.clkbusy) indicates whether the clock is busy or not. this bit is set when the cen bit in the clock register is changed, and cleared when the clock reg- ister can be changed. to change the clock: ? write a zero to clock.cen to disable the clock, without changing clock.cssel ? wait until sr.clkbusy reads as zero ? write the selected value to clock.cssel ? wait until sr.clkbusy reads as zero ? write a one to clock.cen to enable the clock, without changing the clock.cssel ? wait until sr.clkbusy reads as zero 14.5.2 basic operation 14.5.2.1 prescaler when the ast is enabled, the 32-bit pr escaler will increment on the rising edge of clk_ast_prsc. the prescaler value cannot be read or written, but it can be reset by writing a one to the prescaler clear bit in the control register (cr.pclr). the prescaler select field in the control regist er (cr.psel) selects the prescaler bit psel as source clock for the counter (clk_ast_cnt). this results in a counter frequency of: where f prsc is the frequency of the internal prescaler clock clk_ast_prsc. 14.5.2.2 counter operation when enabled, the ast will increment on every 0-to-1 transition of the selected prescaler tap- ping. when the calender bit in the control regi ster (cr.cal) is zero, the counter operates in counter mode. it will increm ent until it reaches the top value of 0xffffffff, and then wrap to 0x00000000. this sets the status bit overflow in the status register (s r.ovf). optionally, the counter can also be reset when an alarm occurs (see section 14.5.3.2 on page 269 . this will also set the ovf bit. the ast counter value can be read from or written to the counter value (cv) register. note that due to synchronization, continuous reading of the cv register with the lowest prescaler setting will skip every third value. in addition, if clk_ast_prsc is as fast as, or faster than, the clk_ast, the prescaler value must be 3 or higher to be able to read the cv without skipping values. f cnt f prsc 2 psel 1 + ---------------------- - =
268 32145c?06/2013 at32uc3l0128/256 14.5.2.3 calendar operation when the cal bit in the control register is one, the counter operates in calendar mode. before this mode is enabled, the prescaler should be set up to give a pulse every second. the date and time can then be read from or written to the calendar value (calv) register. time is reported as seconds, minutes, and hours according to the 24-hour clock format. date is the numeral date of month (starting on 1). month is the numeral month of the year (1 = january, 2 = february, etc.). year is a 6-bit field coun ting the offset from a software-defined leap year (e.g. 2000). the date is automatically compensated for leap years, assuming every year divisible by 4 is a leap year. all peripheral events and interrupts work the same way in calendar mode as in counter mode. however, the alarm register (arn) must be written in time/date format for the alarm to trigger correctly. 14.5.3 interrupts the ast can generate five separate interrupt requests: ?ovf: ovf ? per: per0 , per1 ?alarm: alarm0 , alarm1 ? clkready ? ready this allows the user to allocate separate handlers and priorities to the different interrupt types. the generation of the per interrupt is described in section 14.5.3.1 ., and the generation of the alarm interrupt is described in section 14.5.3.2 . the ovf interrupt is generated when the counter overflows, or when the alarm value is re ached, if the clear on alarm bit in the control register is one. the clkready interrupt is generated when sr.clkbusy has a 1-to-0 transi- tion, and indicates that the clock synchroniz ation is completed. the ready interrupt is generated when sr.busy has a 1-to-0 transition, and indicates that the synchronization described in section 14.5.8 is completed. an interrupt request will be generated if the corresponding bit in the interrupt mask register (imr) is set. bits in imr are set by writing a one to the corresponding bit in the interrupt enable register (ier), and cleared by writing a one to the corresponding bit in the interrupt disable register (idr). the interrupt request remains acti ve until the corresponding bit in sr is cleared by writing a one to the corresponding bit in the status clear register (scr). the ast interrupts can wake the cpu from any sleep mode where the source clock and the interrupt controller is active. 14.5.3.1 periodic interrupt the ast can generate periodic interrupts. if the pern bit in the interrupt mask register (imr) is one, the ast will generate an interrup t request on the 0-to-1 transition of the selected bit in the
269 32145c?06/2013 at32uc3l0128/256 prescaler when the ast is enabled. the bit is selected by the interval select field in the corre- sponding periodic interval register (pirn.insel), resulting in a periodic interrupt frequency of where f cs is the frequency of the selected clock source. the corresponding pern bit in the status register (sr) will be set when the select ed bit in the prescaler has a 0-to-1 transition. because of synchronizat ion, the transfer of the insel va lue will not happen immediately. when changing/setting the insel value, the user must make sure that the prescaler bit number insel will not have a 0-to-1 transition be fore the insel value is transferr ed to the register. in that case, the first periodic interrupt afte r the change will no t be triggered. 14.5.3.2 alarm interrupt the ast can also generate alar m interrupts. if the al armn bit in imr is one, the ast will gen- erate an interrupt request when the counter val ue matches the selected alarm value, when the ast is enabled. the alarm value is selected by writing the value to the value field in the corre- sponding alarm register (arn.value). the corresponding alarmn bit in sr will be set when the counter reaches the selected alarm value. because of synchronization, the transfer of t he alarm value will not happen immediately. when changing/setting the alarm value, the user must make sure that the counter will not count the selected alarm value before the value is transferred to the register. in that case, the first alarm interrupt after the cha nge will not be triggered. if the clear on alarm bit in the control register (cr.can) is one, the corresponding alarm inter- rupt will clear the count er and set the ovf bit in the stat us register. this will generate an overflow interrupt if the ovf bit in imr is set. 14.5.4 peripheral events the ast can generate a number of peripheral events: ?ovf ? per0 ? per1 ?alarm0 ?alarm1 the pern peripheral event(s) is generated the same way as the per interrupt, as described in section 14.5.3.1 . the alarmn peripheral event(s) is generated the same way as the alarm interrupt, as described in section 14.5.3.2 . the ovf peripheral event is generated the same way as the ovf interrupt, as described in section 14.5.3 - f pa f cs 2 insel 1 + ------------------------- =
270 32145c?06/2013 at32uc3l0128/256 the peripheral event will be genera ted if the corresponding bit in the event mask (evm) register is set. bits in evm register are set by writing a one to the corresponding bit in the event enable (eve) register, and cleared by writing a one to the corresponding bit in the event disable (evd) register. 14.5.5 ast wakeup the ast can wake up the cpu directly, without the need to trigger an interrupt. a wakeup can be generated when the counter overflows, when the counter reaches the selected alarm value, or when the selected prescaler bit has a 0-to-1 transition. in this case, the cpu will continue executing from the instruction following the sleep instruction. the ast wakeup is enabled by writing a one to th e corresponding bit in the wake enable regis- ter (wer). when the cpu wakes from sleep, the wake signal must be cleared by writing a one to the corresponding bit in scr to clear the internal wake signal to the sleep controller. if the wake signal is not cleared after waking from sleep, the next sleep instruction will have no effect because the cpu will wake immediatel y after this sl eep instruction. the ast wakeup can wake the cpu from any sleep mode where the source clock is active. the ast wakeup can be configured independently of the interrupt masking. 14.5.6 shutdown mode if the ast is configured to use a clock that is available in shutdown mode, the ast can be used to wake up the system from sh utdown. both the alarm wakeup, periodic wakeup, and overflow wakeup mechanisms can be used in this mode. when waking up from shutdown mode all control registers will have the same value as before the shutdown was entered, except the interrupt mask register (imr). imr will be reset with all interrupts turned off. the software must first reconfigure the interrupt controller and then enable the interrupts in the ast to again receive interrupts from the ast. the cv register will be updated with the current counter value directly after wakeup from shut- down. the sr will show the status of the ast, including the status bits set during shutdown operation. when waking up the system from shutdown the cpu will start executi ng code from the reset start address. 14.5.7 digital tuner the digital tuner adds th e possibility to compensate for a too-sl ow or a too-fast input clock. the add bit in the digital tuner register (dtr.add) selects if the prescaler frequency should be reduced or increased. if add is ?0?, the prescaler frequency is reduced: where f tuned is the tuned frequency, f 0 is the original prescaler frequency, and value and exp are the corresponding fields to be programmed in dtr. note that dtr.exp must be greater than zero. frequency tuning is disabled by programming dtr.value as zero. f tuned f 0 1 1 roundup 256 value ------------------- - ?? ?? 2 exp ?? ? 1 + ------------------------------------------------------------------------------- - ? ?? ?? ?? ?? ?? =
271 32145c?06/2013 at32uc3l0128/256 if add is ?1?, the prescaler frequency is increased: note that for these formulas to be within an error of 0.01%, it is recommended that the prescaler bit that is used as the clock fo r the counter (selected by cr.psel) or to trigger t he periodic inter- rupt (selected by pirn.insel) be bit 6 or higher. 14.5.8 synchronization as the prescaler and counter operate asynchronously from the user interface, the ast needs a few clock cycles to synchronize the values written to the cr, cv, scr, wer, eve, evd, pirn, arn, and dtr registers. the busy bit in the status register (sr.busy) indicates that the syn- chronization is ongoing. during this time, writes to these registers will be discarded and reading will return a ze ro value. note that synchronization takes place also if the prescaler is clocked from clk_ast. f tuned f 0 1 1 roundup 256 value ------------------- - ?? ?? 2 exp ?? ? 1 ? ------------------------------------------------------------------------------- - + ?? ?? ?? ?? ?? =
272 32145c?06/2013 at32uc3l0128/256 14.6 user interface note: 1. the reset values are device specific. please refer to the module configuration section at the end of this chapter. 2. the number of alarm and periodic interval registers are device specific. please refer to the module configuration section at the end of this chapter. table 14-1. ast register memory map offset register register name access reset 0x00 control register cr read/write 0x00000000 0x04 counter value cv read/write 0x00000000 0x08 status register sr read-only 0x00000000 0x0c status clear register scr write-only 0x00000000 0x10 interrupt enable register ier write-only 0x00000000 0x14 interrupt disable register idr write-only 0x00000000 0x18 interrupt mask register imr read-only 0x00000000 0x1c wake enable register wer read/write 0x00000000 0x20 alarm register 0 (2) ar0 read/write 0x00000000 0x24 alarm register 1 (2) ar1 read/write 0x00000000 0x30 periodic interval register 0 (2) pir0 read/write 0x00000000 0x34 periodic interval register 1 (2) pir1 read/write 0x00000000 0x40 clock control register clock read/write 0x00000000 0x44 digital tuner register dtr read/write 0x00000000 0x48 event enable eve write-only 0x00000000 0x4c event disable evd write-only 0x00000000 0x50 event mask evm read-only 0x00000000 0x54 calendar value calv read/write 0x00000000 0xf0 parameter register parameter read-only - (1) 0xfc version register version read-only - (1)
273 32145c?06/2013 at32uc3l0128/256 14.6.1 control register name: cr access type: read/write offset: 0x00 reset value: 0x00000000 when the sr.busy bit is set, writes to this register will be discarded and this re gister will read as zero. ? psel: prescaler select selects prescaler bit psel as source clock for the counter. ? can: clear on alarm n 0: the corresponding alarm will not clear the counter. 1: the corresponding alarm will clear the counter. ? cal: calendar mode 0: the ast operates in counter mode. 1: the ast operates in calendar mode. ? pclr: prescaler clear writing a zero to this bit has no effect. writing a one to this bit clears the prescaler. this bit always reads as zero. ? en: enable 0: the ast is disabled. 1: the ast is enabled. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 - - - psel 15 14 13 12 11 10 9 8 ------ca1ca0 76543210 - - - - - cal pclr en
274 32145c?06/2013 at32uc3l0128/256 14.6.2 counter value name: cv access type: read/write offset: 0x04 reset value: 0x00000000 when the sr.busy bit is set, writes to this register will be discarded and this re gister will read as zero. ? value: ast value the current value of the ast counter. 31 30 29 28 27 26 25 24 value[31:24] 23 22 21 20 19 18 17 16 value[23:16] 15 14 13 12 11 10 9 8 value[15:8] 76543210 value[7:0]
275 32145c?06/2013 at32uc3l0128/256 14.6.3 status register name: sr access type: read-only offset: 0x08 reset value: 0x00000000 ? clkrdy: clock ready this bit is cleared when the corresponding bit in scr is written to one. this bit is set when the sr.clkbusy bit has a 1-to-0 transition. ? clkbusy: clock busy 0: the clock is ready and can be changed. 1: clock.cen has been written and the clock is busy. ? ready: ast ready this bit is cleared when the corresponding bit in scr is written to one. this bit is set when the sr.busy bit has a 1-to-0 transition. ? busy: ast busy 0: the ast accepts writes to cr, cv, scr, wer, eve, evd, arn, pirn, and dtr. 1: the ast is busy and will discard writes to cr, cv, scr, wer, eve, evd, arn, pirn, and dtr. ? pern: periodic n this bit is cleared when the corresponding bit in scr is written to one. this bit is set when the selected bit in the prescaler has a 0-to-1 transition. ? alarmn: alarm n this bit is cleared when the corresponding bit in scr is written to one. this bit is set when the counter reaches the selected alarm value. ? ovf: overflow this bit is cleared when the corresponding bit in scr is written to one. this bit is set when an overflow has occurred. 31 30 29 28 27 26 25 24 - - clkrdy clkbusy - - ready busy 23 22 21 20 19 18 17 16 - - - - - - per1 per0 15 14 13 12 11 10 9 8 - - - - - - alarm1 alarm0 76543210 -------ovf
276 32145c?06/2013 at32uc3l0128/256 14.6.4 status clear register name: scr access type: write-only offset: 0x0c reset value: 0x00000000 when the sr.busy bit is set, writes to this register will be discarded. writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in sr and the corresponding interrupt request. 31 30 29 28 27 26 25 24 - - clkrdy - - - ready - 23 22 21 20 19 18 17 16 - - - - - - per1 per0 15 14 13 12 11 10 9 8 - - - - - - alarm1 alarm0 76543210 -------ovf
277 32145c?06/2013 at32uc3l0128/256 14.6.5 interrupt enable register name: ier access type: write-only offset: 0x10 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will set the corresponding bit in imr. 31 30 29 28 27 26 25 24 - - clkrdy - - - ready - 23 22 21 20 19 18 17 16 - - - - - - per1 per0 15 14 13 12 11 10 9 8 - - - - - - alarm1 alarm0 76543210 -------ovf
278 32145c?06/2013 at32uc3l0128/256 14.6.6 interrupt disable register name: idr access type: write-only offset: 0x14 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in imr. 31 30 29 28 27 26 25 24 - - clkrdy - - - ready - 23 22 21 20 19 18 17 16 - - - - - - per1 per0 15 14 13 12 11 10 9 8 - - - - - - alarm1 alarm0 76543210 -------ovf
279 32145c?06/2013 at32uc3l0128/256 14.6.7 interrupt mask register name: imr access type: read-only offset: 0x18 reset value: 0x00000000 0: the corresponding interrupt is disabled. 1: the corresponding interrupt is enabled. a bit in this register is cleared when the corresponding bit in idr is written to one. a bit in this register is set when the corresponding bit in ier is written to one. 31 30 29 28 27 26 25 24 - - clkrdy - - - ready - 23 22 21 20 19 18 17 16 - - - - - - per1 per0 15 14 13 12 11 10 9 8 - - - - - - alarm1 alarm0 76543210 -------ovf
280 32145c?06/2013 at32uc3l0128/256 14.6.8 wake enable register name: wer access type: read/write offset: 0x1c reset value: 0x00000000 when the sr.busy bit is set writes to this regist er will be discarded and this register will read as zero. this register enables the wakeup signal from the ast. ? pern: periodic n 0: the cpu will not wake up from sleep mode when the se lected bit in the prescaler has a 0-to-1 transition. 1: the cpu will wake up from sleep mode when the sele cted bit in the prescaler has a 0-to-1 transition. ? alarmn: alarm n 0: the cpu will not wake up from sleep mode when the counter reaches the selected alarm value. 1: the cpu will wake up from sleep mode when the counter reaches the selected alarm value. ? ovf: overflow 0: a counter overflow will not wake up the cpu from sleep mode. 1: a counter overflow will wake up the cpu from sleep mode. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 - - - - - - per1 per0 15 14 13 12 11 10 9 8 - - - - - - alarm1 alarm0 76543210 -------ovf
281 32145c?06/2013 at32uc3l0128/256 14.6.9 alarm register 0 name: ar0 access type: read/write offset: 0x20 reset value: 0x00000000 when the sr.busy bit is set writes to this regist er will be discarded and this register will read as zero. ? value: alarm value when the counter reaches this value, an alarm is generated. 31 30 29 28 27 26 25 24 value[31:24] 23 22 21 20 19 18 17 16 value[23:16] 15 14 13 12 11 10 9 8 value[15:8] 76543210 value[7:0]
282 32145c?06/2013 at32uc3l0128/256 14.6.10 alarm register 1 name: ar1 access type: read/write offset: 0x24 reset value: 0x00000000 when the sr.busy bit is set writes to this regist er will be discarded and this register will read as zero. ? value: alarm value when the counter reaches this value, an alarm is generated. 31 30 29 28 27 26 25 24 value[31:24] 23 22 21 20 19 18 17 16 value[23:16] 15 14 13 12 11 10 9 8 value[15:8] 76543210 value[7:0]
283 32145c?06/2013 at32uc3l0128/256 14.6.11 periodic interval register 0 name: pir0 access type: read/write offset: 0x30 reset value: 0x00000000 when the sr.busy bit is set writes to this regist er will be discarded and this register will read as zero. ? insel: interval select the per0 bit in sr will be set when the insel bit in the prescaler has a 0-to-1 transition. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 - - - insel
284 32145c?06/2013 at32uc3l0128/256 14.6.12 periodic interval register 1 name: pir1 access type: read/write offset: 0x34 reset value: 0x00000000 when the sr.busy bit is set writes to this regist er will be discarded and this register will read as zero. ? insel: interval select the per1 bit in sr will be set when the insel bit in the prescaler has a 0-to-1 transition. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 - - - insel
285 32145c?06/2013 at32uc3l0128/256 14.6.13 clock control register name: clock access type: read/write offset: 0x40 reset value: 0x00000000 when writing to this register, follow the sequence in section 14.5.1 on page 266 . ? cssel: clock source selection this field defines the clock source clk_ast_prsc for the prescaler: ? cen: clock enable 0: clk_ast_prsc is disabled. 1: clk_ast_prsc is enabled. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 - - - - - cssel 76543210 -------cen table 14-2. clock source selection cssel clock source 0 system rc oscillator (rcsys) 1 32khz oscillator (osc32k) 2 pb clock 3 generic clock (gclk) 4 1khz clock from 32khz oscillator (clk_1k)
286 32145c?06/2013 at32uc3l0128/256 14.6.14 digital tuner register name: dtr access type: read/write offset: 0x44 reset value: 0x00000000 when the sr.busy bit is set writes to this regist er will be discarded and this register will read as zero. ? value: 0: the frequency is unchanged. 1-255: the frequency will be adjusted according to the formula below. ? add: 0: the resulting frequency is for . 1: the resulting frequency is for . ? exp: the frequency will be adjusted according to the formula above. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 value 76543210 - - add exp f 01 1 roundup 256 value ------------------- - ?? ?? 2 exp ?? ? 1 + ------------------------------------------------------------------------------- ? ?? ?? ?? ?? ?? value 0 ? f 01 1 roundup 256 value ------------------- - ?? ?? 2 exp ?? ? 1 ? ------------------------------------------------------------------------------- + ?? ?? ?? ?? ?? value 0 ?
287 32145c?06/2013 at32uc3l0128/256 14.6.15 event enable register name: eve access type: write-only offset: 0x48 reset value: 0x00000000 when the sr.busy bit is set writes to this regist er will be discarded and this register will read as zero. writing a zero to a bit in this register has no effect. writing a one to a bit in this register will set the corresponding bit in evm. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 - - - - - - per1 per0 15 14 13 12 11 10 9 8 - - - - - - alarm1 alarm0 76543210 -------ovf
288 32145c?06/2013 at32uc3l0128/256 14.6.16 event disable register name: evd access type: write-only offset: 0x4c reset value: 0x00000000 when the sr.busy bit is set writes to this regist er will be discarded and this register will read as zero. writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in evm. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 - - - - - - per1 per0 15 14 13 12 11 10 9 8 - - - - - - alarm1 alarm0 76543210 -------ovf
289 32145c?06/2013 at32uc3l0128/256 14.6.17 event mask register name: evm access type: read-only offset: 0x50 reset value: 0x00000000 0: the corresponding peripheral event is disabled. 1: the corresponding peripheral event is enabled. this bit is cleared when the corresponding bit in evd is written to one. this bit is set when the corresponding bit in eve is written to one. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 - - - - - - per1 per0 15 14 13 12 11 10 9 8 - - - - - - alarm1 alarm0 76543210 -------ovf
290 32145c?06/2013 at32uc3l0128/256 14.6.18 calendar value name: calv access type: read/write offset: 0x54 reset value: 0x00000000 when the sr.busy bit is set writes to this regist er will be discarded and this register will read as zero. ? year: year current year. the year is considered a leap year if year[1:0] = 0. ?month: month 1 = january 2 = february ... 12 = december ?day: day day of month, starting with 1. ? hour: hour hour of day, in 24-hour clock format. legal values are 0 through 23. ? min: minute minutes, 0 through 59. ? sec: second seconds, 0 through 59. 31 30 29 28 27 26 25 24 year month[3:2] 23 22 21 20 19 18 17 16 month[1:0] day hour[4] 15 14 13 12 11 10 9 8 hour[3:0] min[5:2] 76543210 min[1:0] sec
291 32145c?06/2013 at32uc3l0128/256 14.6.19 parameter register name: parameter access type: read-only offset: 0xf0 reset value: - this register gives the configuration used in the specific device. also refe r to the module configuration section. ? dt: digital tuner 0: digital tuner not implemented. 1: digital tuner implemented. ? dtrexpwa: digital tuner exponent writeable 0: digital tuner exponent is a constant value. writes to exp field in dtr will be discarded. 1: digital tuner exponent is chos en by writing to exp field in dtr. ? dtrexpvalue: digital tuner exponent value digital tuner exp onent value if dtexpwa is zero. ? numar: number of alarm comparators 0: zero alarm comparators. 1: one alarm comparator. 2: two alarm comparators. ? numpir: number of periodic comparators 0: one periodic comparator. 1: two periodic comparator. ? pirnwa: periodic interval n writeable 0: periodic interval n prescaler tapping is a constant value. writes to insel field in pirn register will be discarded. 1: periodic interval n prescaler tapping is chos en by writing to insel field in pirn register. ? pernvalue: periodic interval n value periodic interval prescaler n tapping if pirnwa is zero. 31 30 29 28 27 26 25 24 - - - per1value 23 22 21 20 19 18 17 16 - - - per0value 15 14 13 12 11 10 9 8 pir1wa pir0wa - numpir - - numar 76543210 - dtexpvalue dtexpwa dt
292 32145c?06/2013 at32uc3l0128/256 14.6.20 version register name: version access type: read-only offset: 0xfc reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
293 32145c?06/2013 at32uc3l0128/256 14.7 module configuration the specific configuration for each ast instance is listed in the following tables.the module bus clocks listed here are connecte d to the system bus clocks. pleas e refer to the power manager chapter for details. table 14-3. ast configuration feature ast number of alarm comparators 1 number of periodic comparators 1 digital tuner on table 14-4. ast clocks clock name description clk_ast clock for the ast bus interface gclk the generic clock used for the ast is gclk2 pb clock peripheral bus clock from the pba clock domain table 14-5. register reset values register reset value version 0x00000310 parameter 0x00004103
294 32145c?06/2013 at32uc3l0128/256 15. watchdog timer (wdt) rev: 4.1.0.0 15.1 features ? watchdog timer counter with 32-bit counter ? timing window watchdog ? clocked from system rc oscillator or the 32 khz crystal oscillator ? configuration lock ? wdt may be enabled at reset by a fuse 15.2 overview the watchdog timer (wdt) will reset the device unless it is periodically serviced by the soft- ware. this allows the device to recover from a condition that has caused the system to be unstable. the wdt has an internal counter clocked from th e system rc oscillator or the 32 khz crystal oscillator. the wdt counter must be periodically cleared by software to avoid a watchdog reset. if the wdt timer is not cleared correctly, the device will reset and start executing from the boot vector. 15.3 block diagram figure 15-1. wdt block diagram 15.4 product dependencies in order to use this module, other parts of the system must be configured correctly, as described below. clk_cnt watchdog detector 32-bit counter watchdog reset 0 1 rcsys osc32k cssel cen sync clk_cnt domain pb clock domain pb wdtclr window, cleared en, mode, psel, tban ctrl clr sr
295 32145c?06/2013 at32uc3l0128/256 15.4.1 power management when the wdt is enabled, the wdt remains clocked in all sleep modes. it is not possible to enter sleep modes where the source clock of clk_cnt is stopped. atte mpting to do so will result in the device entering the lowest sleep mode where the source clock is running, leaving the wdt operational. please refer to the power manager chapter for details about sleep modes. after a watchdog reset the wdt bit in the reset cause register (rcause) in the power man- ager will be set. 15.4.2 clocks the clock for the wdt bus interface (clk_wdt) is generated by the power manager. this clock is enabled at reset, and can be disabled in the power manager. it is recommended to dis- able the wdt before disabling the clock, to avoid freezing the wdt in an undefined state. there are two possible clock sources for the watchdog timer (clk_cnt): ? system rc oscillator (rcsys): th is oscillator is always enabl ed when selected as clock source for the wdt. please re fer to the power manager chapter for details about the rcsys and sleep modes. please refer to the electrical characteristics chapter for the characteristic frequency of this oscillator. ? 32 khz crystal oscillator (osc32k) : this oscillator has to be en abled in the system control interface before using it as cloc k source for the wdt. the wdt will not be able to detect if this clock is stopped. 15.4.3 debug operation the wdt counter is frozen during debug operation, unless the run in debug bit in the develop- ment control register is set and the bit corresponding to the wdt is set in the peripheral debug register (pdbg). please refer to the on-chip debug chapter in the avr32uc technical refer- ence manual, and the ocd module configuration section, for details. if the wdt counter is not frozen during debug operation it will need periodically clearing to avoid a watchdog reset. 15.4.4 fuses the wdt can be enabled at reset. this is controlled by the wdtauto fuse, see section 15.5.4 for details. please refer to the fuse settings se ction in the flash controller chapter for details about wdtauto and how to program the fuses. 15.5 functional description 15.5.1 basic mode 15.5.1.1 wdt control register access to avoid accidental disabling of the watchdog, the control register (ctrl) must be written twice, first with the key field set to 0x55, t hen 0xaa without changing the other bits. failure to do so will cause the write operat ion to be ignored, an d the value in the ctrl register will not be changed. 15.5.1.2 changing clk_cnt clock source after any reset, except for watchdog reset, clk_cnt will be enabled with the rcsys as source.
296 32145c?06/2013 at32uc3l0128/256 to change the clock for the wdt the following steps need to be taken. note that the wdt should always be disabled before changing the clk_cnt source: 1. write a zero to the clock enable (cen) bit in the ctrl register, leaving the other bits as they are in the ctrl register. this will stop clk_cnt. 2. read back the ctrl register until the cen bit reads zero. the clock has now been stopped. 3. modify the clock source select (cssel) bit in the ctrl register wit h your new clock selec- tion and write it to the ctrl register. 4. write a one to the cen bit, leaving the other bits as they are in the ctrl register. this will enable the clock. 5. read back the ctrl register until the cen bit reads one. the clock has now been enabled. 15.5.1.3 configuring the wdt if the mode bit in the ctrl register is zero, the wdt is in basic mode. the time out prescale select (psel) field in the ctrl regi ster selects the wdt timeout period: t timeout = t psel = 2 (psel+1) / f clk_cnt 15.5.1.4 enabling the wdt to enable the wdt write a one to the enable (en) bit in the ctrl register. due to internal syn- chronization, it will take some time for the ctrl.en bit to read back as one. 15.5.1.5 clearing the wdt counter the wdt counter is cleared by writing a one to the watchdog clear (wdtclr) bit in the clear (clr) register, at any correct write to the ctrl register, or when the counter reaches t timeout and the device is reset. in basic mode the clr.wdtclr can be written at any time when the wdt counter cleared (cleared) bit in the status register (sr) is one. due to internal syn- chronization, clearing the wdt counter takes some time. the sr.cleared bit is cleared when writing to clr.wdtclr bit and set when the clearing is done. any write to the clr.wdtclr bit while sr.cleared is ze ro will not clear the counter. writing to the clr.wdtclr bit has to be done in a particular sequence to be valid. the clr register must be written twice, first with the key field set to 0x 55 and wdtclr set to one, then a second write with the key set to 0xaa without changing the wdtclr bit. writing to the clr register without the correct sequence has no effect. if the wdt counter is periodically cleared within t psel no watchdog rese t will be issued, see fig- ure 15-2 on page 297 .
297 32145c?06/2013 at32uc3l0128/256 figure 15-2. basic mode wdt timing diagram, normal operation. if the wdt counter is not cleared within t psel a watchdog reset will be issued at the end of t psel , see figure 15-3 on page 297 . figure 15-3. basic mode wdt timing diagram, no clear within t psel . 15.5.1.6 watchdog reset a watchdog reset will result in a reset and the code will start executing from the boot vector, please refer to the power manager chapter for details. if the disable after reset (dar) bit in the ctrl register is zero, the wdt counter will restart counting from zero when the watchdog reset is released. if the ctrl.dar bit is one the wdt will be di sabled after a watchdog re set. only the ctrl.en bit will be changed after the watc hdog reset. however, if wdtauto fuse is configured to enable the wdt after a watchdog reset, and the ctrl.fcd bit is zero, writing a one to the ctrl.dar bit will have no effect. 15.5.2 window mode the window mode can protect against tight loops of runaway code. this is obtained by adding a ban period to timeout period. during the ban period clearing the wdt counter is not allowed. if the wdt mode (mode) bit in the ctrl register is one, the wdt is in window mode. note that the ctrl.mode bit can only be change d when the wdt is di sabled (ctrl.en=0). t psel timeout write one to clr.wdtclr watchdog reset t=t 0 t psel timeout write one to clr.wdtclr watchdog reset t=t 0
298 32145c?06/2013 at32uc3l0128/256 the psel and time ban prescale select (tban) fields in the ctrl register selects the wdt timeout period t timeout = t tban + t psel = (2 (tban+1) + 2 (psel+1) ) / f clk_cnt where t tban sets the time period when clearing the wdt counter by writing to the clr.wdtclr bit is not allowed. doing so will result in a watc hdog reset, the device w ill receive a reset and the code will start executing form the boot vector, see figure 15-5 on page 298 . the wdt counter will be cleared. writing a one to the clr.wdtclr bit within the t psel period will clear the wdt counter and the counter starts counting from zero (t=t 0 ), entering t tban , see figure 15-4 on page 298 . if the value in the ctrl register is changed , the wdt counter will be cleared without a watch- dog reset, regardless of if the value in the wdt counter and the tban value. if the wdt counter reaches t timeout , the counter will be cleared, the device will receive a reset and the code will start execut ing form the boot vector. figure 15-4. window mode wdt timing diagram figure 15-5. window mode wdt timing diagram, clearing within t tban , resulting in watchdog reset. t tban t psel timeout write one to clr.wdtclr watchdog reset t=t 0 t tban t psel timeout write one to clr.wdtclr watchdog reset t=t 0
299 32145c?06/2013 at32uc3l0128/256 15.5.3 disabling the wdt the wdt is disabled by writing a zero to the ctrl.en bit. when disabling the wdt no other bits in the ctrl register should be changed until the ctrl.en bit reads back as zero. if the ctrl.cen bit is written to zero, the ctrl.en bit will never read back as zero if changing the value from one to zero. 15.5.4 flash calibration the wdt can be enabled at rese t. this is controlled by the wdtauto fuse. the wdt will be set in basic mode, rcsys is set as source fo r clk_cnt, and psel will be set to a value giving t psel above 100 ms. please refer to the fuse settings chapter for details about wdtauto and how to program the fuses. if the flash calibration done (fcd) bit in the ctrl register is zero at a watchdog reset the flash calibration will be redone, and the ctrl.fcd bi t will be set when the calibration is done. if ctrl.fcd is one at a watchdog reset, the conf iguration of the wdt will not be changed during flash calibration. after any other reset the fl ash calibration will always be done, and the ctrl.fcd bit will be set when the calibration is done. 15.5.5 special considerations care must be taken when selecting the psel/tban va lues so that the time out period is greater than the startup time of the de vice. otherwise a watchdog reset will reset the devi ce before any code has been run. this can also be avoided by writing the ctrl.dar bit to one when configur- ing the wdt. if the store final value (sfv) bit in the ctrl register is one, the ctrl register is locked for further write accesses. all writes to the ctrl register will be i gnored. once the ctrl register is locked, it can only be unlocked by a reset (e.g. por, ocd, and wdt). the ctrl.mode bit can only be changed when the wdt is disabled (ctrl.en=0).
300 32145c?06/2013 at32uc3l0128/256 15.6 user interface note: 1. the reset value for this register is device specific. plea se refer to the module configur ation section at the end of thi s chapter. table 15-1. wdt register memory map offset register register name access reset 0x000 control register ctrl read/write 0x00010080 0x004 clear register clr write-only 0x00000000 0x008 status register sr read-only 0x00000003 0x3fc version register version read-only - (1)
301 32145c?06/2013 at32uc3l0128/256 15.6.1 control register name: ctrl access type: read/write offset: 0x000 reset value: 0x00010080 ?key this field must be written twice, first with key value 0x55, t hen 0xaa, for a write operation to be effective. this field alway s reads as zero. ? tban: time ban prescale select counter bit tban is used as watchdog ?banned? time frame. in this time frame clearing the wdt timer is forbidden, otherwise a watchdog reset is generated and the wdt timer is cleared. ? cssel: clock source select 0: select the system rc oscillator (rcsys) as clock source. 1: select the 32khz crystal osc illator (osc32k) as clock source. ? cen: clock enable 0: the wdt clock is disabled. 1: the wdt clock is enabled. ? psel: time out prescale select counter bit psel is used as watchdog timeout period. ? fcd: flash calibration done this bit is set after any reset. 0: the flash calibration will be redone after a watchdog reset. 1: the flash calibration will not be redone after a watchdog reset. ? sfv: wdt control register store final value 0: wdt control register is not locked. 1: wdt control register is locked. once locked, the control register can not be re-written, only a reset unlocks the sfv bit. ?mode: wdt mode 0: the wdt is in basic mode, only psel time is used. 1: the wdt is in window mode. tota l timeout period is now tban+psel. writing to this bit when the wdt is enabled has no effect. 31 30 29 28 27 26 25 24 key 23 22 21 20 19 18 17 16 - tban cssel cen 15 14 13 12 11 10 9 8 - - - psel 76543210 fcd - - - sfv mode dar en
302 32145c?06/2013 at32uc3l0128/256 ? dar: wdt disable after reset 0: after a watchdog reset, the wdt will still be enabled. 1: after a watchdog reset, the wdt will be disabled. ? en: wdt enable 0: wdt is disabled. 1: wdt is enabled. after writing to this bit the read back value will not change until the wdt is enabled/disabled. this due to internal synchronization.
303 32145c?06/2013 at32uc3l0128/256 15.6.2 clear register name: clr access type: write-only offset: 0x004 reset value: 0x00000000 when the watchdog timer is enabled, this register must be periodically written within the window time frame or within the watchdog timeout period, to prevent a watchdog reset. ?key this field must be written twice, first with key value 0x55, then 0xaa, for a write operation to be effective. ? wdtclr: watchdog clear writing a zero to this bit has no effect. writing a one to this bit clears the wdt counter. 31 30 29 28 27 26 25 24 key 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 -------wdtclr
304 32145c?06/2013 at32uc3l0128/256 15.6.3 status register name: sr access type: read-only offset: 0x008 reset value: 0x00000003 ? cleared: wdt counter cleared this bit is cleared when writing a one to the clr.wdtclr bit. this bit is set when clearing the wdt counter is done. ? window: within window this bit is cleared when the wdt counter is inside the tban period. this bit is set when the wdt counter is inside the psel period. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 ------clearedwindow
305 32145c?06/2013 at32uc3l0128/256 15.6.4 version register name: version access type: read-only offset: 0x3fc reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
306 32145c?06/2013 at32uc3l0128/256 15.7 module configuration the specific configuration for each wdt instance is listed in the following tables.the module bus clocks listed here are connecte d to the system bus clocks. pleas e refer to the power manager chapter for details. table 15-2. wdt clocks clock name description clk_wdt clock for the wdt bus interface table 15-3. register reset values register reset value version 0x00000410
307 32145c?06/2013 at32uc3l0128/256 16. external interrupt controller (eic) rev: 3.0.2.0 16.1 features ? dedicated interrupt requ est for each interrupt ? individually maskable interrupts ? interrupt on rising or falling edge ? interrupt on high or low level ? asynchronous interrupts fo r sleep modes without clock ? filtering of interrupt lines ? non-maskable nmi interrupt 16.2 overview the external interrupt controller (eic) allows pins to be configured as external interrupts. each external interrupt has its own interrupt request and can be individually masked. each external interrupt can generate an interrupt on rising or falling edge, or high or low level. every interrupt input has a configurable filter to remove spikes from the interrupt source. every interrupt pin can also be configured to be asynchronous in order to wake up the part from sleep modes where the clk_sync clock has been disabled. a non-maskable interrupt (nmi) is also supported. this has the same properties as the other external interrupts, but is connected to the nmi request of the cpu, enabling it to interrupt any other interrupt mode. the eic can wake up the part from sleep modes without triggering an interrupt. in this mode, code execution starts from the instruction following the sleep instruction. 16.3 block diagram figure 16-1. eic block diagram edge/level detector mask irqn extintn nmi intn level mode edge ier idr icr ctrl isr im r filter filter polarity control level mode edge asynchronus detector eic_wake enable en dis ctrl clk_sync wake detect async
308 32145c?06/2013 at32uc3l0128/256 16.4 i/o lines description 16.5 product dependencies in order to use this module, other parts of the system must be configured correctly, as described below. 16.5.1 i/o lines the external interrupt pins (extintn and nmi) may be multiplexed with i/o controller lines. the programmer must first program the i/o controller to assign the desired eic pins to their periph- eral function. if i/o lines of the eic are not used by the application, they can be used for other purposes by the i/o controller. it is only required to enable the eic inputs actually in use. if an application requires two external interrupts, then only two i/o lines will be a ssigned to eic inputs. 16.5.2 power management all interrupts are available in all sleep modes as long as the eic module is powered. however, in sleep modes where clk_sync is stopped, the interrupt must be configured to asynchronous mode. 16.5.3 clocks the clock for the eic bus interface (clk_eic) is generated by the power manager. this clock is enabled at reset, and can be disabled in the power manager. the filter and synchronous edge/level detector runs on a clock which is stopped in any of the sleep modes where the system rc osc illator (rcsys) is not running. this clock is referred to as clk_sync. 16.5.4 interrupts the external interrupt request lines are connected to the interrupt controller. using the external interrupts requires the interrupt controller to be programmed first. using the non-maskable interrupt does not require the interrupt controller to be programmed. 16.5.5 debug operation when an external debugger forces the cpu into debug mode, the eic continues normal opera- tion. if the eic is configured in a way that r equires it to be periodically serviced by the cpu through interrupts or similar, improper operation or data loss may result during debugging. 16.6 functional description 16.6.1 external interrupts the external interrupts are not enabled by default, allowing the proper interrupt vectors to be set up by the cpu before the interrupts are enabled. table 16-1. i/o lines description pin name pin description type nmi non-maskable interrupt input extintn external interrupt input
309 32145c?06/2013 at32uc3l0128/256 each external interrupt intn can be configured to prod uce an interrup t on rising or falling edge, or high or low level. external interrupts ar e configured by the mode, edge, and level regis- ters. each interrupt has a bit intn in each of these registers. writing a zero to the intn bit in the mode register enables edge triggered interrupts, while writing a one to the bit enables level trig- gered interrupts. if intn is configured as an edge triggered interrupt, writing a zero to the intn bit in the edge register will cause the interrupt to be triggered on a falling edg e on extintn, while writing a one to the bit will cause the in terrupt to be triggered on a rising edge on extintn. if intn is configured as a leve l triggered interrupt, writing a zero to the intn bit in the level register will cause the interrupt to be triggered on a low level on extintn, while writing a one to the bit will cause the interrupt to be triggered on a high level on extintn. each interrupt has a corresponding bit in each of the interrupt control and status registers. writ- ing a one to the intn bit in the interrupt enabl e register (ier) enables the external interrupt from pin extintn to propagate from the eic to the interrupt controller, while writing a one to intn bit in the interrupt disable register (idr) disables this propagation. the interrupt mask register (imr) can be read to check which interrupts are enabled. when an interrupt triggers, the corresponding bit in th e interrupt status register (isr) will be set. this bit remains set until a one is written to the corresponding bit in the interrupt clear register (icr) or the interrupt is disabled. writing a one to the intn bit in the enable regist er (en) enables the external interrupt on pin extintn, while writing a one to intn bit in the disable register (dis) disables the external inter- rupt. the control register (ctrl) can be read to check which interrupts are enabled. if a bit in the ctrl register is set, but the corresponding bit in imr is not set, an interrupt will not propa- gate to the interrupt controller. however, the corresponding bit in isr will be set, and eic_wake will be set. note that an external inte rrupt should not be en abled before it has been configured correctly. if the ctrl.intn bit is zero, the corresponding bi t in isr will always be ze ro. disabling an exter- nal interrupt by writing a one to the dis.intn bit will clear the corresponding bit in isr. please refer to the module configuration section for the number of external interrupts. 16.6.2 synchronization and filtering of external interrupts in synchronous mode the pin value of the extintn pin is synchronized to clk_sync, so spikes shorter than one clk_sync cycle are not guaranteed to produce an interrupt. the syn- chronization of the extintn to clk_sync will delay the propagation of the interrupt to the interrupt controller by two cycles of clk_sync, see figure 16-2 and figure 16-3 for examples (filter off). it is also possible to apply a filter on extintn by writing a one to the intn bit in the filter reg- ister. this filter is a majority voter, if the condition for an interrupt is true for more than one of the latest three cycles of clk_sync the interrup t will be set. this will additionally delay the propa- gation of the interrupt to the interrupt controller by one or two cycles of clk_sync, see figure 16-2 and figure 16-3 for examples (filter on).
310 32145c?06/2013 at32uc3l0128/256 figure 16-2. timing diagram, synchronous interr upts, high level or rising edge figure 16-3. timing diagram, synchrono us interrupts, low level or falling edge 16.6.3 non-maskable interrupt the nmi supports the same features as the external interrupts, and is accessed through the same registers. th e description in section 16.6.1 should be followed, accessing the nmi bit instead of the intn bits. the nmi is non-maskable within the cpu in the sense that it can interrupt any other execution mode. still, as for the other exte rnal interrupts, the actual nmi input can be enabled an d disabled by accessing the registers in the eic. 16.6.4 asynchronous interrupts each external inte rrupt can be made asynchronous by wr iting a one to intn in the async reg- ister. this will route the interrupt signal through the asynchronous path of the module. all edge interrupts will be interpreted as leve l interrupts and the f ilter is disabled. if an interrupt is config- ured as edge triggered interr upt in asynchronous mode, a zero in edge.intn will be interpreted as low level, and a one in edge.intn will be inte rpreted as high level. eic_wake will be set immediately after the source triggers the interrupt, while the correspond- ing bit in isr and the interrupt to the interrupt controller will be set on the next rising edge of clk_sync. please refer to figure 16-4 on page 311 for details. extintn/nmi clk_sync isr.intn: filter off isr.intn: filter on extintn/nmi clk_sync isr.intn: filter off isr.intn: filter on
311 32145c?06/2013 at32uc3l0128/256 when clk_sync is stopped only asynchronous interrupts remain active, and any short spike on this interrupt will wake up the device . eic_wake will restart clk_sync and isr will be updated on the first rising edge of clk_sync. figure 16-4. timing diagram, asynchronous interrupts 16.6.5 wakeup the external interrupts can be used to wake up the part from sleep modes. the wakeup can be interpreted in two ways. if the corresponding bit in imr is one, then the execution starts at the interrupt handler for this interrupt. if the bit in imr is zero, then the execution starts from the next instruction after the sleep instruction. extintn/nmi clk_sync is r .in t n: rising edge or high level eic_wake: rising edge or high level extintn/nmi clk_sync is r .in tn: rising edge or high level eic_wake: rising edge or high level
312 32145c?06/2013 at32uc3l0128/256 16.7 user interface note: 1. the reset value is device specific. please refer to th e module configuration section at the end of this chapter. table 16-2. eic register memory map offset register register name access reset 0x000 interrupt enable register ier write-only 0x00000000 0x004 interrupt disable register idr write-only 0x00000000 0x008 interrupt mask register imr read-only 0x00000000 0x00c interrupt status register isr read-only 0x00000000 0x010 interrupt clear register icr write-only 0x00000000 0x014 mode register mode read/write 0x00000000 0x018 edge register edge read/write 0x00000000 0x01c level register level read/write 0x00000000 0x020 filter register filter read/write 0x00000000 0x024 test register test read/write 0x00000000 0x028 asynchronous register async read/write 0x00000000 0x030 enable register en write-only 0x00000000 0x034 disable register dis write-only 0x00000000 0x038 control register ctrl read-only 0x00000000 0x3fc version register version read-only - (1)
313 32145c?06/2013 at32uc3l0128/256 16.7.1 interrupt enable register name: ier access type: write-only offset: 0x000 reset value: 0x00000000 ?intn: external interrupt n writing a zero to this bit has no effect. writing a one to this bit will set the corresponding bit in imr. please refer to the module configuration se ction for the number of external interrupts. ? nmi: non-maskable interrupt writing a zero to this bit has no effect. wrting a one to this bit will set the corresponding bit in imr. 31 30 29 28 27 26 25 24 - int30 int29 int28 int27 int26 int25 int24 23 22 21 20 19 18 17 16 int23int22int21int20int19int18int17int16 15 14 13 12 11 10 9 8 int15int14int13int12int11int10 int9 int8 76543210 int7 int6 int5 int4 int3 int2 int1 nmi
314 32145c?06/2013 at32uc3l0128/256 16.7.2 interrupt disable register name: idr access type: write-only offset: 0x004 reset value: 0x00000000 ?intn: external interrupt n writing a zero to this bit has no effect. writing a one to this bit will clear the corresponding bit in imr. please refer to the module configuration se ction for the number of external interrupts. ? nmi: non-maskable interrupt writing a zero to this bit has no effect. writing a one to this bit will clear the corresponding bit in imr. 31 30 29 28 27 26 25 24 - int30 int29 int28 int27 int26 int25 int24 23 22 21 20 19 18 17 16 int23int22int21int20int19int18int17int16 15 14 13 12 11 10 9 8 int15int14int13int12int11int10 int9 int8 76543210 int7 int6 int5 int4 int3 int2 int1 nmi
315 32145c?06/2013 at32uc3l0128/256 16.7.3 interrupt mask register name: imr access type: read-only offset: 0x008 reset value: 0x00000000 ?intn: external interrupt n 0: the corresponding interrupt is disabled. 1: the corresponding interrupt is enabled. this bit is cleared when the corresponding bit in idr is written to one. this bit is set when the corresponding bit in ier is written to one. please refer to the module configuration se ction for the number of external interrupts. ? nmi: non-maskable interrupt 0: the non-maskable in terrupt is disabled. 1: the non-maskable interrupt is enabled. this bit is cleared when the corresponding bit in idr is written to one. this bit is set when the corresponding bit in ier is written to one. 31 30 29 28 27 26 25 24 - int30 int29 int28 int27 int26 int25 int24 23 22 21 20 19 18 17 16 int23int22int21int20int19int18int17int16 15 14 13 12 11 10 9 8 int15int14int13int12int11int10 int9 int8 76543210 int7 int6 int5 int4 int3 int2 int1 nmi
316 32145c?06/2013 at32uc3l0128/256 16.7.4 interrupt status register name: isr access type: read-only offset: 0x00c reset value: 0x00000000 ?intn: external interrupt n 0: an interrupt event has not occurred. 1: an interrupt event has occurred. this bit is cleared by writing a one to the corresponding bit in icr. please refer to the module configuration se ction for the number of external interrupts. ? nmi: non-maskable interrupt 0: an interrupt event has not occurred. 1: an interrupt event has occurred. this bit is cleared by writing a one to the corresponding bit in icr. 31 30 29 28 27 26 25 24 - int30 int29 int28 int27 int26 int25 int24 23 22 21 20 19 18 17 16 int23int22int21int20int19int18int17int16 15 14 13 12 11 10 9 8 int15int14int13int12int11int10 int9 int8 76543210 int7 int6 int5 int4 int3 int2 int1 nmi
317 32145c?06/2013 at32uc3l0128/256 16.7.5 interrupt clear register name: icr access type: write-only offset: 0x010 reset value: 0x00000000 ?intn: external interrupt n writing a zero to this bit has no effect. writing a one to this bit will clear the corresponding bit in isr. please refer to the module configuration se ction for the number of external interrupts. ? nmi: non-maskable interrupt writing a zero to this bit has no effect. writing a one to this bit will clear the corresponding bit in isr. 31 30 29 28 27 26 25 24 - int30 int29 int28 int27 int26 int25 int24 23 22 21 20 19 18 17 16 int23int22int21int20int19int18int17int16 15 14 13 12 11 10 9 8 int15int14int13int12int11int10 int9 int8 76543210 int7 int6 int5 int4 int3 int2 int1 nmi
318 32145c?06/2013 at32uc3l0128/256 16.7.6 mode register name: mode access type: read/write offset: 0x014 reset value: 0x00000000 ?intn: external interrupt n 0: the external interrupt is edge triggered. 1: the external interrupt is level triggered. please refer to the module configuration se ction for the number of external interrupts. ? nmi: non-maskable interrupt 0: the non-maskable interrupt is edge triggered. 1: the non-maskable interrupt is level triggered. 31 30 29 28 27 26 25 24 - int30 int29 int28 int27 int26 int25 int24 23 22 21 20 19 18 17 16 int23int22int21int20int19int18int17int16 15 14 13 12 11 10 9 8 int15int14int13int12int11int10 int9 int8 76543210 int7 int6 int5 int4 int3 int2 int1 nmi
319 32145c?06/2013 at32uc3l0128/256 16.7.7 edge register name: edge access type: read/write offset: 0x018 reset value: 0x00000000 ?intn: external interrupt n 0: the external interrupt triggers on falling edge. 1: the external interrupt triggers on rising edge. please refer to the module configuration se ction for the number of external interrupts. ? nmi: non-maskable interrupt 0: the non-maskable interrupt triggers on falling edge. 1: the non-maskable interrupt triggers on rising edge. 31 30 29 28 27 26 25 24 - int30 int29 int28 int27 int26 int25 int24 23 22 21 20 19 18 17 16 int23int22int21int20int19int18int17int16 15 14 13 12 11 10 9 8 int15int14int13int12int11int10 int9 int8 76543210 int7 int6 int5 int4 int3 int2 int1 nmi
320 32145c?06/2013 at32uc3l0128/256 16.7.8 level register name: level access type: read/write offset: 0x01c reset value: 0x00000000 ?intn: external interrupt n 0: the external interrupt triggers on low level. 1: the external interrupt triggers on high level. please refer to the module configuration se ction for the number of external interrupts. ? nmi: non-maskable interrupt 0: the non-maskable interrupt triggers on low level. 1: the non-maskable interrupt triggers on high level. 31 30 29 28 27 26 25 24 - int30 int29 int28 int27 int26 int25 int24 23 22 21 20 19 18 17 16 int23int22int21int20int19int18int17int16 15 14 13 12 11 10 9 8 int15int14int13int12int11int10 int9 int8 76543210 int7 int6 int5 int4 int3 int2 int1 nmi
321 32145c?06/2013 at32uc3l0128/256 16.7.9 filter register name: filter access type: read/write offset: 0x020 reset value: 0x00000000 ?intn: external interrupt n 0: the external interrupt is not filtered. 1: the external interrupt is filtered. please refer to the module configuration se ction for the number of external interrupts. ? nmi: non-maskable interrupt 0: the non-maskable interrupt is not filtered. 1: the non-maskable interrupt is filtered. 31 30 29 28 27 26 25 24 - int30 int29 int28 int27 int26 int25 int24 23 22 21 20 19 18 17 16 int23int22int21int20int19int18int17int16 15 14 13 12 11 10 9 8 int15int14int13int12int11int10 int9 int8 76543210 int7 int6 int5 int4 int3 int2 int1 nmi
322 32145c?06/2013 at32uc3l0128/256 16.7.10 test register name: test access type: read/write offset: 0x024 reset value: 0x00000000 ? testen: test enable 0: this bit disables external interrupt test mode. 1: this bit enables extern al interrupt test mode. ?intn: external interrupt n writing a zero to this bit will set the input value to intn to zero, if test mode is enabled. writing a one to this bit will set the input value to intn to one, if test mode is enabled. please refer to the module configuration se ction for the number of external interrupts. ? nmi: non-maskable interrupt writing a zero to this bit will set the input value to nmi to zero, if test mode is enabled. writing a one to this bit will set the input value to nmi to one, if test mode is enabled. if testen is 1, the value written to this bit will be the value to the interrupt detector and the value on the pad will be igno red. 31 30 29 28 27 26 25 24 testenint30int29int28int27int26int25int24 23 22 21 20 19 18 17 16 int23int22int21int20int19int18int17int16 15 14 13 12 11 10 9 8 int15int14int13int12int11int10 int9 int8 76543210 int7 int6 int5 int4 int3 int2 int1 nmi
323 32145c?06/2013 at32uc3l0128/256 16.7.11 asynchronous register name: async access type: read/write offset: 0x028 reset value: 0x00000000 ?intn: external interrupt n 0: the external interrupt is synchronized to clk_sync. 1: the external interrupt is asynchronous. please refer to the module configuration se ction for the number of external interrupts. ? nmi: non-maskable interrupt 0: the non-maskable interrupt is synchronized to clk_sync. 1: the non-maskable interrupt is asynchronous. 31 30 29 28 27 26 25 24 - int30 int29 int28 int27 int26 int25 int24 23 22 21 20 19 18 17 16 int23int22int21int20int19int18int17int16 15 14 13 12 11 10 9 8 int15int14int13int12int11int10 int9 int8 76543210 int7 int6 int5 int4 int3 int2 int1 nmi
324 32145c?06/2013 at32uc3l0128/256 16.7.12 enable register name: en access type: write-only offset: 0x030 reset value: 0x00000000 ?intn: external interrupt n writing a zero to this bit has no effect. writing a one to this bit will enable the corresponding external interrupt. please refer to the module configuration se ction for the number of external interrupts. ? nmi: non-maskable interrupt writing a zero to this bit has no effect. writing a one to this bit will enable the non-maskable interrupt. 31 30 29 28 27 26 25 24 - int30 int29 int28 int27 int26 int25 int24 23 22 21 20 19 18 17 16 int23int22int21int20int19int18int17int16 15 14 13 12 11 10 9 8 int15int14int13int12int11int10 int9 int8 76543210 int7 int6 int5 int4 int3 int2 int1 nmi
325 32145c?06/2013 at32uc3l0128/256 16.7.13 disable register name: dis access type: write-only offset: 0x034 reset value: 0x00000000 ?intn: external interrupt n writing a zero to this bit has no effect. writing a one to this bit will disable the corresponding external interrupt. please refer to the module configuration se ction for the number of external interrupts. ? nmi: non-maskable interrupt writing a zero to this bit has no effect. writing a one to this bit will disable the non-maskable interrupt. 31 30 29 28 27 26 25 24 - int30 int29 int28 int27 int26 int25 int24 23 22 21 20 19 18 17 16 int23int22int21int20int19int18int17int16 15 14 13 12 11 10 9 8 int15int14int13int12int11int10 int9 int8 76543210 int7 int6 int5 int4 int3 int2 int1 nmi
326 32145c?06/2013 at32uc3l0128/256 16.7.14 control register name: ctrl access type: read-only offset: 0x038 reset value: 0x00000000 ?intn: external interrupt n 0: the corresponding external interrupt is disabled. 1: the corresponding external interrupt is enabled. please refer to the module configuration se ction for the number of external interrupts. ? nmi: non-maskable interrupt 0: the non-maskable in terrupt is disabled. 1: the non-maskable interrupt is enabled. 31 30 29 28 27 26 25 24 - int30 int29 int28 int27 int26 int25 int24 23 22 21 20 19 18 17 16 int23int22int21int20int19int18int17int16 15 14 13 12 11 10 9 8 int15int14int13int12int11int10 int9 int8 76543210 int7 int6 int5 int4 int3 int2 int1 nmi
327 32145c?06/2013 at32uc3l0128/256 16.7.15 version register name: version access type: read-only offset: 0x3fc reset value: - ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
328 32145c?06/2013 at32uc3l0128/256 16.8 module configuration the specific configuration for each eic instance is listed in the following tables.the module bus clocks listed here are connecte d to the system bus clocks. pleas e refer to the power manager chapter for details. table 16-3. eic configuration feature eic number of external interrupts, including nmi 6 table 16-4. eic clocks clock name description clk_eic clock for the eic bus interface table 16-5. register reset values register reset value version 0x00000302
329 32145c?06/2013 at32uc3l0128/256 17. frequency meter (freqm) rev: 3.1.0.1 17.1 features ? accurately measures a clock frequency ? selectable reference clock ? a selectable clock can be measured ? ratio can be measured with 24-bit accuracy 17.2 overview the frequency meter (freqm) can be used to accurately measure the frequency of a clock by comparing it to a known reference clock. 17.3 block diagram figure 17-1. frequency meter block diagram 17.4 product dependencies in order to use this module, other parts of the system must be configured correctly, as described below. 17.4.1 power management the device can enter a sleep mode while a measurement is ongoing. however, make sure that neither clk_msr nor clk_ref is stopped in the actual sleep mode. freqm interrupts can wake up the device from sleep modes when the measurement is done, but only from sleep modes where clk_freqm is running. please refe r to the power manager chapter for details. counter clk_ref clk_msr refsel refnum, start clksel start value timer trigger isr
330 32145c?06/2013 at32uc3l0128/256 17.4.2 clocks the clock for the freqm bus interface (clk_freqm) is generated by the power manager. this clock is enabled at reset, and can be disabled in the power manager. it is recommended to disable the freqm before disabling the clock, to avoid freezing the freqm ia an undefined state. a set of clocks can be selected as referenc e (clk_ref) and another set of clocks can be selected for measurement (clk_msr). please refer to the clksel and refsel tables in the module configuration section for details. 17.4.3 debug operation when an external debugger forces the cpu in to debug mode, the freqm continues normal operation. if the freqm is configured in a way that requires it to be periodically serviced by the cpu through interrupts or similar, improper operation or data loss may result during debugging. 17.4.4 interrupts the freqm interrupt request line is connected to the internal source of the interrupt controller. using the freqm interrupt requires the interrupt controller to be programmed first. 17.5 functional description the freqm accuratly measures the frequency of a clock by comparing the frequency to a known frequency: f clk_msr = (value/refnum)*f clk_ref 17.5.1 reference clock the reference clock selection (refsel) field in the mode register (mode) selects the clock source for clk_ref. the reference clock is enabled by writing a one to the reference clock enable (refcen) bit in the mode register. this clock should have a known frequency. clk_ref needs to be disabled before switching to another clock. the rclkbusy bit in the status register (sr) indicates whether the clock is busy or not. this bit is set when the mode.refcen bit is written. to change clk_ref: ? write a zero to the mode.refcen bit to disable the clock, without changing the other bits/fields in the mode register. ? wait until the sr.rclkbusy bit reads as zero. ? change the mode.refsel field. ? write a one to the mode.refcen bit to enable the clock, without changing the other bits/fields in the mode register. ? wait until the sr.rclkbusy bit reads as zero. to enable clk_ref: ? write the correct value to the mode.refsel field. ? write a one to the mode.refcen to enable the clock, without changing the other bits/fields in the mode register. ? wait until the sr.rclkbusy bit reads as zero. to disable clk_ref:
331 32145c?06/2013 at32uc3l0128/256 ? write a zero to the mode.refcen to disable he clock, without changing the other bits/fields in the mode register. ? wait until the sr.rclkbusy bit reads as zero. 17.5.1.1 cautionary note note that if clock selected as source for clk_ref is stopped during a measurement, this will not be detected by the freqm. the busy bit in the status register w ill never be cleared, and the done interrupt will never be triggered. if the clock select ed as soruce for clk_ref is stopped, it will not be po ssible to change the source for th e reference clock as long as the selected source is not running. 17.5.2 measurement in the mode register the clock source sele ction (clksel) field selects clk_msr and the number of reference clock cycles (refnum) fi eld selects the duration of the measurement. the duration is given in number of clk_ref periodes. writing a one to the start bit in the control register (ctrl) starts the measurement. the busy bit in sr is cleared when the measurement is done. the result of the measurement can be read from the value register (value). the frequency of the measured clock clk_msr is then: f clk_msr = (value/refnum)*f clk_ref 17.5.3 interrupts the freqm has two interrupt sources: ? done: a frequency measurement is done ? rclkrdy: the reference clock is ready these will generate an interrupt request if the corresponding bit in the interrupt mask register (imr) is set. the interrupt sources are ored together to form one interrupt request. the freqm will generate an interrupt request if at least one of the bits in the interr upt mask register (imr) is set. bits in imr are set by writing a one to the corresponding bit in the interrupt enable register (ier) and cleared by writing a one to this bit in the interrupt disable register (idr). the interrupt request remains active until the corresponding bit in the interrupt status register (isr) is cleared by writing a one to this bit in the interrupt clear register (icr). because all the interrupt sources are ored t ogether, the interrupt reque st from the freqm will rema in active until all the bits in isr are cleared.
332 32145c?06/2013 at32uc3l0128/256 17.6 user interface note: 1. the reset value for this register is device specific. please refer to the module configuration section at the end of thi s chapter. table 17-1. freqm register memory map offset register register name access reset 0x000 control register ctrl write-only 0x00000000 0x004 mode register mode read/write 0x00000000 0x008 status register status read-only 0x00000000 0x00c value register value read-only 0x00000000 0x010 interrupt enable register ier write-only 0x00000000 0x014 interrupt disable register idr write-only 0x00000000 0x018 interrupt mask register imr read-only 0x00000000 0x01c interrupt status register isr read-only 0x00000000 0x020 interrupt clear register icr write-only 0x00000000 0x3fc version register version read-only - (1)
333 32145c?06/2013 at32uc3l0128/256 17.6.1 control register name: ctrl access type: write-only offset: 0x000 reset value: 0x00000000 ?start writing a zero to this bit has no effect. writing a one to this bit will start a measurement. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 -------start
334 32145c?06/2013 at32uc3l0128/256 17.6.2 mode register name: mode access type: read/write offset: 0x004 reset value: 0x00000000 ? refcen: reference clock enable 0: the reference clock is disabled 1: the reference clock is enabled ? clksel: clock source selection selects the source for clk_msr. see table in module configuration chapter for details. ? refnum: number of reference clock cycles selects the duration of a measurement, given in number of clk_ref cycles. ? refsel: reference clock selection selects the source for clk_ref. see table in module configuration chapter for details. 31 30 29 28 27 26 25 24 refcen------- 23 22 21 20 19 18 17 16 - - - clksel 15 14 13 12 11 10 9 8 refnum 76543210 - - - - - refsel
335 32145c?06/2013 at32uc3l0128/256 17.6.3 status register name: status access type: read-only offset: 0x008 reset value: 0x00000000 ? rclkbusy: freqm reference clock status 0: the freqm ref clk is ready, so a measurement can start. 1: the freqm ref clk is not ready, so a measurement should not be started. ? busy: freqm status 0: the frequency meter is idle. 1: frequency measurement is on-going. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 ------rclkbusybusy
336 32145c?06/2013 at32uc3l0128/256 17.6.4 value register name: value access type: read-only offset: 0x00c reset value: 0x00000000 ? value: result from measurement. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 value[23:16] 15 14 13 12 11 10 9 8 value[15:8] 76543210 value[7:0]
337 32145c?06/2013 at32uc3l0128/256 17.6.5 interrupt enable register name: ier access type: write-only offset: 0x010 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will set the corresponding bit in imr. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 ------rclkrdydone
338 32145c?06/2013 at32uc3l0128/256 17.6.6 interrupt disable register name: idr access type: write-only offset: 0x014 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in imr. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 ------rclkrdydone
339 32145c?06/2013 at32uc3l0128/256 17.6.7 interrupt mask register name: imr access type: read-only offset: 0x018 reset value: 0x00000000 0: the corresponding interrupt is disabled. 1: the corresponding interrupt is enabled. a bit in this register is cleared when the corresponding bit in idr is written to one. a bit in this register is set when the corresponding bit in ier is written to one. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 ------rclkrdydone
340 32145c?06/2013 at32uc3l0128/256 17.6.8 interrupt status register name: isr access type: read-only offset: 0x01c reset value: 0x00000000 0: the corresponding interrupt is cleared. 1: the corresponding interrupt is pending. a bit in this register is set when the corresponding bit in status has a one to zero transition. a bit in this register is cleared when the corresponding bit in icr is written to one. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 ------rclkrdydone
341 32145c?06/2013 at32uc3l0128/256 17.6.9 interrupt clear register name: icr access type: write-only offset: 0x020 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in isr and the corresponding interrupt request. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 ------rclkrdydone
342 32145c?06/2013 at32uc3l0128/256 17.6.10 version register name: version access type: read-only offset: 0x3fc reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
343 32145c?06/2013 at32uc3l0128/256 17.7 module configuration the specific configuration for each freqm instance is listed in the following tables. the module bus clocks listed here are connected to the syst em bus clocks. please refer to the power man- ager chapter for details. table 17-2. freqm clock name module name clock name description freqm clk_freqm bus interface clock clk_msr measured clock clk_ref reference clock table 17-3. register reset values register reset value version 0x00000310 table 17-4. clock sources for clk_msr clksel clock/oscillator description 0 clk_cpu the clock the cpu runs on 1 clk_hsb high speed bus clock 2 clk_pba peripheral bus a clock 3 clk_pbb peripheral bus b clock 4 osc0 output clock from oscillator 0 5 osc32k output clock from osc32k 6 rcsys output clock from rcsys oscillator 7 dfll0 output clock from dfll0 8 reserved 9-18 gclk0-9 generic clock 0 through 9 19 rc120m aw clock output clock from rc120m to aw 20 rc120m output clock from rc120m to main clock mux 21 rc32k output clock from rc32k 22-31 reserved table 17-5. clock sources for clk_ref refsel clock/oscillator description 0 rcsys system rc oscillator clock
344 32145c?06/2013 at32uc3l0128/256 1 osc32k output clock form osc32k 2 gclk9 generic clock 9 3-7 reserved table 17-5. clock sources for clk_ref refsel clock/oscillator description
345 32145c?06/2013 at32uc3l0128/256 18. general-purpose input/ output controller (gpio) rev: 2.1.3.5 18.1 features ? configurable pin-change, rising-edge, or falling-edge interrupt ? configurable peripheral event generator ? glitch filter providing rejection of pulses shorter than one clock cycle ? input visibility and output control ? multiplexing of peripheral functions on i/o pins ? programmable internal pull-up resistor ? optional locking of configuration to avoid accidental reconfiguration 18.2 overview the general purpose input/output controller (gpio) controls the i/o pins of the microcontroller. each gpio pin may be used as a general-purpose i/o or be assigned to a function of an embed- ded peripheral. the gpio is configured using the peripheral bus (pb). some registers can also be configured using the low latency cpu local bus. see section 18.6.2.7 for details. 18.3 block diagram figure 18-1. gpio block diagram interrupt controller power manager embedded peripheral general purpose input/output - gpio gpio interrupt request clk_gpio pin control signals pin pin pin pin pin mcu i/o pins configuration interface
346 32145c?06/2013 at32uc3l0128/256 18.4 i/o lines description 18.5 product dependencies in order to use this module, other parts of the system must be configured correctly, as described below. 18.5.1 power management if the cpu enters a sleep mode that disables cl ocks used by the gpio, the gpio will stop func- tioning and resume operation after the system wakes up from sleep mode. if a peripheral function is configured for a gpio pin, the peripheral will be able to control the gpio pin even if the gpio clock is stopped. 18.5.2 clocks the gpio is connected to a peripheral bus clock (clk_gpio). this clock is generated by the power manager. clk_gpio is enabled at reset, and can be disabled by writing to the power manager. clk_gpio must be enabled in order to access the configuration registers of the gpio or to use the gpio interrupts. after configuring the gpio, the clk_gpio can be disabled by writing to the power manager if interrupts are not used. if the cpu local bus is used to access the c onfiguration interface of the gpio, the clk_gpio must be equal to the cpu clock to avoid data loss. 18.5.3 interrupts the gpio interrupt request lines are connected to the interrupt controller. using the gpio inter- rupts requires the interrupt controller to be programmed first. 18.5.4 peripheral events the gpio peripheral events are connected via the peripheral event system. refer to the peripheral event system chapter for details. 18.5.5 debug operation when an external debugger forces the cpu into debug mode, the gpio continues normal oper- ation. if the gpio is configured in a way that requires it to be periodically serviced by the cpu through interrupts or similar, improper operation or data loss may result during debugging. pin name description type gpion gpio pin n digital
347 32145c?06/2013 at32uc3l0128/256 18.6 functional description the gpio controls the i/o pins of the microcontr oller. the control logic associated with each pin is shown in the figure below. figure 18-2. overview of the gpio 0 1 0 1 0 1 gper 1 0 ovr oder pmrn periph. func. a periph.func. b periph. func. c pin puer* pvr 0 1 glitch filter gfer edge detector 1 0 interrupt request imr1 imr0 ier pullup .... output output enable input *) register value is overrided if a peripheral function that support this function is enabled ifr
348 32145c?06/2013 at32uc3l0128/256 18.6.1 basic operation 18.6.1.1 module configuration the gpio user interface registers are organiz ed into ports and each port controls 32 different gpio pins. most of the registers supports bit wise access operations such as set, clear and tog- gle in addition to the standard word access. for details regarding interface registers, refer to section 18.7 . 18.6.1.2 available features the gpio features implemented are device dependent, and not all functions are implemented on all pins. the user must refer to the module configuration section and the gpio function mul- tiplexing section in the package and pinout chapter for the device specific settings used in the at32uc3l0128/256. device specific settings includes: ? number of gpio pins ? functions implemented on each pin ? peripheral function(s) multiplexed on each gpio pin ? reset state of registers 18.6.1.3 inputs the level on each gpio pin can be read through the pin value register (pvr). this register indicates the level of the gpio pins regardless of the pins being driven by the gpio or by an external component. note that due to power saving measures , the pvr register will only be updated when the corresponding bit in gper is one or if an interrupt is enabled for the pin, i.e. ier is one for the corresponding pin. 18.6.1.4 output control when the gpio pin is assigned to a peripheral function, i.e. the corresponding bit in gper is zero, the peripheral determines whether the pin is driven or not. when the gpio pin is controlled by the gpio, the value of output driver enable register (oder) determines whether the pin is driven or not. when a bit in this register is one, the corre- sponding gpio pin is driven by the gpio. when the bit is zero, the gpio does not drive the pin. the level driven on a gpio pin can be determined by writing the value to the corresponding bit in the output value register (ovr). 18.6.1.5 peripheral muxing the gpio allows a single gpio pin to be shared by multiple peripheral pins and the gpio itself. peripheral pins sharing the same gpio pin are arranged into peripheral functions that can be selected one at a time. peripheral functions are configured by writing the selected function value to the peripheral mux registers (pmrn). to allow a peripheral pin access to the shared gpio pin, gpio control must be disabled for that pin, i.e. the corresponding bit in gper must read zero. a peripheral function value is set by writing bit zero to pmr0 and bit one to the same index posi- tion in pmr1 and so on. in a system with 4 peripheral functions a,b,c, and d, peripheral function c for gpio pin four is selected by writi ng a zero to bit four in pmr0 and a one to the same bit index in pmr1. refer to the gpio function multiplexing chapter for details regarding pin function configuration for each gpio pin.
349 32145c?06/2013 at32uc3l0128/256 18.6.2 advanced operation 18.6.2.1 peripheral i/o pin control when a gpio pin is assigned to a peripheral function, i.e. the corresponding bit in gper is zero, output and output enable is controlled by the selected peripheral pin. in addition the peripheral may control some or all of the other gpio pin functions listed in table 18-1 , if the peripheral sup- ports those features. all pin features not controlled by the selected peripheral is controlled by the gpio. refer to the module configuration section for details regarding implemented gpio pin functions and to the peripheral chapter for details regarding i/o pin function control. 18.6.2.2 pull-up resistor control pull-up can be configured for each gpio pin. pull-up allows the pin and any connected net to be pulled up to vdd if the net is not driven. pull-up is useful for detecting if a pin is unconnected or if a mechanical button is pressed, for var- ious communication protocols and to keep unconnected pins from floating. pull-up can be enabled and disabled by writing a one and a zero respectively to the correspond- ing bit in the pull-up enable register (puer). 18.6.2.3 output pin timings figure 18-3 shows the timing of the gpio pin when writing to the output value register (ovr). the same timing applies when performing a ?set? or ?clear? access, i.e. writing to ovrs or ovrc. the timing of pvr is also shown. figure 18-3. output pin timings 18.6.2.4 interrupts the gpio can be configured to generate an inte rrupt when it detects a change on a gpio pin. interrupts on a pin are enabled by writing a one to the corresponding bit in the interrupt enable table 18-1. i/o pin function control function name gpio mode peripheral mode output ovr peripheral output enable oder peripheral pull-up puer peripheral if supported, else gpio pb access pb access clk_gpio write ovr to 1 write ovr to 0 ovr / i/o line pvr
350 32145c?06/2013 at32uc3l0128/256 register (ier). the module can be configured to generate an interrupt whenever a pin changes value, or only on rising or falling edges. this is controlled by the interrupt mode registers (imrn). interrupts on a pin can be enabled rega rdless of the gpio pin being controlled by the gpio or assigned to a peripheral function. an interrupt can be generated on each gpio pin. these interrupt generators are further grouped into groups of eight and connected to the interrupt controller. an interrupt request from any of the gpio pin generators in the group will result in an interrupt request from that group to the inter- rupt controller if the corresponding bit for the gp io pin in the ier is set. by grouping interrupt generators into groups of eight, four different interrupt handlers can be installed for each gpio port. the interrupt flag register (ifr) can be read by software to determine which pin(s) caused the interrupt. the interrupt flag must be manually cleared by writing a zero to the corresponding bit in ifr. gpio interrupts will only be gener ated when clk_gpio is enabled. 18.6.2.5 input glitch filter input glitch filters can be enabled on each gpio pi n. when the glitch filter is enabled, a glitch with duration of less than 1 clk_gpio cycle is automatically rejected, while a pulse with dura- tion of 2 clk_gpio cycles or more is accepted. for pulse durations between 1 and 2 clk_gpio cycles, the pulse may or may not be taken into account, depending on the precise timing of its occurrence. thus for a pulse to be guaranteed visible it must exceed 2 clk_gpio cycles, whereas for a glitch to be reliably filtered out, its duration must not exceed 1 clk_gpio cycle. the filter introduces 2 clock cycles latency. the glitch filters are controlled by the glitch filter enable register (gfer). when a bit in gfer is one, the glitch filter on the corresponding pin is enabled. the glitch filter affects only interrupt inputs. inputs to peripherals or the value read through pvr are not affected by the glitch filters. 18.6.2.6 interrupt timings figure 18-4 shows the timing for rising edge (or pin-change) interrupts when the glitch filter is disabled. for the pulse to be registered, it must be sampled at the rising edge of the clock. in this example, this is not the case for the first pulse. the second pulse is sampled on a rising edge and will trigger an interrupt request. figure 18-4. interrupt timing with glitch filter disabled figure 18-5 shows the timing for rising edge (or pin-change) interrupts when the glitch filter is enabled. for the pulse to be registered, it must be sampled on two subsequent rising edges. in the example, the first pulse is rejected while th e second pulse is accepted and causes an inter- rupt request. clk_gpio pin level ifr
351 32145c?06/2013 at32uc3l0128/256 figure 18-5. interrupt timing with g litch filter enabled 18.6.2.7 cpu local bus the cpu local bus can be used for application where low latency read and write access to the output value register (ovr) and output drive enable register (oder) is required. the cpu local bus allows the cpu to configure the mentioned gpio registers directly, bypassing the shared peripheral bus (pb). to avoid data loss when using the cpu local bus, the clk_gpio must run at the same fre- quency as the clk_cpu. see section 18.5.2 for details. the cpu local bus is mapped to a different base address than the gpio but the over and oder offsets are the same. see the cpu local bus mapping section in the memories chapter for details. 18.6.2.8 peripheral events peripheral events allow direct peripheral to peripheral communication of specified events. see the peripheral event system chapter for more information. the gpio can be programmed to output peripheral events whenever an interrupt condition is detected. the peripheral events configuration depends on the interrupt configuration. an event will be generated on the same condition as the interrupt (pin change, rising edge, or falling edge). the interrupt configuration is controlled by the imr register. peripheral event on a pin is enabled by writing a one to the corresponding bit in the event enable register (ever). the peripheral event trigger mode is shared with the interrupt trigger and is configured by writing to the imr0 and imr1 registers. interrupt does not need to be enabled on a pin when peripheral events are enabled. peripheral events are also affected by the input glitch filter settings. see section 18.6.2.5 for more information. a peripheral event can be generated on each gpio pin. each port can then have up to 32 peripheral event generators. gro ups of eight peripheral event g enerators in each port are ored together to form a peripheral event line, so that each port has four peripheral event lines con- nected to the peripheral event system. clk_gpio pin level ifr
352 32145c?06/2013 at32uc3l0128/256 18.7 user interface the gpio controller manages all the gpio pins on the 32-bit avr microcontroller. the pins are managed as 32-bit ports that are configurable through a peripheral bus (pb) interface. each port has a set of configuration registers. the ov erall memory map of the gpio is shown below. the number of pins and hence the number of ports is product specific. figure 18-6. port configuration registers in the peripheral muxing table in the package and pinout chapter each gpio pin has a unique number. note that the pa, pb, pc, and px ports do not necessarily directly correspond to the gpio ports. to find the corresponding port and pin the following formulas can be used: gpio port = floor((gpio number) / 32), example: floor((36)/32) = 1 gpio pin = gpio number % 32, example: 36 % 32 = 4 table 18-2 shows the configuration registers for one port. addresses shown are relative to the port address offset. the specific address of a co nfiguration register is found by adding the regis- ter offset and the port offset to the gpio start address. one bit in each of the configuration registers corresponds to a gpio pin. 18.7.1 access types most configuration register can be accessed in four different ways. the first address location can be used to write the register directly. this address can also be used to read the register value. the following addresses facilitate th ree different types of write acce ss to the regi ster. performing a ?set? access, all bits written to one will be se t. bits written to zero will be unchanged by the operation. performing a ?clear? a ccess, all bits written to one will be cleared. bits written to zero will be unchanged by the operation. finally, a toggle access will toggle the value of all bits writ- port 0 configuration registers port 1 configuration registers port 2 configuration registers port n configuration registers 0x0000 0x0200 0x0400 n*0x200 ?.
353 32145c?06/2013 at32uc3l0128/256 ten to one. again all bits written to zero remain unchanged. note that for some registers (e.g. ifr), not all access methods are permitted. note that for ports with less than 32 bits, the corresponding control registers will have unused bits. this is also the case for features that are not implemented for a specific pin. writing to an unused bit will have no effect. reading unused bits will always return 0. 18.7.2 configuration protection in order to protect the configuration of individual gpio pins from software failure, configuration bits for individual gpio pins may be locked by writing a one to the corresponding bit in the lock register. while this bit is one, any write to the same bit position in any lockable gpio register using the peripheral bus (pb) will not have an effect. the cpu local bus is not checked and thus allowed to write to all bits in a cpu local bus mapped register no mather the lock value. the registers required to clear bits in the lock register are protected by the access protection mechanism described in section 18.7.3 , ensuring the lock mechanis m itself is robust against software failure. 18.7.3 access protection in order to protect critical registers from software failure, some registers are protected by a key protection mechanism. these registers can only be changed by first writing the unlock regis- ter, then the protected register. protected registers are indicated in table 18-2 . the unlock register contains a key field which must always be written to 0xaa, and an offset field corre- sponding to the offset of the register to be modified. the next write operation resets the unlock register, so if the register is to be modified again, the unlock register must be written again. attempting to write to a protected register without first writing the unlock register results in the write operation being discarded, and the access error bit in the access status register (asr.ae) will be set. table 18-2. gpio register memory map offset register function register name access reset config. protection access protection 0x000 gpio enable register read/write gper read/write - (1) yn 0x004 gpio enable register set gpers write-only y n 0x008 gpio enable register clear gperc write-only y n 0x00c gpio enable register toggle gpert write-only y n 0x010 peripheral mux register 0 read/write pmr0 read/write - (1) y n 0x014 peripheral mux register 0 set pmr0s write-only y n 0x018 peripheral mux register 0 clear pmr0c write-only y n 0x01c peripheral mux register 0 toggle pmr0t write-only y n 0x020 peripheral mux register 1 read/write pmr1 read/write - (1) yn 0x024 peripheral mux register 1 set pmr1s write-only y n 0x028 peripheral mux register 1 clear pmr1c write-only y n
354 32145c?06/2013 at32uc3l0128/256 0x02c peripheral mux register 1 toggle pmr1t write-only y n 0x030 peripheral mux register 2 read/write pmr2 read/write - (1) y n 0x034 peripheral mux register 2 set pmr2s write-only y n 0x038 peripheral mux register 2 clear pmr2c write-only y n 0x03c peripheral mux register 2 toggle pmr2t write-only y n 0x040 output driver enable register read/write oder read/write - (1) yn 0x044 output driver enable register set oders write-only y n 0x048 output driver enable register clear oderc write-only y n 0x04c output driver enable register toggle odert write-only y n 0x050 output value register read/write ovr read/write - (1) n n 0x054 output value register set ovrs write-only n n 0x058 output value register clear ovrc write-only n n 0x05c output value register toggle ovrt write-only n n 0x060 pin value register read pvr read-only depe nding on pin states nn 0x064 pin value register - - - n n 0x068 pin value register - - - n n 0x06c pin value register - - - n n 0x070 pull-up enable register read/write puer read/write - (1) y n 0x074 pull-up enable register set puers write-only y n 0x078 pull-up enable register clear puerc write-only y n 0x07c pull-up enable register toggle puert write-only y n 0x090 interrupt enable register read/write ier read/write - (1) n n 0x094 interrupt enable register set iers write-only n n 0x098 interrupt enable register clear ierc write-only n n 0x09c interrupt enable register toggle iert write-only n n 0x0a0 interrupt mode register 0 read/write imr0 read/write - (1) nn 0x0a4 interrupt mode register 0 set imr0s write-only n n 0x0a8 interrupt mode register 0 clear imr0c write-only n n 0x0ac interrupt mode register 0 toggle imr0t write-only n n 0x0b0 interrupt mode register 1 read/write imr1 read/write - (1) n n 0x0b4 interrupt mode register 1 set imr1s write-only n n 0x0b8 interrupt mode register 1 clear imr1c write-only n n 0x0bc interrupt mode register 1 toggle imr1t write-only n n table 18-2. gpio register memory map offset register function register name access reset config. protection access protection
355 32145c?06/2013 at32uc3l0128/256 note: 1. the reset values for these registers are device specific. please refer to the module configuration section at the end of this chapter. 0x0c0 glitch filter enable register read/write gfer read/write - (1) nn 0x0c4 glitch filter enable register set gfers write-only n n 0x0c8 glitch filter enable register clear gferc write-only n n 0x0cc glitch filter enable register toggle gfert write-only n n 0x0d0 interrupt flag register read ifr read-only - (1) n n 0x0d4 interrupt flag register - - - n n 0x0d8 interrupt flag register clear ifrc write-only n n 0x0dc interrupt flag register - - - n n 0x180 event enable register read ever read/write - (1) n n 0x184 event enable register set evers write-only n n 0x188 event enable register clear everc write-only n n 0x18c event enable register toggle evert write-only n n 0x1a0 lock register read/write lock read/write - (1) n y 0x1a4 lock register set locks write-only n n 0x1a8 lock register clear lockc write-only n y 0x1ac lock register toggle lockt write-only n y 0x1e0 unlock register read/write unlock write-only n n 0x1e4 access status register read/write asr read/write n 0x1f8 parameter register read pa r a m e t e r read-only - (1) n n 0x1fc version register read version read-only - (1) n n table 18-2. gpio register memory map offset register function register name access reset config. protection access protection
356 32145c?06/2013 at32uc3l0128/256 18.7.4 gpio enable register name: gper access: read/write, set, clear, toggle offset : 0x000, 0x004, 0x008, 0x00c reset value: - ? p0-p31: gpio enable 0: a peripheral function controls the corresponding pin. 1: the gpio controls the corresponding pin. 31 30 29 28 27 26 25 24 p31 p30 p29 p28 p27 p26 p25 p24 23 22 21 20 19 18 17 16 p23 p22 p21 p20 p19 p18 p17 p16 15 14 13 12 11 10 9 8 p15 p14 p13 p12 p11 p10 p9 p8 76543210 p7 p6 p5 p4 p3 p2 p1 p0
357 32145c?06/2013 at32uc3l0128/256 18.7.5 peripheral mux register 0 name: pmr0 access: read/write, set, clear, toggle offset : 0x010, 0x014, 0x018, 0x01c reset value: - ? p0-31: peripheral mult iplexer select bit 0 31 30 29 28 27 26 25 24 p31 p30 p29 p28 p27 p26 p25 p24 23 22 21 20 19 18 17 16 p23 p22 p21 p20 p19 p18 p17 p16 15 14 13 12 11 10 9 8 p15 p14 p13 p12 p11 p10 p9 p8 76543210 p7 p6 p5 p4 p3 p2 p1 p0
358 32145c?06/2013 at32uc3l0128/256 18.7.6 peripheral mux register 1 name: pmr1 access: read/write, set, clear, toggle offset : 0x020, 0x024, 0x028, 0x02c reset value: - ? p0-31: peripheral mult iplexer select bit 1 31 30 29 28 27 26 25 24 p31 p30 p29 p28 p27 p26 p25 p24 23 22 21 20 19 18 17 16 p23 p22 p21 p20 p19 p18 p17 p16 15 14 13 12 11 10 9 8 p15 p14 p13 p12 p11 p10 p9 p8 76543210 p7 p6 p5 p4 p3 p2 p1 p0
359 32145c?06/2013 at32uc3l0128/256 18.7.7 peripheral mux register 2 name: pmr2 access: read/write, set, clear, toggle offset : 0x030, 0x034, 0x038, 0x03c reset value: - ? p0-31: peripheral mult iplexer select bit 2 31 30 29 28 27 26 25 24 p31 p30 p29 p28 p27 p26 p25 p24 23 22 21 20 19 18 17 16 p23 p22 p21 p20 p19 p18 p17 p16 15 14 13 12 11 10 9 8 p15 p14 p13 p12 p11 p10 p9 p8 76543210 p7 p6 p5 p4 p3 p2 p1 p0 {pmr2, pmr1, pmr0} selected peripheral function 000 a 001 b 010 c 011 d 100 e 101 f 110 g 111 h
360 32145c?06/2013 at32uc3l0128/256 18.7.8 output driver enable register name: oder access: read/write, set, clear, toggle offset : 0x040, 0x044, 0x048, 0x04c reset value: - ? p0-31: output driver enable 0: the output driver is disabled for the corresponding pin. 1: the output driver is enabled for the corresponding pin. 31 30 29 28 27 26 25 24 p31 p30 p29 p28 p27 p26 p25 p24 23 22 21 20 19 18 17 16 p23 p22 p21 p20 p19 p18 p17 p16 15 14 13 12 11 10 9 8 p15 p14 p13 p12 p11 p10 p9 p8 76543210 p7 p6 p5 p4 p3 p2 p1 p0
361 32145c?06/2013 at32uc3l0128/256 18.7.9 output value register name: ovr access: read/write, set, clear, toggle offset : 0x050, 0x054, 0x058, 0x05c reset value: - ? p0-31: output value 0: the value to be driven on the gpio pin is 0. 1: the value to be driven on the gpio pin is 1. 31 30 29 28 27 26 25 24 p31 p30 p29 p28 p27 p26 p25 p24 23 22 21 20 19 18 17 16 p23 p22 p21 p20 p19 p18 p17 p16 15 14 13 12 11 10 9 8 p15 p14 p13 p12 p11 p10 p9 p8 76543210 p7 p6 p5 p4 p3 p2 p1 p0
362 32145c?06/2013 at32uc3l0128/256 18.7.10 pin value register name: pvr access: read-only offset : 0x060, 0x064, 0x068, 0x06c reset value: depending on pin states ? p0-31: pin value 0: the gpio pin is at level zero. 1: the gpio pin is at level one. note that the level of a pin can only be read when the corresponding pin in gper is one or interrupt is enabled for the pin. 31 30 29 28 27 26 25 24 p31 p30 p29 p28 p27 p26 p25 p24 23 22 21 20 19 18 17 16 p23 p22 p21 p20 p19 p18 p17 p16 15 14 13 12 11 10 9 8 p15 p14 p13 p12 p11 p10 p9 p8 76543210 p7 p6 p5 p4 p3 p2 p1 p0
363 32145c?06/2013 at32uc3l0128/256 18.7.11 pull-up enable register name: puer access: read/write, set, clear, toggle offset : 0x070, 0x074, 0x078, 0x07c reset value: - ? p0-31: pull-up enable writing a zero to a bit in this register will disable pull-up on the corresponding pin. writing a one to a bit in this register will enable pull-up on the corresponding pin. 31 30 29 28 27 26 25 24 p31 p30 p29 p28 p27 p26 p25 p24 23 22 21 20 19 18 17 16 p23 p22 p21 p20 p19 p18 p17 p16 15 14 13 12 11 10 9 8 p15 p14 p13 p12 p11 p10 p9 p8 76543210 p7 p6 p5 p4 p3 p2 p1 p0
364 32145c?06/2013 at32uc3l0128/256 18.7.12 interrupt enable register name: ier access: read/write, set, clear, toggle offset : 0x090, 0x094, 0x098, 0x09c reset value: - ? p0-31: interrupt enable 0: interrupt is disabled for the corresponding pin. 1; interrupt is enabled for the corresponding pin. 31 30 29 28 27 26 25 24 p31 p30 p29 p28 p27 p26 p25 p24 23 22 21 20 19 18 17 16 p23 p22 p21 p20 p19 p18 p17 p16 15 14 13 12 11 10 9 8 p15 p14 p13 p12 p11 p10 p9 p8 76543210 p7 p6 p5 p4 p3 p2 p1 p0
365 32145c?06/2013 at32uc3l0128/256 18.7.13 interrupt mode register 0 name: imr0 access: read/write, set, clear, toggle offset : 0x0a0, 0x0a4, 0x0a8, 0x0ac reset value: - ? p0-31: interrupt mode bit 0 31 30 29 28 27 26 25 24 p31 p30 p29 p28 p27 p26 p25 p24 23 22 21 20 19 18 17 16 p23 p22 p21 p20 p19 p18 p17 p16 15 14 13 12 11 10 9 8 p15 p14 p13 p12 p11 p10 p9 p8 76543210 p7 p6 p5 p4 p3 p2 p1 p0
366 32145c?06/2013 at32uc3l0128/256 18.7.14 interrupt mode register 1 name: imr1 access: read/write, set, clear, toggle offset : 0x0b0, 0x0b4, 0x0b8, 0x0bc reset value: - ? p0-31: interrupt mode bit 1 31 30 29 28 27 26 25 24 p31 p30 p29 p28 p27 p26 p25 p24 23 22 21 20 19 18 17 16 p23 p22 p21 p20 p19 p18 p17 p16 15 14 13 12 11 10 9 8 p15 p14 p13 p12 p11 p10 p9 p8 76543210 p7 p6 p5 p4 p3 p2 p1 p0 {imr1, imr0} interrupt mode 00 pin change 01 rising edge 10 falling edge 11 reserved
367 32145c?06/2013 at32uc3l0128/256 18.7.15 glitch filter enable register name: gfer access: read/write, set, clear, toggle offset : 0x0c0, 0x0c4, 0x0c8, 0x0cc reset value: - ? p0-31: glitch filter enable 0: glitch filter is disabled for the corresponding pin. 1: glitch filter is enabled for the corresponding pin. note! the value of this register should only be changed when the corresponding bit in ier is zero. updating gfer while interrupt on the corresponding pin is enabled can cause an unintentional interrupt to be triggered. 31 30 29 28 27 26 25 24 p31 p30 p29 p28 p27 p26 p25 p24 23 22 21 20 19 18 17 16 p23 p22 p21 p20 p19 p18 p17 p16 15 14 13 12 11 10 9 8 p15 p14 p13 p12 p11 p10 p9 p8 76543210 p7 p6 p5 p4 p3 p2 p1 p0
368 32145c?06/2013 at32uc3l0128/256 18.7.16 interrupt flag register name: ifr access: read, clear offset :0x0d0, 0x0d8 reset value: - ? p0-31: interrupt flag 0: no interrupt condition has been detected on the corresponding pin. 1: an interrupt condition has been detected on the corresponding pin. the number of interrupt request lines depe nds on the number of gpio pins on the mcu. refer to the product specific data for details. note also that a bit in the interrupt flag register is only valid if the corresponding bit in ier is one. 31 30 29 28 27 26 25 24 p31 p30 p29 p28 p27 p26 p25 p24 23 22 21 20 19 18 17 16 p23 p22 p21 p20 p19 p18 p17 p16 15 14 13 12 11 10 9 8 p15 p14 p13 p12 p11 p10 p9 p8 76543210 p7 p6 p5 p4 p3 p2 p1 p0
369 32145c?06/2013 at32uc3l0128/256 18.7.17 event enable register name: ever access: read/write, set, clear, toggle offset : 0x180, 0x184, 0x188, 0x18c reset value: - ? p0-31: event enable 0: peripheral event is disabled for the corresponding pin. 1: peripheral event is enabled for the corresponding pin. 31 30 29 28 27 26 25 24 p31 p30 p29 p28 p27 p26 p25 p24 23 22 21 20 19 18 17 16 p23 p22 p21 p20 p19 p18 p17 p16 15 14 13 12 11 10 9 8 p15 p14 p13 p12 p11 p10 p9 p8 76543210 p7 p6 p5 p4 p3 p2 p1 p0
370 32145c?06/2013 at32uc3l0128/256 18.7.18 lock register name: lock access: read/write, set, clear, toggle offset : 0x1a0, 0x1a4, 0x1a8, 0x1ac reset value: - ? p0-31: lock state 0: pin is unlocked. the corresponding bit can be changed in any gpio register for this port. 1: pin is locked. the corresponding bit can not be changed in any gpio register for this port. the value of lock determines which bits are locked in the lockable registers. the lock, lockc, and lockt registers are protected, which means they can only be written immediately after a write to the unlock register with the proper key and offset. locks is not protected, and can be written at any time. 31 30 29 28 27 26 25 24 p31 p30 p29 p28 p27 p26 p25 p24 23 22 21 20 19 18 17 16 p23 p22 p21 p20 p19 p18 p17 p16 15 14 13 12 11 10 9 8 p15 p14 p13 p12 p11 p10 p9 p8 76543210 p7 p6 p5 p4 p3 p2 p1 p0
371 32145c?06/2013 at32uc3l0128/256 18.7.19 unlock register name: unlock access: write-only offset :0x1e0 reset value: - ? offset: register offset this field must be written with the offset va lue of the lock, lockc or lockt register to unlock. this offset must also include the port offset for the register to unlock. locks can not be locke d so no unlock is required before writing to this register. ?key: unlocking key this bitfield must be written to 0xaa for a wr ite to this register to have an effect. this register always reads as zero. 31 30 29 28 27 26 25 24 key 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 ------ offset 76543210 offset
372 32145c?06/2013 at32uc3l0128/256 18.7.20 access status register name: asr access: read/write offset :0x1e4 reset value: - ? ae: access error this bit is set when a write to a locked register occurs. this bit can be written to 0 by software. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 -------ae
373 32145c?06/2013 at32uc3l0128/256 18.7.21 parameter register name: parameter access type: read-only offset :0x1f8 reset value: - ? parameter: 0: the corresponding pin is not implemented in this gpio port. 1: the corresponding pin is implemented in this gpio port. there is one parameter register per gpio port. each bit in the parameter register indica tes whether the corresponding gper bit is implemented. 31 30 29 28 27 26 25 24 pa r a m e t e r 23 22 21 20 19 18 17 16 pa r a m e t e r 15 14 13 12 11 10 9 8 pa r a m e t e r 76543210 pa r a m e t e r
374 32145c?06/2013 at32uc3l0128/256 18.7.22 version register name: version access type: read-only offset :0x1fc reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 ---- version[11:8] 76543210 version[7:0]
375 32145c?06/2013 at32uc3l0128/256 18.8 module configuration the specific configuration for each gpio instance is listed in the following tables. the module bus clocks listed here are conn ected to the system bus clocks. refer to the power manager chapter for details. the reset values for all gpio registers are zero, with the following exceptions: table 18-3. gpio configuration feature gpio number of gpio ports 2 number of peripheral functions 8 table 18-4. implemented pin functions pin function implemented notes pull-up on all pins controlled by puer or peripheral table 18-5. gpio clocks module name clock name description gpio clk_gpio clock for the gpio bus interface table 18-6. register reset values port register reset value 0 gper 0x004dff5f 0 pmr0 0x00320020 0 pmr1 0x00020080 0 pmr2 0x00100800 0 puer 0x00000001 0 gfer 0x007fffff 0 parameter 0x007fffff 0 version 0x00000213 1 gper 0x0fffffcf 1 pmr0 0x00000030 1 pmr1 0x00000030 1 gfer 0x0fffffff 1 parameter 0x0fffffff 1 version 0x00000213
376 32145c?06/2013 at32uc3l0128/256 19. universal synchronous asynchrono us receiver transmitter (usart) rev: 4.4.0.6 19.1 features ? configurable baud rate generator ? 5- to 9-bit full-duplex, synchronous and asynchronous, se rial communication ? 1, 1.5, or 2 stop bits in asynchrono us mode, and 1 or 2 in synchronous mode ? parity generation and error detection ? framing- and overrun error detection ? msb- or lsb-first ? optional break generation and detection ? receiver frequency over-sampling by 8 or 16 times ? optional rts-cts hardware handshaking ? receiver time-out and transmitter timeguard ? optional multidrop mode with address generation and detection ? spi mode ? master or slave ? configurable serial clock phase and polarity ? clk spi serial clock frequency up to a quarte r of the clk_usart internal clock frequency ? lin mode ? compliant with lin 1.3 and lin 2.0 sp ecifications ? master or slave ? processing of frames with up to 256 data bytes ? configurable response data length, optiona lly defined automatica lly by the identifier ? self synchronization in slave node configuration ? automatic processing and verification of the ?break field? and ?sync field? ? the ?break field? is detected even if it is partially superimposed with a data byte ? optional, automatic iden tifier parity management ? optional, automatic checksum management ? supports both ?classic? an d ?enhanced? checksum types ? full lin error checking and reporting ? frame slot mode: the master allocates slots to scheduled frames automatically. ? wakeup signal generation ? test modes ? automatic echo, remote- and local loopback ? supports two peripheral dma controller channels ? buffer transfers without processor intervention 19.2 overview the universal synchronous asynchronous receiver transceiver (usart) provides a full duplex, universal, synchronous/asynchronous serial link. data frame format is widely configu- rable, including basic length, parity, and stop bit settings, maximizing standards support. the receiver implements parity-, framing-, and overrun error detection, and can handle un-fixed frame lengths with the time-out feature. the usart supports several operating modes, provid- ing an interface to, lin, and spi buses and infrared transceivers. communication with slow and remote devices is eased by the timeguard. d uplex multidrop communication is supported by address and data differentiation through the parity bit. the hardware handshaking feature enables an out-of-band flow control, automatically managing rts and cts pins. the peripheral dma controller connection enables memory transactions, and the usart supports chained
377 32145c?06/2013 at32uc3l0128/256 buffer management without processor intervention. automatic echo, remote-, and local loopback -test modes are also supported. 19.3 block diagram figure 19-1. usart block diagram peripheral dma controller channel channel interrupt controller power manager div receiver transmitter user interface i/o controller rxd rts txd cts clk baudrate generator usart interrupt clk_usart clk_usart/div usart peripheral bus table 19-1. spi operating mode pin usart spi slave spi master rxd rxd mosi miso txd txd miso mosi rts rts ? cs cts cts cs ?
378 32145c?06/2013 at32uc3l0128/256 19.4 i/o lines description 19.5 product dependencies 19.5.1 i/o lines the usart pins may be multiplexed with the i/o controller lines. the user must first configure the i/o controller to assign these pins to t heir peripheral functions. unused i/o lines may be used for other purposes. to prevent the txd line from falling when the usart is di sabled, the use of an internal pull up is required. if the hardware handshaking feature or modem mode is used, the internal pull up on txd must also be enabled. 19.5.2 clocks the clock for the usart bus interface (clk_usart) is generated by the power manager. this clock is enabled at reset, and can be disabled in the power manager. it is recommended to dis- able the usart before disabling the clock, to avoid freezing the usart in an undefined state. 19.5.3 interrupts the usart interrupt request line is connected to the interrupt controller. using the usart interrupt requires the interrupt controller to be programmed first. table 19-2. i/o lines description name description type active level clk serial clock i/o txd transmit serial data or master out slave in (mosi) in spi master mode or master in slave out (miso) in spi slave mode output rxd receive serial data or master in slave out (miso) in spi master mode or master out slave in (mosi) in spi slave mode input cts clear to send or slave select (nss) in spi slave mode input low rts request to send or slave select (nss) in spi master mode output low
379 32145c?06/2013 at32uc3l0128/256 19.6 functional description 19.6.1 baud rate generator the baud rate generator provides the bit period clock named the baud rate clock to both receiver and transmitter. it is based on a 16-bit di vider, which is specified in the clock divider field in the baud rate generator register (brgr.cd). a non-zero value enables the generator, and if cd is one, the divider is bypassed and i nactive. the clock selection field in the mode register (mr.usclks) selects clock source between: ? clk_usart (internal clock) ? clk_usart/div (a divided clk_usart, refer to module configuration section) ? clk (external clock, available on the clk pin) if the external clk clock is selected, the duration of the low and high levels of the signal pro- vided on the clk pin must be at least 4.5 times longer than those provided by clk_usart. figure 19-2. baud rate generator 19.6.1.1 baud rate in asynchronous mode if the usart is configured to operate in an asynchronous mode, the selected clock is divided by the cd value before it is provided to the rece iver as a sampling clock. depending on the over- sampling mode bit (mr.over) value, the clock is then divided by either 8 (over=1), or 16 (over=0). the baud rate is calculated with the following formula: this gives a maximum baud rate of clk_usar t divided by 8, assuming that clk_usart is the fastest clock possible, and that over is one. 19.6.1.2 baud rate calculation example table 19-3 shows calculations based on the cd field to obtain 38400 baud from different source clock frequencies. this table also shows the actual resulting baud rate and error. 16-bit counter cd usclks cd clk_usart clk_usart/div reserved clk sync sync usclks= 3 fidi over sampling divider baudrate clock sampling clock 1 0 0 clk 0 1 2 3 >1 1 1 0 0 baudrate selectedclock 82 over ? ?? cd ?? ----------------------------------------------- - =
380 32145c?06/2013 at32uc3l0128/256 the baud rate is calculated with the following formula (over=0): the baud rate error is calculated with the following formula. it is not recommended to work with an error higher than 5%. 19.6.1.3 fractional baud rate in asynchronous mode the baud rate generator has a limitation: the source frequency is always a multiple of the baud rate. an approach to this problem is to integrate a high resolution fractional n clock generator, outputting fractional multiples of the reference source clock. this fractional part is selected with the fractional part field (brgr.fp), and is activa ted by giving it a non-zero value. the resolu- tion is one eighth of cd. the resulting baud rate is calculated using the following formula: the modified architecture is presented below: table 19-3. baud rate example (over=0) source clock (hz) expected baud rate (bit/s) calculation result cd actual baud rate (bit/s) error 3 686 400 38 400 6.00 6 38 400.00 0.00% 4 915 200 38 400 8.00 8 38 400.00 0.00% 5 000 000 38 400 8.14 8 39 062.50 1.70% 7 372 800 38 400 12.00 12 38 400.00 0.00% 8 000 000 38 400 13.02 13 38 461.54 0.16% 12 000 000 38 400 19.53 20 37 500.00 2.40% 12 288 000 38 400 20.00 20 38 400.00 0.00% 14 318 180 38 400 23.30 23 38 908.10 1.31% 14 745 600 38 400 24.00 24 38 400.00 0.00% 18 432 000 38 400 30.00 30 38 400.00 0.00% 24 000 000 38 400 39.06 39 38 461.54 0.16% 24 576 000 38 400 40.00 40 38 400.00 0.00% 25 000 000 38 400 40.69 40 38 109.76 0.76% 32 000 000 38 400 52.08 52 38 461.54 0.16% 32 768 000 38 400 53.33 53 38 641.51 0.63% 33 000 000 38 400 53.71 54 38 194.44 0.54% 40 000 000 38 400 65.10 65 38 461.54 0.16% 50 000 000 38 400 81.38 81 38 580.25 0.47% 60 000 000 38 400 97.66 98 38 265.31 0.35% baudrate clkusart ?? cd 16 ? ?? ? = error 1 expectedbaudrate actualbaudrate -------------------------------------------------- - ?? ?? ? = baudrate selectedclock 82 over ? ?? cd fp 8 ------- + ?? ?? ?? ?? -------------------------------------------------------------------- =
381 32145c?06/2013 at32uc3l0128/256 figure 19-3. fractional baud rate generator 19.6.1.4 baud rate in synchronous and spi mode if the usart is configured to operate in synchronous mode, the selected clock is divided by the brgr.cd field. this does not apply when clk is selected. when clk is selected the external frequency must be at least 4.5 times lower than the system clock, and when either clk or clk_usart/div are selected, cd must be even to ensure a 50/50 duty cycle. if clk_usart is selected, the generator ensures this regardless of value. 19.6.2 receiver and transmitter control after a reset, the transceiver is disabled. the re ceiver/transmitter is enabled by writing a one to either the receiver enable, or transmitter enab le bit in the control register (cr.rxen, or cr.txen). they may be enabled together and can be configured both before and after they have been enabled. the user can reset the usart receiver/transmitter at any time by writing a one to either the reset receiver (cr.rstrx), or reset transmitter (cr.rsttx) bit. this soft- ware reset clears status bits and resets internal state machines, immediately halting any communication. the user interface configur ation registers will re tain their values. the user can disable the receiver/transmitter by writing a one to either the receiver disable, or transmitter disable bit (cr.rxdis, or cr.txdis). if the receiver is disabled during a character reception, the usart will wait for the current character to be received before disabling. if the transmitter is disabled during transmission, the usart will wa it until both the current character and the character stored in the transmitter holding register (thr) are transmitted before dis- abling. if a timeguard has been implemented it will remain func tional during the transaction. usclks cd modulus control fp fp cd glitch-free logic 16-bit counter over fidi sync sampling divider clk_usart clk_usart/div reserved clk clk baudrate clock sampling clock sync usclks = 3 >1 1 2 3 0 0 1 0 1 1 0 0 baudrate selectedclock cd ------------------------------------- - =
382 32145c?06/2013 at32uc3l0128/256 19.6.3 synchronous and asynchronous modes 19.6.3.1 transmitter operations the transmitter performs equally in both syn chronous and asynchronous operating modes (mr.sync). one start bit, up to 9 data bits, an optional parity bit, and up to two stop bits are successively shifted out on the txd pin at each falling edge of the serial clock. the number of data bits is selected by th e character length field (mr.chrl) and the mr.mode9 bit. nine bits are selected by writing a one to mode9, overridi ng any value in chrl. the parity bit configura- tion is selected in the mr.par field. the most significant bit first bit (mr.msbf) selects which data bit to send first. the number of stop bits is selected by the mr.nbstop field. the 1.5 stop bit configuration is only supported in asynchronous mode. figure 19-4. character transmit the characters are sent by writing to the c haracter to be transmitted field (thr.txchr). the transmitter reports status with the transm itter ready (txrdy) and transmitter empty (txempty) bits in the channel status register (csr). txrdy is set when thr is empty. txempty is set when both thr and the transmit shift register are empty (transmission com- plete). both txrdy and txempty are cleared wh en the transmitter is disabled. writing a character to thr while txrdy is zero has no effect and the written character will be lost. figure 19-5. transmitter status 19.6.3.2 asynchronous receiver if the usart is configured in an asynchronous operating mode (mr.sync = 0), the receiver will oversample the rxd input line by either 8 or 16 times the baud rate clock, as selected by the oversampling mode bit (mr.over). if the line is ze ro for half a bit period (four or eight consecu- tive samples, respective ly), a start bit will be assumed, and the following 8th or 16th sample will determine the logical value on the line, in effect resulting in bit values being determined at the middle of the bit period. d0 d1 d2 d3 d4 d5 d6 d7 txd start bit parity bit stop bit example: 8-bit, parity enabled one stop baud rate clock d0 d1 d2 d3 d4 d5 d6 d7 txd start bit parity bit stop bit baud rate clock start bit write thr d0 d1 d2 d3 d4 d5 d6 d7 parity bit stop bit txrdy txempty
383 32145c?06/2013 at32uc3l0128/256 the number of data bits, endianess, parity mode, and stop bits are selected by the same bits and fields as for the transmitter (mr.chrl, mode9, msbf, par, and nbstop). the synchro- nization mechanism will only c onsider one stop bit, regardless of the used protocol, and when the first stop bit has been sampled, the receiver will automatically begin looking for a new start bit, enabling resynchronization even if there is a protocol miss-match. figure 19-6 and figure 19-7 illustrate start bit detection and char acter reception in asynchronous mode. figure 19-6. asynchronous star t bit detection figure 19-7. asynchronous character reception 19.6.3.3 synchronous receiver in synchronous mode (sync=1), the receiver samples the rxd signal on each rising edge of the baud rate clock. if a low level is detected, it is considered as a start bit. configuration bits and fields are the same as in asynchronous mode. sampling clock (x16) rxd start detection sampling baud rate clock rxd start rejection sampling 12345678 12345670 1234 12345678 9 10111213141516 d0 sampling d0 d1 d2 d3 d4 d5 d6 d7 rxd parity bit stop bit example: 8-bit, parity enabled baud rate clock start detection 16 samples 16 samples 16 samples 16 samples 16 samples 16 samples 16 samples 16 samples 16 samples 16 samples
384 32145c?06/2013 at32uc3l0128/256 figure 19-8. synchronous mode character reception 19.6.3.4 receiver operations when a character reception is completed, it is transferred to the received character field in the receive holding register (rhr.rxchr), and th e receiver ready bit in the channel status register (csr.rxrdy) is set. if rxrdy is alre ady set, rhr will be overwritten and the overrun error bit (csr.ovre) is set. reading rhr will clear rxrdy, and writing a one to the reset status bit in the control register (cr.rststa) will clear ovre. figure 19-9. receiver status 19.6.3.5 parity the usart supports five parity modes selected by mr.par. the par field also enables the multidrop mode, see ?multidrop mode? on page 385 . if even parity is selected, the parity bit will be a zero if there is an even number of ones in the data character, and if there is an odd number it will be a one. for odd pa rity the reverse applies. if space or mark parity is chosen, the parity bit will always be a zero or one, respectively. see table 19-4 . d0 d1 d2 d3 d4 d5 d6 d7 rxd start sampling parity bit stop bit example: 8-bit, parity enabled 1 stop baud rate clock d0 d1 d2 d3 d4 d5 d6 d7 rxd start bit parity bit stop bit baud rate clock write cr rxrdy ovre d0 d1 d2 d3 d4 d5 d6 d7 start bit parity bit stop bit rststa = 1 read rhr table 19-4. parity bit examples alphanum character hex bin parity mode odd even mark space none a 0x41 0100 0001 1 0 1 0 - v 0x56 0101 0110 1 0 1 0 - r 0x52 0101 0010 0 1 1 0 -
385 32145c?06/2013 at32uc3l0128/256 the receiver will report parity errors in csr.par e, unless parity is disabled. writing a one to cr.rststa will clear pare. see figure 19-10 figure 19-10. parity error 19.6.3.6 multidrop mode if par is either 0x6 or 0x7, the usart runs in multidrop mode. this mode differentiates data and address characters. data has the parity bit zero and addresses have a one. by writing a one to the send address bit (cr.senda) the user will cause the next character written to thr to be transmitted as an address. re ceiving a character with a one as parity bit will set pare. 19.6.3.7 transmitter timeguard the timeguard feature enables the usart to interface slow devices by inserting an idle state on the txd line in between two characters. this idle state corresponds to a long stop bit, whose duration is selected by the timeguard value field in the transmitter timeguard register (ttgr.tg). the transmitter will hold the txd line high for tg bit periods, in addition to the num- ber of stop bits. as illustrated in figure 19-11 , the behavior of txrdy and txempty is modified when tg has a non-zero value. if a pending character has been written to thr, the txrdy bit will not be set until this characters start bit has been sent. txempty will remain low until the timeguard transmission has completed. figure 19-11. timeguard operation d0 d1 d2 d3 d4 d5 d6 d7 rxd start bit bad parity bit stop bit baud rate clock write cr pare rxrdy rststa = 1 d0 d1 d2 d3 d4 d5 d6 d7 txd start bit parity bit stop bit baud rate clock start bit tg = 4 write thr d0 d1 d2 d3 d4 d5 d6 d7 parity bit stop bit txrdy txempty tg = 4
386 32145c?06/2013 at32uc3l0128/256 table 19-5. maximum baud rate dependent timeguard durations 19.6.3.8 receiver time-out the time-out value field in the receiver time-o ut register (rtor.to) enables handling of vari- able-length frames by detection of selectable idle durations on the rxd line. the value written to to is loaded to a decr emental counter, and unless it is ze ro, a time-out will occur when the amount of inactive bit periods match the initial counter value. if a time-out has not occurred, the counter will reload and restart ever y time a new character arrives. a time-out sets the timeout bit in csr. clearing timeout can be done in two ways: ? writing a one to the start time-out bit (cr.sttto). this also aborts count down until the next character has been received. ? writing a one to the reload and start time-out bit (cr.retto). this also reloads the counter and restarts count down immediately. figure 19-12. receiver time-o ut block diagram table 19-6. maximum time-out period baud rate (bit/sec) bit time (s) timeguard (ms) 1 200 833 212.50 9 600 104 26.56 14400 69.4 17.71 19200 52.1 13.28 28800 34.7 8.85 33400 29.9 7.63 56000 17.9 4.55 57600 17.4 4.43 115200 8.7 2.21 baud rate (bit/sec) bit time (s) time-out (ms) 600 1 667 109 225 1 200 833 54 613 2 400 417 27 306 4 800 208 13 653 16-bit time-out counter 0 to timeout baud rate clock = character received retto load clock 16-bit value sttto dq 1 clear
387 32145c?06/2013 at32uc3l0128/256 19.6.3.9 framing error the receiver is capable of detecting framing erro rs. a framing error has occurred if a stop bit reads as zero. this can occur if the transmitter and receiver are not synchronized. a framing error is reported by csr.frame as soon as the error is detected, at the middle of the stop bit. figure 19-13. framing error status 19.6.3.10 transmit break when txrdy is set, the user can request the transmitter to generate a break condition on the txd line by writing a one to the start break bit (cr.sttbrk). the break is treated as a normal 0x00 character transmission, clearing txrdy and txempty, but with zeroes for preambles, start, parity, stop, and ti me guard bits. writing a one to the stop break bit (cr.stbrk) will stop the generation of new break characters, and send ones for tg duration or at least 12 bit periods, ensuring that the receiver detects end of break, before resuming normal operation. figure 19-14 illustrates sttbrk and stpbrk effect on the txd line. writing to sttbrk and stpbrk simultaneously can lead to unpredictable results. writes to thr before a pending break has started will be ignored. 9 600 104 6 827 14400 69 4 551 19200 52 3 413 28800 35 2 276 33400 30 1 962 56000 18 1 170 57600 17 1 138 200000 5 328 baud rate (bit/sec) bit time (s) time-out (ms) d0 d1 d2 d3 d4 d5 d6 d7 rxd start bit parity bit stop bit baud rate clock write cr frame rxrdy rststa = 1
388 32145c?06/2013 at32uc3l0128/256 figure 19-14. break transmission 19.6.3.11 receive break a break condition is assumed when incoming data, parity, and stop bits are zero. this corre- sponds to a framing error, but frame will remain zero while the break received/end of break bit (csr.rxbrk) is set. writing a one to cr.rst sta will clear rxbrk. an end of break will also set rxbrk, and is assumed when tx is high for at least 2/16 of a bit period in asynchro- nous mode, or when a high level is sampled in synchronous mode. 19.6.3.12 hardware handshaking the usart features an out-of-band hardwar e handshaking flow control mechanism, imple- mentable by connecting the rts and cts pins with the remote device, as shown in figure 19- 15 . figure 19-15. connection with a remote device for hardware handshaking writing 0x2 to the mr.mode field configures th e usart to operate in this mode. the receiver will drive its rts pin high when disabled or when the reception buffer full bit (csr.rxbuff) is set by the buffer full signal from the peripheral dma controller. if the re ceivers rts pin is high, the transmitters cts pin will also be high and only th e active character transactions will be com- pleted. allocating a new buffer to the dma contro ller by clearing rxbuff, will drive the rts pin low, allowing the transmitter to resume transmission. detected level changes on the cts pin can trigger interrupts, and are reported by the cts input change bit in the channel status reg- ister (csr.ctsic). figure 19-16 illustrates receiver functionality, and figure 19-17 illustrates transmitter functionality. d0 d1 d2 d3 d4 d5 d6 d7 txd start bit parity bit stop bit baud rate clock write cr txrdy txempty stpbrk = 1 sttbrk = 1 break transmission end of break usart txd cts remote device rxd txd rxd rts rts cts
389 32145c?06/2013 at32uc3l0128/256 figure 19-16. receiver behavior when operating with hardware handshaking figure 19-17. transmitter behavior when operating with hardware handshaking figure 19-18. 19.6.4 spi mode the usart features a serial peripheral interface (spi) link compliant mode, supporting syn- chronous, full-duplex communication, in both master and slave mode. writing 0xe (master) or 0xf (slave) to mr.mode will enable this mode. a spi in master mo de controls th e data flow to and from the other spi devices, who are in slave mode. it is possible to let devices take turns being masters (aka multi-master protocol), and one master may shift data simultaneously into several slaves, but only one slave may respond at a time. a slave is selected when its slave select (nss) signal has been raised by the master. the usart can only generate one nss sig- nal, and it is possible to use standard i/o lines to address more than one slave. 19.6.4.1 modes of operation the spi system consists of two data lines and two control lines: ? master out slave in (mosi): this line supplies the data shifted from master to slave. in master mode this is connected to txd, and in slave mode to rxd. ? master in slave out (miso): this line supplies the data shifted from slave to master. in master mode this is connected to rxd, and in slave mode to txd. ? serial clock (clk): this is controlled by the master. one period per bit transmission. in both modes this is connected to clk. ? slave select (nss): this control line allows the master to select or deselect a slave. in master mode this is connected to rts, and in slave mode to cts. changing spi mode after initial configuration has to be followed by a transceiver software reset in order to avoid unpredictable behavior. 19.6.4.2 baud rate the baud rate generator operates as described in ?baud rate in synchronous and spi mode? on page 381 , with the following requirements: in spi master mode: rts rxbuff write cr rxen = 1 rxd rxdis = 1 cts txd
390 32145c?06/2013 at32uc3l0128/256 ? the clock selection field (mr.usclks) must not equal 0x3 (external clock, clk). ? the clock output select bit (mr.clko) must be one. ? the brgr.cd field must be at least 0x4. ? if usclks is one (internal divided clock, clk_usart/div), the value in cd has to be even, ensuring a 50:50 duty cycle. cd can be odd if usclks is zero (internal clock, clk_usart). in spi slave mode: ? clk frequency must be at least four times lower than the system clock. 19.6.4.3 data transfer ? up to nine data bits are successively shifted out on the txd pin at each edge. there are no start, parity, or stop bits, and msb is always sent first. the spi clock polarity (mr.cpol), and spi clock phase (mr.cpha) bits configure clk by selecting the edges upon which bits are shifted and sampled, resulting in four non-interoperable protocol modes see table 19-7 . a master/slave pair must use the same configuration, and the master must be reconfigured if it is to communicate with slaves using different configurations. see figures 19-19 and 19-20 . figure 19-19. spi transfer format (cpha=1, 8 bits per transfer) table 19-7. spi bus protocol modes spi bus protocol mode cpol cpha 001 100 211 310 clk cycle (for reference) clk (cpol= 1) mosi spi master ->txd spi slave ->rxd miso spi master ->rxd spi slave ->txd nss spi master ->rts spi slave ->cts msb msb 1 clk (cpol= 0) 3 5 6 78 lsb 1 2 3 4 6 65 5 43 21 lsb 24
391 32145c?06/2013 at32uc3l0128/256 figure 19-20. spi transfer format (cpha=0, 8 bits per transfer) 19.6.4.4 receiver and transmitter control see ?transmitter operations? on page 382 , and ?receiver operations? on page 384 . 19.6.4.5 character transmission and reception in spi master mode, the slave select line (nss) is asserted low one bit period before the start of transmission, and released high one bit period after every character transmission. a delay for at least three bit periods is always inserted in between characters. in order to address slave devices supporting the chip select active after transf er (csaat) mode, nss can be forced low by writing a one to the force spi chip sele ct bit (cr.rtsen/fcs). releasing nss when fcs is one, is only possible by writing a one to the release spi chip sele ct bit (cr.rtsdis/rcs). in spi slave mode, a low level on nss for at least one bit period will allow the slave to initiate a transmission or reception. the underrun error bi t (csr.unre) is set if a character must be sent while thr is empty, and txd will be high during character transmission , as if 0xff was being sent. if a new character is written to thr it will be sent correctly during the next transmission slot. writing a one to cr.rststa will clear unre. to ensure correct behav ior of the receiver in spi slave mode, the master device sending the frame must ensure a minimum delay of one bit period in between each character transmission. 19.6.4.6 receiver time-out receiver time-out?s are not possible in spi mode as the baud rate clock is only active during data transfers. 19.6.5 lin mode the usart features a lin (local interconnect network) 1.3 and 2.0 compliant mode, embed- ding full error checking and reporting, automatic frame processing with up to 256 data bytes, clk cycle (for reference) clk (cpol= 0) clk (cpol= 1) mosi spi master -> txd spi slave -> rxd miso spi master -> rxd spi slave -> txd nss spi master -> rts spi slave -> cts msb 6 5 msb 6 5 4 43 32 21 1 lsb lsb 8 7 6 5 4 3 2 1
392 32145c?06/2013 at32uc3l0128/256 customizable response data lengths, and requires minimal cpu resources. writing 0xa (master) or 0xb (slave) to mr.mode enables this mode. 19.6.5.1 modes of operation changing lin mode after initial configuration has to be followed by a transceiver software reset in order to avoid unpredictable behavior. 19.6.5.2 receiver and transmitter control see section ?19.6.2? on page 381. 19.6.5.3 baud rate configuration the lin nodes baud rate is configured in the baud rate generator register (brgr), see sec- tion ?19.6.1.1? on page 379. 19.6.5.4 character transmission and reception see ?transmitter operations? on page 382 , and ?receiver operations? on page 384 . 19.6.5.5 header transmission (master node configuration) all lin frames start with a header sent by the master. as soon as the identifier has been written to the identifier character field in the lin id entifier register (linir.idchr), txrdy is cleared and the header is sent. the header consists of a break, sync, and identifier field. txrdy is set when the identifier has been transferred into the transmitters shift register. the break field consists of 13 dominant bits, th e break, and one recessive bit, the break delim- iter. the sync field is the character 0x55. the i dentifier field contains the identifier as written to idchr. the identifier parity bits ca n be genera ted automatically (see section 19.6.5.8 ). figure 19-21. header transmission 19.6.5.6 header reception (slave node configuration) the usart stays idle until it detects a break field, consisting of at least 11 consecutive domi- nant bits (zeroes) on the bus. a received break will set t he lin break bit (csr.linbk). the sync field is used to synchronize the baud rate (see section 19.6.5.7 ). idchr is updated and the lin identifier bit (csr.linid) is set when the identifier has been received. the identifier parity bits can be automatically checked (see section 19.6.5.8 ). writing a one to rststa will clear linbk and linid. txd baud rate clock start bit write linir 10101010 txrdy stop bit start bit id0 id1 id2 id3 id4 id5 id6 id7 break field 13 dominant bits (at 0) stop bit break delimiter 1 recessive bit (at 1) synch byte = 0x55 linir id
393 32145c?06/2013 at32uc3l0128/256 figure 19-22. header reception 19.6.5.7 slave node synchronization synchronization is only done by the slave. if t he sync field is not 0x55, an inconsistent sync field error (csr.linisfe) is gen erated. the time betw een falling edges is measured by a 19-bit counter, driven by the sampling clock (see section 19.6.1 ). figure 19-23. sync field the counter starts when the sync field start bit is detected, and continues for eight bit periods. the 16 most significant bits (counter value divided by 8) becomes the new clock divider (brgr.cd), and the three least significant bits (the remainder) becomes the new fractional part (brgr.fp). figure 19-24. slave node synchronization the synchronization accuracy depends on: ? the theoretical slave node clock frequency; nominal clock frequency (f nom ) ? the baud rate break field 13 dominant bits (at 0) break delimiter 1 recessive bit (at 1) start bit 10101010 stop bit start bit id0 id1 id2 id4 id3 id6 id5 id7 stop bit synch byte = 0x55 baud rate clock rxd write us_cr with rststa=1 us_linir linid start bit stop bit synch field 8 tbit 2 tbit 2 tbit 2 tbit 2 tbit rxd baud rate clock linidrx synchro counter 000_0011_0001_0110_1101 brgr clcok divider (cd) 0000_0110_0010_1101 brgr fractional part (fp) 101 initial cd initial fp reset start bit 10101010 stop bit start bit id0 id1 id2 id3 id4 id5 id6 id7 break field 13 dominant bits (at 0) stop bit break delimiter 1 recessive bit (at 1) synch byte = 0x55
394 32145c?06/2013 at32uc3l0128/256 ? the oversampling mode (over=0 => 16x, or over=1 => 8x) the following formula is used to calculate synchronization deviation, where f slave is the real slave node clock frequency, and f tol_unsync is the difference between f nom and f slave accord- ing to the lin specification, f tol_unsynch may not exceed 15%, and the bit rates between two nodes must be within 2% of each other, resulting in a maximal baudrate_deviation of 1%. minimum nominal clock frequency with a fractional part: examples: ? baud rate = 20 kbit/s, over=0 (oversampling 16x) => f nom (min) = 2.64 mhz ? baud rate = 20 kbit/s, over=1 (oversampling 8x) => f nom (min) = 1.47 mhz ? baud rate = 1 kbit/s, over=0 (oversampling 16x) => f nom (min) = 132 khz ? baud rate = 1 kbit/s, over=1 (oversampling 8x) => f nom (min) = 74 khz if the fractional part is not used, the synchroniza tion accuracy is much lower. the 16 most signif- icant bits, added with the first least significant bit, becomes the new clock divider (cd). the equation of the baud rate deviation is the same as above, but the constants are: minimum nominal clock frequency without a fractional part: examples: ? baud rate = 20 kbit/s, over=0 (oversampling 16x) => f nom (min) = 19.12 mhz ? baud rate = 20 kbit/s, over=1 (oversampling 8x) => f nom (min) = 9.71 mhz ? baud rate = 1 kbit/s, over=0 (oversampling 16x) => f nom (min) = 956 khz ? baud rate = 1 kbit/s, over=1 (oversampling 8x) => f nom (min) = 485 khz 19.6.5.8 identifier parity an identifier field consists of two sub-fields; the identifier and its parity. bits 0 to 5 are assigned to the identifier, while bits 6 and 7 are assigned to parity. automatic parity management is dis- abled by writing a one to the parity disable bit in the lin mode register (linmr.pardis). baudrate_deviation 100 ? ? 8 2 over ? ??? + ? baudrate ? ? ? 8f slave ? -------------------------------------------------------------------------------------------------- - ? ?? ?? % = baudrate_deviation 100 ? ? 8 2 over ? ??? + ? baudrate ? ? ? 8 f tol_unsync 100 ----------------------------------- - ?? ?? xf nom ? -------------------------------------------------------------------------------------------------- - ? ?? ?? ?? ?? ?? % = 0.5 ? ? +0.5 -1 ? +1 ?? ?? ?? 100 0.5 8 2 over ? ?? ? ? 1 + ?? baudrate ? 8 15 ? 100 --------- - 1 + ?? ?? ? 1% ? ------------------------------------------------------------------------------------------------------ - ? ?? ?? ?? ?? ?? hz = 4 ? ? +4 -1 ? +1 ?? ?? ?? 100 48 2over ? ?? ? ? 1 + ?? baudrate ? 8 15 ? 100 --------- - 1 + ?? ?? ? 1% ? ----------------------------------------------------------------------------------------------- ? ?? ?? ?? ?? ?? hz =
395 32145c?06/2013 at32uc3l0128/256 ? pardis=0: during header transmission, the parity bits are computed and in the shift register they replace bits six and seven from idchr. du ring header reception, the parity bits are checked and can generate a lin identifier parity error (see section 19.6.6 ). bits six and seven in idchr read as zero when receiving. ? pardis=1: during header transmission, all th e bits in idchr are sent on the bus. during header reception, all the bits in idchr are updated with the received identifier. 19.6.5.9 node action after an identifier transaction, a lin response mode has to be selected. this is done in the node action field (linmr.nact). below are some response modes exemplified in a small lin cluster: ? response, from master to slave1: master: nact=publish slave1: nact=subscribe slave2: nact=ignore ? response, from slave1 to master: master: nact=subscribe slave1: nact=publish slave2: nact=ignore ? response, from slave1 to slave2: master: nact=ignore slave1: nact=publish slave2: nact=subscribe 19.6.5.10 lin response data length the response data length is the number of data fields (bytes), excluding the checksum. figure 19-25. response data length the response data length can be configured, either by the user, or automatically by bits 4 and 5 in the identifier (idchr), in acco rdance to lin 1.1. the user sele cts mode by writing to the data length mode bit (linmr.dml): ? dlm=0: the response data length is configured by the user by writing to the 8-bit data length control field (linmr.dlc). the response data length equals dlc + 1 bytes. user configuration: 1 - 256 data fields (dlc+1) identifier configuration: 2/4/8 data fields sync break sync field identifier field checksum field data field data field data field data field
396 32145c?06/2013 at32uc3l0128/256 ? dlm=1: the response data length is defined by the identifier bits according to the table below. 19.6.5.11 checksum the last frame field is the checksum. it is configured by the checksum type (linmr.chktyp), and the checksum disable (linmr.chkdis) bits. txrdy will not be set after the last thr data write if enabled. writing a one to chkdis will disable the automatic checksum genera- tion/checking, and the user may send/check this last byte manually, disguised as a normal data. the checksum is an inverted 8-bit sum with carry, either: ? over all data bytes, called a classic checksum. this is used for lin 1.3 compliant slaves, and automatically managed when chkdis=0, and chktyp=1. ? over all data bytes and the protected identifier, called an enhanced checksum. this is used for lin 2.0 compliant slaves, and automatically managed when chkdis=0, and chktyp=0. 19.6.5.12 frame slot mode a lin master can be configured to use frame slots with a pre-defined minimum length. writing a one to the frame slot mode disable bit (linmr.f sdis) disables this mode. this mode will not allow txrdy to be set after a frame transfer until the entire frame slot duration has elapsed, in effect preventing the master from sending a new header. the lin transfer complete bit (csr.lintc) will still be set after the checksum has been s ent. writing a one to cr.rstst clears lintc. figure 19-26. frame slot mode with automatic checksum the minimum frame slot size is determined by tframe_maximum, and calculated below (all val- ues in bit periods): ? theader_nominal = 34 table 19-8. response data length if dlm = 1 idchr[5] idchr[4] response data length [bytes] 00 2 01 2 10 4 11 8 break synch protected identifier data n checksum header inter- frame space response space frame frame slot = tframe_maximum response txrdy write thr write linid data 1 data 2 data 3 data3 data n-1 data n frame slot mode disabled frame slot mode enabled lintc data 1
397 32145c?06/2013 at32uc3l0128/256 ? tframe_maximum = 1.4 x (theader_nominal + tresponse_nominal + 1) (note:) note: the term ?+1? leads to an integer re sult for tframe_max (lin specification 1.3) if the checksum is sent (chkdis=0): ? tresponse_nominal = 10 x (ndata + 1) ? tframe_maximum = 1.4 x (34 + 10 x (dlc + 1 + 1) + 1) ? tframe_maximum = 77 + 14 x dlc if the checksum is not sent (chkdis=1): ? tresponse_nominal = 10 x ndata ? tframe_maximum = 1.4 x (34 + 10 x (dlc + 1) + 1) ? tframe_maximum = 63 + 14 x dlc 19.6.6 lin errors these error bits are cleared by writing a one to csr.rststa. 19.6.6.1 slave not responding error (csr.linsnre) this error is generated if no valid message appears within the tframe_maximum time frame slot, while the usart is expecting a response from another node (nact=subscribe). 19.6.6.2 checksum error (csr.lince) this error is generated if the received checksum is wrong. this error can only be generated if the checksum feature is enabled (chkdis=0). 19.6.6.3 identifier parity error (csr.linipe) this error is generated if the identifier parity is wrong. this error can only be generated if parity is enabled (pardis=0). 19.6.6.4 inconsistent sync field error (csr.linisfe) this error is generated in slave mode if the sy nc field character receiv ed is not 0x55. synchro- nization procedure is aborted. 19.6.6.5 bit error (csr.linbe) this error is generated if the value transmitted by the usart on tx differs from the value sam- pled on rx. if a bit error is detected, the tran smission is aborted at the next byte border. 19.6.7 lin frame handling 19.6.7.1 master node configuration ? write a one to cr.txen and cr.rxen to enable both transmitter and receiver ? select lin mode and master node by writing to mr.mode ? configure the baud rate by writing to cd and fp in brgr ? configure the frame transfer by writing to nact, pardis, chkdis, chktype, dlcm, fsdis, and dlc in linmr ? check that csr.txrdy is one ? send the header by wr iting to linir.idchr the following procedure depends on the nact setting:
398 32145c?06/2013 at32uc3l0128/256 ? case 1: nact=publish, the usart sends a response ? wait until txrdy is a one ? send a byte by writing to thr.txchr ? repeat the two previous steps until there is no more data to send ? wait until csr.lintc is a one ? check for lin errors ? case 2: nact=subscribe, the usart receives a response ? wait until rxrdy is a one ? read rhr.rxchr ? repeat the two previous steps until there is no more data to read ? wait until lintc is a one ? check for lin errors ? case 3: nact=ignore, the usart is not concerned by a response ? wait until lintc is a one ? check for lin errors figure 19-27. master node configur ation, nact=publish frame break synch protected identifier data 1 data n checksum txrdy write thr write linir data 1 data 2 data 3 data n-1 data n rxrdy header inter- frame space response space frame slot = tframe_maximum response data3 lintc fsdis=1 fsdis=0
399 32145c?06/2013 at32uc3l0128/256 figure 19-28. master node configur ation, nact=subscribe figure 19-29. master node configuration, nact=ignore 19.6.7.2 slave node configuration this is identical to the master node configuration above, except for: ? lin mode selected in mr.mode is slave ? when the baud rate is configured, wait until csr.linid is a one, then; ? check for linisfe and linpe errors, clear errors and linidby writing a one to rststa ? read idchr ? configure the frame transfer by writing to nact, pardis, chkdis, chktype, dlcm, and dlc in linmr important : if nact=publish, and this field is alre ady correct, the linmr register must still be written with this value in order to set txrdy, and to request the corresponding peripheral dma controller write transfer. the different nact settings result in the same procedure as for the master node, see page 397 . break synch protected identifier data 1 data n checksum txrdy read rhr write linir data 1 data n-1 data n-1 rxrdy data n data n-2 header inter- frame space response space frame frame slot = tframe_maximum response data3 lintc fsdis=0 fsdis=1 txrdy write linir rxrdy lintc break synch protected identifier data 1 data n checksum data n-1 header inter- frame space response space frame frame slot = tframe_maximum response data3 fsdis=1 fsdis=0
400 32145c?06/2013 at32uc3l0128/256 figure 19-30. slave node configuration, nact=publish figure 19-31. slave node configuration, nact=subscribe figure 19-32. slave node configuration, nact=ignore 19.6.8 lin frame handling with the peripheral dma controller the usart can be used together with the peripheral dma controller in order to transfer data without processor intervention. the dma controller uses the txrdy and rxrdy bits, to trigger one byte writes or reads. it always writes to thr, and it always reads rhr. break synch protected identifier data 1 data n checksum txrdy write thr read linid data 1 data 3 data n-1 data n rxrdy linidrx data 2 lintc txrdy read rhr read linid rxrdy linidrx lintc break synch protected identifier data 1 data n checksum data 1 data n-1 data n-1 data n data n-2 txrdy read rhr read linid rxrdy linidrx lintc break synch protected identifier data 1 data n checksum data n-1
401 32145c?06/2013 at32uc3l0128/256 19.6.8.1 master node configuration the peripheral dma controller mode bit (linmr.pdc m) allows the user to select configuration: ? pdcm=0: lin configuration must be written to li nmr, it is not stored in the write buffer. ? pdcm=1: lin configuration is written by the dma controller to thr, and is stored in the write buffer. since data transfer size is a byte, the transfer is split into two accesses. the first writes the nact, pardis, chkdis, chktyp, dl m and fsdis bits, while the second writes the dlc field. if nact=publish, the writ e buffer will also contain the identifier. when nact=subscribe, the read buffer contains the data. figure 19-33. master node with periphera l dma controller (pdcm=0) figure 19-34. master node with periphera l dma controller (pdcm=1) | | | | rxrdy txrdy peripheral bus usart lin controller data 0 data n | | | | read buffer node action = publish node action = subscribe peripheral dma controller rxrdy peripheral bus data 0 data 1 data n write buffer peripheral dma controller usart lin controller | | | | | | | | nact pardis chkdis chktyp dlm fsdis dlc identifier data 0 data n write buffer rxrdy peripheral bus dlc identifier data 0 data n write buffer rxrdy read buffer node action = publish node action = subscribe peripheral dma controller peripheral dma controller usart lin controller nact pardis chkdis chktyp dlm fsdis usart lin controller txrdy peripheral bus
402 32145c?06/2013 at32uc3l0128/256 19.6.8.2 slave node configuration in this mode, the peripheral dma controller transfers only data. the user reads the identifier from linir, and selects lin mode by writing to linmr. when na ct=publish the data is in the write buffer, while the read buffer contains the data when nact=subscribe. important: if in slave mode, nact is alread y configured correctly as publish, the linmr register must still be written with this value in order to set txrdy, and to request the corre- sponding peripheral dma controller write transfer. figure 19-35. slave node with peripheral dma controller 19.6.9 wake-up request any node in a sleeping lin cluster may request a wake-up. by writing to the wakeup signal type bit (linmr.wkuptyp), the user can choose to send either a lin 1.3 (wkuptyp=1), or a lin 2.0 (wkuptyp=0) compliant wakeup request. writing a one to the send lin wakeup sig- nal bit (cr.linwkup), transmits a wake up, and when completed sets lintc. according to lin 1.3, the wakeup request should be generated with the character 0x80 in order to impose eight successive dominant bits. according to lin 2.0, the wakeup request is issued by forcing the bus into the dominant state for 250s to 5ms. sending the character 0xf0 does this, regardless of baud rate. ? baud rate max = 20 kbit/s -> one bit period = 50s -> five bit periods = 250s ? baud rate min = 1 kbit/s -> one bit period = 1ms -> five bit periods = 5ms 19.6.10 bus idle time-out lin bus inactivity should eventually cause slaves to time-out and enter sleep mode. lin 1.3 specifies this to 25000 bit periods, whilst lin 2.0 specifies 4seconds. for the time-out counter operation see section 19.6.3.8 ?receiver time-out? on page 386 . | | | | | | | | data 0 data n rxrdy per ipheral bus read buffer nact = subscribe data 0 data n txrdy per ipheral bus write buffer usart lin controller usart lin controller peripheral dma controller peripheral dma controller table 19-9. receiver time-out values (rtor.to) lin specification baud rate time-out period to 2.0 1 000 bit/s 4s 4 000 2 400 bit/s 9 600 9 600 bit/s 38 400 19 200 bit/s 76 800 20 000 bit/s 80 000 1.3 - 25 000 bit periods 25 000
403 32145c?06/2013 at32uc3l0128/256 19.6.11 test modes the internal loopback feature enables on-board diagnostics, and allows the usart to operate in three different test modes, with reconfigured pin functionality, as shown below. 19.6.11.1 normal mode during normal operation, a receivers rxd pin is connected to a transmitters txd pin. figure 19-36. normal mode configuration 19.6.11.2 automatic echo mode automatic echo mode allows bit-by-bit retransmission. when a bit is received on the rxd pin, it is also sent to the txd pin, as shown in figure 19-37 . transmitter configuration has no effect. figure 19-37. automatic echo mode configuration 19.6.11.3 local loopback mode local loopback mode c onnects the output of the transmitter directly to the input of the receiver, as shown in figure 19-38 . the txd and rxd pins are not used. the rxd pin has no effect on the receiver and the txd pin is continuously driven high, as in idle state. figure 19-38. local loopback mode configuration 19.6.11.4 remote loopback mode remote loopback mode connects the rxd pin to the txd pin, as shown in figure 19-39 . the transmitter and the receiver are disabled and have no effect. this mode allows bit-by-bit retransmission. receiver transmitter rxd txd receiver transmitter rxd txd receiver transmitter rxd txd 1
404 32145c?06/2013 at32uc3l0128/256 figure 19-39. remote loopback mode configuration 19.6.12 write protection registers to prevent single software errors from corrupting usart behavior, certain address spaces can be write-protected by writing the correct write protect key and a one to the write protect enable bit in the write protect mode regi ster (wpmr.wpkey, and wpmr.wpen). disabling the write protection is done by writing the correct key, and a zero to wpen. write attempts to a write protected register are detected and the write protect violation status bit in the write protect status register (wpsr.wpvs) is set, wh ile the write protect violation source field (wpsr.wpvsrc) indicates the tar geted register. writing the correct key to the write protect key bit (wpmr.wpkey) clears wpvsrc and wpvs. the protected registers are: ? ?mode register? on page 408 ? ?baud rate generator register? on page 418 ? ?receiver time-out register? on page 419 ? ?transmitter timeguard register? on page 420 receiver transmitter rxd txd 1
405 32145c?06/2013 at32uc3l0128/256 19.7 user interface note: 1. values in the version register vary wi th the version of the ip block implementation. table 19-10. usart register memory map offset register name access reset 0x0000 control register cr write-only 0x00000000 0x0004 mode register mr read-write 0x00000000 0x0008 interrupt enable register ier write-only 0x00000000 0x000c interrupt disable register idr write-only 0x00000000 0x0010 interrupt mask register imr read-only 0x00000000 0x0014 channel status register csr read-only 0x00000000 0x0018 receiver holding register rhr read-only 0x00000000 0x001c transmitter holding register thr write-only 0x00000000 0x0020 baud rate generator register brgr read-write 0x00000000 0x0024 receiver time-out register rtor read-write 0x00000000 0x0028 transmitter timeguard register ttgr read-write 0x00000000 0x0054 lin mode register linmr read-write 0x00000000 0x0058 lin identifier register linir read-write 0x00000000 0x00e4 write protect mode register wpmr read-write 0x00000000 0x00e8 write protect status register wpsr read-only 0x00000000 0x00fc version register version read-only 0x? (1)
406 32145c?06/2013 at32uc3l0128/256 19.7.1 control register name: cr access type: write-only offset: 0x0 reset value: 0x00000000 ? linwkup: send lin wakeup signal writing a zero to this bit has no effect. writing a one to this bit will sends a wakeup signal on the lin bus. ? linabt: abort lin transmission writing a zero to this bit has no effect. writing a one to this bit will abort the current lin transmission. ? rtsdis/rcs: request to send disable/release spi chip select writing a zero to this bit has no effect. writing a one to this bit when usart is not in spi master mode drives rts pin high. writing a one to this bit when usart is in spi master mode releases nss (rts pin). ? rtsen/fcs: request to send enable/force spi chip select writing a zero to this bit has no effect. writing a one to this bit when usart is not in spi master mode drives rts low. writing a one to this bit when usart is in spi master mode when; fcs=0: has no effect. fcs=1: forces nss (rts pin) low, even if usart is not transmitting, in order to address spi slave devices supporting the csaat mode (chip select active after transfer). ? retto: rearm time-out writing a zero to this bit has no effect. writing a one to this bit reloads the time-out counter and clears csr.timeout. ? rstnack: reset non acknowledge writing a zero to this bit has no effect. writing a one to this bit clears csr.nack. ? senda: send address writing a zero to this bit has no effect. writing a one to this bit will in multidrop mode se nd the next character written to thr as an address. ? sttto: start time-out writing a zero to this bit has no effect. writing a one to this bit will abort any current time-out count down, and trigger a new count down when the next character has been received. csr.timeout is also cleared. 31 30 29 28 27 26 25 24 ???????? 23 22 21 20 19 18 17 16 ? ? linwkup linabt rtsdis/rcs rtsen/fcs ? ? 15 14 13 12 11 10 9 8 retto rstnack ? senda sttto stpbrk sttbrk rststa 76543210 txdis txen rxdis rxen rsttx rstrx ? ?
407 32145c?06/2013 at32uc3l0128/256 ? stpbrk: stop break writing a zero to this bit has no effect. writing a one to this bit will stop the generation of break sign al characters, and then send ones for ttgr.tg duration, or at l east 12 bit periods. no effect if no break is being transmitted. ? sttbrk: start break writing a zero to this bit has no effect. writing a one to this bit will start transmission of break charac ters when current characters present in thr and the transmit s hift register have been sent. no effect if a break signal is already being generated. ? rststa: reset status bits writing a zero to this bit has no effect. writing a one to this bit will clear the following bits in csr: pare, frame, ovre, linbe, linsfe, linipe, lince, linsnre, and rxbrk. ? txdis: transmitter disable writing a zero to this bit has no effect. writing a one to this bit disables the transmitter. ? txen: transmitter enable writing a zero to this bit has no effect. writing a one to this bit enables the transmitter if txdis is zero. ? rxdis: receiver disable writing a zero to this bit has no effect. writing a one to this bit disables the receiver. ? rxen: receiver enable writing a zero to this bit has no effect. writing a one to this bit enables the receiver if rxdis is zero. ? rsttx: reset transmitter writing a zero to this bit has no effect. writing a one to this bit will reset the transmitter. ? rstrx: reset receiver writing a zero to this bit has no effect. writing a one to this bit will reset the receiver.
408 32145c?06/2013 at32uc3l0128/256 19.7.2 mode register name: mr access type: read-write offset: 0x4 reset value: 0x00000000 this register can only be written if the wpen bit is cleared in the write protect mode register. ? inack: inhibit non acknowledge 0: the nack is generated. 1: the nack is not generated. ? over: oversampling mode 0: oversampling at 16 times the baud rate. 1: oversampling at 8 times the baud rate. ? clko: clock output select 0: the usart does not drive the clk pin. 1: the usart drives the clk pin unless usclks selects the external clock. ? mode9: 9-bit character length 0: chrl defines character length. 1: 9-bit character length. ? msbf/cpol: bit order or spi clock polarity if usart does not operate in spi mode: msbf=0: least significant bit is sent/received first. msbf=1: most significant bit is sent/received first. if usart operates in spi mode, cpol is used with cpha to produce the required clock/data relationship between devices. cpol=0: the inactive state value of clk is logic level zero. cpol=1: the inactive state value of clk is logic level one. 31 30 29 28 27 26 25 24 ????? ? 23 22 21 20 19 18 17 16 ? ? ? inack over clko mode9 msbf/cpol 15 14 13 12 11 10 9 8 chmode nbstop par sync/cpha 76543210 chrl usclks mode
409 32145c?06/2013 at32uc3l0128/256 ? chmode: channel mode ? nbstop: number of stop bits ? par: parity type ? sync/cpha: synchronous mode select or spi clock phase if usart does not operate in spi mode (mode is ? 0xe and 0xf): sync = 0: usart operates in asynchronous mode. sync = 1: usart operates in synchronous mode. if usart operates in spi mode, cpha determines which edge of clk causes data to change and which edge causes data to be captured. cpha is used with cpol to produce the requ ired clock/data relationship between master and slave devices. cpha = 0: data is changed on the leading edge of clk and captured on the following edge of clk. cpha = 1: data is captured on the leading edge of clk and changed on the following edge of clk. table 19-11. chmode mode description 0 0 normal mode 0 1 automatic echo. receiver input is connected to the txd pin. 1 0 local loopback. transmitter output is connected to the receiver input. 1 1 remote loopback. rxd pin is internally connected to the txd pin. table 19-12. nbstop asynchronous (sync=0) synchronous (sync=1) 0 0 1 stop bit 1 stop bit 0 1 1.5 stop bits reserved 1 0 2 stop bits 2 stop bits 1 1 reserved reserved table 19-13. par parity type 0 0 0 even parity 001odd parity 0 1 0 parity forced to 0 (space) 0 1 1 parity forced to 1 (mark) 1 0 x no parity 1 1 x multidrop mode
410 32145c?06/2013 at32uc3l0128/256 ? chrl: character length. ? usclks: clock selection note: 1. the value of div is device dependent. please refer to the module configuration secti on at the end of this chapter. ?mode table 19-14. chrl character length 0 0 5 bits 0 1 6 bits 1 0 7 bits 1 1 8 bits table 19-15. usclks selected clock 0 0 clk_usart 0 1 clk_usart/div (1) 10reserved 11 clk table 19-16. mode mode of the usart 0000normal 0010hardware handshaking 1010lin master 1011lin slave 1110spi master 1111spi slave others reserved
411 32145c?06/2013 at32uc3l0128/256 19.7.3 interrupt enable register name: ier access type: write-only offset: 0x8 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will set the corresponding bit in imr. 31 30 29 28 27 26 25 24 ? ? linsnre lince linipe linisfe linbe ? 23 22 21 20 19 18 17 16 ????ctsic? ? ? 15 14 13 12 11 10 9 8 lintc linid nack/linbk rxbuff ? iter/unre txempty timeout 76543210 pare frame ovre ? ? rxbrk txrdy rxrdy
412 32145c?06/2013 at32uc3l0128/256 19.7.4 interrupt disable register name: idr access type: write-only offset: 0xc reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in imr. 31 30 29 28 27 26 25 24 ? ? linsnre lince linipe linisfe linbe ? 23 22 21 20 19 18 17 16 ????ctsic? ? ? 15 14 13 12 11 10 9 8 lintc linid nack/linbk rxbuff ? iter/unre txempty timeout 76543210 pare frame ovre ? ? rxbrk txrdy rxrdy
413 32145c?06/2013 at32uc3l0128/256 19.7.5 interrupt mask register name: imr access type: read-only offset: 0x10 reset value: 0x00000000 0: the corresponding interrupt is disabled. 1: the corresponding interrupt is enabled. a bit in this register is cleared when the corresponding bit in idr is written to one. a bit in this register is set when the corresponding bit in ier is written to one. 31 30 29 28 27 26 25 24 ? ? linsnre lince linipe linisfe linbe ? 23 22 21 20 19 18 17 16 ????ctsic? ? ? 15 14 13 12 11 10 9 8 lintc linid nack/linbk rxbuff ? iter/unre txempty timeout 76543210 pare frame ovre ? ? rxbrk txrdy rxrdy
414 32145c?06/2013 at32uc3l0128/256 19.7.6 channel status register name: csr access type: read-only offset: 0x14 reset value: 0x00000000 ? linsnre: lin slave not responding error 0: no lin slave not responding error has been detected since the last rststa. 1: a lin slave not responding error has been detected since the last rststa. ? lince: lin checksum error 0: no lin checksum error has been detected since the last rststa. 1: a lin checksum error has been detected since the last rststa. ? linipe: lin identi fier parity error 0: no lin identifier parity error has been detected since the last rststa. 1: a lin identifier parity error has been detected since the last rststa. ? linisfe: lin inconsistent sync field error 0: no lin inconsistent sync field erro r has been detected since the last rststa 1: the usart is configured as a slave node and a lin inconsiste nt sync field error has been detected since the last rststa. ? linbe: lin bit error 0: no bit error has been detected since the last rststa. 1: a bit error has been detected since the last rststa. ? cts: image of cts input 0: cts is low. 1: cts is high. ? ctsic: clear to send input change flag 0: no change has been detected on the cts pin since the last csr read. 1: at least one change has been detected on the cts pin since the last csr read. ? lintc: lin transfer completed 0: the usart is either idle or a lin transfer is ongoing. 1: a lin transfer has been comp leted since the last rststa. ? linid: lin identifier 0: no lin identifier has been sent or received. 1: a lin identifier has been s ent (master) or received (slave), since the last rststa. ? nack: non acknowledge 0: no non acknowledge has been detected since the last rstnack. 1: at least one non acknowledge has been detected since the last rstnack. ? rxbuff: reception buffer full 0: the buffer full signal from the peripheral dma controller channel is inactive. 31 30 29 28 27 26 25 24 ? ? linsnre lince linipe linisfe linbe ? 23 22 21 20 19 18 17 16 cts ? ? ? ctsic ? ? ? 15 14 13 12 11 10 9 8 lintc linid nack/linbk rxbuff ? iter/unre txempty timeout 76543210 pare frame ovre ? ? rxbrk txrdy rxrdy
415 32145c?06/2013 at32uc3l0128/256 1: the buffer full signal from the peripheral dma controller channel is active. ? iter/unre: max number of repetitions reached or spi underrun error if usart does not operate in spi slave mode: iter=0: maximum number of repetitions has not been reached since the last rststa. iter=1: maximum number of repetitions has been reached since the last rststa. if usart operates in spi slave mode: unre=0: no spi underrun error has occurred since the last rststa. unre=1: at least one spi underrun error has occurred since the last rststa. ? txempty: transmitter empty 0: the transmitter is either disabled or there are ch aracters in thr, or in the transmit shift register. 1: there are no characters in neither th r, nor in the transmit shift register. ? timeout: receiver time-out 0: there has not been a time-out since the last star t time-out command (cr.sttto), or rtor.to is zero. 1: there has been a time-out since the last start time-out command. ? pare: parity error 0: either no parity error has been detect ed, or the parity bit is a zero in multidrop mode, since the last rststa. 1: either at least one parity error has been detected, or the pa rity bit is a one in multidrop mode, since the last rststa. ? frame: framing error 0: no stop bit has been found as low since the last rststa. 1: at least one stop bit has been f ound as low since the last rststa. ? ovre: overrun error 0: no overrun error has occurred since the last rststa. 1: at least one overrun error has occurred since the last rststa. ? rxbrk: break received/end of break 0: no break received or end of break detected since the last rststa. 1: break received or end of break detected since the last rststa. ? txrdy: transmitter ready 0: the transmitter is either disabled, or a character in thr is waiting to be transferred to the transmit shift register, or an sttbrk command has been requested. as soon as th e transmitter is enabled, txrdy becomes one. 1: there is no character in the thr. ? rxrdy: receiver ready 0: the receiver is either disabled, or no complete character has been received since the last read of rhr. if characters were being received when the receiver was disabled, rxrdy changes to 1 when the receiver is enabled. 1: at least one complete character has been received and rhr has not yet been read.
416 32145c?06/2013 at32uc3l0128/256 19.7.7 receiver holding register name: rhr access type: read-only offset: 0x18 reset value: 0x00000000 ? rxchr: received character last received character. 31 30 29 28 27 26 25 24 ???????? 23 22 21 20 19 18 17 16 ???????? 15 14 13 12 11 10 9 8 ???????rxchr[8] 76543210 rxchr[7:0]
417 32145c?06/2013 at32uc3l0128/256 19.7.8 transmitter holding register name: thr access type: write-only offset: 0x1c reset value: 0x00000000 ? txchr: character to be transmitted if txrdy is zero this field contains the next character to be transmitted. 31 30 29 28 27 26 25 24 ???????? 23 22 21 20 19 18 17 16 ???????? 15 14 13 12 11 10 9 8 ???????tx chr[8] 76543210 txchr[7:0]
418 32145c?06/2013 at32uc3l0128/256 19.7.9 baud rate generator register name: brgr access type: read-write offset: 0x20 reset value: 0x00000000 this register can only be written to if write protection is disabled, see ?write protect mode register? on page 424 . ? fp: fractional part 0: fractional divider is disabled. 1 - 7: baud rate resolution, defined by fp x 1/8. ? cd: clock divider 31 30 29 28 27 26 25 24 ???????? 23 22 21 20 19 18 17 16 ????? fp 15 14 13 12 11 10 9 8 cd[15:8] 76543210 cd[7:0] table 19-17. cd sync = 0 sync = 1 or mode = spi (master or slave) over = 0 over = 1 0 baud rate clock disabled 1 to 65535 baud rate = selected clock/16/cd baud rate = selected clock/8/cd baud rate = selected clock /cd
419 32145c?06/2013 at32uc3l0128/256 19.7.10 receiver time-out register name: rtor access type: read-write offset: 0x24 reset value: 0x00000000 this register can only be written to if write protection is disabled, see ?write protect mode register? on page 424 . ? to: time-out value 0: the receiver time-out is disabled. 1 - 131071: the receiver time-out is enabled and the time-out delay is to x bit period. note that the size of the to counter is devi ce dependent, see the module configuration section. 31 30 29 28 27 26 25 24 ???????? 23 22 21 20 19 18 17 16 ???????to[16] 15 14 13 12 11 10 9 8 to[15:8] 76543210 to[7:0]
420 32145c?06/2013 at32uc3l0128/256 19.7.11 transmitter timeguard register name: ttgr access type: read-write offset: 0x28 reset value: 0x00000000 this register can only be written to if write protection is disabled, see ?write protect mode register? on page 424 . ? tg: timeguard value 0: the transmitter timeguard is disabled. 1 - 255: the transmitter timeguard is enabled and the timeguard delay is tg x bit period. 31 30 29 28 27 26 25 24 ???????? 23 22 21 20 19 18 17 16 ???????? 15 14 13 12 11 10 9 8 ???????? 76543210 tg
421 32145c?06/2013 at32uc3l0128/256 19.7.12 lin mode register name: linmr access type: read-write offset: 0x54 reset value: 0x00000000 ? pdcm: peripheral dma controller mode 0: the lin mode register is not writ ten by the peripheral dma controller. 1: the lin mode register is, except for this bit, written by the peripheral dma controller. ? dlc: data length control 0 - 255: if dlm=0 this field defines the response data length to dlc+1 bytes. ? wkuptyp: wakeup signal type 0: writing a one to cr.linwkup will send a lin 2.0 wakeup signal. 1: writing a one to cr.linwkup will send a lin 1.3 wakeup signal. ? fsdis: frame slot mode disable 0: the frame slot mode is enabled. 1: the frame slot mode is disabled. ? dlm: data length mode 0: the response data length is defined by dlc. 1: the response data length is defined by bi ts 4 and 5 of the identifier (linir.idchr). ? chktyp: checksum type 0: lin 2.0 ?enhanced? checksum 1: lin 1.3 ?classic? checksum ? chkdis: checksum disable 0: checksum is automatically computed and sent when master, and checked when slave. 1: checksum is not computed and sent, nor checked. ? pardis: parity disable 0: identifier parity is automatically computed and sent when master, and checked when slave. 1: identifier parity is not computed and sent, nor checked. ? nact: lin node action 31 30 29 28 27 26 25 24 ???????? 23 22 21 20 19 18 17 16 ???????pdcm 15 14 13 12 11 10 9 8 dlc 76543210 wkuptyp fsdis dlm chktyp chkdis pardis nact table 19-18. nact mode description 0 0 publish: the usart transmits the response.
422 32145c?06/2013 at32uc3l0128/256 0 1 subscribe: the usart receives the response. 1 0 ignore: the usart does not transmit and does not receive the response. 11reserved table 19-18.
423 32145c?06/2013 at32uc3l0128/256 19.7.13 lin identifier register name: linir access type: read-write or read-only offset: 0x58 reset value: 0x00000000 ? idchr: identifier character if usart is in lin master mode, the idchr field is read-write, and its value is the identifier character to be transmitted. if usart is in lin slave mode, the idchr field is read-only , and its value is the last received identifier character. 31 30 29 28 27 26 25 24 ???????? 23 22 21 20 19 18 17 16 ???????? 15 14 13 12 11 10 9 8 ???????? 76543210 idchr
424 32145c?06/2013 at32uc3l0128/256 19.7.14 write protect mode register register name: wpmr access type: read-write offset: 0xe4 reset value: see table 19-10 ? wpkey: write protect key has to be written to 0x555341 (?usa? in ascii) in order to successfully write wpen. always reads as zero. ? wpen: write protect enable 0 = write protection disabled. 1 = write protection enabled. protects the registers: ? ?mode register? on page 408 ? ?baud rate generator register? on page 418 ? ?receiver time-out register? on page 419 ? ?transmitter timeguard register? on page 420 31 30 29 28 27 26 25 24 wpkey[23:16] 23 22 21 20 19 18 17 16 wpkey[15:8] 15 14 13 12 11 10 9 8 wpkey[7:0] 76543210 ???????wpen
425 32145c?06/2013 at32uc3l0128/256 19.7.15 write protect status register register name: wpsr access type: read-only offset: 0xe8 reset value: see table 19-10 ? wpvsrc: write protect violation source if wpvs=1 this field indicates wh ich write-protected register was unsuccessfully wr itten to, either by ad dress offset or code. ? wpvs: write protect violation status 0= no write protect violation has occurred since the last wpsr read. 1= a write protect violation has occurred since the last wpsr read. note: reading wpsr automatically clears all fields. 31 30 29 28 27 26 25 24 ???????? 23 22 21 20 19 18 17 16 wpvsrc[15:8] 15 14 13 12 11 10 9 8 wpvsrc[7:0] 76543210 ???????wpvs
426 32145c?06/2013 at32uc3l0128/256 19.7.16 version register name: version access type: read-only offset: 0xfc reset value: - ?mfn reserved. no functionality associated. ? version version of the module. no functionality associated. 31 30 29 28 27 26 25 24 ???????? 23 22 21 20 19 18 17 16 ???? mfn 15 14 13 12 11 10 9 8 ???? version[11:8] 76543210 version[7:0]
427 32145c?06/2013 at32uc3l0128/256 19.8 module configuration the specific configuration for each usart instance is listed in the following tables.the module bus clocks listed here are connected to the syst em bus clocks. please refer to the power man- ager chapter for details. table 19-19. usart configuration feature usart0 usart1 usart2 usart3 receiver time-out counter size (size of the rtor.to field) 17 bit 17 bit 17 bit 17 bit div value for divided clk_usart 8 8 8 8 table 19-20. usart clocks module name clock name description usart0 clk_usart0 clock for the usart0 bus interface usart1 clk_usart1 clock for the usart1 bus interface usart2 clk_usart2 clock for the usart2 bus interface usart3 clk_usart3 clock for the usart3 bus interface table 19-21. register reset values register reset value version 0x00000440
428 32145c?06/2013 at32uc3l0128/256 20. serial peripheral interface (spi) rev: 2.1.1.3 20.1 features ? compatible with an embedded 32-bit microcontroller ? supports communication with serial external devices ? four chip selects with extern al decoder support allow co mmunication with up to 15 peripherals ? serial memories, such as da taflash and 3-wire eeproms ? serial peripherals, such as adcs, dacs, l cd controllers, can controllers and sensors ? external co-processors ? master or slave serial peripheral bus interface ? 4 - to 16-bit programmable da ta length per chip select ? programmable phase and polarity per chip select ? programmable transfer delays between consecutive transfers and between clock and data per chip select ? programmable delay between consecutive transfers ? selectable mode fault detection ? connection to peripheral dma controller ch annel capabilities opti mizes data transfers ? one channel for the receiver, one channel for the transmitter ? next buffer support ? four character fifo in reception 20.2 overview the serial peripheral interface (spi) circuit is a synchronous serial data link that provides com- munication with external devices in master or slave mode. it also enables communication between processors if an external processor is connected to the system. the serial peripheral interface is essentially a shift register that serially transmits data bits to other spis. during a data transfer, one spi syste m acts as the ?master?' which controls the data flow, while the other devices act as ?slaves'' whic h have data shifted into and out by the master. different cpus can take turn being masters (multiple master protocol opposite to single master protocol where one cpu is always the master while all of the others are always slaves) and one master may simultaneously shift da ta into multiple slaves. howeve r, only one slave may drive its output to write data back to the master at any given time. a slave device is selected when the master asse rts its nss signal. if multiple slave devices exist, the master generates a separate slav e select signal for each slave (npcs). the spi system consists of two data lines and two control lines: ? master out slave in (mosi): this data line supplies the output data from the master shifted into the input(s) of the slave(s). ? master in slave out (miso): this data line supplies the output data from a slave to the input of the master. there may be no more than one slave transmitting data during any particular transfer. ? serial clock (spck): this contro l line is driven by the master and regulates the flow of the data bits. the master may transmit data at a variety of baud rates; the spck line cycles once for each bit that is transmitted. ? slave select (nss): this control line allows slaves to be turned on and off by hardware.
429 32145c?06/2013 at32uc3l0128/256 20.3 block diagram figure 20-1. spi block diagram 20.4 application block diagram figure 20-2. application block diagram: single master/multiple slave implementation spi interface interrupt control peripheral dma controller i/o controller clk_spi peripheral bus spi interrupt spck npcs3 npcs2 npcs1 npcs0/nss mosi miso slave 0 slave 2 slave 1 spck npcs3 npcs2 npcs1 npcs0 mosi miso spi master spck nss mosi miso spck nss mosi miso spck nss mosi miso nc
430 32145c?06/2013 at32uc3l0128/256 20.5 i/o lines description 20.6 product dependencies in order to use this module, other parts of the system must be configured correctly, as described below. 20.6.1 i/o lines the pins used for interfacing the compliant external devices may be multiplexed with i/o lines. the user must first configure the i/o controll er to assign the spi pins to their peripheral functions. 20.6.2 clocks the clock for the spi bus interface (clk_spi) is generated by the power manager. this clock is enabled at reset, and can be disabled in the power manager. it is recommended to disable the spi before disabling the clock, to avoid freezing the spi in an undefined state. 20.6.3 interrupts the spi interrupt request line is connected to the interrupt controller. using the spi interrupt requires the interrupt controller to be programmed first. 20.7 functional description 20.7.1 modes of operation the spi operates in master mode or in slave mode. operation in master mode is configured by writing a one to the master/slave mode bit in the mode register (mr.mstr). the pins npcs0 to npcs3 are all configured as outputs, the spck pin is driven, the miso line is wired on the rece iver input and the mosi line driven as an output by the transmitter. if the mr.mstr bit is written to zero, the spi ope rates in slave mode. the miso line is driven by the transmitter output, the mosi line is wired on the receiver input, the spck pin is driven by the transmitter to synchronize the receiver. the npcs0 pin becomes an input, and is used as a slave select signal (nss). the pins npcs1 to npcs3 are not driven and can be used for other purposes. the data transfers are identically programmable for both modes of operations. the baud rate generator is activated only in master mode. table 20-1. i/o lines description pin name pin description type master slave miso master in slave out input output mosi master out slave in output input spck serial clock output input npcs1-npcs3 peripheral chip selects output unused npcs0/nss peripheral chip select/slave select output input
431 32145c?06/2013 at32uc3l0128/256 20.7.2 data transfer four combinations of polarity and phase are available for data transfers. the clock polarity is configured with the clock polarity bit in the chip select registers (csrn.cpol). the clock phase is configured with the clock phase bit in the csrn registers (csrn.ncpha). these two bits determine the edges of the clock signal on which data is driven and sampled. each of the two bits has two possible states, resulting in four possible combinations that are incompatible with one another. thus, a master/slave pair must use the same parameter pair values to com- municate. if multiple slaves are used and fix ed in different configurations, the master must reconfigure itself each time it needs to communicate with a different slave. table 20-2 on page 431 shows the four modes and corresponding parameter settings. figure 20-3 on page 431 and figure 20-4 on page 432 show examples of data transfers. figure 20-3. spi transfer format (ncpha = 1, 8 bits per transfer) table 20-2. spi modes spi mode cpol ncpha 001 100 211 310 14 3 2 58 7 6 spck cycle (for reference) spck (cpol = 0) nss (to slave) miso (from slave) mosi (from master) spck (cpol = 1) msb 6 4 5lsb 1 2 3 msb 6 *** lsb 1 2 3 4 5 *** not defined, but normaly msb of previous character received
432 32145c?06/2013 at32uc3l0128/256 figure 20-4. spi transfer format (ncpha = 0, 8 bits per transfer) 20.7.3 master mode operations when configured in master mode, the spi uses the internal programmable baud rate generator as clock source. it fully controls the data transfers to and from the slave(s) connected to the spi bus. the spi drives the chip select line to the slave and the serial clock signal (spck). the spi features two holding registers, the tr ansmit data register (tdr) and the receive data register (rdr), and a single shi ft register. the holding registers maintain the data flow at a constant rate. after enabling the spi, a data transfer begins when the processor writes to the tdr register. the written data is immediately transferred in the shift register and transfer on the spi bus starts. while the data in the shift register is shifted on the mosi line, the miso line is sampled and shifted in the shift register. transmission cannot occur without reception. before writing to the tdr, the peripheral chip select field in tdr (tdr.pcs) must be written in order to select a slave. if new data is written to tdr during the transfer, it stays in it until the current transfer is com- pleted. then, the received data is transferred from the shift register to rdr, the data in tdr is loaded in the shift register and a new transfer starts. the transfer of a data written in tdr in the shift register is indicated by the transmit data reg- ister empty bit in the status register (sr.tdre). when new data is written in tdr, this bit is cleared. the sr.tdre bit is used to trigger the transmit peripheral dma controller channel. the end of transfer is indicated by the transmission registers empty bit in the sr register (sr.txempty). if a transfer delay (csrn.dlybct) is greater than zero for the last transfer, sr.txempty is set after the completion of said delay. the clk_spi can be switched off at this time. during reception, received data are transferred from the shift register to the reception fifo. the fifo can contain up to 4 characters (both receive data and peripheral chip select fields). while a character of the fifo is unread, the receive data register full bit in sr remains high (sr.rdrf). characters are read through the rdr re gister. if the four char acters stored in the fifo are not read and if a new character is stored, this sets the overrun error status bit in the sr register (sr.ovres). the procedure to follow in such a case is described in section 20.7.3.8 . 14 3 2 58 7 6 spck cycle (for reference) spck (cpol = 0) nss (to slave) miso (from slave) mosi (from master) spck (cpol = 1) msb 6 4 5lsb 1 2 3 6 lsb 1 2 3 4 5 *** not defined, but normaly lsb of previous character transmitted msb ***
433 32145c?06/2013 at32uc3l0128/256 figure 20-5 on page 433 shows a block diagram of the spi when operating in master mode. fig- ure 20-6 on page 434 shows a flow chart describing how transfers are handled. 20.7.3.1 master mode block diagram figure 20-5. master mode block diagram baud rate generator rxfifoen 4 ? character fifo shift register tdre rxfifoen 4 ? character fifo ps pcsdec current peripheral modf modfdis mstr scbr csr0..3 csr0..3 cpol ncpha bits rdr rd rdrf ovres td tdr rdr csaat csnaat csr0..3 pcs mr pcs tdr spck clk_spi miso mosi msb lsb npcs1 npcs2 npcs3 npcs0 spi clock 0 1 0 1 0 1 npcs0
434 32145c?06/2013 at32uc3l0128/256 20.7.3.2 master mode flow diagram figure 20-6. master mode flow diagram spi enable csaat ? ps ? 1 0 0 1 1 npcs = tdr(pcs) npcs = mr(pcs) delay dlybs serializer = tdr(td) tdre = 1 data transfer rdr(rd) = serializer rdrf = 1 tdre ? npcs = 0xf delay dlybcs fixed peripheral variable peripheral delay dlybct 0 1 csaat ? 0 tdre ? 1 0 ps ? 0 1 tdr(pcs) = npcs ? no yes mr(pcs) = npcs ? no npcs = 0xf delay dlybcs npcs = tdr(pcs) npcs = 0xf delay dlybcs npcs = mr(pcs), tdr(pcs) fixed peripheral variable peripheral - npcs defines the current chip select - csaat, dlybs, dlybct refer to the fields of the chip select register corresponding to the current chip select - when npcs is 0xf, csaat is 0.
435 32145c?06/2013 at32uc3l0128/256 20.7.3.3 clock generation the spi baud rate clock is generated by dividing the clk_spi , by a value between 1 and 255. this allows a maximum operating baud rate at up to clk_spi and a minimum operating baud rate of clk_spi divided by 255. writing the serial clock baud rate field in the csrn registers (csrn.scbr) to zero is forbid- den. triggering a transfer while csrn.scbr is zero can lead to unpredictable results. at reset, csrn.scbr is zero and the user has to configure it at a valid value before performing the first transfer. the divisor can be defined independently for each chip select, as it has to be configured in the csrn.scbr field. this allows the spi to au tomatically adapt the baud rate for each interfaced peripheral without reprogramming. 20.7.3.4 transfer delays figure 20-7 on page 435 shows a chip select transfer change and consecutive transfers on the same chip select. three delays can be configured to modify the transfer waveforms: ? the delay between chip selects, programmable only once for all the chip selects by writing to the delay between chip selects field in the mr register (mr. dlybcs). allows insertion of a delay between release of one chip select and before assertion of a new one. ? the delay before spck, independently programmable for each chip select by writing the delay before spck field in the csrn register s (csrn.dlybs). allows the start of spck to be delayed after the chip select has been asserted. ? the delay between consecutive transfers, independently programmable for each chip select by writing the delay between consecutive transfers field in the csrn registers (csrn.dlybct). allows insertion of a delay be tween two transfers occurring on the same chip select these delays allow the spi to be adapted to the interfaced peripherals and their speed and bus release time. figure 20-7. programmable delays dlybcs dlybs dlybct dlybct chip select 1 chip select 2 spck
436 32145c?06/2013 at32uc3l0128/256 20.7.3.5 peripheral selection the serial peripherals are selected through the assertion of the npcs0 to npcs3 signals. by default, all the npcs signals are high before and after each transfer. the peripheral selection can be performed in two different ways: ? fixed peripheral select: spi exchanges data with only one peripheral ? variable peripheral select: data can be exchanged with more than one peripheral fixed peripheral select is activated by writing a ze ro to the peripheral select bit in mr (mr.ps). in this case, the current peripheral is defined by the mr.pcs field and the tdr.pcs field has no effect. variable peripheral select is activated by writing a one to the mr.ps bit . the tdr.pcs field is used to select the current peripheral. this means that the peripheral selection can be defined for each new data. the fixed peripheral selection allows buffer transfers with a single peripheral. using the periph- eral dma controller is an optimal means, as the size of the data transfer between the memory and the spi is either 4 bits or 16 bits. however, changing the peripheral selection requires the mode register to be reprogrammed. the variable peripheral selection allows buffer transfers with multiple peripherals without repro- gramming the mr register. data written to tdr is 32-bits wide and defines the real data to be transmitted and the peripheral it is destined to. using the peripheral dma controller in this mode requires 32-bit wide buffers, with the data in the lsbs and the pcs and lastxfer fields in the msbs, however the spi still controls the number of bits (8 to16) to be transferr ed through miso and mosi lines with the csrn registers. this is not the optimal means in term of memory size for the buffers, but it provides a very effective means to exchange data with several peripherals without any intervention of the processor. 20.7.3.6 peripheral chip select decoding the user can configure the spi to operate with up to 15 peripherals by decoding the four chip select lines, npcs0 to npcs3 with an external logic. this can be enabled by writing a one to the chip select decode bit in the mr register (mr.pcsdec). when operating without decoding, the spi makes sure that in any case only one chip select line is activated, i.e. driven low at a time. if two bits are defined low in a pcs field, only the lowest numbered chip select is driven low. when operating with decoding, the spi directly outputs the value defined by the pcs field of either the mr register or the tdr register (depending on ps). as the spi sets a default value of 0xf on the chip select lines (i.e. all chip select lines at one) when not processing any transfer, only 15 peripherals can be decoded. the spi has only four chip select registers, not 15. as a result, when decoding is activated, each chip select defines the characteristics of up to four peripherals. as an example, the crs0 register defines the characteristics of the exter nally decoded peripherals 0 to 3, corresponding to the pcs values 0x0 to 0x3. thus, the user has to make sure to connect compatible peripherals on the decoded chip select lines 0 to 3, 4 to 7, 8 to 11 and 12 to 14. 20.7.3.7 peripheral deselection when operating normally, as soon as the transfer of the last data written in tdr is completed, the npcs lines all rise. this might lead to runtime error if the processor is too long in responding
437 32145c?06/2013 at32uc3l0128/256 to an interrupt, and thus might lead to difficulties for interfacing with some serial peripherals requiring the chip select li ne to remain active during a full set of transfers. to facilitate interfacing with such devices, the csrn registers can be configur ed with the chip select active after transfer bit written to one (csrn.csaat) . this allows the chip select lines to remain in their current state (low = active) until transfer to another peripheral is required. when the csrn.csaat bit is written to qero, the npcs does not rise in all cases between two transfers on the same peripheral. during a transfe r on a chip select, the sr.tdre bit rises as soon as the content of the tdr is transferred into the internal shifter. when this bit is detected the tdr can be reloaded. if this reload occurs before the end of the current transfer and if the next transfer is performed on the same chip select as the current transfer, the chip select is not de-asserted between the two transfers. this migh t lead to difficulties fo r interfacing with some serial peripherals requiring the chip select to be de-asserted after each transfer. to facilitate interfacing with such devices, the csrn register s can be configured with the chip select not active after transfer bit (csrn.csnaat) written to one. this allows to de-assert systematically the chip select lines during a time dlybcs. (the value of th e csrn.csnaat bit is taken into account only if the csrn.csaat bit is written to zero for the same chip select). figure 20-8 on page 438 shows different peripheral deselection cases and the effect of the csrn.csaat and csrn.csnaat bits. 20.7.3.8 fifo management a fifo has been implemented in reception fifo (both in master and in slave mode), in order to be able to store up to 4 characters without causing an overrun error. if an attempt is made to store a fifth character, an overrun error rises. if such an event occurs, the fifo must be flushed. there are two ways to flush the fifo: ? by performing four read accesses of the rdr (the data read must be ignored) ? by writing a one to the flush fifo command bit in the cr register (cr.flushfifo). after that, the spi is able to receive new data.
438 32145c?06/2013 at32uc3l0128/256 figure 20-8. peripheral deselection figure 20-8 on page 438 shows different peripheral deselection cases and the effect of the csrn.csaat and csrn.csnaat bits. 20.7.3.9 mode fault detection the spi is capable of detecting a mode fault when it is configured in master mode and npcs0, mosi, miso, and spck are configured as open dr ain through the i/o controller with either internal or external pullup resistors. if t he i/o controller does not have open-drain capability, mode fault detection must be disabled by writing a one to the mode fault detection bit in the mr a npcs[0..3] write tdr tdre npcs[0..3] write tdr tdre npcs[0..3] write tdr tdre dlybcs pcs = a dlybcs dlybct a pcs = b b dlybcs pcs = a dlybcs dlybct a pcs = b b dlybcs dlybct pcs=a a dlybcs dlybct a pcs = a a a dlybct aa csaat = 0 and csnaat = 0 dlybct aa csaat = 1 and csnaat= 0 / 1 a dlybcs pcs = a dlybct aa csaat = 0 and csnaat = 1 npcs[0..3] write tdr tdre pcs = a dlybct aa csaat = 0 and csnaat = 0
439 32145c?06/2013 at32uc3l0128/256 register (mr.modfdis). in systems with open-drain i/o lines, a mode fault is detected when a low level is driven by an external master on the npcs0/nss signal. when a mode fault is detected, the mode fault error bit in the sr (sr.modf) is set until the sr is read and the spi is automatically disabled until re-enabled by writing a one to the spi enable bit in the cr register (cr.spien). by default, the mode fault detection circuitry is enabled. the user can disable mode fault detec- tion by writing a one to the mode fault dete ction bit in the mr re gister (mr.modfdis). 20.7.4 spi slave mode when operating in slave mode, the spi processes data bits on the clock provided on the spi clock pin (spck). the spi waits for nss to go active before receiving the serial clock from an external master. when nss falls, the clock is validated on the serializer, which processes the number of bits defined by the bits per transfer field of the chip select register 0 (csr0.bits). these bits are processed following a phase and a polarity defined respectively by the csr0.ncpha and csr0.cpol bits. note that the bits, cpol, and ncp ha bits of the other chip select registers have no effect when the spi is configured in slave mode. the bits are shifted out on the miso line and sampled on the mosi line. when all the bits are processed, the received data is transferred in the receive data register and the sr.rdrf bit rises. if the rdr register has not been read before new data is received, the sr.ovres bit is set. data is loaded in rdr even if this flag is set. th e user has to read the sr register to clear the sr.ovres bit. when a transfer starts, the data shifted out is the data present in the shift register. if no data has been written in the tdr register, the last data received is transferred. if no data has been received since the last reset, all bits are transmi tted low, as the shift register resets to zero. when a first data is written in tdr, it is trans ferred immediately in the shift register and the sr.tdre bit rises. if new data is written, it rema ins in tdr until a transfer occurs, i.e. nss falls and there is a valid clock on the spck pin. w hen the transfer occurs, the last data written in tdr is transferred in the shift register and the sr.tdre bit rises. this enables frequent updates of critical variables with single transfers. then, a new data is loaded in the shift register from the tdr. in case no character is ready to be transmitted, i.e. no character has been written in tdr since the last load from tdr to the shift register, the shift register is not modified and the last received character is retransmitted. in this case the underrun error status bit is set in sr (sr.undes). figure 20-9 on page 440 shows a block diagram of the spi when operating in slave mode.
440 32145c?06/2013 at32uc3l0128/256 figure 20-9. slave mode functional block diagram shift register spck spiens lsb msb nss mosi spi clock tdre tdr td rdrf ovres csr0 cpol ncpha bits spien spidis miso undes rdr rd 4 - character fifo 0 1 rxfifoen
441 32145c?06/2013 at32uc3l0128/256 20.8 user interface note: 1. the reset values are device specific. please refer to the module configuration section at the end of this chapter. table 20-3. spi register memory map offset register register name access reset 0x00 control register cr write-only 0x00000000 0x04 mode register mr read/write 0x00000000 0x08 receive data register rdr read-only 0x00000000 0x0c transmit data register tdr write-only 0x00000000 0x10 status register sr read-only 0x00000000 0x14 interrupt enable register ier write-only 0x00000000 0x18 interrupt disable register idr write-only 0x00000000 0x1c interrupt mask register imr read-only 0x00000000 0x30 chip select register 0 csr0 read/write 0x00000000 0x34 chip select register 1 csr1 read/write 0x00000000 0x38 chip select register 2 csr2 read/write 0x00000000 0x3c chip select register 3 csr3 read/write 0x00000000 0x e4 write protection control register wpcr read/write 0x00000000 0xe8 write protection status register wpsr read-only 0x00000000 0xf8 features register features read-only - (1) 0xfc version register version read-only - (1)
442 32145c?06/2013 at32uc3l0128/256 20.8.1 control register name: cr access type: write-only offset: 0x00 reset value: 0x00000000 ? lastxfer: last transfer 1: the current npcs will be deasserted after the character wr itten in td has been transferred. when csrn.csaat is one, this allows to close the communication with the current serial peri pheral by raising the corresponding npcs line as soon as td transfer has completed. 0: writing a zero to this bit has no effect. ? flushfifo: flush fifo command 1: if the fifo mode is enabled (mr.fifoen written to one) and if an overrun error has been detected, this command allows to empty the fifo. 0: writing a zero to this bit has no effect. ? swrst: spi software reset 1: writing a one to this bit will reset the spi. a software-trigger ed hardware reset of the spi interface is performed. the spi is in slave mode after software reset. peripheral dma controller channels are not affected by software reset. 0: writing a zero to this bit has no effect. ? spidis: spi disable 1: writing a one to this bit will disable the spi. as soon as spi dis is written to one, the spi finishes its transfer, all pins are set in input mode and no data is received or transmitted. if a transfe r is in progress, the transfer is finished before the spi is disabled. if both spien and spidis are equal to one when the cr register is written, the spi is disabled. 0: writing a zero to this bit has no effect. ? spien: spi enable 1: writing a one to this bit will enable the spi to transfer and receive data. 0: writing a zero to this bit has no effect. 31 30 29 28 27 26 25 24 -------lastxfer 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------flushfifo 76543210 swrst - - - - - spidis spien
443 32145c?06/2013 at32uc3l0128/256 20.8.2 mode register name: mr access type: read/write offset: 0x04 reset value: 0x00000000 ? dlybcs: delay between chip selects this field defines the delay from npcs inactive to the activation of another npcs. the dlybcs time guarantees non- overlapping chip selects and solves bus contentions in case of peripherals having long data float times. if dlybcs is less than or equal to six, six clk_spi periods will be inserted by default. otherwise, the following equation determines the delay: ? pcs: peripheral chip select this field is only used if fixed peripheral select is active (ps = 0). if pcsdec = 0: pcs = xxx0npcs[3:0] = 1110 pcs = xx01npcs[3:0] = 1101 pcs = x011npcs[3:0] = 1011 pcs = 0111npcs[3:0] = 0111 pcs = 1111forbidden (no peripheral is selected) (x = don?t care) if pcsdec = 1: npcs[3:0] output signals = pcs. ? llb: local loopback enable 1: local loopback path enabled. llb controls the local loopback on the data serializer for testing in master mode only (miso is internally connected on mosi). 0: local loopback path disabled. ? rxfifoen: fifo in reception enable 1: the fifo is used in reception (f our characters can be stored in the spi). 31 30 29 28 27 26 25 24 dlybcs 23 22 21 20 19 18 17 16 ---- pcs 15 14 13 12 11 10 9 8 -------- 76543210 llb rxfifoen - modf dis - pcsdec ps mstr delay between chip selects dlybcs clkspi ---------------------- - =
444 32145c?06/2013 at32uc3l0128/256 0: the fifo is not used in reception (onl y one character can be stored in the spi). ? modfdis: mode fault detection 1: mode fault detection is disabled. if the i/o controller does not have o pen-drain capability, mode fault detection must be disabled for proper operation of the spi. 0: mode fault detection is enabled. ? pcsdec: chip select decode 0: the chip selects are directly connected to a peripheral device. 1: the four chip select lines are co nnected to a 4- to 16-bit decoder. when pcsdec equals one, up to 15 chip select signals can be generated with the four lines using an external 4- to 16-bit decoder. the csrn registers define the characteristics of the 15 chip selects according to the following rules: csr0 defines peripheral chip select signals 0 to 3. csr1 defines peripheral chip select signals 4 to 7. csr2 defines peripheral chip select signals 8 to 11. csr3 defines peripheral chip select signals 12 to 14. ? ps: peripheral select 1: variable peripheral select. 0: fixed peripheral select. ? mstr: master/slave mode 1: spi is in master mode. 0: spi is in slave mode.
445 32145c?06/2013 at32uc3l0128/256 20.8.3 receive data register name: rdr access type: read-only offset: 0x08 reset value: 0x00000000 ? rd: receive data data received by the spi interface is stored in this register right-justified. unused bits read zero. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 rd[15:8] 76543210 rd[7:0]
446 32145c?06/2013 at32uc3l0128/256 20.8.4 transmit data register name: tdr access type: write-only offset: 0x0c reset value: 0x00000000 ? lastxfer: last transfer 1: the current npcs will be deasserted after the character wr itten in td has been transferred. when csrn.csaat is one, this allows to close the communication with the current serial peri pheral by raising the corresponding npcs line as soon as td transfer has completed. 0: writing a zero to this bit has no effect. this field is only used if variable peripheral select is active (mr.ps = 1). ? pcs: peripheral chip select if pcsdec = 0: pcs = xxx0npcs[3:0] = 1110 pcs = xx01npcs[3:0] = 1101 pcs = x011npcs[3:0] = 1011 pcs = 0111npcs[3:0] = 0111 pcs = 1111forbidden (no peripheral is selected) (x = don?t care) if pcsdec = 1: npcs[3:0] output signals = pcs this field is only used if variable peripheral select is active (mr.ps = 1). ? td: transmit data data to be transmitted by the spi interface is stored in this re gister. information to be transmitted must be written to the td r register in a right-justified format. 31 30 29 28 27 26 25 24 -------lastxfer 23 22 21 20 19 18 17 16 ---- pcs 15 14 13 12 11 10 9 8 td[15:8] 76543210 td[7:0]
447 32145c?06/2013 at32uc3l0128/256 20.8.5 status register name: sr access type: read-only offset: 0x10 reset value: 0x00000000 ? spiens: spi enable status 1: this bit is set when the spi is enabled. 0: this bit is cleared when the spi is disabled. ? undes: underrun error st atus (slave mode only) 1: this bit is set when a transfer begins whereas no data has been loaded in the tdr register. 0: this bit is cleared when the sr register is read. ? txempty: transmission registers empty 1: this bit is set when tdr and internal shifter are empty. if a transfer delay has been defined, txempty is set after the completion of such delay. 0: this bit is cleared as soon as data is written in tdr. ? nssr: nss rising 1: a rising edge occurred on nss pin since last read. 0: this bit is cleared when the sr register is read. ? ovres: overrun error status 1: this bit is set when an overrun has occurred. an overrun occu rs when rdr is loaded at least twice from the serializer since the last read of the rdr. 0: this bit is cleared when the sr register is read. ? modf: mode fault error 1: this bit is set when a mode fault occurred. 0: this bit is cleared when the sr register is read. ? tdre: transmit data register empty 1: this bit is set when the last data written in the tdr register has been transferred to the serializer. 0: this bit is cleared when data has been written to tdr and not yet transferred to the serializer. tdre equals zero when the spi is disabled or at re set. the spi enable command sets this bit to one. ? rdrf: receive data register full 1: data has been received and the received data has been transferr ed from the serializer to rdr since the last read of rdr. 0: no data has been received since the last read of rdr 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------spiens 15 14 13 12 11 10 9 8 - - - - - undes txempty nssr 76543210 - - - - ovres modf tdre rdrf
448 32145c?06/2013 at32uc3l0128/256 20.8.6 interrupt enable register name: ier access type: write-only offset: 0x14 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will set the corresponding bit in imr. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 - - - - - undes txempty nssr 76543210 - - - - ovres modf tdre rdrf
449 32145c?06/2013 at32uc3l0128/256 20.8.7 interrupt disable register name: idr access type: write-only offset: 0x18 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in imr. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 - - - - - undes txempty nssr 76543210 - - - - ovres modf tdre rdrf
450 32145c?06/2013 at32uc3l0128/256 20.8.8 interrupt mask register name: imr access type: read-only offset: 0x1c reset value: 0x00000000 0: the corresponding interrupt is disabled. 1: the corresponding interrupt is enabled. a bit in this register is cleared when the corresponding bit in idr is written to one. a bit in this register is set when the corresponding bit in ier is written to one. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 - - - - - undes txempty nssr 76543210 - - - - ovres modf tdre rdrf
451 32145c?06/2013 at32uc3l0128/256 20.8.9 chip select register 0 name: csr0 access type: read/write offset: 0x30 reset value: 0x00000000 ? dlybct: delay between consecutive transfers this field defines the delay between two consecutive transfers wi th the same peripheral without removing the chip select. the delay is always inserted after each transfer and before removing the chip select if needed. when dlybct equals zero, no delay between consecutive transfers is inserted and the clock keeps its duty cycle over the character transfers. otherwise, the following equation determines the delay: ? dlybs: delay before spck this field defines the delay from npcs valid to the first valid spck transition. when dlybs equals zero, the npcs valid to spck transition is 1/2 the spck clock period. otherwise, the following equat ions determine the delay: ? scbr: serial clock baud rate in master mode, the spi interface uses a modulus counter to de rive the spck baud rate from the clk_spi. the baud rate is selected by writing a value from 1 to 255 in the scbr field. the following equations determine the spck baud rate: writing the scbr field to zero is forbidden. triggering a transfer while scbr is zero can lead to unpredictable results. at reset, scbr is zero and the user has to write it to a valid value before performing the first transfer. if a clock divider (scbrn) field is set to one and the other scbr fields differ from one, access on csn is correct but no corre ct access will be possible on other cs. 31 30 29 28 27 26 25 24 dlybct 23 22 21 20 19 18 17 16 dlybs 15 14 13 12 11 10 9 8 scbr 76543210 bits csaat csnaat ncpha cpol delay between consecutive transfers 32 dlybct ? clkspi ------------------------------------ = delay before spck dlybs clkspi --------------------- = spck baudrate clkspi scbr --------------------- =
452 32145c?06/2013 at32uc3l0128/256 ? bits: bits per transfer the bits field determines the number of data bits transferred. reserved values should not be used. ? csaat: chip select active after transfer 1: the peripheral chip select does not ri se after the last transfer is achieved. it remains active until a new transfer is requ ested on a different chip select. 0: the peripheral chip select line rises as soon as the last transfer is achieved. ? csnaat: chip select not active afte r transfer (ignored if csaat = 1) 0: the peripheral chip select does not rise between two transfers if the tdr is reloaded before the end of the first transfer a nd if the two transfers occur on the same chip select. 1: the peripheral chip select rises system atically between each transfe r performed on the same slave for a minimal duration of: (if dlybct field is different from 0) (if dlybct field equals 0) ? ncpha: clock phase 1: data is captured after the leading (inactive-to-active) e dge of spck and changed on the trailing (active-to-inactive) edge o f spck. 0: data is changed on the leading (inactive-to-active) edge of spck and captured after the trailing (active-to-inactive) edge o f spck. ncpha determines which edge of spck causes data to change and which edge causes data to be captured. ncpha is used with cpol to produce the requir ed clock/data relationship between master and slave devices. ?cpol: clock polarity 1: the inactive state value of spck is logic level one. 0: the inactive state value of spck is logic level zero. bits bits per transfer 0000 8 0001 9 0010 10 0011 11 0100 12 0101 13 0110 14 0111 15 1000 16 1001 4 1010 5 1011 6 1100 7 1101 reserved 1110 reserved 1111 reserved dlybcs clkspi ---------------------- - dlybcs 1 + clkspi -------------------------------- -
453 32145c?06/2013 at32uc3l0128/256 cpol is used to determine the inactive state value of the seri al clock (spck). it is used wit h ncpha to produce the required clock/data relationship between master and slave devices.
454 32145c?06/2013 at32uc3l0128/256 20.8.10 chip select register 1 name: csr1 access type: read/write offset: 0x34 reset value: 0x00000000 ? dlybct: delay between consecutive transfers this field defines the delay between two consecutive transfers wi th the same peripheral without removing the chip select. the delay is always inserted after each transfer and before removing the chip select if needed. when dlybct equals zero, no delay between consecutive transfers is inserted and the clock keeps its duty cycle over the character transfers. otherwise, the following equation determines the delay: ? dlybs: delay before spck this field defines the delay from npcs valid to the first valid spck transition. when dlybs equals zero, the npcs valid to spck transition is 1/2 the spck clock period. otherwise, the following equat ions determine the delay: ? scbr: serial clock baud rate in master mode, the spi interface uses a modulus counter to de rive the spck baud rate from the clk_spi. the baud rate is selected by writing a value from 1 to 255 in the scbr field. the following equations determine the spck baud rate: writing the scbr field to zero is forbidden. triggering a transfer while scbr is zero can lead to unpredictable results. at reset, scbr is zero and the user has to write it to a valid value before performing the first transfer. if a clock divider (scbrn) field is set to one and the other scbr fields differ from one, access on csn is correct but no corre ct access will be possible on other cs. 31 30 29 28 27 26 25 24 dlybct 23 22 21 20 19 18 17 16 dlybs 15 14 13 12 11 10 9 8 scbr 76543210 bits csaat csnaat ncpha cpol delay between consecutive transfers 32 dlybct ? clkspi ------------------------------------ = delay before spck dlybs clkspi --------------------- = spck baudrate clkspi scbr --------------------- =
455 32145c?06/2013 at32uc3l0128/256 ? bits: bits per transfer the bits field determines the number of data bits transferred. reserved values should not be used. ? csaat: chip select active after transfer 1: the peripheral chip select does not ri se after the last transfer is achieved. it remains active until a new transfer is requ ested on a different chip select. 0: the peripheral chip select line rises as soon as the last transfer is achieved. ? csnaat: chip select not active afte r transfer (ignored if csaat = 1) 0: the peripheral chip select does not rise between two transfers if the tdr is reloaded before the end of the first transfer a nd if the two transfers occur on the same chip select. 1: the peripheral chip select rises system atically between each transfe r performed on the same slave for a minimal duration of: (if dlybct field is different from 0) (if dlybct field equals 0) ? ncpha: clock phase 1: data is captured after the leading (inactive-to-active) e dge of spck and changed on the trailing (active-to-inactive) edge o f spck. 0: data is changed on the leading (inactive-to-active) edge of spck and captured after the trailing (active-to-inactive) edge o f spck. ncpha determines which edge of spck causes data to change and which edge causes data to be captured. ncpha is used with cpol to produce the requir ed clock/data relationship between master and slave devices. ?cpol: clock polarity 1: the inactive state value of spck is logic level one. 0: the inactive state value of spck is logic level zero. bits bits per transfer 0000 8 0001 9 0010 10 0011 11 0100 12 0101 13 0110 14 0111 15 1000 16 1001 4 1010 5 1011 6 1100 7 1101 reserved 1110 reserved 1111 reserved dlybcs clkspi ---------------------- - dlybcs 1 + clkspi -------------------------------- -
456 32145c?06/2013 at32uc3l0128/256 cpol is used to determine the inactive state value of the seri al clock (spck). it is used wit h ncpha to produce the required clock/data relationship between master and slave devices.
457 32145c?06/2013 at32uc3l0128/256 20.8.11 chip select register 2 name: csr2 access type: read/write offset: 0x38 reset value: 0x00000000 ? dlybct: delay between consecutive transfers this field defines the delay between two consecutive transfers wi th the same peripheral without removing the chip select. the delay is always inserted after each transfer and before removing the chip select if needed. when dlybct equals zero, no delay between consecutive transfers is inserted and the clock keeps its duty cycle over the character transfers. otherwise, the following equation determines the delay: ? dlybs: delay before spck this field defines the delay from npcs valid to the first valid spck transition. when dlybs equals zero, the npcs valid to spck transition is 1/2 the spck clock period. otherwise, the following equat ions determine the delay: ? scbr: serial clock baud rate in master mode, the spi interface uses a modulus counter to de rive the spck baud rate from the clk_spi. the baud rate is selected by writing a value from 1 to 255 in the scbr field. the following equations determine the spck baud rate: writing the scbr field to zero is forbidden. triggering a transfer while scbr is zero can lead to unpredictable results. at reset, scbr is zero and the user has to write it to a valid value before performing the first transfer. if a clock divider (scbrn) field is set to one and the other scbr fields differ from one, access on csn is correct but no corre ct access will be possible on other cs. 31 30 29 28 27 26 25 24 dlybct 23 22 21 20 19 18 17 16 dlybs 15 14 13 12 11 10 9 8 scbr 76543210 bits csaat csnaat ncpha cpol delay between consecutive transfers 32 dlybct ? clkspi ------------------------------------ = delay before spck dlybs clkspi --------------------- = spck baudrate clkspi scbr --------------------- =
458 32145c?06/2013 at32uc3l0128/256 ? bits: bits per transfer the bits field determines the number of data bits transferred. reserved values should not be used. ? csaat: chip select active after transfer 1: the peripheral chip select does not ri se after the last transfer is achieved. it remains active until a new transfer is requ ested on a different chip select. 0: the peripheral chip select line rises as soon as the last transfer is achieved. ? csnaat: chip select not active afte r transfer (ignored if csaat = 1) 0: the peripheral chip select does not rise between two transfers if the tdr is reloaded before the end of the first transfer a nd if the two transfers occur on the same chip select. 1: the peripheral chip select rises system atically between each transfe r performed on the same slave for a minimal duration of: (if dlybct field is different from 0) (if dlybct field equals 0) ? ncpha: clock phase 1: data is captured after the leading (inactive-to-active) e dge of spck and changed on the trailing (active-to-inactive) edge o f spck. 0: data is changed on the leading (inactive-to-active) edge of spck and captured after the trailing (active-to-inactive) edge o f spck. ncpha determines which edge of spck causes data to change and which edge causes data to be captured. ncpha is used with cpol to produce the requir ed clock/data relationship between master and slave devices. ?cpol: clock polarity 1: the inactive state value of spck is logic level one. 0: the inactive state value of spck is logic level zero. bits bits per transfer 0000 8 0001 9 0010 10 0011 11 0100 12 0101 13 0110 14 0111 15 1000 16 1001 4 1010 5 1011 6 1100 7 1101 reserved 1110 reserved 1111 reserved dlybcs clkspi ---------------------- - dlybcs 1 + clkspi -------------------------------- -
459 32145c?06/2013 at32uc3l0128/256 cpol is used to determine the inactive state value of the seri al clock (spck). it is used wit h ncpha to produce the required clock/data relationship between master and slave devices.
460 32145c?06/2013 at32uc3l0128/256 20.8.12 chip select register 3 name: csr3 access type: read/write offset: 0x3c reset value: 0x00000000 ? dlybct: delay between consecutive transfers this field defines the delay between two consecutive transfers wi th the same peripheral without removing the chip select. the delay is always inserted after each transfer and before removing the chip select if needed. when dlybct equals zero, no delay between consecutive transfers is inserted and the clock keeps its duty cycle over the character transfers. otherwise, the following equation determines the delay: ? dlybs: delay before spck this field defines the delay from npcs valid to the first valid spck transition. when dlybs equals zero, the npcs valid to spck transition is 1/2 the spck clock period. otherwise, the following equat ions determine the delay: ? scbr: serial clock baud rate in master mode, the spi interface uses a modulus counter to de rive the spck baud rate from the clk_spi. the baud rate is selected by writing a value from 1 to 255 in the scbr field. the following equations determine the spck baud rate: writing the scbr field to zero is forbidden. triggering a transfer while scbr is zero can lead to unpredictable results. at reset, scbr is zero and the user has to write it to a valid value before performing the first transfer. if a clock divider (scbrn) field is set to one and the other scbr fields differ from one, access on csn is correct but no corre ct access will be possible on other cs. 31 30 29 28 27 26 25 24 dlybct 23 22 21 20 19 18 17 16 dlybs 15 14 13 12 11 10 9 8 scbr 76543210 bits csaat csnaat ncpha cpol delay between consecutive transfers 32 dlybct ? clkspi ------------------------------------ = delay before spck dlybs clkspi --------------------- = spck baudrate clkspi scbr --------------------- =
461 32145c?06/2013 at32uc3l0128/256 ? bits: bits per transfer the bits field determines the number of data bits transferred. reserved values should not be used. ? csaat: chip select active after transfer 1: the peripheral chip select does not ri se after the last transfer is achieved. it remains active until a new transfer is requ ested on a different chip select. 0: the peripheral chip select line rises as soon as the last transfer is achieved. ? csnaat: chip select not active afte r transfer (ignored if csaat = 1) 0: the peripheral chip select does not rise between two transfers if the tdr is reloaded before the end of the first transfer a nd if the two transfers occur on the same chip select. 1: the peripheral chip select rises system atically between each transfe r performed on the same slave for a minimal duration of: (if dlybct field is different from 0) (if dlybct field equals 0) ? ncpha: clock phase 1: data is captured after the leading (inactive-to-active) e dge of spck and changed on the trailing (active-to-inactive) edge o f spck. 0: data is changed on the leading (inactive-to-active) edge of spck and captured after the trailing (active-to-inactive) edge o f spck. ncpha determines which edge of spck causes data to change and which edge causes data to be captured. ncpha is used with cpol to produce the requir ed clock/data relationship between master and slave devices. ?cpol: clock polarity 1: the inactive state value of spck is logic level one. 0: the inactive state value of spck is logic level zero. bits bits per transfer 0000 8 0001 9 0010 10 0011 11 0100 12 0101 13 0110 14 0111 15 1000 16 1001 4 1010 5 1011 6 1100 7 1101 reserved 1110 reserved 1111 reserved dlybcs clkspi ---------------------- - dlybcs 1 + clkspi -------------------------------- -
462 32145c?06/2013 at32uc3l0128/256 cpol is used to determine the inactive state value of the seri al clock (spck). it is used wit h ncpha to produce the required clock/data relationship between master and slave devices.
463 32145c?06/2013 at32uc3l0128/256 20.8.13 write protection control register register name: wpcr access type: read-write offset: 0xe4 reset value: 0x00000000 ? spiwpkey: spi write protection key password if a value is written in spiwpen, the value is taken into acc ount only if spiwpkey is written wit h ?spi? (spi written in ascii code, i.e. 0x535049 in hexadecimal). ? spiwpen: spi write protection enable 1: the write protection is enabled 0: the write protection is disabled 31 30 29 28 27 26 25 24 spiwpkey[23:16] 23 22 21 20 19 18 17 16 spiwpkey[15:8] 15 14 13 12 11 10 9 8 spiwpkey[7:0] 76543210 -------spiwpen
464 32145c?06/2013 at32uc3l0128/256 20.8.14 write protection status register register name: wpsr access type: read-only offset: 0xe8 reset value: 0x00000000 ? spiwpvsrc: spi write protection violation source this field indicates the peripheral bus offset of th e register concerned by the violation (mr or csrx) ? spiwpvs: spi write prot ection violation status 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 spiwpvsrc 76543210 - - - - - spiwpvs spiwpvs value violation type 1 the write protection has blocked a write access to a protected register (since the last read). 2 software reset has been performed while write protection was enabled (since the last read or since the last write acce ss on mr, ier, idr or csrx). 3 both write protection violation and software reset with write protection enabled have occurred since the last read. 4 write accesses have been detected on mr (while a chip select was active) or on csri (while the chip select ?i? was active) since the last read. 5 the write protection has blocked a write acce ss to a protected register and write accesses have been detected on mr (while a chip select was active) or on csri (while the chip select ?i? was active) since the last read. 6 software reset has been performed while write protection was enabled (since the last read or since the last write access on mr, ier, idr or csrx) and some write accesses have been detected on mr (while a chip select was active) or on csri (while the chip select ?i? was active) since the last read. 7 - the write protection has blocked a write access to a protected register. and - software reset has been performed while write protection was enabled. and - write accesses have been detected on mr (while a chip select was active) or on csri (while the chip select ?i? was active) since the last read.
465 32145c?06/2013 at32uc3l0128/256 20.8.15 features register register name: features access type: read-only offset: 0xf8 reset value: ? ? swimpl: spurious write protection implemented 0: spurious write protec tion is not implemented. 1: spurious write prot ection is implemented. ? fiforimpl: fifo in reception implemented 0: fifo in receptio n is not implemented. 1: fifo in recept ion is implemented. ? brpbhsb: bridge type is pb to hsb 0: bridge type is not pb to hsb. 1: bridge type is pb to hsb. ? csnaatimpl: csnaat features implemented 0: csnaat (chip select not active after transfer) features are not implemented. 1: csnaat features are implemented. ? extdec: external decoder true 0: external decoder capability is not implemented. 1: external decoder capability is implemented. ? lennconf: character length if not configurable if the character length is not configurable, th is field specifies the fixed character length. ? lenconf: character length configurable 0: the character length is not configurable. 1: the character length is configurable. ? phznconf: phase is zero if phase not configurable 0: if phase is not config urable, phase is non-zero. 1: if phase is not configurable, phase is zero. ? phconf: phase configurable 0: phase is not configurable. 1: phase is configurable. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 - - - swimpl fiforimpl brpbhsb csnaatimpl extdec 15 14 13 12 11 10 9 8 lennconf lenconf 76543210 phznconf phconf ppnconf pconf ncs
466 32145c?06/2013 at32uc3l0128/256 ? ppnconf: polarity positive if polarity not configurable 0: if polarity is not configurable, polarity is negative. 1: if polarity is not configurable, polarity is positive. ? pconf: polarity configurable 0: polarity is not configurable. 1: polarity is configurable. ? ncs: number of chip selects this field indicates the number of chip selects implemented.
467 32145c?06/2013 at32uc3l0128/256 20.8.16 version register register name: version access type: read-only offset: 0xfc reset value: ? ? mfn reserved. no functionality associated. ? version version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- mfn 15 14 13 12 11 10 9 8 version[11:8] 76543210 version[7:0]
468 32145c?06/2013 at32uc3l0128/256 20.9 module configuration the specific configuration for each spi instance is listed in the following tables.the module bus clocks listed here are connecte d to the system bus clocks. pleas e refer to the power manager chapter for details. table 20-4. spi clock name module name clock name description spi clk_spi clock for the spi bus interface table 20-5. register reset value features 0x001f0154 version 0x00000211
469 32145c?06/2013 at32uc3l0128/256 21. two-wire master interface (twim) rev.: 1.1.0.1 21.1 features ? compatible with i 2 c standard ? multi-master support ? transfer speeds of 100 and 400 kbit/s ? 7- and 10-bit and general call addressing ? compatible with smbus standard ? hardware packet error checking (crc) ge neration and verification with ack control ? smbus alert interface ? 25 ms clock low timeout delay ? 10 ms master cumulative clock low extend time ? 25 ms slave cumulative clock low extend time ? compatible with pmbus ? compatible with atmel two-wi re interface serial memories ? dma interface for reducing cpu load ? arbitrary transfer lengths, including 0 data bytes ? optional clock stretching if transmit or receive buffers not ready for data transfer 21.2 overview the atmel two-wire master interface (twim) in terconnects components on a unique two-wire bus, made up of one clock line and one data line with speeds of up to 400 kbit/s, based on a byte-oriented transfer format. it can be used with any atmel two-wire interface bus serial eeprom and i2c compatible device such as a real time clock (rtc), dot matrix/graphic lcd controller, and temperature sensor, to name a few. the twim is always a bus master and can transfer sequential or single byte s. multiple master capability is supported. arbitration of the bus is performed internally and relin quishes the bus automatically if the bus arbitration is lost. a configurable baud rate generator permits the output data rate to be adapted to a wide range of core clock frequencies. table 21-1 lists the compatibilit y level of the atmel two-wire interface in master mode and a full i2c compatible device. note: 1. start + b000000001 + ack + sr table 21-1. atmel twim compatib ility with i2c standard i2c standard atmel twim standard-mode (100 kbit/s) supported fast-mode (400 kbit/s) supported fast-mode plus (1 mbit/s) supported 7- or 10-bits slave addressing supported start byte (1) not supported repeated start (sr) condition supported ack and nack management supported slope control and input filtering (fast mode) supported clock stretching supported
470 32145c?06/2013 at32uc3l0128/256 table 21-2 lists the compatibility level of the atmel two-wire master interface and a full smbus compatible master. 21.3 list of abbreviations 21.4 block diagram figure 21-1. block diagram table 21-2. atmel twim compatibilit y with smbus standard smbus standard atmel twim bus timeouts supported address resolution protocol supported alert supported host functionality supported packet error checking supported table 21-3. abbreviations abbreviation description twi two-wire interface a acknowledge na non acknowledge pstop sstart sr repeated start sadr slave address adr any address except sadr r read wwrite peripheral bus bridge two-wire interface i/o controller twck twd intc twi interrupt power manager clk_twim twalm
471 32145c?06/2013 at32uc3l0128/256 21.5 application block diagram figure 21-2. application block diagram 21.6 i/o lines description 21.7 product dependencies in order to use this module, other parts of the system must be configured correctly, as described below. 21.7.1 i/o lines twd and twck are bidirectional lines, connected to a positive supply voltage via a current source or pull-up resistor (see figure 21-4 on page 473 ). when the bus is free, both lines are high. the output stages of devices connected to the bus must have an open-drain or open-col- lector to perform the wired-and function. twalm is used to implement the optional smbus smbalert signal. the twalm, twd, and twck pins may be multip lexed with i/o controller lines. to enable the twim, the user must perform the following steps: ? program the i/o controller to: ? dedicate twd, twck, and optionally twalm as peripheral lines. ? define twd, twck, and optionally twalm as open-drain. 21.7.2 power management if the cpu enters a sleep mode th at disables clocks used by t he twim, the twim will stop func- tioning and resume operation after the system wakes up from sleep mode. twi master twd twck atmel twi serial eeprom i 2 c rtc i 2 c lcd controller i 2 c temp sensor slave 2 slave 3 slave 4 vdd rp: pull-up value as given by the i2c standard twalm slave 1 rp rp rp table 21-4. i/o lines description pin name pin description type twd two-wire serial data input/output twck two-wire serial clock input/output twalm smbus smbalert input/output
472 32145c?06/2013 at32uc3l0128/256 21.7.3 clocks the clock for the twim bus interface (clk_twim) is generated by the power manager. this clock is enabled at reset, and can be disabled in the power manager. it is recommended to dis- able the twim before disabling the clock, to avoid freezing the twim in an undefined state. 21.7.4 dma the twim dma handshake interface is connected to the peripheral dma controller. using the twim dma functionality requires the peripheral dma controller to be programmed after setting up the twim. 21.7.5 interrupts the twim interrupt request lines are connected to the interrupt controller. using the twim inter- rupts requires the interrupt controller to be programmed first. 21.7.6 debug operation when an external debugger forces the cpu into debug mode, the twim continues normal oper- ation. if the twim is configured in a way that requires it to be periodically serviced by the cpu through interrupts or similar, improper operation or data loss may result during debugging.
473 32145c?06/2013 at32uc3l0128/256 21.8 functional description 21.8.1 transfer format the data put on the twd line must be 8 bits long. data is transferred msb first; each byte must be followed by an acknowledgement. the number of bytes per transfer is unlimited (see figure 21-4 ). each transfer begins with a start condition and terminates with a stop condition (see figure 21-4 ). ? a high-to-low transition on the twd line while twck is high defines the start condition. ? a low-to-high transition on the twd line while twck is high defines a stop condition. figure 21-3. start and stop conditions figure 21-4. transfer format 21.8.2 operation the twim has two modes of operation: ? master transmitter mode ? master receiver mode the master is the device which starts and stops a transfer and generates the twck clock. these modes are described in the following chapters. twd twck start stop twd twck start address r/w ack data ack data ack stop
474 32145c?06/2013 at32uc3l0128/256 21.8.2.1 clock generation the clock waveform generator register (cwgr) is used to control the waveform of the twck clock. cwgr must be written so that the desired twi bus timings are generated. cwgr describes bus timings as a function of cycles of a prescaled clock. the clock prescaling can be selected through the clock presca ler field in cwgr (cwgr.exp). cwgr has the following fields: low: prescaled clock cycles in cl ock low count. used to time t low and t buf . high: prescaled clock cycles in cl ock high count. used to time t high . stasto: prescaled clock cycles in cl ock high count. used to time t hd_sta , t su_sta , t su_sto . data: prescaled clock cycles for data setu p and hold count. used to time t hd_dat , t su_dat . exp: specifies the clo ck prescaler setting. note that the total clock low time generated is the sum of t hd_dat + t su_dat + t low . any slave or other bus master taking part in the transfer may extend the twck low period at any time. the twim hardware monitors the state of the tw ck line as required by the i2c specification. the clock generation counters are started when a high/low level is detected on the twck line, not when the twim hardware releases/drives the twck line. this means that the cwgr set- tings alone do not determine the twck frequency. the cwgr settings determine the clock low time and the clock high time, but the twck rise a nd fall times are determined by the external cir- cuitry (capacitive load, etc.). figure 21-5. bus timing diagram f prescaler f clk_twim 2 exp 1 + ?? ------------------------- - = s t hd:sta t low t su:dat t high t hd:dat t low p t su:sto sr t su:sta t su:dat
475 32145c?06/2013 at32uc3l0128/256 21.8.2.2 setting up and performing a transfer operation of the twim is mainly controlled by the control register (cr) and the command reg- ister (cmdr). twim status is pr ovided in the status register (sr). the following list presents the main steps in a typical communication: 1. before any transfers can be performed, bus timings must be configured by writing to the clock waveform generator register (cwgr). if operating in smbus mode, the smbus timing register (smbtr) register must also be configured. 2. if the peripheral dma controller is to be used for the transfers, it must be set up. 3. cmdr or ncmdr must be written with a value describing the transfer to be performed. the interrupt system can be set up to give interrupt requests on specific events or error condi- tions in the sr, for example when the transfer is complete or if arbitration is lost. the interrupt enable register (ier) and interrupt disable register (idr) can be written to specify which bits in the sr will generate interrupt requests. the sr.busfree bit is set when activity is co mpleted on the two-wire bus. the sr.crdy bit is set when cmdr and/or ncmdr is read y to receive one or more commands. the controller will refuse to start a new trans fer while anak, dnak, or arblst in the status register (sr) is one. this is necessary to avoid a race when the software issues a continuation of the current transfer at the same time as one of these errors happen. also, if anak or dnak occurs, a stop condition is sent automatically. the user will have to restart the transmission by clearing the error bits in sr after resolving the cause for the nack. after a data or address nack from the slave, a stop will be transmitted automatically. note that the valid bit in cmdr is not cleared in this case. if this transfer is to be discarded, the valid bit can be cleared manually allowing any command in ncmdr to be copied into cmdr. when a data or address nack is returned by the slave while the master is transmitting, it is pos- sible that new data has already been written to th e thr register. this data will be tran sferred out as the first data byte of the next transfer. if this behavior is to be avoided, the safest approach is to perform a software reset of the twim. 21.8.3 master transmitter mode a start condition is transmitted and master trans mitter mode is initiated when the bus is free and cmdr has been written with start=1 and read=0. start and sadr+w will then be transmitted. during the address acknowledge clock pulse (9th pulse), the master releases the data line (high), enabling the slave to pull it down in order to acknowledge the address. the master polls the data line during this clock pulse and sets the address not acknowledged bit (anak ) in the status register if no slave acknowledges the address. after the address phase, the following is repeated: while (nbytes>0) 1. wait until thr contains a valid data byte, stretching low period of twck. sr.txrdy indicates the state of thr. software or the peripheral dma controller must write the data byte to thr. 2. transmit this data byte 3. decrement nbytes 4. if (nbytes==0) and stop=1, transmit stop condition writing cmdr with start=stop=1 and nbytes= 0 will generate a transm ission with no data bytes, ie start, sadr+w, stop.
476 32145c?06/2013 at32uc3l0128/256 twi transfers require the slave to acknowle dge each received data byte. during the acknowl- edge clock pulse (9th pulse), the master releases the data line (high), enabling the slave to pull it down in order to generate the acknowledge. t he master polls the data line during this clock pulse and sets the data acknowledge bit (dnack) in the status register if the slave does not acknowledge the data byte. as with the other status bits, an interrupt can be generated if enabled in the interrupt enable register (ier). txrdy is used as transmit ready for the pe ripheral dma controller transmit channel. the end of a command is marked when the twim sets the sr.ccomp bit. see figure 21-6 and figure 21-7 . figure 21-6. master write with one data byte figure 21-7. master write with mu ltiple data bytes 21.8.4 master receiver mode a start condition is transmitted and master receiver mode is initiated when the bus is free and cmdr has been written with start=1 and read=1. start and sadr+r will then be trans- mitted. during the address acknowledge clock puls e (9th pulse), the master releases the data line (high), enabling the slave to pull it down in order to acknowledge the address. the master polls the data line during this clock pulse and sets the address not acknowledged bit (anak) in the status register if no slave acknowledges the address. after the address phase, the following is repeated: while (nbytes>0) twd sr.idle txrdy write thr (data) nbytes set to 1 stop sent automatically (ack received and nbytes=0) s dadr w a data a p twd sr.idle txrdy write thr (datan) nbytes set to n stop sent automatically (ack received and nbytes=0) s dadr w a datan a datan+5 aa datan+m p write thr (datan+1) write thr (datan+m) last data sent
477 32145c?06/2013 at32uc3l0128/256 1. wait until rhr is empty, stretching low period of twck. sr.rxrdy indicates the state of rhr. software or the peripheral dma controller must read any data byte present in rhr. 2. release twck generating a clock that the slave uses to transmit a data byte. 3. place the received data byte in rhr, set rxrdy. 4. if nbytes=0, generate a nak after the data byte, otherwise generate an ack. 5. decrement nbytes 6. if (nbytes==0) and stop=1, transmit stop condition. writing cmdr with start=stop=1 and nbytes= 0 will generate a transm ission with no data bytes, ie start, dadr+r, stop the twi transfers require the master to a cknowledge each received data byte. during the acknowledge clock pulse (9th pulse), the slave releases the data line (high), enabling the mas- ter to pull it down in order to generate the acknowledge. all data bytes except the last are acknowledged by the master. not acknowledging the last byte informs the slave that the transfer is finished. rxrdy is used as receive ready for the peripheral dma controller receive channel. figure 21-8. master read with one data byte figure 21-9. master read with mu ltiple data bytes twd sr.idle rxrdy write start & stop bit nbytes set to 1 read rhr s dadr r a data n p twd sr.idle rxrdy write start + stop bit nbytes set to m sdadr r a datan a datan+m-1 an datan+m p read rhr datan datan+1 read rhr datan+m-2 read rhr datan+m-1 read rhr datan+m send stop when nbytes=0
478 32145c?06/2013 at32uc3l0128/256 21.8.5 using the peripheral dma controller the use of the peripheral dma controller signifi cantly reduces the cpu load. the user can set up ring buffers for the peripheral dma controller, containing data to transmit or free buffer space to place received data. to assure correct behavior, respect the following programming sequences: 21.8.5.1 data transmit with the peripheral dma controller 1. initialize the transmit peripheral dma controller (memory pointers, size, etc.). 2. configure the twim (adr, nbytes, etc.). 3. start the transfer by enabling the peripheral dma controller to transmit. 4. wait for the peripheral dma controller end-of-transmit flag. 5. disable the peripheral dma controller. 21.8.5.2 data receive with th e peripheral dma controller 1. initialize the receive peripheral dma controller (memory pointers, size, etc.). 2. configure the twim (adr, nbytes, etc.). 3. start the transfer by enabling the peripheral dma controller to receive. 4. wait for the peripheral dma controller end-of-receive flag. 5. disable the peripheral dma controller. 21.8.6 multi-master mode more than one master may access the bus at the same time without data corruption by using arbitration. arbitration starts as soon as two or more masters place information on the bus at the same time, and stops (arbitration is lost) for the master that intends to send a logical one while the other master sends a logical zero. as soon as arbitration is lost by a master, it st ops sending data and listens to the bus in order to detect a stop. the sr.arblst flag will be set. when the stop is detected, the master who lost arbitration may reinitiate the data transfer. arbitration is illustrated in figure 21-11 . if the user starts a transfer and if the bus is busy, the twim automatically waits for a stop con- dition on the bus before initiating the transfer (see figure 21-10 ). note: the state of the bus (busy or free) is not indicated in the user interface.
479 32145c?06/2013 at32uc3l0128/256 figure 21-10. user sends data while the bus is busy figure 21-11. arbitration cases 21.8.7 combined transfers cmdr and ncmdr may be used to generate longer sequences of connected transfers, since generation of start and/or stop conditions is programmable on a per-command basis. writing ncmdr with start=1 when the previous transfer was written with stop=0 will cause a repeated start on the bus. the ability to gener ate such connected tr ansfers allows arbi- trary transfer lengths, since it is legal to write cmdr with both start=0 and stop=0. if this is done in master receiver mode, the cmdr.acklast bit must also be controlled. twck twd data sent by a master stop sent by the master start sent by the twi data sent by the twi bus is busy bus is free a transfer is programmed (dadr + w + start + write thr) transfer is initiated twi data transfer transfer is kept bus is considered as free twck bus is busy bus is free a transfer is programmed (dadr + w + start + write thr) transfer is initiated twi data transfer transfer is kept bus is considered as free data from a master data from twi s 0 s 0 0 1 1 1 arblst s 0 s 0 0 1 1 1 twd s 0 0 1 1 1 1 1 arbitration is lost twi stops sending data p s 0 1 p 0 1 1 1 1 data from the master data from the twi arbitration is lost the master stops sending data transfer is stopped transfer is programmed again (dadr + w + start + write thr) twck twd
480 32145c?06/2013 at32uc3l0128/256 as for single data transfers, the txrdy and rxrd y bits in the status register indicates when data to transmit can be written to thr, or when received data can be read from rhr. transfer of data to thr and from rhr can also be done automatically by dma, see section 21.8.5 21.8.7.1 write followed by write consider the fo llowing transfer: start, dadr+w, data+a, data+a, repstart, dadr+w, data+a, data+a, stop. to generate this transfer: 1. write cmdr with start=1, stop=0, dadr, nbytes=2 and read=0. 2. write ncmdr with start=1, stop=1, dadr, nbytes=2 and read=0. 3. wait until sr.txrdy==1, then write first data byte to transfer to thr. 4. wait until sr.txrdy==1, then write second data byte to transfer to thr. 5. wait until sr.txrdy==1, then write third data byte to transfer to thr. 6. wait until sr.txrdy==1, then write fourth data byte to transfer to thr. 21.8.7.2 read followed by read consider the fo llowing transfer: start, dadr+r, data+a, data+na, repstart, dadr+r, data+a, data+na, stop. to generate this transfer: 1. write cmdr with start=1, stop=0, dadr, nbytes=2 and read=1. 2. write ncmdr with start=1, stop=1, dadr, nbytes=2 and read=1. 3. wait until sr.rxrdy==1, then read first data byte received from rhr. 4. wait until sr.rxrdy==1, then read second data byte received from rhr. 5. wait until sr.rxrdy==1, then read third data byte received from rhr. 6. wait until sr.rxrdy==1, then read fourth data byte received from rhr. if combining several tr ansfers, without any stop or repeated start between them, remem- ber to write a one to the acklast bit in cmdr to keep from ending each of the partial transfers with a nack. 21.8.7.3 write followed by read consider the fo llowing transfer: start, dadr+w, data+a, data+a, repstart, dadr+r, data+a, data+na, stop.
481 32145c?06/2013 at32uc3l0128/256 figure 21-12. combining a write and read transfer to generate this transfer: 1. write cmdr with start=1, stop=0, dadr, nbytes=2 and read=0. 2. write ncmdr with start=1, stop=1, dadr, nbytes=2 and read=1. 3. wait until sr.txrdy==1, then write first data byte to transfer to thr. 4. wait until sr.txrdy==1, then write second data byte to transfer to thr. 5. wait until sr.rxrdy==1, then read first data byte received from rhr. 6. wait until sr.rxrdy==1, then read second data byte received from rhr. 21.8.7.4 read followed by write consider the fo llowing transfer: start, dadr+r, data+a, data+na, repstart, dadr+w, data+a, data+a, stop. figure 21-13. combining a read and write transfer to generate this transfer: 1. write cmdr with start=1, stop=0, dadr, nbytes=2 and read=1. 2. write ncmdr with start=1, stop=1, dadr, nbytes=2 and read=0. 3. wait until sr.rxrdy==1, then read first data byte received from rhr. 4. wait until sr.rxrdy==1, then read second data byte received from rhr. 5. wait until sr.txrdy==1, then write first data byte to transfer to thr. 6. wait until sr.txrdy==1, then write second data byte to transfer to thr. twd sr.idle txrdy sdadr wa data0 a data1 na sr dadr r a data2 a data3 a p data0 data1 thr rxrdy 1 rhr data3 data2 twd sr.idle txrdy s sadr r a data0 a data1 sr dadr w a data2 a data3 na p data2 thr rxrdy rhr data3 data0 a 1 2 data3 read twi_rhr
482 32145c?06/2013 at32uc3l0128/256 21.8.8 ten bit addressing writing a one to cmdr.tenbit enables 10-bit addressing in hardware. performing transfers with 10-bit addressing is similar to transfers with 7-bit addresses, except that bits 9:7 of cmdr.sadr must be written appropriately. in figure 21-14 and figure 21-15 , the grey boxes represent signals driven by the master, the white boxes are driven by the slave. 21.8.8.1 master transmitter to perform a master transmitter transfer: 1. write cmdr with tenbit=1, repsame=0, read=0, start=1, stop=1 and the desired address and nbytes value. figure 21-14. a write transfer with 10-bit addressing 21.8.8.2 master receiver when using master receiver mode with 10-bit addressing, cm dr.repsame must also be con- trolled. cmdr.repsame must be written to one when the address phase of the transfer should consist of only 1 address byte (the 11110xx byte) and not 2 address bytes. the i2c standard specifies that such addressing is required wh en addressing a slave for reads using 10-bit addressing. to perform a master receiver transfer: 1. write cmdr with tenbit=1, repsam e=0, read=0, start=1, stop=0, nbytes=0 and the desired address. 2. write ncmdr with tenbit=1, repsame=1, read=1, start=1, stop=1 and the desired address and nbytes value. figure 21-15. a read transfer with 10-bit addressing 21.8.9 smbus mode smbus mode is enabled and disabled by writing to the smen and smdis bits in cr. smbus mode operation is similar to i2c ope ration with the following exceptions: ? only 7-bit addressing can be used. ? the smbus standard describes a set of timeout values to ensure progress and throughput on the bus. these timeout values must be written into smbtr. ? transmissions can optionally include a crc byte, called packet error check (pec). ? a dedicated bus line, smbalert, allows a slave to get a master?s attention. ? a set of addresses have been reserved for protocol handling, such as alert response address (ara) and host header (hh) address. s slave address 1st 7 bits p a data rw a1 a2 slave address 2nd byte aa data 11110xx0 s slave address 1st 7 bits p a data rw a1 a2 slave address 2nd byte a data 11110xx0 sr slave address 1st 7 bits rw a3 11110xx1
483 32145c?06/2013 at32uc3l0128/256 21.8.9.1 packet error checking each smbus transfer can optionally end with a crc byte, called the pec byte. writing a one to cmdr.pecen enables automatic pec handling in the current transfer. transfers with and with- out pec can freely be intermixed in the same system, since some slaves may not support pec. the pec lfsr is always updated on every bit transmitted or received, so that pec handling on combined transfers will be correct. in master transmitter mode, the master calculates a pec value and transmits it to the slave after all data bytes have been transm itted. upon reception of this pec byte, the slave will compare it to the pec value it has computed itself. if the values match, the data was received correctly, and the slave will return an ack to th e master. if the pec values differ, data was corrupted, and the slave will return a nack value. the dnak bit in sr reflects the state of the last received ack/nack value. some slaves may not be able to check the received pec in time to return a nack if an error occurred. in this case, the slave should always retu rn an ack after the pec byte, and some other mechanism must be implemented to verify that the transmission was received correctly. in master receiver mode, the slave calculates a pec value and transmits it to the master after all data bytes have been transmitted. upon reception of this pec by te, the master will compare it to the pec value it has computed itself. if the values match, the data was received correctly. if the pec values differ, data was corrupted, and sr.pecerr is set. in master receiver mode, the pec byte is always followed by a nack transmitted by the master, since it is the last byte in the transfer. the pec byte is automatically inserted in a master transmitter transmission if pec is enabled when nbytes reaches zero. the pec byte is identified in a master receiver transmission if pec is enabled when nbytes reaches zero. nbytes must therefore be written with the total number of data bytes in the transmission, including the pec byte. in combined transfers, the pecen bit should only be written to one in the last of the combined transfers. consider the following transfer: s, adr+w, command_byte, ack, sr, adr+ r, data_byte, ack, pec_byte, nack, p this transfer is generated by writing two commands to the command registers. the first com- mand is a write with nbytes=1 and pecen=0, and the second is a read with nbytes=2 and pecen=1. writing a one to the stop bit in cr will pl ace a stop condition on the bus after the current byte. no pec byte will be sent in this case. 21.8.9.2 timeouts the tlows and tlowm fields in smbtr confi gure the smbus timeout values. if a timeout occurs, the master will transmit a stop conditi on and leave t he bus. the sr.tout bit is set. 21.8.9.3 smbus alert signal a slave can get the master?s at tention by pulling the twalm line low. the twim will then set the sr.smbalert bit. this can be set up to trigger an interrupt, and software can then take the appropriate action, as defined in the smbus standard.
484 32145c?06/2013 at32uc3l0128/256 21.8.10 identifying bus events this chapter lists the different bus events, and how they affect bits in the twim registers. this is intended to help writin g drivers for the twim. table 21-5. bus events event effect master transmitter has sent a data byte sr.thr is cleared. master receiver has received a data byte sr.rhr is set. start+sadr sent, no ack received from slave sr.anak is set. sr.ccomp not set. cmdr.valid remains set. stop automatically transmitted on bus. data byte sent to slave, no ack received from slave sr.dnak is set. sr.ccomp not set. cmdr.valid remains set. stop automatically transmitted on bus. arbitration lost sr.arblst is set. sr.ccomp not set. cmdr.valid remains set. twck and twd immediately released to a pulled-up state. smbus alert received sr.smbalert is set. smbus timeout received sr.smbtout is set. sr.ccomp not set. cmdr.valid remains set. stop automatically transmitted on bus. master transmitter receives smbus pec error sr.dnak is set. sr.ccomp not set. cmdr.valid remains set. stop automatically transmitted on bus. master receiver discovers smbus pec error sr.pecerr is set. sr.ccomp not set. cmdr.valid remains set. stop automatically transmitted on bus. cr.stop is written by user sr.stop is set. sr.ccomp set. cmdr.valid remains set. stop transmitted on bus after curr ent byte transfer has finished.
485 32145c?06/2013 at32uc3l0128/256 21.9 user interface note: 1. the reset values for these registers are device specific. please refer to the module configuration section at the end of this chapter. table 21-6. twim register memory map offset register register name access reset 0x00 control register cr write-only 0x00000000 0x04 clock waveform generator register cwgr read/write 0x00000000 0x08 smbus timing register smbtr read/write 0x00000000 0x0c command register cmdr read/write 0x00000000 0x10 next command register ncmdr read/write 0x00000000 0x14 receive holding register rhr read-only 0x00000000 0x18 transmit holding register thr write-only 0x00000000 0x1c status register sr read-only 0x00000002 0x20 interrupt enable register ier write-only 0x00000000 0x24 interrupt disable register idr write-only 0x00000000 0x28 interrupt mask register imr read-only 0x00000000 0x2c status clear register scr write-only 0x00000000 0x30 parameter register pr read-only - (1) 0x34 version register vr read-only - (1)
486 32145c?06/2013 at32uc3l0128/256 21.9.1 control register name: cr access type: write-only offset: 0x00 reset value: 0x00000000 ? stop: stop the current transfer writing a one to this bit terminates the current transfer, sendi ng a stop condition after the shifter has become idle. if there are additional pending transfers, they will have to be explicitly re started by software after the stop condition has been successfu lly sent. writing a zero to this bit has no effect. ? swrst: software reset if the twim master interface is enabled, writing a one to this bit resets the twim. all transfers are halted immediately, possi bly violating the bus semantics. if the twim master interface is not enabled, it must first be enabled before writing a one to this bit. writing a zero to this bit has no effect. ? smdis: smbus disable writing a one to this bit disables smbus mode. writing a zero to this bit has no effect. ? smen: smbus enable writing a one to this bit enables smbus mode. writing a zero to this bit has no effect. ? mdis: master disable writing a one to this bit disables the master interface. writing a zero to this bit has no effect. ? men: master enable writing a one to this bit enables the master interface. writing a zero to this bit has no effect. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------stop 76543210 swrst - smdis smen - - mdis men
487 32145c?06/2013 at32uc3l0128/256 21.9.2 clock waveform generator register name: cwgr access type: read/write offset: 0x04 reset value: 0x00000000 ? exp: clock prescaler used to specify how to prescale the twck clock. coun ters are prescaled according to the following formula ? data: data setu p and hold cycles clock cycles for data setup and hold count. prescaled by cwgr.exp . used to time t hd_dat , t su_dat . ? stasto: start and stop cycles clock cycles in clock high count. prescaled by cwgr.exp. used to time t hd_sta , t su_sta , t su_sto ? high: clock high cycles clock cycles in clock high count. prescaled by cwgr.exp. used to time t high . ? low: clock low cycles clock cycles in clock low count. pre scaled by cwgr.exp. used to time t low , t buf . 31 30 29 28 27 26 25 24 - exp data 23 22 21 20 19 18 17 16 stasto 15 14 13 12 11 10 9 8 high 76543210 low f prescaler f clk_twim 2 exp 1 + ?? ------------------------- - =
488 32145c?06/2013 at32uc3l0128/256 21.9.3 smbus timing register name: smbtr access type: read/write offset: 0x08 reset value: 0x00000000 ? exp: smbus timeout clock prescaler used to specify how to prescale the ti m and tlowm counters in smbtr. counters are prescaled according to the following formula ? thmax: clock high maximum cycles clock cycles in clock high maximum count. prescaled by smbtr.exp. used for bus free de tection. used to time t high:max . note: uses the prescaler specified by cw gr, not the prescaler specified by smbtr. ? tlowm: master clock stretch maximum cycles clock cycles in master maximum clock stretch count. prescaled by smbtr.exp. used to time t low:mext ? tlows: slave clock stretch maximum cycles clock cycles in slave maximum clock stretch coun t. prescaled by smbtr.exp. used to time t low:sext . 31 30 29 28 27 26 25 24 exp ---- 23 22 21 20 19 18 17 16 thmax 15 14 13 12 11 10 9 8 tlowm 76543210 tlows f prescaled smbus ? f clktwim 2 exp 1 + ?? ------------------------ =
489 32145c?06/2013 at32uc3l0128/256 21.9.4 command register name: cmdr access type: read/write offset: 0x0c reset value: 0x00000000 ? acklast: ack last master rx byte 0: causes the last byte in master re ceive mode (when nbytes has reached 0) to be nacked. this is the standard way of ending a master receiver transfer. 1: causes the last byte in master re ceive mode (when nbytes has reached 0) to be acked. used for performing linked transfers in master receiver mode with no stop or repeated start between the subtransfers. this is needed when more than 255 bytes are to be received in one single transmission. ? pecen: packet error checking enable 0: causes the transfer not to use pec byte verification. the pec lfsr is still updated for every bit transmitted or received. m ust be used if smbus mode is disabled. 1: causes the transfer to use pec. pec byte generation (if master transmitter) or pec byte verifi cation (if master receiver) wi ll be performed. ? nbytes: number of data bytes in transfer the number of data bytes in the transfer. after the specified number of bytes have been transferred, a stop condition is transmitted if cmdr.stop is one. in smbus mode, if pec is used, nbytes includes the pec byte, i.e. there are nbytes-1 data bytes and a pec byte. ? valid: cmdr valid 0: indicates that cmdr does not contain a valid command. 1: indicates that cmdr contains a valid command. this bit is cleared when the command is finished. ? stop: send stop condition 0: do not transmit a stop condition af ter the data bytes have been transmitted. 1: transmit a stop condition after t he data bytes have been transmitted. ? start: send start condition 0: the transfer in cmdr should not commen ce with a start or repeated start condition. 1: the transfer in cmdr should commence with a start or re peated start condition. if the bus is free when the command is executed, a start condition is used. if the bus is busy, a repeated start is used. ? repsame: transfer is to same address as previous address only used in 10-bit addressing mode, always write to 0 in 7-bit addressing mode. 31 30 29 28 27 26 25 24 - - - - acklast pecen 23 22 21 20 19 18 17 16 nbytes 15 14 13 12 11 10 9 8 valid stop start repsame tenbit sadr[9:7] 76543210 sadr[6:0] read
490 32145c?06/2013 at32uc3l0128/256 write this bit to one if the command in cmdr performs a repeated start to the same slave address as addressed in the previous transfer in order to enter master receiver mode. write this bit to zero otherwise. ? tenbit: ten bit addressing mode 0: use 7-bit addressing mode. 1: use 10-bit addressing mode. must not be used when the twim is in smbus mode. ? sadr: slave address address of the slave involved in the transfer. bi ts 9-7 are don?t care if 7-bit addressing is used. ? read: transfer direction 0: allow the master to transmit data. 1: allow the master to receive data.
491 32145c?06/2013 at32uc3l0128/256 21.9.5 next command register name: ncmdr access type: read/write offset: 0x10 reset value: 0x00000000 this register is identical to cmdr. when the valid bit in cmdr becomes 0, the content of ncmdr is copied into cmdr, clearing the valid bit in ncmdr. if the valid bit in cmdr is cleared when ncmdr is written, the content is copied immediately. 31 30 29 28 27 26 25 24 - - - - acklast pecen 23 22 21 20 19 18 17 16 nbytes 15 14 13 12 11 10 9 8 valid stop start repsame tenbit sadr[9:7] 76543210 sadr[6:0] read
492 32145c?06/2013 at32uc3l0128/256 21.9.6 receive holding register name: rhr access type: read-only offset: 0x14 reset value: 0x00000000 ? rxdata: received data when the rxrdy bit in the status regist er (sr) is one, this field contains a byte received from the twi bus. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 rxdata
493 32145c?06/2013 at32uc3l0128/256 21.9.7 transmit holding register name: thr access type: write-only offset: 0x18 reset value: 0x00000000 ? txdata: data to transmit write data to be transf erred on the twi bus here. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 txdata
494 32145c?06/2013 at32uc3l0128/256 21.9.8 status register name: sr access type: read-only offset: 0x1c reset value: 0x00000002 ? menb: master interface enable 0: master interface is disabled. 1: master interface is enabled. ? stop: stop request accepted this bit is one when a stop request caused by writing a one to cr.stop has been accepted, and transfer has stopped. this bit is cleared by writing 1 to the corresp onding bit in the status clear register (scr). ? pecerr: pec error this bit is one when a smbus pec error occurred. this bit is cleared by writing 1 to the corresp onding bit in the status clear register (scr). ?tout: timeout this bit is one when a smbus timeout occurred. this bit is cleared by writing 1 to the corresp onding bit in the status clear register (scr). ? smbalert: smbus alert this bit is one when an smbus alert was received. this bit is cleared by writing 1 to the corresp onding bit in the status clear register (scr). ? arblst: arbitration lost this bit is one when the actual state of the sda line did not correspond to the data driven onto it, indicating a higher-priori ty transmission in progress by a different master. this bit is cleared by writing 1 to the corresp onding bit in the status clear register (scr). ? dnak: nak in data phase received this bit is one when no ack was received form slave during data transmission. this bit is cleared by writing 1 to the corresp onding bit in the status clear register (scr). ? anak: nak in address phase received this bit is one when no ack was received from slave during address phase this bit is cleared by writing 1 to the correspo nding bit in the status clear register (scr). ? busfree: two-wire bus is free this bit is one when activity has completed on the two-wire bus. otherwise, this bit is cleared. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------menb 15 14 13 12 11 10 9 8 - stop pecerr tout smbalert arblst dnak anak 76543210 - - busfree idle ccomp crdy txrdy rxrdy
495 32145c?06/2013 at32uc3l0128/256 ? idle: master interface is idle this bit is one when no command is in progress, and no command waiting to be issued. otherwise, this bit is cleared. ? ccomp: command complete this bit is one when the current command has completed successfully. this bit is zero if the command failed due to conditions such as a nak receved from slave. this bit is cleared by writing 1 to the corresp onding bit in the status clear register (scr). ? crdy: ready for more commands this bit is one when cmdr and/or ncmdr is ready to receive one or more commands. this bit is cleared when this is no longer true. ? txrdy: thr data ready this bit is one when thr is ready for one or more data bytes. this bit is cleared when this is no longer tr ue (i.e. thr is full or transmission has stopped). ? rxrdy: rhr data ready this bit is one when rx data are ready to be read from rhr. this bit is cleared when this is no longer true.
496 32145c?06/2013 at32uc3l0128/256 21.9.9 interrupt enable register name: ier access type: write-only offset: 0x20 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will set the corresponding bit in imr 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 - stop pecerr tout smbalert arblst dnak anak 76543210 - - busfree idle ccomp crdy txrdy rxrdy
497 32145c?06/2013 at32uc3l0128/256 21.9.10 interrupt disable register name: idr access type: write-only offset: 0x24 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in imr 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 - stop pecerr tout smbalert arblst dnak anak 76543210 - - busfree idle ccomp crdy txrdy rxrdy
498 32145c?06/2013 at32uc3l0128/256 21.9.11 interrupt mask register name: imr access type: read-only offset: 0x28 reset value: 0x00000000 0: the corresponding interrupt is disabled. 1: the corresponding interrupt is enabled. this bit is cleared when the corresponding bit in idr is written to one. this bit is set when the corresponding bit in ier is written to one. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 - stop pecerr tout smbalert arblst dnak anak 76543210 - - busfree idle ccomp crdy txrdy rxrdy
499 32145c?06/2013 at32uc3l0128/256 21.9.12 status clear register name: scr access type : write-only offset: 0x2c reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in sr and the corresponding interrupt request. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 - stop pecerr tout smbalert arblst dnak anak 76543210 ----ccomp---
500 32145c?06/2013 at32uc3l0128/256 21.9.13 parameter register (pr) name: pr access type: read-only offset: 0x30 reset value: - 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 --------
501 32145c?06/2013 at32uc3l0128/256 21.9.14 version register (vr) name: vr access type: read-only offset: 0x34 reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 ---- version [11:8] 76543210 version [7:0]
502 32145c?06/2013 at32uc3l0128/256 21.10 module configuration the specific configuration for each twim inst ance is listed in the following tables.the module bus clocks listed here are connected to the syst em bus clocks. please refer to the power man- ager chapter for details. table 21-7. module clock name module name clock name description twim0 clk_twim0 clock for the twim0 bus interface twim1 clk_twim1 clock for the twim1 bus interface table 21-8. register reset values register reset value version 0x00000110 parameter 0x00000000
503 32145c?06/2013 at32uc3l0128/256 22. two-wire slave interface (twis) rev.: 1.2.0.1 22.1 features ? compatible with i 2 c standard ? transfer speeds of 100 and 400 kbit/s ? 7 and 10-bit and general call addressing ? compatible with smbus standard ? hardware packet error checking (crc) ge neration and verification with ack response ? smbalert interface ? 25 ms clock low timeout delay ? 25 ms slave cumulative clock low extend time ? compatible with pmbus ? dma interface for reducing cpu load ? arbitrary transfer lengths, including 0 data bytes ? optional clock stretching if transmit or receive buffers not ready for data transfer ? 32-bit peripheral bus interface fo r configuration of the interface 22.2 overview the atmel two-wire slave interface (twis) interconnects components on a unique two-wire bus, made up of one clock line and one data line with speeds of up to 400 kbit/s, based on a byte-oriented transfer format. it can be used with any atmel two-wire interface bus, i2c, or smbus-compatible master. the twis is always a bus slave and can transfer sequential or sin- gle bytes. below, table 22-1 lists the compatibility leve l of the atmel two-wire slav e interface and a full i2c compatible device. note: 1. start + b000000001 + ack + sr table 22-1. atmel twis compatibility with i2c standard i2c standard atmel twis standard-mode (100 kbit/s) supported fast-mode (400 kbit/s) supported 7 or 10 bits slave addressing supported start byte (1) not supported repeated start (sr) condition supported ack and nak management supported slope control and input filtering (fast mode) supported clock stretching supported
504 32145c?06/2013 at32uc3l0128/256 below, table 22-2 lists the compatibility level of the atmel two-wire slave interface and a full smbus compatible device. 22.3 list of abbreviations 22.4 block diagram figure 22-1. block diagram table 22-2. atmel twis compatibilit y with smbus standard smbus standard atmel twis bus timeouts supported address resolution protocol supported alert supported packet error checking supported table 22-3. abbreviations abbreviation description twi two-wire interface a acknowledge na non acknowledge pstop sstart sr repeated start sadr slave address adr any address except sadr r read wwrite peripheral bus bridge two-wire interface i/o controller twck twd interrupt controller twi interrupt power manager clk_twis twalm
505 32145c?06/2013 at32uc3l0128/256 22.5 application block diagram figure 22-2. application block diagram 22.6 i/o lines description 22.7 product dependencies in order to use this module, other parts of the system must be configured correctly, as described below. 22.7.1 i/o lines twdand twck are bidirectional lines, connected to a positive supply voltage via a current source or pull-up resistor (see figure 22-5 on page 507 ). when the bus is free, both lines are high. the output stages of devices connected to the bus must have an open-drain or open-col- lector to perform the wired-and function. twalm is used to implement the optional smbus smbalert signal. twalm, twd, and twck pins may be multiple xed with i/o controller lines. to enable the twis, the user must perform the following steps: ? program the i/o controller to: ? dedicate twd, twck, and optionally twalm as peripheral lines. ? define twd, twck, and optionally twalm as open-drain. host with twi interface twd twck atmel twi serial eeprom i2c rtc i2c lcd controller slave 1 slave 2 slave 3 vdd i2c temp. sensor slave 4 rp: pull up value as given by the i2c standard rp rp table 22-4. i/o lines description pin name pin description type twd two-wire serial data input/output twck two-wire serial clock input/output twalm smbus smbalert input/output
506 32145c?06/2013 at32uc3l0128/256 22.7.2 power management if the cpu enters a sleep mode that disables cl ocks used by the twis, the twis will stop func- tioning and resume operation after the system wakes up from sleep mode. the twis is able to wake the system from sleep mode upon address match, see section 22.8.8 on page 514 . 22.7.3 clocks the clock for the twis bus interface (clk_twis) is generated by the power manager. this clock is enabled at reset, and can be disabled in the power manager. it is recommended to dis- able the twis before disabling the clock, to avoid freezing the twis in an undefined state. 22.7.4 dma the twis dma handshake interface is connected to the peripheral dma controller. using the twis dma functionality requires the peripheral dma controller to be programmed after setting up the twis. 22.7.5 interrupts the twis interrupt request lines are connected to the interrupt controller. using the twis inter- rupts requires the interrupt controller to be programmed first. 22.7.6 debug operation when an external debugger forces the cpu into debug mode, the twis continues normal oper- ation. if the twis is configured in a way that requires it to be periodically serviced by the cpu through interrupts or similar, improper operation or data loss may result during debugging. 22.8 functional description 22.8.1 transfer format the data put on the twd line must be 8 bits long. data is transferred msb first; each byte must be followed by an acknowledgement. the number of bytes per transfer is unlimited (see figure 22-4 on page 507 ). each transfer begins with a start condition and terminates with a stop condition (see figure 22-3 ). ? a high-to-low transition on the twd line while twck is high defines the start condition. ? a low-to-high transition on the twd line while twck is high defines a stop condition. figure 22-3. start and stop conditions twd twck start stop
507 32145c?06/2013 at32uc3l0128/256 figure 22-4. transfer format 22.8.2 operation the twis has two modes of operation: ? slave transmitter mode ? slave receiver mode a master is a device which starts and stops a transfer and generates the twck clock. a slave is assigned an address and responds to requests from the master. these modes are described in the following chapters. figure 22-5. typical application block diagram 22.8.2.1 bus timing the timing register (tr) is used to control the timing of bus signals driven by the twis. tr describes bus timings as a function of cycles of the prescaled clk_twis. the clock prescaling can be selected through tr.exp. tr has the following fields: tlows: prescaled clock cycles used to time smbus timeout t low:sext . twd twck start address r/w ack data ack data ack stop host with twi interface twd twck atmel twi serial eeprom i2c rtc i2c lcd controller slave 1 slave 2 slave 3 vdd i2c temp. sensor slave 4 rp: pull up value as given by the i2c standard rp rp f prescaled f clk_twis 2 exp 1 + ?? ------------------------ - =
508 32145c?06/2013 at32uc3l0128/256 ttout: prescaled clock cycles us ed to time smbus timeout t timeout . sudat: non-prescaled clock cycles for data setup and hold count. used to time t su_dat . exp: specifies the clock prescaler setting used for the smbus timeouts. figure 22-6. bus timing diagram 22.8.2.2 setting up and performing a transfer operation of the twis is mainly controlled by the control register (cr) . the following list pres- ents the main steps in a typical communication: 3. before any transfers can be performed, bus timings must be configured by writing to the timing register (tr).if the peripheral dma controller is to be used for the transfers, it must be set up. 4. the control register (cr) must be configured with information such as the slave address, smbus mode, packet error checking (pec), number of bytes to transfer, and which addresses to match. the interrupt system can be set up to generate interrupt request on specific events or error con- ditions, for example when a byte has been received. the nbytes register is only used in smbus mode, when pec is enabled. in i2c mode or in smbus mode when pec is disabled, the nbytes register is not used, and should be written to zero. nbytes is updated by hardware, so in order to avoid hazards, software updates of nbytes can only be done through writes to the nbytes register. 22.8.2.3 address matching the twis can be set up to match several different addresses. more than one address match may be enabled simultaneously, allowing the twis to be assigned to several addresses. the address matching phase is initiated after a st art or repeated start condition. when the twis receives an address that generates an addr ess match, an ack is automatically returned to the master. s t hd:sta t low t su:dat t high t hd:dat t low p t su:sto sr t su:sta t su:dat
509 32145c?06/2013 at32uc3l0128/256 in i2c mode: ? the address in cr.adr is checked fo r address match if cr.smatch is one. ? the general call address is checked fo r address match if cr.gcmatch is one. in smbus mode: ? the address in cr.adr is checked fo r address match if cr.smatch is one. ? the alert response address is checked for address match if cr.smal is one. ? the default address is checked for address match if cr.smda is one. ? the host header address is checked for address match if cr.smhh is one. 22.8.2.4 clock stretching any slave or bus master taking part in a transfer may extend the twck low period at any time. the twis may extend the twck low period afte r each byte transfer if cr.stren is one and: ? module is in slave transmitter mode, data should be transmitted, but thr is empty, or ? module is in slave receiver mode, a byte has been received and placed into the internal shifter, but the receive holdin g register (rhr) is full, or ? stretch-on-address-match bit cr.soam=1 and slave was addressed. bus clock remains stretched until all address match bits in the status register (sr) have been cleared. if cr.stren is zero and: ? module is in slave transmitter mode, data shou ld be transmitted but thr is empty: transmit the value present in thr (the last transmitted byte or reset value), and set sr.urun. ? module is in slave receiver mode, a byte has been received and placed into the internal shifter, but rhr is full: discard the received byte and set sr.orun. 22.8.2.5 bus errors if a bus error (misplaced start or stop) cond ition is detected, the sr.buserr bit is set and the twis waits for a new start condition. 22.8.3 slave transmitter mode if the twis matches an address in which the r/w bit in the twi address phase transfer is set, it will enter slave transmitter mode and set the sr.tra bit (note that sr.tra is set one clk_twis cycle after the relevant address match bit in the same register is set). after the address phase, the following actions are performed: 1. if smbus mode and pec is used, nbytes must be set up with the number of bytes to transmit. this is necessary in order to know when to transmit the pec byte. nbytes can also be used to count the number of bytes received if using dma. 2. byte to transmit depends on i2c/smbus mode and cr.pec: ? if in i2c mode or cr.pec is zero or n bytes is non-zero: the twis waits until thr contains a valid data byte, possibly stretching the low period of twck. after thr contains a valid data byte, the data byte is transferred to a shifter, and then sr.txrdy is changed to one be cause the thr is empty again. ? smbus mode and cr.pec is one: if nbytes is zero, the generated pec byte is automatically transmitted in stead of a data byte from thr. twck will not be stretched by the twis. 3. the data byte in the shifter is transmitted.
510 32145c?06/2013 at32uc3l0128/256 4. nbytes is updated. if cr.cup is one, nbyt es is incremented, otherwise nbytes is decremented. 5. after each data byte has been transmitted, the master transmits an ack (acknowledge) or nak (not acknowledge) bit. if a nak bit is received by the twis, the sr.nak bit is set. note that this is done two clk_twis cycles after twck has been sampled by the twis to be high (see figure 22-9 ). the nak indicates that the transfer is finished, and the twis will wait for a stop or repeated start. if an ack bit is received, the sr.nak bit remains low. the ack indicates that more data should be transmitted, jump to step 2. at the end of the ack/nak clock cycle, the byte transfer finished (sr.btf) bit is set. note that this is done two clk_twis cycles after twck has been sampled by the twis to be low (see figure 22-9 ). also note that in the event that sr.nak bit is set, it must not be cleared be fore the sr.btf bit is set to ensure correct twis behavior. 6. if stop is received, sr.tcomp and sr.sto will be set. 7. if repeated start is received, sr.rep will be set. the twi transfers require the receiver to acknowledge each received data byte. during the acknowledge clock pulse (9th pulse), the slave releases the data line (high), enabling the mas- ter to pull it down in order to generate the acknowledge. the slave polls the data line during this clock pulse and sets the nak bit in sr if the ma ster does not acknowledge the data byte. a nak means that the master does not wish to receive additional data bytes. as with the other status bits, an interrupt can be generated if enabled in the interrupt enable register (ier). sr.txrdy is used as transmit ready for the peripheral dma controller transmit channel. the end of the complete transfer is marked by the sr.tcomp bit changing from zero to one. see figure 22-7 and figure 22-8 . figure 22-7. slave transmitter with one data byte tcomp txrdy write thr (data) stop sent by master twd adatan sdadrr p nbytes set to 1
511 32145c?06/2013 at32uc3l0128/256 figure 22-8. slave transmitter with multiple data bytes figure 22-9. timing relationship between twck, sr.nak, and sr.btf 22.8.4 slave receiver mode if the twis matches an address in which the r/w bit in the twi address phase transfer is cleared, it will enter slave receiver mode and cl ear sr.tra (note that sr.tra is cleared one clk_twis cycle after the relevant address match bit in the same register is set). after the address phase, the following is repeated: 1. if smbus mode and pec is used, nbytes must be set up with the number of bytes to receive. this is necessary in order to know which of the received bytes is the pec byte. nbytes can also be used to count the number of bytes received if using dma. 2. receive a byte. set sr.btf when done. 3. update nbytes. if cr.cup is written to one, nbytes is incremented, otherwise nbytes is decremented. nbytes is usually configured to count downwards if pec is used. 4. after a data byte has been received, the slave transmits an ack or nak bit. for ordi- nary data bytes, the cr.ack field controls if an ack or nak should be returned. if pec is enabled and the last byte received was a pec byte (indicated by nbytes equal to zero), the twis will automatically return an ack if the pec value was correct, other- wise a nak will be returned. 5. if stop is received, sr.tcomp will be set. 6. if repeated start is received, sr.rep will be set. the twi transfers require the receiver to acknowledge each received data byte. during the acknowledge clock pulse (9th pulse), the master releases the data line (high), enabling the adata na sdadrr data n+5a p data n+m n tco m p txrdy write thr (data n) nbytes set to m stop sent by master twd write thr (data n+1) write thr (data n+m) last data sent data (lsb) n p twck sr.nak sr.btf t 1 t 1 t 1 : (clk_twis period) x 2 twd
512 32145c?06/2013 at32uc3l0128/256 slave to pull it down in order to generate the acknowledge. the master polls the data line during this clock pulse. the sr.rxrdy bit indicates that a data byte is available in the rhr. the rxrdy bit is also used as receive ready for the peripheral dma controller receive channel. figure 22-10. slave receiver with one data byte figure 22-11. slave receiver with multiple data bytes 22.8.5 interactive acking received data bytes when implementing a register interface over twi, it may sometimes be necessary or just useful to report reads and writes to invalid register addresses by sending a nak to the host. to be able to do this, one must first receive the register address from the twi bus, and then tell the twis whether to ack or nak it. in normal operation of the twis, this is not possible because the con- troller will automatically ack the byte at about the same time as the rxrdy bit changes from zero to one. writing a one to the stretch on data byte received bit (cr.sodr) will stretch the clock allowing the user to update cr.ack bit before returning the desired value. after the last bit in the data byte is received, the twi bus clock is stretched, the received data byte is transferred to the rhr register, and sr.btf is set. at this time, the user can examine the received byte and write the desired ack or nack value to cr.a ck. when the user clears sr.btf, the desired ack value is transferred on the twi bus. this makes it possible to look at the byte received, determine if it is valid, and then decide to ack or nak it. 22.8.6 using the peripheral dma controller the use of the peripheral dma controller signifi cantly reduces the cpu load. the user can set up ring buffers for the peripheral dma controller, containing data to transmit or free buffer space to place received data. by initializing nbytes to zero before a transfer, and writing a one to cr.cup, nbytes is incremented by one each time a data has been transmitted or received. this allows the user to detect how much data was actually transferred by the dma system. a sdadrw data ap tcomp rx r dy read rhr twd a a sdadrw data na a data (n+1) a data (n+m) dat a (n+m)-1 p twd tco m p rx r dy read rhr data n read rhr dat a (n+1) read rhr dat a (n+m)-1 read rhr data (n+m)
513 32145c?06/2013 at32uc3l0128/256 to assure correct behavior, respect the following programming sequences: 22.8.6.1 data transmit with the peripheral dma controller 1. initialize the transmit peripheral dma controller (memory pointers, size, etc.). 2. configure the twis (adr, nbytes, etc.). 3. start the transfer by enabling the peripheral dma controller to transmit. 4. wait for the peripheral dma controller end-of-transmit flag. 5. disable the peripheral dma controller. 22.8.6.2 data receive with th e peripheral dma controller 1. initialize the receive peripheral dma controller (memory pointers, size - 1 , etc.). 2. configure the twis (adr, nbytes, etc.). 3. start the transfer by enabling the peripheral dma controller to receive. 4. wait for the peripheral dma controller end-of-receive flag. 5. disable the peripheral dma controller. 22.8.7 smbus mode smbus mode is enabled by writing a one to the smbus mode enable (smen) bit in cr. smbus mode operation is similar to i2c ope ration with the following exceptions: ? only 7-bit addressing can be used. ? the smbus standard describes a set of timeout values to ensure progress and throughput on the bus. these timeout values must be written to tr. ? transmissions can optionally include a crc byte, called packet error check (pec). ? a dedicated bus line, smbalert, allows a slave to get a master?s attention. ? a set of addresses have been reserved for protocol handling, such as alert response address (ara) and host header (hh) addres s. address matching on these addresses can be enabled by configuring cr appropriately. 22.8.7.1 packet error checking (pec) each smbus transfer can optionally end with a crc byte, called the pec byte. writing a one to the packet error checking enable (pecen) bit in cr enables automatic pec handling in the current transfer. the pec generator is always updated on every bit transmitted or received, so that pec handling on following linked transfers will be correct. in slave receiver mode, the master calculates a pec value and transmits it to the slave after all data bytes have been tr ansmitted. upon reception of this pec byte, the slave will compare it to the pec value it has computed itself. if the values match, the data was received correctly, and the slave will return an ack to th e master. if the pec values differ, data was corrupted, and the slave will return a nak value. the sr.smbpece rr bit is set automatically if a pec error occurred. in slave transmitter mode, the slave calculates a pec value and transmits it to the master after all data bytes have been transm itted. upon reception of this pec byte, the master will compare it to the pec value it has computed itself. if the values match, the data was received correctly. if the pec values differ, data was corrupted, and the master must take appropriate action. the pec byte is automatically inserted in a slave transmitter transmission if pec enabled when nbytes reaches zero. the pec byte is identified in a slave receiver transmission if pec
514 32145c?06/2013 at32uc3l0128/256 enabled when nbytes reaches zero. nbytes must therefore be set to the total number of data bytes in the transmission, including the pec byte. 22.8.7.2 timeouts the timing register (tr) configures the smbus timeout values. if a timeout occurs, the slave will leave the bus. the sr.smbtout bit is also set. 22.8.7.3 smbalert a slave can get the master?s att ention by pulling the smbalert line lo w. this is done by writing a one to the smbus alert (smbalert) bit in cr. this will also enable address match on the alert response address (ara). 22.8.8 wakeup from sleep modes by twi address match the twis is able to wake the device up from a sleep mode upon an address match, including sleep modes where clk_twis is stopped. after detecting the start condition on the bus, the twis will stretch twck until clk_twis has star ted. the time required for starting clk_twis depends on which sleep mode the device is in. after clk_twis has started, the twis releases its twck stretching and receives one byte of data on the bus. at this time, only a limited part of the device, including the twis, receives a cl ock, thus saving power. the twis goes on to receive the slave address. if the address phase causes a twis address match, the entire device is wakened and normal twis address matching actions are performed. normal twi transfer then follows. if the twis is not addres sed, clk_twis is automatically stopped and the device returns to its original sleep mode. 22.8.9 identifying bus events this chapter lists the different bus events, and how these affects the bits in the twis registers. this is intended to help wr iting drivers for the twis. table 22-5. bus events event effect slave transmitter has sent a data byte sr.thr is cleared. sr.btf is set. the value of the ack bit sent immediately after the data byte is given by cr.ack. slave receiver has received a data byte sr.rhr is set. sr.btf is set. sr.nak updated according to value of ack bit received from master. start+sadr on bus, but address is to another slave none. start+sadr on bus, current slave is addressed, but address match enable bit in cr is not set none. start+sadr on bus, current slave is addressed, corresponding address match enable bit in cr set correct address match bit in sr is set. sr.tra updated according to transfer direction (updating is done one clk_twis cycle af ter address match bit is set) slave enters appropriate transfer direction mode and data transfer can commence.
515 32145c?06/2013 at32uc3l0128/256 start+sadr on bus, current slave is addressed, corresponding address match enable bit in cr set, sr.stren and sr.soam are set. correct address match bit in sr is set. sr.tra updated according to transfer direction (updating is done one clk_twis cycle af ter address match bit is set). slave stretches twck immediately after transmitting the address ack bit. twck remains stretched until all address match bits in sr have been cleared. slave enters appropriate transfer direction mode and data transfer can commence. repeated start received after being addressed sr.rep set. sr.tcomp unchanged. stop received after being addressed sr.sto set. sr.tcomp set. start, repeated start, or stop received in illegal position on bus sr.buserr set. sr.sto and sr.tcomp may or may not be set depending on the exact position of an illegal stop. data is to be received in slave receiver mode, sr.stren is set, and rhr is full twck is stretched until rhr has been read. data is to be transmitted in slave receiver mode, sr.stren is set, and thr is empty twck is stretched until thr has been written. data is to be received in slave receiver mode, sr.stren is cleared, and rhr is full twck is not stretched, read data is discarded. sr.orun is set. data is to be transmitted in slave receiver mode, sr.stren is cleared, and thr is empty twck is not stretched, previous c ontents of thr is written to bus. sr.urun is set. smbus timeout received sr.smbtout is set. twck and twd are immediately released. slave transmitter in smbus pec mode has transmitted a pec byte, that was not identical to the pec calculated by the master receiver. master receiver will transmit a nak as usual after the last byte of a master receiver transfer. master receiver will retry th e transfer at a later time. slave receiver discovers smbus pec error sr.smbpecerr is set. nak returned after the data byte. table 22-5. bus events event effect
516 32145c?06/2013 at32uc3l0128/256 22.9 user interface note: 1. the reset values for these registers are device specific. please refer to the module configuration section at the end of this chapter. table 22-6. twis register memory map offset register register name access reset 0x00 control register cr read/write 0x00000000 0x04 nbytes register nbytes read/write 0x00000000 0x08 timing register tr read/write 0x00000000 0x0c receive holding register rhr read-only 0x00000000 0x10 transmit holding register thr write-only 0x00000000 0x14 packet error check register pecr read-only 0x00000000 0x18 status register sr read-only 0x00000002 0x1c interrupt enable register ier write-only 0x00000000 0x20 interrupt disable register idr write-only 0x00000000 0x24 interrupt mask register imr read-only 0x00000000 0x28 status clear register scr write-only 0x00000000 0x2c parameter register pr read-only - (1) 0x30 version register vr read-only - (1)
517 32145c?06/2013 at32uc3l0128/256 22.9.1 control register name: cr access type: read/write offset :0x00 reset value: 0x00000000 ? tenbit: ten bit address match 0: disables ten bit address match. 1: enables ten bit address match. ? adr: slave address slave address used in slave address match. bits 9: 0 are used if in 10-bit mode, bits 6:0 otherwise. ? sodr: stretch clock on data byte reception 0: does not stretch bus clock immediately before acking a received data byte. 1: stretches bus clock immediately before acking a received data byte. ? soam: stretch clock on address match 0: does not stretch bus clock after address match. 1: stretches bus clock after address match. ? cup: nbytes count up 0: causes nbytes to count down (decrement) per byte transferred. 1: causes nbytes to count up (increment) per byte transferred. ? ack: slave receiver data phase ack value 0: causes a low value to be returned in the ack cycle of the data phase in slave receiver mode. 1: causes a high value to be returned in the ack cycle of the data phase in slave receiver mode. ? pecen: packet error checking enable 0: disables smbus pec (crc) generation and check. 1: enables smbus pec (crc) generation and check. ? smhh: smbus host header 0: causes the twis not to ackn owledge the smbus host header. 1: causes the twis to acknowledge the smbus host header. ? smda: smbus default address 0: causes the twis not to ackn owledge the smbus default address. 1: causes the twis to acknowledge the smbus default address. ? smbalert: smbus alert 0: causes the twis to release the smbalert line and no t to acknowledge the smbus alert response address (ara). 1: causes the twis to pull down the smbalert line and to acknowledge the smbus alert response address (ara). 31 30 29 28 27 26 25 24 -----tenbit adr[9:8] 23 22 21 20 19 18 17 16 adr[7:0] 15 14 13 12 11 10 9 8 sodr soam cup ack pecen smhh smda smbalert 76543210 swrst - - stren gcmatch smatch smen sen
518 32145c?06/2013 at32uc3l0128/256 ? swrst: software reset this bit will always read as 0. writing a zero to this bit has no effect. writing a one to this bit resets the twis. ? stren: clock stretch enable 0: disables clock stretching if rhr/thr bu ffer full/empty. may cause over/underrun. 1: enables clock stretching if rhr/thr buffer full/empty. ? gcmatch: general call address match 0: causes the twis not to ackn owledge the general call address. 1: causes the twis to ackno wledge the general call address. ? smatch: slave address match 0: causes the twis not to acknowledge the slave address. 1: causes the twis to acknowledge the slave address. ? smen: smbus mode enable 0: disables smbus mode. 1: enables smbus mode. ? sen: slave enable 0: disables the slave interface. 1: enables the slave interface.
519 32145c?06/2013 at32uc3l0128/256 22.9.2 nbytes register name: nbytes access type: read/write offset :0x04 reset value: 0x00000000 ? nbytes: number of bytes to transfer writing to this field updates the nbytes counter. the field can also be read to learn the progress of the transfer. nbytes can be incremented or decremented automatically by hardware. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 nbytes
520 32145c?06/2013 at32uc3l0128/256 22.9.3 timing register name: tr access type: read/write offset :0x08 reset value: 0x00000000 ? exp: clock prescaler used to specify how to prescale the smbus tlows counter. the counter is prescaled according to the following formula: ? sudat: data setup cycles non-prescaled clock cycles for data setup count. used to time t su_dat . data is driven sudat cycles after twck low detected. this timing is used for timing the ack/nak bits, and any data bits driven in slave transmitter mode. ?ttout: smbus t timeout cycles prescaled clock cycles used to time smbus t timeout . ?tlows: smbus t low:sext cycles prescaled clock cycles used to time smbus t low:sext . 31 30 29 28 27 26 25 24 exp ---- 23 22 21 20 19 18 17 16 sudat 15 14 13 12 11 10 9 8 ttout 76543210 tlows f prescaled f clk_twis 2 exp 1 + ?? ------------------------ - =
521 32145c?06/2013 at32uc3l0128/256 22.9.4 receive holding register name: rhr access type: read-only offset :0x0c reset value: 0x00000000 ? rxdata: received data byte when the rxrdy bit in the status regist er (sr) is one, this field contains a byte received from the twi bus. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 rxdata
522 32145c?06/2013 at32uc3l0128/256 22.9.5 transmit holding register name: thr access type: write-only offset :0x10 reset value: 0x00000000 ? txdata: data byte to transmit write data to be transf erred on the twi bus here. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 txdata
523 32145c?06/2013 at32uc3l0128/256 22.9.6 packet error check register name: pecr access type: read-only offset :0x14 reset value: 0x00000000 ? pec: calculated pec value the calculated pec value. updated automatical ly by hardware after each byte has been transferred. reset by hardware after a stop condition. provided if the user manu ally wishes to control when the pec byte is transmitted, or wishes to access the pec value for other reasons. in ordinary operation, the pec handling is done automatically by hardware. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 pec
524 32145c?06/2013 at32uc3l0128/256 22.9.7 status register name: sr access type: read-only offset :0x18 reset value: 0x000000002 ? btf: byte transfer finished this bit is cleared when the corresponding bit in scr is written to one. this bit is set when byte transfer has completed. ? rep: repeated start received this bit is cleared when the corresponding bit in scr is written to one. this bit is set when a repeate d start condition is received. ? sto: stop received this bit is cleared when the corresponding bit in scr is written to one. this bit is set when the stop condition is received. ? smbdam: smbus default address match this bit is cleared when the corresponding bit in scr is written to one. this bit is set when the received address matched the smbus default address. ? smbhhm: smbus host header address match this bit is cleared when the corresponding bit in scr is written to one. this bit is set when the received address matched the smbus host header address. ? smbalertm: smbus alert response address match this bit is cleared when the corresponding bit in scr is written to one. this bit is set when the received address matched the smbus alert response address. ? gcm: general call match this bit is cleared when the corresponding bit in scr is written to one. this bit is set when the received address matched the general call address. ? sam: slave address match this bit is cleared when the corresponding bit in scr is written to one. this bit is set when the received address matched the slave address. ? buserr: bus error this bit is cleared when the corresponding bit in scr is written to one. this bit is set when a misplaced start or stop condition has occurred. 31 30 29 28 27 26 25 24 -- - ----- 23 22 21 20 19 18 17 16 btf rep sto smbdam smbhhm smbalertm gcm sam 15 14 13 12 11 10 9 8 - buserr smbpecerr smbtout - - - nak 76 5 43210 orun urun tra - tcomp sen txrdy rxrdy
525 32145c?06/2013 at32uc3l0128/256 ? smbpecerr: smbus pec error this bit is cleared when the corresponding bit in scr is written to one. this bit is set when a smbus pec error has occurred. ? smbtout: smbus timeout this bit is cleared when the corresponding bit in scr is written to one. this bit is set when a smbus timeout has occurred. ? nak: nak received this bit is cleared when the corresponding bit in scr is written to one. this bit is set when a nak was received from the master during slave transmitter operation. ? orun: overrun this bit is cleared when the corresponding bit in scr is written to one. this bit is set when an overrun has occurred in slave receiver mode. can only occur if cr.stren is zero. ? urun: underrun this bit is cleared when the corresponding bit in scr is written to one. this bit is set when an underrun has occurred in slave transmitter mode. can only occur if cr.stren is zero. ? tra: transmitter mode 0: the slave is in slave receiver mode. 1: the slave is in slave transmitter mode. ? tcomp: transmis sion complete this bit is cleared when the corresponding bit in scr is written to one. this bit is set when transmission is complete. set after receiving a stop after being addressed. ? sen: slave enabled 0: the slave interface is disabled. 1: the slave interface is enabled. ? txrdy: tx buffer ready 0: the tx buffer is full and should not be written to. 1: the tx buffer is empty, and can accept new data. ? rxrdy: rx buffer ready 0: no rx data ready in rhr. 1: rx data is ready to be read from rhr.
526 32145c?06/2013 at32uc3l0128/256 22.9.8 interrupt enable register name: ier access type: write-only offset :0x1c reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will write a one to the corresponding bit in imr. 31 30 29 28 27 26 25 24 -- - ----- 23 22 21 20 19 18 17 16 btf rep sto smbdam smbhhm smbalertm gcm sam 15 14 13 12 11 10 9 8 - buserr smbpecerr smbtout - - - nak 76 5 43210 orun urun - - tcomp - txrdy rxrdy
527 32145c?06/2013 at32uc3l0128/256 22.9.9 interrupt disable register name: idr access type: write-only offset :0x20 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in imr. 31 30 29 28 27 26 25 24 -- - ----- 23 22 21 20 19 18 17 16 btf rep sto smbdam smbhhm smbalertm gcm sam 15 14 13 12 11 10 9 8 - buserr smbpecerr smbtout - - - nak 76 5 43210 orun urun - - tcomp - txrdy rxrdy
528 32145c?06/2013 at32uc3l0128/256 22.9.10 interrupt mask register name: imr access type: read-only offset :0x24 reset value: 0x00000000 0: the corresponding interrupt is disabled. 1: the corresponding interrupt is enabled. this bit is cleared when the corresponding bit in idr is written to one. this bit is set when the corresponding bit in ier is written to one. 31 30 29 28 27 26 25 24 -- - ----- 23 22 21 20 19 18 17 16 btf rep sto smbdam smbhhm smbalertm gcm sam 15 14 13 12 11 10 9 8 - buserr smbpecerr smbtout - - - nak 76 5 43210 orun urun - - tcomp - txrdy rxrdy
529 32145c?06/2013 at32uc3l0128/256 22.9.11 status clear register name: scr access type: write-only offset :0x28 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in sr and the corresponding interrupt request. 31 30 29 28 27 26 25 24 -- - ----- 23 22 21 20 19 18 17 16 btf rep sto smbdam smbhhm smbalertm gcm sam 15 14 13 12 11 10 9 8 - buserr smbpecerr smbtout - - - nak 76 5 43210 orun urun - - tcomp - - -
530 32145c?06/2013 at32uc3l0128/256 22.9.12 parameter register name: pr access type: read-only offset :0x2c reset value: - 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 --------
531 32145c?06/2013 at32uc3l0128/256 22.9.13 version register (vr) name: vr access type: read-only offset: 0x30 reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 ---- version [11:8] 76543210 version [7:0]
532 32145c?06/2013 at32uc3l0128/256 22.10 module configuration the specific configuration for each twis instance is listed in the following tables.the module bus clocks listed here are connected to the syst em bus clocks. please refer to the power man- ager chapter for details. table 22-7. module clock name module name clock name description twis0 clk_twis0 clock for the twis0 bus interface twis1 clk_twis1 clock for the twis1 bus interface table 22-8. register reset values register reset value version 0x00000120 parameter 0x00000000
533 32145c?06/2013 at32uc3l0128/256 23. pulse width modulation controller (pwma) rev: 2.0.1.0 23.1 features ? left-aligned non-inverted 8-bit pwm ? common 8-bit timebase counter ? asynchronous clock source supported ? spread-spectrum counter to allow a constantly varying duty cycle ? separate 8-bit duty cycle register per channel ? synchronized channel updates ? no glitches when changing the duty cycles ? interlinked operation supported ? up to 32 channels can be updated with the same duty cycle value at a time ? up to 4 channels can be updated with different duty cycle values at a time ? interrupt on pwm timebase overflow ? incoming peripheral events supported ? pre-defined channels support incoming (i ncrease/decrease) peripheral events from the peripheral event system ? incoming increase/decrease event can either increase or decrease the duty cycle by one ? one output peripheral event supported ? connected to channel 0 and ass erted when the common timeb ase counter is equal to the programmed duty cycle for channel 0 ? output pwm waveforms ? support normal waveform output for each channel ? support composite waveform generati on (xor?ed) for each pair channels ? open drain driving on selected pins for 5v pwm operation 23.2 overview the pulse width modulation controller (pwma) controls several pulse width modulation (pwm) channels. the number of channels is specific to the device. each channel controls one square output pwm waveform. characteristics of the output pwm waveforms such as period and duty cycle are configured through the user interface. all user interface registers are mapped on the peripheral bus. the duty cycle value for each channel can be set independently, while the period is determined by a common timebase counter (tc). the timebase for the counter is selected by using the allo- cated asynchronous generic clock (gclk). the user interface for the pwma contains handshake and synchronizing logic to ensure that no glitches occur on the output pwm wave- forms while changing the duty cycle values. pwma duty cycle values can be changed using two approaches, either an interlinked single- value mode or an interlinked multi-value mode. in the interlinked single-value mode, any set of channels, up to 32 channels, can be updated simultaneously with the same value while the other channels remain unchanged. there is also an inte rlinked multi-value mode, where the 8 least significant bits of up to 4 ch annels can be updated with 4 different values while the other chan- nels remain unchanged. some pins can be driven in open drain mode, allowing the pwma to generate a 5v waveform using an external pullup resistor.
534 32145c?06/2013 at32uc3l0128/256 23.3 block diagram figure 23-1. pwma block diagram 23.4 i/o lines description each channel outputs one pwm waveform on one external i/o line. 23.5 product dependencies in order to use this module, other parts of the system must be configured correctly, as described below. pwm blocks channel m channel 1 channel 0 duty cycle register comp pwma[m:0] interrupt handling irq pb top timebase counter spread adjust tofl ready channel_0 clk_pwma gclk domain pb clock domain spread spectrum counter sync gclk etv control duty cycle channel select wavexor cwg cherr tclr table 23-1. i/o line description pin name pin description type pwma[n] output pwm waveform for one channel n output pwmmod[n] output pwm waveform for one channel n, open drain mode output
535 32145c?06/2013 at32uc3l0128/256 23.5.1 i/o lines the pins used for interfacing the pwma may be multiplexed with i/o controller lines. the pro- grammer must first program the i/o controller to assign the desired pwma pins to their peripheral function. it is only required to enable the pwma outputs actually in use. 23.5.2 power management if the cpu enters a sleep mode that disables clocks used by the pwma, the pwma will stop functioning and resume operation after the system wakes up from sleep mode. 23.5.3 clocks the clock for the pwma bus interface (clk_pwma) is controlled by the power manager. this clock is enabled at reset, and can be disabled in the power manager. it is recommended to dis- able the pwma before disabling the clock, to avoid freezing the pwma in an undefined state. additionally, the pwma depends on a dedicated ge neric clock (gclk). the gclk can be set to a wide range of frequencies and clock sources and must be enabled in the system control interface (scif) before the pwma can be used. 23.5.4 interrupts the pwma interrupt request lines are connected to the interrupt controller. using the pwma interrupts requires the interrupt controller to be programmed first. 23.5.5 peripheral events the pwma peripheral events are connected via the peripheral event system. refer to the peripheral event system chapter for details. 23.5.6 debug operation when an external debugger forces the cpu into debug mode, the pwma continues normal operation. if the pwma is configured in a way t hat requires it to be periodically serviced by the cpu through interrupts, improper operation or data loss may result during debugging. 23.6 functional description the pwma embeds a number of pwm channel submodules, each providing an output pwm waveform. each pwm channel contains a duty cycle register and a comparator. a common timebase counter for all channels determines the frequency and the period for all the pwm waveforms. 23.6.1 enabling the pwma once the gclk has been enabled, the pwma is enabled by writing a one to the en bit in the control register (cr). 23.6.2 timebase counter the top value of the timebase counter defines the period of the pwma output waveform. the timebase counter starts at zero when the pwma is enabled and counts upwards until it reaches its effective top value (etv). the effective top value is defined by specifying the desired number of gclk clock cycles in the top field of top value register (tvr.top) in normal operation (the
536 32145c?06/2013 at32uc3l0128/256 spread field of cr (c r.spread) is zero). when the timebase counter reaches its effective top value, it restarts counting from zero. the period of the pwma output waveform is then: the timebase counter can be reset by writing a one to the timebase clear bit in cr (cr.tclr). note that this can cause a glitch to the output pwm waveforms in use. 23.6.3 spread spectrum counter the spread spectrum counter allows the generati on of constantly varying duty cycles on the out- put pwm waveforms. this is achieved by varying the effective top value of the timebase counter in a range defined by the spread spectrum counter value. when cr.spread is not zero, the spread spectrum counter is enabled. its range is defined by cr.spread. it starts to count from -cr.spr ead when the pwma is enabled or after reset and counts upwards. when it reaches cr.spread, it restarts to count from -cr.spread again. the spread spectrum counter will cause the effective top value to vary from top- spread to top+spread. figure 23-2 on page 536 illustrates this. this leads to a constantly varying duty cycle on the pwm output waveforms though the duty cycle values stored are unchanged. figure 23-2. pwma adjusting top value for timebase counter 23.6.3.1 special considerations the maximum value of the timeba se counter is 0xff. if spread is written to a value that will cause the etv to exceed this va lue, the spread spectrum counte r?s range will be limited to pre- vent the timebase counter to exceed its maximum value. if spread is written to a value causing (top- spread) to be below zero, the spread spectrum counter?s range will be limited to prevent the timebase c ounter to coun t below zero. in both cases, the spread valu e read from the control register will be the same value as writ- ten to the spread field. t pwma etv 1 + ?? t gclk ? = 0x0 0x0fff duty cycle -spread spread top adjusting top value range for the timerbase counter
537 32145c?06/2013 at32uc3l0128/256 when writing a one to cr.tclr, the timebase c ounter and the spread spectrum counter are reset at their lower limit values and the effective top value of the timebase counter will also be reset. 23.6.4 duty cycle and waveform properties each pwm channel has its own duty cycle value (dcv) which is write-only and cannot be read out. the duty cycle value can be changed in two approaches as described in section23.6.6 . when the duty cycle value is zero, the pwm output is zero. otherwise, the pwm output is set when the timebase counter is zero, and cleared when the timebase counter reaches the duty cycle value. this is summarized as: note that when increasing the duty cycle value for one channel from 0 to 1, the number of gclk cycles when the pwm waveform is high will jump from 0 to 2. when incrementing the duty cycle value by one for any other values, the number of gclk cycle when the waveform is high will increase by one. this is summarized in table 23-2 . 23.6.5 waveform output pwma waveforms are output to i/o lines. the output waveform properties are controlled by composite waveform generation (cwg) register(s). if this register is cleared (by default), the channel waveforms are out directly to the i/o line s. to avoid too many i/o toggling simultane- ously on the output i/o lines, every other output pwm waveform toggles on the negative edge of the gclk instead of the positive edge. in cwg mode, all channels are paired and their outputs are xor?ed together if the correspond- ing bit of cwg register is set. the even number of output is the xor?ed output and the odd number of output is the inverse of its. each bit of cwg register controls one pair channels and the least significant bit refers to the lowest number of pair channels. 23.6.6 updating duty cycle values 23.6.6.1 interlinked single value pwm operation the pwm channels can be interlinked to allow multiple channels to be updated simultaneously with the same duty cycle value. this value must be written to the interlinked single value duty table 23-2. pmw waveform duty cycles duty cycle value #clock cycles when waveform is high #clock cycles when waveform is low 00 etv+1 12 etv-1 23 etv-2 ... ... ... etv-1 etv 1 etv etv+1 0 pwm waveform = low when dcv 0 or tc dcv ? = high when tc dcv and dcv 0 ? ? ? ? ?
538 32145c?06/2013 at32uc3l0128/256 (isduty) register. each channel has a corresponding enabling bit in the interlinked single value channel set (ischset) register(s). when a bit is written to one in the ischset register, the duty cycle register for the corresponding channel will be updated with the value stored in the isduty register. it can only be updated when t he ready bit in the status register (sr.ready) is one, indicating that the pwma is ready for writing. figure 23-3 on page 538 shows the writing procedure. it is thus possible to update the duty cycle values for up to 32 pwm channels within one ischset register at a time. figure 23-3. interlinked single value pwm operation flow 23.6.6.2 interlinked multip le value pwm operation the interlinked multip le value pwm operation allows up to four channels to be updated simulta- neously with different duty cycle values. the four duty cycle values are required to be written to the four fields imduty.duty3, imduty .duty2, imduty.duty1 and imduty.duty0, respectively. the index number of the four channels to be updated is written to the four sel fields in the interlinked multiple value channel select (imchsel) register (imchsel.sel). when the imchsel register is written, the values stored in imduty register are synchronized to the duty cycle registers for the channels selected by the sel fields. figure 23-4 on page 538 shows the writing procedure. note that only writes to the implemented channels will be ef fective. if one of the imchsel.sel fields points to a non-existing channel, the corresponding valu e in the imduty.dutyx field will not be written. if the same channel is specifi ed multiple times in the imchsel.sel fields, the channel will be update d with the value referred by the upper imchsel.sel field. figure 23-4. interlinked multiple value pwm operation flow ischsetm ... write enable channeln duty channel1 duty channel0 duty isduty channel2 duty duty3/2/1/0 imduty imchsel channeln duty ... mux channel1 duty channel0 duty
539 32145c?06/2013 at32uc3l0128/256 23.6.7 open drain mode some pins can be used in open drain mode, allowing the pwma waveform to toggle between 0v and up to 5v on these pins. in this mode the pwma will drive th e pin to zero or leave the out- put open. an external pullup can be used to pull the pin up to the desired voltage. to enable open drain mode on a pin the pwmaod function must be selected instead of the pwma function in the i/o controller. please refer to the module configuration chapter for infor- mation about which pins are available in open drain mode. 23.6.8 synchronization both the timebase counter and the spread spectrum counter can be reset and the duty cycle registers can be written through the user interface of the module. this requires a synchroniza- tion between the pb and gclk clock domains, wh ich takes a few clock cycles of each clock domain. the busy bit in sr indicates when the synchronization is ongoing. writing to the mod- ule while the busy bit is set will re sult in discarding the new value. note that the duty cycle regi sters will not be updated with the new values until the timebase counter reaches its top value, in order to avoid glitches. the busy bit in sr will always be set during this updating and synchronization period. 23.6.9 interrupts when the timebase counter overflows, the tim ebase overflow bit in the status register (sr.tofl) is set. if the corresponding bit in the interrupt mask register (imr) is set, an interrupt request will be generated. since the user needs to wait until the user inte rface is available between each write due to syn- chronization, a ready bit is provided in sr, which can be used to generate an interrupt request. the interrupt request will be generated if th e corresponding bit in imr is set. bits in imr are set by writing a one to the corresponding bit in the interrupt enable register (ier), and cleared by writing a one to the corresponding bit in the in terrupt disable register (idr). the interrupt request remains active until the corresponding bit in sr is cleared by writing a one to the corre- sponding bit in the status clear register (scr). 23.6.10 peripheral events 23.6.10.1 input peripheral events the pre-defined channels support input peripheral events from the peripheral event system. input peripheral events must be enabled by writing a one to the corresponding bit in the channel event enable registers (cheers) before peripheral events can be used to control the duty cycle value. each bit in the register correspo nds to one channel, wh ere bit 0 corresponds to channel 0 and so on. both the increase and decrease events are enabled for the corresponding channel when a bit in the cheer register is set. an increase or decrease event (event_incr/event_de cr) can either increase or decrease the duty cycle value by one in a pwm period. the events are taken into account when the common time- base counter reaches its top. the behavior is defined by the channel event response register (cherr). each bit in the register corresponds to one channel, where bit 0 corresponds to chan- nel 0 and so on. if the bit in cherr is set to 0 (default) for a channel, the increase event will increase the duty cycle value and the decrease event will decrease the duty cycle value for that channel. if the bit is set to 1, the increase and decrea se event will have reve rse function so that
540 32145c?06/2013 at32uc3l0128/256 the increase event will decrease the duty cycle value and decrea se event will increase the duty cycle value. if both the increase event and the decrease event occur at the same time for a channel, the duty cycle value will not be changed. the number of channels supporting input peripheral events is device specific. please refer to the module configuration section at the end of this chapter for details. 23.6.10.2 output peripheral event the pwma also supports one output peripheral event (event_ch0) to the peripheral event sys- tem. this output peripheral event is connec ted to channel 0 and will be asserted when the timebase counter reaches the duty cycle valu e for channel 0. this output event is always enabled.
541 32145c?06/2013 at32uc3l0128/256 23.7 user interface note: 1. the reset values are device specific. please refer to the module configuration section at the end of this chapter. table 23-3. pwma register memory map offset register register name access reset 0x00 control register cr read/write 0x00000000 0x04 interlinked single value duty register isduty write-only 0x00000000 0x08 interlinked multiple value duty register imduty write-only 0x00000000 0x0c interlinked multiple value channel select imchsel write-only 0x00000000 0x10 interrupt enable register ier write-only 0x00000000 0x14 interrupt disable register idr write-only 0x00000000 0x18 interrupt mask register imr read-only 0x00000000 0x1c status register sr read-only 0x00000000 0x20 status clear register scr write-only 0x00000000 0x24 parameter register parameter read-only - (1) 0x28 version register version read-only - (1) 0x2c top value register tvr read/write 0x00000000 0x30+m*0x10 interlinked single value channel set m ischsetm write-only 0x00000000 0x34+m*0x10 channel event response register m cherrm read/write 0x00000000 0x38+m*0x10 channel event enable register m cheerm read/write 0x00000000 0x3c+k*0x10 cwg register cwgk read/write 0x00000000
542 32145c?06/2013 at32uc3l0128/256 23.7.1 control register name: cr access type: read/write offset :0x00 reset value: 0x00000000 ? spread: spread spectrum limit value the spread spectrum limit value, together with the top field, defines the range for the spread spectrum counter. it is introduc ed in order to achieve constant varying duty cycl es on the output pwm waveforms. refer to section23.6.3 for more information. ? top: timebase counter top value the top value for the timebase counter. the value written to this field will update the 8 bits of the tvr.top field. in fact, cr.top and tvr.top are just equavilent. when the tvr.top field is written, this cr.top field will also be updated with tvr.top field. ? tclr: timebase clear writing a zero to this bit has no effect. writing a one to this bit will clear the timebase counter. this bit is always read as zero. ? en: module enable 0: the pwma is disabled 1: the pwma is enabled 31 30 29 28 27 26 25 24 -------reserved 23 22 21 20 19 18 17 16 reserved spread[4:0] 15 14 13 12 11 10 9 8 top 76543210 ------tclren
543 32145c?06/2013 at32uc3l0128/256 23.7.2 interlinked single value duty register name: isduty access type: write-only offset :0x04 reset value: 0x00000000 ? duty: duty cycle value the duty cycle value written to this field is written simultaneously to all channels selected in the ischsetm register. if the value zero is written to duty all affected channels will be disabled. in this state the output waveform will be zero all the time. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 ---- reserved 76543210 duty[7:0]
544 32145c?06/2013 at32uc3l0128/256 23.7.3 interlinked multiple value duty register name: imduty access type: write-only offset :0x08 reset value: 0x00000000 ? dutyn: duty cycle the value written to duty field n will be updated for the selected channels. which channel is selected for updating is defined by the corresponding sel field in the imchsel register. if the value zero is written to duty all affected channels will be disabled. in this state the output waveform will be zero all the time. 31 30 29 28 27 26 25 24 duty3 23 22 21 20 19 18 17 16 duty2 15 14 13 12 11 10 9 8 duty1 76543210 duty0
545 32145c?06/2013 at32uc3l0128/256 23.7.4 interlinked multiple value channel select name: imchsel access type: write-only offset :0x0c reset value: 0x00000000 ? seln: channel select the duty cycle of the pwma channel seln will be updated with the value stored in t he imduty.dutyn field when imchsel is written. if seln points to a non-implem ented channel, the write will be discarded. note: the duty registers will be updated with the value stored in the imduty.duty3, imdut y.duty2, imduty.duty1 and imduty.duty0 fields when the imchsel re gister is written. synchronization ta kes place immediately when an imchsel reg- ister is written. the duty cycle registers will, however, no t be updated until the synchronization is completed and the timebas e counter reaches its top value in order to avoid glitches. 31 30 29 28 27 26 25 24 sel3 23 22 21 20 19 18 17 16 sel2 15 14 13 12 11 10 9 8 sel1 76543210 sel0
546 32145c?06/2013 at32uc3l0128/256 23.7.5 interrupt enable register name: ier access type: write-only offset :0x10 reset value: 0x00000000 writing a zero to a bit in this register has no effect writing a one to a bit in this register will set the corresponding bit in imr. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 -----ready-tofl
547 32145c?06/2013 at32uc3l0128/256 23.7.6 interrupt disable register name: idr access type: write-only offset :0x14 reset value: 0x00000000 writing a zero to a bit in this register has no effect writing a one to a bit in this register will clear the corresponding bit in imr. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 -----ready-tofl
548 32145c?06/2013 at32uc3l0128/256 23.7.7 interrupt mask register name: imr access type: read-only offset :0x18 reset value: 0x00000000 0: the corresponding interrupt is disabled. 1: the corresponding interrupt is enabled. a bit in this register is cleared when the corresponding bit in idr is written to one. a bit in this register is set when the corresponding bit in ier is written to one. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 -----ready-tofl
549 32145c?06/2013 at32uc3l0128/256 23.7.8 status register name: sr access type: read-only offset :0x1c reset value: 0x00000000 ? busy: interface busy this bit is automatically cleared when the interface is no longer busy. this bit is set when the user interface is busy and will not respond to new write operations. ? ready: interface ready this bit is cleared by writing a one to the corresponding bit in the scr register. this bit is set when the busy bit has a 1-to-0 transition. ? tofl: timebase overflow this bit is cleared by writing a one to corresponding bit in the scr register. this bit is set when the timebase counter has wrapped at its top value. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 - - - - busy ready - tofl
550 32145c?06/2013 at32uc3l0128/256 23.7.9 status clear register name: scr access type: write-only offset :0x20 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in sr and the corresponding interrupt request. this register always reads as zero. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 1 10 9 8 -------- 76543210 -----ready-tofl
551 32145c?06/2013 at32uc3l0128/256 23.7.10 parameter register name: parameter access type: read-only offset :0x24 reset value: - ? channels: channels implemented this field contains the number of channels implemented on the device. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 channels
552 32145c?06/2013 at32uc3l0128/256 23.7.11 version register name: version access type: read-only offset :0x28 reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 ---- version[11:8] 76543210 version[7:0]
553 32145c?06/2013 at32uc3l0128/256 23.7.12 top value register name: tvr access type: read/write offset :0x2c reset value: 0x00000000 ? top: timebase counter top value the top value for the timebase counter. the value written to t he cr.top field will automatically be written tothis field. when this register is written, it will also automatically update the cr.top field . the effective top value of the tim ebase counter is defined by both tvr.top and the cr.spread. refer to section23.6.2 for more information. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 ---- reserved 76543210 top[7:0]
554 32145c?06/2013 at32uc3l0128/256 23.7.13 interlinked single value channel set name: ischsetm access type: write-only offset : 0x30+m*0x10 reset value: 0x00000000 ? set: single value channel set if the bit n in set is one, the duty cycle of pwma channel n will be updated with the va lue written to isduty. if more than one ischset register is present, ischset0 contro ls channels 31 to 0 and ischset1 controls channels 63 to 32. note: the duty registers will be updated with the value stored in the isduty register when any ischsetm register is written. syn - chronization takes place immediately when an ischset register is written. the duty cycle r egisters will, however, not be updated until the synchronization is completed and the timebase counter reaches its top value in order to avoid glitches. 31 30 29 28 27 26 25 24 set 23 22 21 20 19 18 17 16 set 15 14 13 12 11 10 9 8 set 76543210 set
555 32145c?06/2013 at32uc3l0128/256 23.7.14 channel event response register name: cherrm access type: read/write offset : 0x34+m*0x10 reset value: 0x00000000 ? cher: channel event response 0: the increase event will increase the duty cycle value by one in a pwm period for the corresponding channel and the decrease event will decrease the duty cycle value by one. 1: the increase event will decrease the duty cycle value by one in a pwm period for the corresponding channel and the decrease event will increase the duty cycle value by one. the events are taken into account when the common timebase counter reaches its top. if more than one cherr register is present, cherr0 controls channels 31-0 and cherr1 controls channels 64-32 and so on. 31 30 29 28 27 26 25 24 cher 23 22 21 20 19 18 17 16 cher 15 14 13 12 11 10 9 8 cher 76543210 cher
556 32145c?06/2013 at32uc3l0128/256 23.7.15 channel event enable register name: cheerm access type: read/write offset : 0x38+m*0x10 reset value: 0x00000000 ? chee: channel event enable 0: the input peripheral event for the corresponding channel is disabled. 1: the input peripheral event for the corresponding channel is enabled. both increase and decrease events for channel n are enabled if bit n is one. if more than one cheer register is present, cheer0 controls channels 31-0 and cheer1 controls channels 64-32 and so on. 31 30 29 28 27 26 25 24 chee 23 22 21 20 19 18 17 16 chee 15 14 13 12 11 10 9 8 chee 76543210 chee
557 32145c?06/2013 at32uc3l0128/256 23.7.16 composite waveform generation name: cwg access type: read/write offset : 0x3c+k*0x10 reset value: 0x00000000 ? xor: pair waveform xor?ed if the bit n in xor field is one, the pair of pwma output waveform s will be xored before output. the even number output will be the xor?ed output and the odd number output will be reverse of it. for example, if bit 0 in xor is one, the pair of pwma output waveforms for channel 0 and 1 will be xored together. if bit n in xor is zero, normal waveforms are output for that pair. note that if more than one cwg register is present, cwg0 controls the first 32 pairs, corresponding to channels 63 downto 0, and cwg1 controls the second 32 pairs, corresponding to channels 127 downto 64. 31 30 29 28 27 26 25 24 xor 23 22 21 20 19 18 17 16 xor 15 14 13 12 11 10 9 8 xor 76543210 xor
558 32145c?06/2013 at32uc3l0128/256 23.8 module configuration the specific configuration for each pwma instan ce is listed in the following tables. the module bus clocks listed here are connected to the syst em bus clocks. please refer to the power man- ager chapter for details. table 23-4. pwma configuration feature pwma number of pwm channels 36 channels supporting incoming peripheral events 0, 6, 8, 9, 11, 14, 19, and 20 pwma channels with open drain mode 21, 27, and 28 table 23-5. pwma clocks clock name descripton clk_pwma clock for the pwma bus interface gclk the generic clock used for the pwma is gclk3 table 23-6. register reset values register reset value version 0x00000201 parameter 0x00000024
559 32145c?06/2013 at32uc3l0128/256 24. timer/counter (tc) rev: 2.2.3.1.3 24.1 features ? three 16-bit timer counter channels ? a wide range of functions including: ? frequency measurement ? event counting ? interval measurement ? pulse generation ?delay timing ? pulse width modulation ? up/down capabilities ? each channel is user-conf igurable and contains: ? three external clock inputs ? five internal clock inputs ? two multi-purpose input/output signals ? internal interrupt signal ? two global registers that act on all three tc channels ? peripheral event input on all a lines in capture mode 24.2 overview the timer counter (tc) includes three identical 16-bit timer counter channels. each channel can be independently programmed to perform a wide range of functions including frequency measurement, event counting, interval measurement, pulse generation, delay timing, and pulse width modulation. each channel has three external clock inputs, fi ve internal clock inputs, and two multi-purpose input/output signals which can be configured by the user. each channel drives an internal inter- rupt signal which can be programmed to generate processor interrupts. the tc block has two global registers which act upon all three tc channels. the block control register (bcr) allows the th ree channels to be started simultaneously with the same instruction. the block mode register (bmr) defines the ex ternal clock inputs for each channel, allowing them to be chained.
560 32145c?06/2013 at32uc3l0128/256 24.3 block diagram figure 24-1. tc block diagram 24.4 i/o lines description 24.5 product dependencies in order to use this module, other parts of the system must be configured correctly, as described below. 24.5.1 i/o lines the pins used for interfacing the compliant external devices may be multiplexed with i/o lines. the user must first program the i/o controller to assign the tc pins to their peripheral functions. i/o contr oller tc2xc2s int0 int1 int2 tioa0 tioa1 tioa2 tiob0 tiob1 tiob2 xc2 tclk0 tclk1 tclk2 tclk0 tclk1 tclk2 tclk0 tclk1 tclk2 tioa1 tioa2 tioa0 tioa2 tioa1 interrupt controller clk0 clk1 clk2 a0 b0 a1 b1 a2 b2 timer count er tiob tioa tiob sync timer_clock1 tioa sync sync tioa tiob timer_clock2 timer_clock3 timer_clock4 timer_clock5 xc1 xc0 xc0 xc2 xc1 xc0 xc1 xc2 timer/counter channel 2 timer/counter channel 1 timer/counter channel 0 tc1xc1s tc0xc0s tioa0 table 24-1. i/o lines description pin name description type clk0-clk2 external clock input input a0-a2 i/o line a input/output b0-b2 i/o line b input/output
561 32145c?06/2013 at32uc3l0128/256 when using the tioa lines as inputs the user must make sure that no peripheral events are gen- erated on the line. refer to the peripheral event system chapter for details. 24.5.2 power management if the cpu enters a sleep mode that disables cl ocks used by the tc, the tc will stop functioning and resume operation after the system wakes up from sleep mode. 24.5.3 clocks the clock for the tc bus interface (clk_tc) is generated by the power manager. this clock is enabled at reset, and can be disabled in the power manager. it is recommended to disable the tc before disabling the clock, to avoid freezing the tc in an undefined state. 24.5.4 interrupts the tc interrupt request line is connected to the interrupt controller. using the tc interrupt requires the interrupt controller to be programmed first. 24.5.5 peripheral events the tc peripheral events are connected via the pe ripheral event system. refer to the periph- eral event system chapter for details. 24.5.6 debug operation the timer counter clocks are frozen du ring debug operation, unless the ocd system keeps peripherals running in debug operation. 24.6 functional description 24.6.1 tc description the three channels of the timer counter are independent and identical in operation. the regis- ters for channel programming are listed in figure 24-3 on page 576 . 24.6.1.1 channel i/o signals as described in figure 24-1 on page 560 , each channel has the following i/o signals. 24.6.1.2 16-bit counter each channel is organized around a 16-bit counter. the value of the counter is incremented at each positive edge of the selected clock. when the counter has reached the value 0xffff and passes to 0x0000, an overflow occurs and the counter overflow status bit in the channel n sta- tus register (srn.covfs) is set. table 24-2. channel i/o signals description block/channel sign al name description channel signal xc0, xc1, xc2 external clock inputs tioa capture mode: timer counter input waveform mode: timer counter output tiob capture mode: timer counter input waveform mode: timer counter input/output int interrupt signal output sync synchronization input signal
562 32145c?06/2013 at32uc3l0128/256 the current value of the counter is accessible in real time by reading the channel n counter value register (cvn). the counter can be reset by a trigger. in this case, the counter value passes to 0x0000 on the next valid edge of the selected clock. 24.6.1.3 clock selection at block level, input clock signals of each channel can either be connected to the external inputs tclk0, tclk1 or tclk2, or be connected to the configurable i/o signals a0, a1 or a2 for chaining by writing to the bmr register. see figure 24-2 on page 562 . each channel can independently select an internal or external clock source for its counter: ? internal clock signals: timer_cl ock1, timer_clock2, timer_clock3, timer_clock4, timer_clock5. see the module configuration chapter for details about the connection of these clock sources. ? external clock signals: xc0, xc1 or xc2. see the module configuration chapter for details about the connection of these clock sources. this selection is made by the clock sele ction field in the channel n mode register (cmrn.tcclks). the selected clock can be inverted with the clock invert bit in cmrn (cmrn.clki). this allows counting on the opposite edges of the clock. the burst function allows the clock to be valida ted when an external signal is high. the burst signal selection field in the cmrn regi ster (cmrn.burst) defines this signal. note: in all cases, if an external clock is used, the du ration of each of its leve ls must be longer than the clk_tc period. the external clock frequency must be at least 2.5 times lower than the clk_tc. figure 24-2. clock selection timer_clock5 xc2 tcclks clki burst 1 selected clock xc1 xc0 timer_clock4 timer_clock3 timer_clock2 timer_clock1
563 32145c?06/2013 at32uc3l0128/256 24.6.1.4 clock control the clock of each counter can be controlled in two different ways: it can be enabled/disabled and started/stopped. see figure 24-3 on page 563 . ? the clock can be enabled or disabled by the user by writing to the counter clock enable/disable command bits in the channel n clock contro l register (ccrn.clken and ccrn.clkdis). in capture mode it can be disabled by an rb load event if the counter clock disable with rb loading bit in cmrn is written to one (cmrn.ldbdis). in waveform mode, it can be disabled by an rc compare event if the counter clock disable with rc compare bit in cmrn is written to one (cmrn.cpcdis). when disabled, the start or the stop actions have no effect: only a clken command in ccrn can re-enable the clock. when the clock is enabled, the clock enabling status bit is set in srn (srn.clksta). ? the clock can also be started or stopped: a trigger (software, synchro, external or compare) always starts the clock. in capture mode the clock can be stopped by an rb load event if the counter clock stopped with rb loading bit in cmrn is written to one (cmrn.ldbstop). in waveform mode it can be stopped by an rc compare event if the counter clock stopped with rc compare bit in cmrn is written to one (cmrn.cpcstop). the start and the stop commands have effect only if the clock is enabled. figure 24-3. clock control 24.6.1.5 tc operating modes each channel can independently operate in two different modes: ? capture mode provides measurement on signals. ? waveform mode provides wave generation. the tc operating mode selection is done by writing to the wave bit in the ccrn register (ccrn.wave). in capture mode, tioa and tiob are configured as inputs. qs r s r q clksta clken clkdis stop event disable counter clock selected clock trigger event
564 32145c?06/2013 at32uc3l0128/256 in waveform mode, tioa is always configured to be an output and tiob is an output if it is not selected to be the external trigger. 24.6.1.6 trigger a trigger resets the counter and starts the counter clock. three types of triggers are common to both modes, and a fourth external trigger is available to each mode. the following triggers are common to both modes: ? software trigger: each channel has a software trigger, available by writing a one to the software trigger command bi t in ccrn (ccrn.swtrg). ? sync: each channel has a synchronization signal sync. when asserted, this signal has the same effect as a software trigger. the sync signals of all channels are asserted simultaneously by writing a one to the synchro command bit in the bcr register (bcr.sync). ? compare rc trigger: rc is implemented in each channel and can provide a trigger when the counter value matches the rc value if the rc compare trigger enable bit in cmrn (cmrn.cpctrg) is written to one. the channel can also be configured to have an external trigger. in capture mode, the external trigger signal can be selected between tioa and tiob. in waveform mode, an external event can be programmed to be one of the following si gnals: tiob, xc0, xc1, or xc2. this external event can then be programmed to perform a trigger by writing a one to the external event trig- ger enable bit in cmrn (cmrn.enetrg). if an external trigger is used, the duration of the pulses must be longer than the clk_tc period in order to be detected. regardless of the trigger used, it will be taken into account at the following active edge of the selected clock. this means that the counter value can be read differently from zero just after a trigger, especially when a low frequency signal is selected as the clock. 24.6.1.7 peripheral events on tioa inputs the tioa input lines are ored internally with peripheral events from the peripheral event sys- tem. to capture using events the user must ens ure that the corresponding pin functions for the tioa line are disabled. when capturing on the ex ternal tioa pin the user must ensure that no peripheral events are generated on this pin. 24.6.2 capture operating mode this mode is entered by writin g a zero to the cmrn.wave bit. capture mode allows the tc channel to perform measurements such as pulse timing, fre- quency, period, duty cycle and phase on tioa and tiob sig nals which are considered as inputs. figure 24-4 on page 566 shows the configuration of the tc channel when programmed in cap- ture mode. 24.6.2.1 capture registers a and b registers a and b (ra and rb) are used as capture registers. this means that they can be loaded with the counter value when a progr ammable event occurs on the signal tioa.
565 32145c?06/2013 at32uc3l0128/256 the ra loading selection field in cmrn (cmrn.ldra) defines the tioa edge for the loading of the ra register, and the rb loading selection fi eld in cmrn (cmrn.ldrb) defines the tioa edge for the loading of the rb register. ra is loaded only if it has not been loaded since the last trigger or if rb has been loaded since the last loading of ra. rb is loaded only if ra has been loaded sinc e the last trigger or t he last loading of rb. loading ra or rb before the read of the last value loaded sets the load overrun status bit in srn (srn.lovrs). in this case, the old value is overwritten. 24.6.2.2 trigger conditions in addition to the sync signal, the software trigger and the rc compare trigger, an external trig- ger can be defined. the tioa or tiob external trigger selection bit in cmrn (cmrn.abetrg) selects tioa or tiob input signal as an external trigger. the external trigger edge selection bit in cmrn (cmrn.etredg) defines the edge (rising, falling or both) detected to generate an external trig- ger. if cmrn.etrgedg is zero (none), the external trigger is disabled.
566 32145c?06/2013 at32uc3l0128/256 figure 24-4. capture mode timer_clock1 xc0 xc1 xc2 tcclks clki qs r s r q clksta clken clkdis burst tiob capture register a compare rc = 16-bit counter abetrg swtrg etrgedg cpctrg imr trig ldrbs ldras etrgs sr lovrs covfs sync 1 mti ob tioa mti oa ld r a ldbstop if ra is not loaded or rb is loaded if ra is loaded ldbdis cpcs int edge det ect o r ldrb clk ovf reset timer/counter channel edge detector edge detector capture register b register c timer_clock2 timer_clock3 timer_clock4 timer_clock5
567 32145c?06/2013 at32uc3l0128/256 24.6.3 waveform operating mode waveform operating mode is entered by writing a one to the cmrn.wave bit. in waveform operating mode the tc channel generates one or two pwm signals with the same frequency and independently programmable duty cy cles, or generates different types of one- shot or repetitive pulses. in this mode, tioa is configured as an output and tiob is defined as an output if it is not used as an external event. figure 24-5 on page 568 shows the configuration of the tc channel when programmed in waveform operating mode. 24.6.3.1 waveform selection depending on the waveform se lection field in cmrn (cmrn. wavsel), the behavior of cvn varies. with any selection, ra, rb and rc can all be used as compare registers. ra compare is used to control the tioa output, rb compare is used to control the tiob output (if correctly configured) and rc compare is used to control tioa and/or tiob outputs.
568 32145c?06/2013 at32uc3l0128/256 figure 24-5. waveform mode tcclks clki qs r s r q clksta clken clkdis cpcdis burst tiob register a compare rc = cpcstop 16-bit counter e e vt e e vtedg sync swtrg en e tr g wavsel imr t rig acpc acpa aeevt aswtrg bcpc bcpb beevt bswtrg tioa mti oa tiob mtiob cpas covfs etrgs sr cpcs cpbs clk ovf reset output contr oller o utput controller int 1 ed g e det ect o r timer/counter channel timer_clock1 xc0 xc1 xc2 wavsel register b register c compare rb = compare ra = timer_clock2 timer_clock3 timer_clock4 timer_clock5
569 32145c?06/2013 at32uc3l0128/256 24.6.3.2 wavsel = 0 when cmrn.wavsel is zero, the value of cv n is incremented fr om 0 to 0xffff. once 0xffff has been reached, the value of cvn is re set. incrementation of cvn starts again and the cycle continues. see figure 24-6 on page 569 . an external event trigger or a software trigger can reset the value of cvn. it is important to note that the trigger may occur at any time. see figure 24-7 on page 570 . rc compare cannot be programmed to generate a trigger in this configuration. at the same time, rc compare can stop the counter clock (cmrn.cpcstop = 1) and/or disable the counter clock (cmrn.cpcdis = 1). figure 24-6. wavsel= 0 without trigger time counter value rc rb ra tiob tioa counter cleared by compare match with 0xffff 0xffff waveform examples
570 32145c?06/2013 at32uc3l0128/256 figure 24-7. wavsel= 0 with trigger 24.6.3.3 wavsel = 2 when cmrn.wavsel is two, the value of cvn is incremented from zero to the value of rc, then automatically reset on a rc compare. once the value of cvn has been reset, it is then incremented and so on. see figure 24-8 on page 571 . it is important to note that cvn can be reset at any time by an external event or a software trig- ger if both are programmed correctly. see figure 24-9 on page 571 . in addition, rc compare can stop the counter clock (cmrn.cpcstop) and/or disable the counter clock (cmrn.cpcdis = 1). time counter value rc rb ra tiob tioa counter cleared by compare match with 0xffff 0xffff waveform examples counter cleared by trigger
571 32145c?06/2013 at32uc3l0128/256 figure 24-8. wavsel = 2 without trigger figure 24-9. wavsel = 2 with trigger 24.6.3.4 wavsel = 1 when cmrn.wavsel is one, the value of cvn is incremented from 0 to 0xffff. once 0xffff is reached, the value of cvn is decremented to 0, then re-incremented to 0xffff and so on. see figure 24-10 on page 572 . time counter value rc rb ra tiob tioa counter cleared by compare match with rc 0xffff waveform examples time counter value r c r b r a tiob tioa counter cleared by compare match with rc 0xffff waveform examples counter cleared by trigger
572 32145c?06/2013 at32uc3l0128/256 a trigger such as an external event or a software trigger can modify cvn at any time. if a trigger occurs while cvn is incrementing, cvn then decrements. if a trigger is received while cvn is decrementing, cvn then increments. see figure 24-11 on page 572 . rc compare cannot be programmed to generate a trigger in this configuration. at the same time, rc compare can stop the counter clock (cmrn.cpcstop = 1) and/or dis- able the counter clock (cmrn.cpcdis = 1). figure 24-10. wavsel = 1 without trigger figure 24-11. wavsel = 1 with trigger time counter value rc rb ra tiob tioa counter decremented by compare match with 0xffff 0xffff waveform examples time counter value tiob tioa counter decremented by compare match with 0xffff 0xffff waveform examples counter decremented by trigger rc rb ra counter incremented by trigger
573 32145c?06/2013 at32uc3l0128/256 24.6.3.5 wavsel = 3 when cmrn.wavsel is three, the value of cvn is incremented from ze ro to rc. once rc is reached, the value of cvn is decremented to zero, then re-incremented to rc and so on. see figure 24-12 on page 573 . a trigger such as an external event or a software trigger can modify cvn at any time. if a trigger occurs while cvn is incrementing, cvn then decrements. if a trigger is received while cvn is decrementing, cvn then increments. see figure 24-13 on page 574 . rc compare can stop the counter clock (cmrn.cp cstop = 1) and/or disable the counter clock (cmrn.cpcdis = 1). figure 24-12. wavsel = 3 without trigger time counter value rc rb ra tiob tioa counter cleared by compare match with rc 0xffff waveform examples
574 32145c?06/2013 at32uc3l0128/256 figure 24-13. wavsel = 3 with trigger 24.6.3.6 external event/trigger conditions an external event can be programmed to be detected on one of the clock sources (xc0, xc1, xc2) or tiob. the external event selected can then be used as a trigger. the external event selection field in cmrn (cmrn.eevt) selects the external trigger. the external event edge selection field in cmrn (cmrn.eevtedg) defines the trigger edge for each of the possible external triggers (rising, fa lling or both). if cmrn.eevtedg is written to zero, no external event is defined. if tiob is defined as an ex ternal event signal (cmrn.eevt = 0), tiob is no longer used as an output and the compare register b is not used to generate waveforms and subsequently no irqs. in this case the tc channel can only generate a waveform on tioa. when an external event is defined, it can be used as a trigger by writing a one to the cmrn.enetrg bit. as in capture mode, the sync signal and the softw are trigger are also available as triggers. rc compare can also be used as a trigger depending on the cmrn.wavsel field. 24.6.3.7 output controller the output controller defines the output level changes on tioa and tiob following an event. tiob control is used only if tiob is defin ed as output (not as an external event). the following events control tioa and tiob: ? software trigger ? external event ? rc compare ra compare controls tioa and rb compare controls tiob. each of these events can be pro- grammed to set, clear or toggle the output as defined in the following fields in cmrn: ? rc compare effect on tiob (cmrn.bcpc) time counter value tiob tioa counter decremented by compare match with rc 0xffff waveform examples rc rb ra counter decremented by trigger counter incremented by trigger
575 32145c?06/2013 at32uc3l0128/256 ? rb compare effect on tiob (cmrn.bcpb) ? rc compare effect on tioa (cmrn.acpc) ? ra compare effect on tioa (cmrn.acpa)
576 32145c?06/2013 at32uc3l0128/256 24.7 user interface table 24-3. tc register memory map offset register register name access reset 0x00 channel 0 control register ccr0 write-only 0x00000000 0x04 channel 0 mode register cmr0 read/write 0x00000000 0x10 channel 0 counter value cv0 read-only 0x00000000 0x14 channel 0 register a ra0 read/write (1) 0x00000000 0x18 channel 0 register b rb0 read/write (1) 0x00000000 0x1c channel 0 register c rc0 read/write 0x00000000 0x20 channel 0 status register sr0 read-only 0x00000000 0x24 interrupt enable register ier0 write-only 0x00000000 0x28 channel 0 interrupt disable register idr0 write-only 0x00000000 0x2c channel 0 interrupt mask register imr0 read-only 0x00000000 0x40 channel 1 control register ccr1 write-only 0x00000000 0x44 channel 1 mode register cmr1 read/write 0x00000000 0x50 channel 1 counter value cv1 read-only 0x00000000 0x54 channel 1 register a ra1 read/write (1) 0x00000000 0x58 channel 1 register b rb1 read/write (1) 0x00000000 0x5c channel 1 register c rc1 read/write 0x00000000 0x60 channel 1 status register sr1 read-only 0x00000000 0x64 channel 1 interrupt enable register ier1 write-only 0x00000000 0x68 channel 1 interrupt disable register idr1 write-only 0x00000000 0x6c channel 1 interrupt mask register imr1 read-only 0x00000000 0x80 channel 2 control register ccr2 write-only 0x00000000 0x84 channel 2 mode register cmr2 read/write 0x00000000 0x90 channel 2 counter value cv2 read-only 0x00000000 0x94 channel 2 register a ra2 read/write (1) 0x00000000 0x98 channel 2 register b rb2 read/write (1) 0x00000000 0x9c channel 2 register c rc2 read/write 0x00000000 0xa0 channel 2 status register sr2 read-only 0x00000000 0xa4 channel 2 interrupt enable register ier2 write-only 0x00000000 0xa8 channel 2 interrupt disable register idr2 write-only 0x00000000 0xac channel 2 interrupt mask register imr2 read-only 0x00000000 0xc0 block control register bcr write-only 0x00000000 0xc4 block mode register bmr read/write 0x00000000 0xf8 features register features read-only - (2) 0xfc version register version read-only - (2)
577 32145c?06/2013 at32uc3l0128/256 notes: 1. read-only if cmrn.wave is zero. 2. the reset values are device specific. please re fer to the module config uration section at the end of this chapter.
578 32145c?06/2013 at32uc3l0128/256 24.7.1 channel control register name: ccr access type: write-only offset: 0x00 + n * 0x40 reset value: 0x00000000 ? swtrg: software trigger command 1: writing a one to this bit will perform a software tr igger: the counter is reset and the clock is started. 0: writing a zero to this bit has no effect. ? clkdis: counter cl ock disable command 1: writing a one to this bit will disable the clock. 0: writing a zero to this bit has no effect. ? clken: counter clock enable command 1: writing a one to this bit will enab le the clock if clkdis is not one. 0: writing a zero to this bit has no effect. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 - - - - - swtrg clkdis clken
579 32145c?06/2013 at32uc3l0128/256 24.7.2 channel mode register: capture mode name: cmr access type: read/write offset: 0x04 + n * 0x40 reset value: 0x00000000 ? ldrb: rb loading selection ? ldra: ra loading selection ?wave 1: capture mode is disabled (waveform mode is enabled). 0: capture mode is enabled. ? cpctrg: rc compare trigger enable 1: rc compare resets the counter and starts the counter clock. 0: rc compare has no effect on the counter and its clock. ? abetrg: tioa or tiob external trigger selection 1: tioa is used as an external trigger. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 - - - - ldrb ldra 15 14 13 12 11 10 9 8 wave cpctrg - - - abetrg etrgedg 76543210 ldbdis ldbstop burst clki tcclks ldrb edge 0 none 1 rising edge of tioa 2 falling edge of tioa 3 each edge of tioa ldra edge 0 none 1 rising edge of tioa 2 falling edge of tioa 3 each edge of tioa
580 32145c?06/2013 at32uc3l0128/256 0: tiob is used as an external trigger. ? etrgedg: external trigger edge selection ? ldbdis: counter clock disable with rb loading 1: counter clock is disabled when rb loading occurs. 0: counter clock is not disabled when rb loading occurs. ? ldbstop: counter clock stopped with rb loading 1: counter clock is stopped when rb loading occurs. 0: counter clock is not stopped when rb loading occurs. ? burst: burst signal selection ? clki: clock invert 1: the counter is incremented on falling edge of the clock. 0: the counter is incremented on rising edge of the clock. ? tcclks: clock selection etrgedg edge 0 none 1 rising edge 2 falling edge 3 each edge burst burst signal selection 0 the clock is not gated by an external signal 1 xc0 is anded with the selected clock 2 xc1 is anded with the selected clock 3 xc2 is anded with the selected clock tcclks clock selected 0timer_clock1 1timer_clock2 2timer_clock3 3timer_clock4 4timer_clock5 5xc0 6xc1 7xc2
581 32145c?06/2013 at32uc3l0128/256 24.7.3 channel mode register: waveform mode name: cmr access type: read/write offset: 0x04 + n * 0x40 reset value: 0x00000000 ? bswtrg: software trigger effect on tiob ? beevt: external event effect on tiob 31 30 29 28 27 26 25 24 bswtrg beevt bcpc bcpb 23 22 21 20 19 18 17 16 aswtrg aeevt acpc acpa 15 14 13 12 11 10 9 8 wave wavsel enetrg eevt eevtedg 76543210 cpcdis cpcstop burst clki tcclks bswtrg effect 0 none 1set 2clear 3 toggle beevt effect 0 none 1set 2clear 3 toggle
582 32145c?06/2013 at32uc3l0128/256 ? bcpc: rc compare effect on tiob ? bcpb: rb compare effect on tiob ? aswtrg: software trigger effect on tioa ? aeevt: external event effect on tioa ? acpc: rc compare effect on tioa bcpc effect 0 none 1set 2clear 3 toggle bcpb effect 0 none 1set 2clear 3 toggle aswtrg effect 0 none 1set 2clear 3 toggle aeevt effect 0 none 1set 2clear 3 toggle acpc effect 0 none 1set 2clear 3 toggle
583 32145c?06/2013 at32uc3l0128/256 ? acpa: ra compare effect on tioa ?wave 1: waveform mode is enabled. 0: waveform mode is disabled (capture mode is enabled). ? wavsel: waveform selection ? enetrg: external event trigger enable 1: the external event resets the counter and starts the counter clock. 0: the external event has no effect on the counter and its clock. in this case, the selected external event only controls the t ioa output. ? eevt: external event selection note: 1. if tiob is chosen as the external event signal, it is conf igured as an input and no longer generates waveforms and subse- quently no irqs . ? eevtedg: external ev ent edge selection ? cpcdis: counter clock disable with rc compare 1: counter clock is disabled when counter reaches rc. 0: counter clock is not disabled when counter reaches rc. acpa effect 0 none 1set 2clear 3 toggle wavsel effect 0 up mode without automatic trigger on rc compare 1 updown mode without automat ic trigger on rc compare 2 up mode with automatic trigger on rc compare 3 updown mode with automatic trigger on rc compare eevt signal selected as exte rnal event tiob direction 0 tiob input (1) 1 xc0 output 2 xc1 output 3 xc2 output eevtedg edge 0none 1 rising edge 2 falling edge 3 each edge
584 32145c?06/2013 at32uc3l0128/256 ? cpcstop: counter clock stopped with rc compare 1: counter clock is stopped when counter reaches rc. 0: counter clock is not stopped when counter reaches rc. ? burst: burst signal selection ? clki: clock invert 1: counter is incremented on falling edge of the clock. 0: counter is incremented on rising edge of the clock. ? tcclks: clock selection burst burst signal selection 0 the clock is not gated by an external signal. 1 xc0 is anded with the selected clock. 2 xc1 is anded with the selected clock. 3 xc2 is anded with the selected clock. tcclks clock selected 0timer_clock1 1timer_clock2 2timer_clock3 3timer_clock4 4timer_clock5 5xc0 6xc1 7xc2
585 32145c?06/2013 at32uc3l0128/256 24.7.4 channel counter value register name: cv access type: read-only offset: 0x10 + n * 0x40 reset value: 0x00000000 ?cv: counter value cv contains the counter value in real time. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 cv[15:8] 76543210 cv[7:0]
586 32145c?06/2013 at32uc3l0128/256 24.7.5 channel register a name: ra access type: read-only if cmrn.wave = 0, read/write if cmrn.wave = 1 offset: 0x14 + n * 0x40 reset value: 0x00000000 ? ra: register a ra contains the register a value in real time. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 ra[15:8] 76543210 ra[7:0]
587 32145c?06/2013 at32uc3l0128/256 24.7.6 channel register b name: rb access type: read-only if cmrn.wave = 0, read/write if cmrn.wave = 1 offset: 0x18 + n * 0x40 reset value: 0x00000000 ? rb: register b rb contains the register b value in real time. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 rb[15:8] 76543210 rb[7:0]
588 32145c?06/2013 at32uc3l0128/256 24.7.7 channel register c name: rc access type: read/write offset: 0x1c + n * 0x40 reset value: 0x00000000 ? rc: register c rc contains the register c value in real time. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 rc[15:8] 76543210 rc[7:0]
589 32145c?06/2013 at32uc3l0128/256 24.7.8 channel status register name: sr access type: read-only offset: 0x20 + n * 0x40 reset value: 0x00000000 note: reading the status register will also clear th e interrupt bit for the co rresponding interrupts. ? mtiob: tiob mirror 1: tiob is high. if cmrn.wave is zero, this means that tiob pi n is high. if cmrn.wave is one, this means that tiob is driven high. 0: tiob is low. if cmrn.wave is zero, this means that tiob pi n is low. if cmrn.wave is one, this means that tiob is driven low. ? mtioa: tioa mirror 1: tioa is high. if cmrn.wave is zero, this means that tioa pin is high. if cmrn.wave is one, th is means that tioa is driven high. 0: tioa is low. if cmrn.wave is zero, this means that tioa pin is low. if cmrn.wave is one, this means that tioa is driven low. ? clksta: clock enabling status 1: this bit is set when the clock is enabled. 0: this bit is cleared when the clock is disabled. ? etrgs: external trigger status 1: this bit is set when an external trigger has occurred. 0: this bit is cleared when the sr register is read. ? ldrbs: rb loading status 1: this bit is set when an rb load has occurred and cmrn.wave is zero. 0: this bit is cleared when the sr register is read. ? ldras: ra loading status 1: this bit is set when an ra load has occurred and cmrn.wave is zero. 0: this bit is cleared when the sr register is read. ? cpcs: rc compare status 1: this bit is set when an rc compare has occurred. 0: this bit is cleared when the sr register is read. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 - - - - - mtiob mtioa clksta 15 14 13 12 11 10 9 8 -------- 76543210 etrgs ldrbs ldras cpcs cpbs cpas lovrs covfs
590 32145c?06/2013 at32uc3l0128/256 ? cpbs: rb compare status 1: this bit is set when an rb compare has occurred and cmrn.wave is one. 0: this bit is cleared when the sr register is read. ? cpas: ra compare status 1: this bit is set when an ra compare has occurred and cmrn.wave is one. 0: this bit is cleared when the sr register is read. ? lovrs: load overrun status 1: this bit is set when ra or rb have been loaded at l east twice without any read of the corresponding register and cmrn.wave is zero. 0: this bit is cleared when the sr register is read. ? covfs: counter overflow status 1: this bit is set when a counter overflow has occurred. 0: this bit is cleared when the sr register is read.
591 32145c?06/2013 at32uc3l0128/256 24.7.9 channel interrupt enable register name: ier access type: write-only offset: 0x24 + n * 0x40 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will set the corresponding bit in imr. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 etrgs ldrbs ldras cpcs cpbs cpas lovrs covfs
592 32145c?06/2013 at32uc3l0128/256 24.7.10 channel interrupt disable register name: idr access type: write-only offset: 0x28 + n * 0x40 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in imr. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 etrgs ldrbs ldras cpcs cpbs cpas lovrs covfs
593 32145c?06/2013 at32uc3l0128/256 24.7.11 channel interrupt mask register name: imr access type: read-only offset: 0x2c + n * 0x40 reset value: 0x00000000 0: the corresponding interrupt is disabled. 1: the corresponding interrupt is enabled. a bit in this register is cleared when the corresponding bit in idr is written to one. a bit in this register is set when the corresponding bit in ier is written to one. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 etrgs ldrbs ldras cpcs cpbs cpas lovrs covfs
594 32145c?06/2013 at32uc3l0128/256 24.7.12 block control register name: bcr access type: write-only offset: 0xc0 reset value: 0x00000000 ? sync: synchro command 1: writing a one to this bit asserts the sync signal which generates a software trigger simultaneously for each of the channels . 0: writing a zero to this bit has no effect. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 -------sync
595 32145c?06/2013 at32uc3l0128/256 24.7.13 block mode register name: bmr access type: read/write offset: 0xc4 reset value: 0x00000000 ? tc2xc2s: external clock signal 2 selection ? tc1xc1s: external clock signal 1 selection 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 - - tc2xc2s tc1xc1s tc0xc0s tc2xc2s signal connected to xc2 0tclk2 1none 2tioa0 3tioa1 tc1xc1s signal connected to xc1 0tclk1 1none 2tioa0 3tioa2
596 32145c?06/2013 at32uc3l0128/256 ? tc0xc0s: external cloc k signal 0 selection tc0xc0s signal connected to xc0 0tclk0 1none 2tioa1 3tioa2
597 32145c?06/2013 at32uc3l0128/256 24.7.14 features register name: features access type: read-only offset: 0xf8 reset value: - ? brpbhsb: bridge type is pb to hsb 1: bridge type is pb to hsb. 0: bridge type is not pb to hsb. ? updnimpl: up/down is implemented 1: up/down counter capability is implemented. 0: up/down counter capability is not implemented. ? ctrsize: counter size this field indicates the size of the counter in bits. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ------- 15 14 13 12 11 10 9 8 - - - - - - brpbhsb updnimpl 76543210 ctrsize
598 32145c?06/2013 at32uc3l0128/256 24.7.15 version register name: version access type: read-only offset: 0xfc reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
599 32145c?06/2013 at32uc3l0128/256 24.8 module configuration the specific configuration for each timer/counter instance is listed in the following tables.the module bus clocks listed here are c onnected to the system bus cloc ks. please refer to the power manager chapter for details. 24.8.1 clock connections there are two timer/counter modules, tc0 and tc1, with three channels each, giving a total of six timer/counter channels. each timer/counter channel can independently select an internal or external clock source for its counter: table 24-4. tc bus interface clocks module name clock name description tc0 clk_tc0 clock for the tc0 bus interface tc1 clk_tc1 clock for the tc1 bus interface table 24-5. timer/counter clock connections module source name connection tc0 internal timer_clock1 32 khz oscillator clock (clk_32k) timer_clock2 pba clock / 2 timer_clock3 pba clock / 8 timer_clock4 pba clock / 32 timer_clock5 pba clock / 128 external xc0 see section 3.2.1 on page 8 xc1 xc2 tc1 internal timer_clock1 32 khz oscillator clock (clk_32k) timer_clock2 pba clock / 2 timer_clock3 pba clock / 8 timer_clock4 pba clock / 32 timer_clock5 pba clock / 128 external xc0 see section 3.2.1 on page 8 xc1 xc2
600 32145c?06/2013 at32uc3l0128/256 25. peripheral event system rev: 1.0.0.1 25.1 features ? direct peripheral to peri pheral communication system ? allows peripherals to receive, react to, and send peripheral events without cpu intervention ? cycle deterministic event communication ? asynchronous interrupts allow advanced peri pheral operation in low power sleep modes 25.2 overview several peripheral modules can be configured to emit or respond to signals known as peripheral events. the exact condition to trigger a peripheral event, or the action taken upon receiving a peripheral event, is specific to each module. pe ripherals that respond to peripheral events are called peripheral event users and peripherals that emit peripheral events are called peripheral event generators. a single module can be both a peripheral event generator and user. the peripheral event generators and users are interconnected by a network known as the peripheral event system. this allows low latency peripheral-to-peripheral signaling without cpu intervention, and without cons uming system resources such as bus or ram bandwidth. this offloads the cpu and system resources compared to a traditional interrupt-based software driven system. 25.3 peripheral event sy stem block diagram figure 25-1. peripheral event system block diagram 25.4 functional description 25.4.1 configuration the peripheral event system in the at32uc3l0128/256 has a fixed mapping of peripheral events between generators and users, as described in table 25-1 to table 25-4 . thus, the user does not need to configure the interconnection between the modules, although each peripheral event can be enabled or disabled at the generator or user side as described in the peripheral chapter for each module. peripheral event system generator generator user generator/ user
601 32145c?06/2013 at32uc3l0128/256 table 25-1. peripheral event mapping from acifb to pwma generator generated event user effect asynchronous acifb channel 0 ac0 v inp > ac0 v inn pwma channel 0 pwma duty cycle value increased by one no ac0 v inn > ac0 v inp pwma duty cycle value decreased by one acifb channel 1 ac1 v inp > ac1 v inn pwma channel 6 pwma duty cycle value increased by one ac1 v inn > ac1 v inp pwma duty cycle value decreased by one acifb channel 2 ac2 v inp > ac2 v inn pwma channel 8 pwma duty cycle value increased by one ac2 v inn > ac2 v inp pwma duty cycle value decreased by one acifb channel 3 ac3 v inp > ac3 v inn pwma channel 9 pwma duty cycle value increased by one ac3 v inn > ac3 v inp pwma duty cycle value decreased by one acifb channel 4 ac4 v inp > ac4 v inn pwma channel 11 pwma duty cycle value increased by one ac4 v inn > ac4 v inp pwma duty cycle value decreased by one acifb channel 5 ac5 v inp > ac5 v inn pwma channel 14 pwma duty cycle value increased by one ac5 v inn > ac5 v inp pwma duty cycle value decreased by one acifb channel 6 ac6 v inp > ac6 v inn pwma channel 19 pwma duty cycle value increased by one ac6 v inn > ac6 v inp pwma duty cycle value decreased by one acifb channel 7 ac7 v inp > ac7 v inn pwma channel 20 pwma duty cycle value increased by one ac7 v inn > ac7 v inp pwma duty cycle value decreased by one acifb channel n acn v inn > acn v inp cat automatically used by the cat when performing qmatrix acquisition. no table 25-2. peripheral event mapping from gpio to tc generator generated event user effect asynchronous gpio pin change on pa00-pa07 tc0 a0 capture no pin change on pa08-pa15 a1 capture pin change on pa16-pa23 a2 capture pin change on pb00-pb07 tc1 a1 capture pin change on pb08-pb15 a2 capture
602 32145c?06/2013 at32uc3l0128/256 25.4.2 peripheral event connections each generated peripheral event is connected to one or more users. if a peripheral event is con- nected to multiple users, the peripheral event can trigger actions in multiple modules. a peripheral event user can likewise be connected to one or more peripheral event generators. if a peripheral event user is connected to multiple generators, the peripheral events are or?ed together to a single peripheral event. this means that peripheral events from either one of the generators will result in a pe ripheral event to the user. to configure a peripheral event, the peripheral event must be enabled at both the generator and user side. even if a generator is connected to multiple users, only the users with the peripheral event enabled will trigger on the peripheral event. 25.4.3 low power operation as the peripheral events do not require cpu intervention, they are available in idle mode. they are also available in deeper sleep modes if both the generator and user remain clocked in that mode. certain events are known as asynchronous peripheral events, as identified in table 25-1 to table 25-4 . these can be issued even when the system clock is stopped, and revive unclocked user peripherals. the clock will be restarted for this m odule only, without wa king the system from sleep mode. the clock remains active only as long as required by the triggered function, before being switched off again, and the system remain s in the original sleep mode. the cpu and sys- table 25-3. peripheral event mapping from ast generator generated even t user effect asynchronous ast overflow event acifb comparison is triggered if the acifb.confn register is written to 11 (event triggered single measurement mode) and the eventen bit in the acifb.ctrl register is written to 1. ye s periodic event alarm event overflow event adcifb conversion is triggered if the trgmod bit in the adcifb.trgr register is written to 111 (peripheral event trigger). periodic event alarm event overflow event cat trigger one iteration of autonomous touch detection. periodic event alarm event table 25-4. peripheral event mapping from pwma generator generated even t user effect asynchronous pwma channel 0 timebase counter reaches the duty cycle value. acifb comparison is triggered if the acifb.confn register is written to 11 (event triggered single measurement mode) and the eventen bit in the acifb.ctrl register is written to 1. no adcifb conversion is triggered if the trgmod bit in the adcifb.trgr register is written to 111 (peripheral event trigger).
603 32145c?06/2013 at32uc3l0128/256 tem will only be woken up if the user peripheral generates an interrupt as a result of the operation. this concept is know n as sleepwalking and is described in further detail in the power manager chapter. note that asynchronous peripheral events may be associated with a delay due to the need to restart the system clock source if this has been stopped in the sleep mode. 25.5 application example this application example shows how the peripheral event system can be used to program the adc interface to perform adc conversions at selected intervals. conversions of the active analog channels are st arted with a software or a hardware trigger. one of the possible hardware triggers is a peripheral event trigger, allowing the peripheral event system to synchronize conversion with some configured peripheral event source. from table 25-3 and table 25-4 , it can be read that this peripheral event source can be either an ast peripheral event, or an event from the pwm controller. the ast can generate periodic periph- eral events at selected intervals, among other types of peripheral events. the peripheral event system can then be used to set up the adc interface to sample an analog signal at regular intervals. the user must enable peripheral events in the ast and in the adc interface to accomplish this. the periodic peripheral event in the ast is enabl ed by writing a one to the corresponding bit in the ast event enable register (eve). to select the peripheral event trigger for the adc inter- face, the user must write the value 0x7 to the trigger mode (trgmod) field in the adc interface trig ger register (trgr). when the peripheral ev ents are enabled, the ast will gener- ate peripheral events at the selected intervals, and the peripheral event system will route the peripheral events to the adc interface, which will perform adc conversions at the selected intervals. figure 25-2. application example since the ast peripheral event is asynchronous , the description above will also work in sleep modes where the adc clock is stopped. in this case, the adc clock (and clock source, if needed) will be restarted during the adc conversion. after the conversion, the adc clock and clock source will return to the sleep state, unless the adc generates an interrupt, which in turn will wake up the system. using asynchronous inte rrupts thus allows adc operation in much lower power states than would otherwise be possible. peripheral event system ast adc interface trigger conversion periodic peripheral event
604 32145c?06/2013 at32uc3l0128/256 26. adc interface (adcifb) rev:1.1.0.1 26.1 features ? multi-channel analog-to-digital converter with up to 12-bit resolution ? enhanced resolution mode ? 11-bit resolution obtained by interpolating 4 samples ? 12-bit resolution obtained by interpolating 16 samples ? glueless interface with resisti ve touch screen panel, allowing ? resistive touch screen position measurement ? pen detection and pen loss detection ? integrated enhanced sequencer ? adc mode ? resistive touch screen mode ? numerous trigger sources ? software ? embedded 16-bit timer for periodic trigger ? pen detect trigger ? continuous trigger ? external trigger, rising, falling, or any-edge trigger ? peripheral event trigger ? adc sleep mode for low power adc applications ? programmable adc timings ? programmable adc clock ? programmable startup time 26.2 overview the adc interface (adcifb) converts analog input voltages to digital values. the adcifb is based on a successive approximation register (sar) 10-bit analog-to-digital converter (adc). the conversions extend from 0v to advrefp. the adcifb supports 8-bit and 10-bit resolution mode, in addition to enhanced resolution mode with 11-bit and 12-bit resolution. conversion results are reported in a common register for all channels. the 11-bit and 12-bit resolution modes are obtained by interpolating multiple samples to acquire better accuracy. for 11-bit mode 4 samples are used, which gives an effective sample rate of 1/4 of the actual sample frequency. for 12-bit mode 16 samples are used, giving a effective sample rate of 1/16 of actual. this arrangement allows conversion speed to be traded for better accuracy. conversions can be started for all enabled channels, either by a software trigger, by detection of a level change on the external trigger pin (trigger), or by an integrated programmable timer. when the resistive touch screen mode is enabled, an integrated sequencer automatically con- figures the pad control signals and performs resistive touch screen conversions. the adcifb also integrates an adc sleep mode, a pen-detect mode, and an analog compare mode, and connects with one peripheral dma controller channel. these features reduce both power consumption and processor intervention.
605 32145c?06/2013 at32uc3l0128/256 26.3 block diagram figure 26-1. adcifb block diagram advrefp analog multiplexer successive approximation register analog-to-digital converter trigger adc control logic timer user interface ad0 ad1 ad3 adn ad2 resisitve touch screen sequencer clk_adcifb .... adcifb adp0 adp1 i/o controller trigger peripheral bus dma request interrupt request clk_adc
606 32145c?06/2013 at32uc3l0128/256 26.4 i/o lines description 26.5 product dependencies in order to use this module, other parts of the system must be configured correctly, as described below. 26.5.1 i/o lines the analog input pins can be multiplexed with i/o controller lines. the user must make sure the i/o controller is configured correctly to allow the adcifb access to the ad pins before the adcifb is instructed to start converting data. if the user fails to do this the converted data may be wrong. the number of analog inputs is device dependent, please refer to the adcifb module configu- ration chapter for the number of available ad inputs on the current device. the advrefp pin must be connect ed correctly prior to using the adcifb. failing to do so will result in invalid adc operatio n. see the electrical characteristics chapter for details. if the trigger, adp0, and adp1 pins are to be used in the application, the user must config- ure the i/o controller to assign the needed pins to the adcifb function. 26.5.2 power management if the cpu enters a sleep mode that disables clocks used by the adcifb, the adcifb will stop functioning and resume operation after the system wakes up from sleep mode. if the peripheral event system is configur ed to send asynchronous peripheral events to the adcifb and the clock used by the adcifb is stopped, a local and tem porary clock will automat- ically be requested so the event can be processed. refer to section 26.6.13 , section 26.6.12 , and the peripheral event system chapter for details. before entering a sleep mode where the clock to the adcifb is stopped, make sure the analog- to-digital conver ter cell is put in an inactive state. refer to section 26.6.13 for more information. 26.5.3 clocks the clock for the adcifb bus interface (clk _adcifb) is generated by the power manager. this clock is enabled at reset, and can be disabled in the power manager. it is recommended to disable the adcifb before disabling the clock, to avoid freezing the adcifb in an undefined state. table 26-1. i/o lines description pin name description type advrefp reference voltage analog trigger external trigger digital adp0 drive pin 0 for resistive touch screen top channel (xp) digital adp1 drive pin 1 for resistive touch screen right channel (yp) digital ad0-adn analog input channels 0 to n analog
607 32145c?06/2013 at32uc3l0128/256 26.5.4 dma the adcifb dma handshake interface is connected to the peripheral dma controller. using the adcifb dma functionality requires the peripheral dma controller to be programmed first. 26.5.5 interrupts the adcifb interrupt request line is connected to the interrupt controller. using the adcifb interrupt request functionality requires the interrupt controller to be programmed first. 26.5.6 peripheral events the adcifb peripheral events are connected via the peripheral event system. refer to the peripheral event system chapter for details 26.5.7 debug operation when an external debugger forces the cpu into debug mode, this module continues normal operation. if this module is configured in a way that requires it to be peri odically serviced by the cpu through interrupt requests or similar, improper operation or data loss may result during debugging. 26.6 functional description the adcifb embeds a successive approximation register (sar) analog-to-digital converter (adc). the adc supports 8-bit or 10-bit resolution, which can be extended to 11 or 12 bits by the enhanced resolution mode. the conversion is performed on a full range between 0v and the reference voltage pin advrefp. analog inputs between these voltages converts to digital values (codes) based on a linear conversion. this linear conversion is described in the expression below where m is the number of bits used to represent the analog value, vin is the voltage of the analog value to con- vert, vref is the maximum voltage, and code is the converted digital value. 26.6.1 initializing the adcifb the adc interface is enabled by writing a one to the enable bit in the control register (cr.en). after the adc interface is enabled, the adc timings needs to be configured by writing the cor- rect values to the res, prescal, and startu p fields in the adc configuration register (acr). see section 26.6.5 , and section 26.6.7 for details. before the adcifb can be used, the i/o controller must be configured correctly and the reference voltage (advrefp) signal must be connected. refer to section 26.5.1 for details. 26.6.2 basic operation to convert analog values to digital values the user must first initialize the adcifb as described in section 26.6.1 . when the adcifb is initialized the channels to convert must be enabled by writing a one the corresponding bits in the c hannel enable register (cher). enabling channel n instructs the adcifb to convert the analog voltage applied to ad pin n at each conversion sequence. multiple channels can be enabled resu lting in multiple ad pins being converted at each conversion sequence. code 2 m v in ? v ref ------------------- - =
608 32145c?06/2013 at32uc3l0128/256 to start converting data the user can either manually start a conversion sequence by writing a one to the start bit in the control register (cr.start) or configure an automatic trigger to initiate the conversions. the automatic trigger can be configured to trig on many different condi- tions. refer to section 26.8.1 for details. the result of the conversion is stored in the last converted data register (lcdr) as they become available, overwriting the result from the previous conversion. to avoid data loss if more than one channel is enabled, the user must read the conversion results as they become avail- able either by using an interrupt handler or by using a peripheral dma channel to copy the results to memory. failing to do so will result in an overrun er ror condition, indicated by the ovre bit in the status register (sr). to use an interrupt h andler the user must enable the data ready (drdy) interrupt request by writing a one to the corresponding bit in the interrupt enable register (ier). to clear the inter- rupt after the conversion result is read, the user must write a one to the corresponding bit in the interrupt clear register (icr). see section 26.6.11 for details. to use a peripheral dma controller channel the user must configure the peripheral dma con- troller appropriately. the peripheral dma controller will, when configured, automatically read converted data as they become available. ther e is no need to manually clear any bits in the interrupt status register as this is performed by the hardware. if an overrun error condition hap- pens during dma operat ion, the ovre bit in the sr will be set. 26.6.3 adc resolution the analog-to-digital converter cell supports 8-bit or 10-bit resolution, which can be extended to 11-bit and 12-bit with the enhanced resolution mode. the resolution is selected by writing the selected resolution value to the res field in the adc configuration register (acr). see section 26.9.3 . by writing a zero to the res field, the adc swit ches to the lowest resolution and the conversion results can be read in the eight lowest significa nt bits of the last converted data register (lcdr). the four highest bits of the last co nverted data (l data) field in the lcdr register reads as zero. writing a one to the res field enables 10-bit resolution, the optimal resolution for both sampling speed and accuracy. writing two or three automatically enables enhanced reso- lution mode with 11-bit or 12-bit resolution, see section 26.6.4 for details. when a peripheral dma controller channel is con nected to the adcifb in 10-bit, 11-bit, or 12- bit resolution mode, a transfer size of 16 bits must be used. by writing a zero to the res field, the destination buffers can be optimized for 8-bit transfers. 26.6.4 enhanced resolution mode the enhanced resolution mode is automatically enabled when 11-bit or 12-bit mode is selected in the adc configuration register (acr). in this mode the a dcifb will trade conversion perfor- mance for accuracy by aver aging multiple samples. to be able to increase the accuracy by averaging multiple samples it is important that some noise is present in the input signal. the noise level should be between one and two lsb peak- to-peak to get good averaging performance. the performance cost of enabling 11-bit mode is 4 adc samples, which reduces the effective adc performance by a factor 4. for 12-bit mode this factor is 16. for 12-bit mode the effective sample rate is maximum adc sample rate divided by 16.
609 32145c?06/2013 at32uc3l0128/256 26.6.5 adc clock the adcifb generates an internal clock named clk_adc that is used by the analog-to-digital converter cell to perform conversions. the clk_ adc frequency is selected by writing to the prescal field in the adc configuration register (acr). the clk_adc range is between clk_adcifb/2, if prescal is 0, and clk_ adcifb/128, if prescal is 63 (0x3f). a sensible prescal value must be used in orde r to provide an adc clock frequency according to the maximum sampling rate parameter given in the electrical characteristics section. failing to do so may result in incorrect analog-to-digital converter operation. 26.6.6 adc sleep mode the adc sleep mode maximizes power saving by automatically deactivating the analog-to-dig- ital converter cell when it is not being used for conversions. the adc sleep mode is enabled by writing a one to the sleep bit in the adc configuration register (acr). when a trigger occurs while the adc sleep mode is enabled, the analog-to-digital converter cell is automatically activated. as the analog cell requires a startup time, the logic waits during this time and then starts the conversion of the enabled channels. when conversions of all enabled channels are complete, the adc is deactivated until the next trigger. 26.6.7 startup time the analog-to-digital converter cell has a minimal startup time when the cell is activated. this startup time is given in the electrical characteristics chapter and must be written to the startup field in the adc configuration register (acr) to get correct conversion results. the startup field expects the startup time to be represented as the number of clk_adc cycles between 8 and 1024 and in steps of 8 that is needed to cover the adc startup time as specified in the electrical characteristics chapter. the analog-to-digital converter cell is activated at the first conversion after reset and remains active if acr.sleep is zero. if acr.sleep is one , the analog-to-digital converter cell is auto- matically deactivated when idle and thus each conversion se quence will have a initial startup time delay. 26.6.8 sample and hold time a minimal sample and hold time is necessary for the adcifb to guarantee the best converted final value when switching between adc channels. this time depends on the input impedance of the analog input, but also on the output impedance of the driver providing the signal to the analog input, as there is no input buffer amplifier. the sample and hold time has to be programmed through the shtim field in the adc configu- ration register (acr). this field can def ine a sample and hold time between 1 and 16 clk_adc cycles. 26.6.9 adc conversion adc conversions are performed on all enabled c hannels when a trigger condition is detected. for details regarding trigger conditions see section 26.8.1 . the term channel is used to identify a specific analog input pin so it can be included or excluded in an analog-to-digital conversion sequence and to identify which ad pin was used to convert the current value in the last con- verted data register (lcdr). channel number n corresponding to ad pin number n.
610 32145c?06/2013 at32uc3l0128/256 channels are enabled by writing a one to the corresponding bit in the channel enable register (cher), and disabled by writing a one to the co rresponding bit in the channel disable register (chdr). active channels are listed in the channel status register (chsr). when a conversion sequence is started, all enabled channels will be c onverted in one sequence and the result will be placed in the last conv erted data register (l cdr) with the channel num- ber used to produce the result. it is important to read out the results while the conversion sequence is ongoi ng, as new values will auto matically overwrite any ol d value and the old value will be lost if not previously read by the user. if the analog-to-digital converter cell is inactive when starting a conversion sequence, the con- version logic will wait a configurable number of clk_adc cycles as defined in the startup time field in the adc configuration register (acr). af ter the cell is activated all enabled channels is converted one by one until no more enabled ch annels exist. the conversion sequence converts each enabled channel in order starting with the channel with the lowest channel number. if the acr.sleep bit is one, the analog- to-digital converter cell is d eactivated after the conversion sequence has finished. for each channel converted, the adcifb waits a sample and hold number of clk_adc cycles as defined in the shtim field in acr, and then instructs the analog-to-digital converter cell to start converting the analog voltage. the adc ce ll requires 10 clk_adc cycles to actually con- vert the value, so the total time to convert a channel is sample and hold + 10 clk_adc cycles. 26.6.10 analog compare mode the adcifb can test if the converted values, as they become availabl e, are below, above, or inside a specified range and generate interrupt requests based on this information. this is useful for applications where the user wants to moni tor some external analog signal and only initiate actions if the value is above, below, or inside some specified range. the analog compare mode is enabled by writing a one to the analog compare enable (ace) bit in the mode register (mr). the values to compare must be written to the low value (lv) field and the high value (hv) field in the compare value register (cvr). the analog compare mode will, when enabled, check all enabled channels against the pre-programmed high and low values and set status bits. to generate an interrupt request if a converted value is below a limit, write the limit to the cvr.lv field and enable interrupt request on th e compare lesser than (clt) bit by writing a one to the corresponding bit in the interrupt enable register (ier). to generate an interrupt request if a converted value is above a limit, writ e the limit to the cvr.hv field and enable inter- rupt for compare greater than (cgt) bit. to generate an interrupt request if a converted value is inside a range, write the low and high limit to the lv and hv fields and enable the compare else (celse) interrupt. to generate an interrupt request if a value is outside a range, write the lv and hv fields to the low and high limits of the range and enable cgt and clt interrupts. note that the values written to lv and hv must match the resolution selected in the adc config- uration register (acr). 26.6.11 interrupt operation interrupt requests are enabled by writing a one to the corresponding bit in the interrupt enable register (ier) and disabled by writing a one to the corresponding bit in the interrupt disable register (idr). enabled interrupts can be read fr om the interrupt mask register (imr). active interrupt requests, but potentially masked, are visible in the interrupt status register (isr). to
611 32145c?06/2013 at32uc3l0128/256 clear an active interrupt request, write a one to the corresponding bit in the interrupt clear reg- ister (icr). the source for the interrupt requests are the status bits in the status register (sr). the sr shows the adcifb status at the time the register is read. the interrupt status register (isr) shows the status since the last write to the interrupt clear register. the combination of isr and sr allows the user to react to status change co nditions but also allows the user to read the cur- rent status at any time. 26.6.12 peripheral events the peripheral event system can be used together with the adcifb to allow any peripheral event generator to be used as a trigger source. to enable peripheral events to trigger a conver- sion sequence the user must write the peripheral event trigger value (0x7) to the trigger mode (trgmod) field in the trigger register (trgr). refer to table 26-4 on page 623 . the user must also configure a peripheral event generator to emit peripheral events for the adcifb to trigger on. refer to the peripheral event system chapter for details. 26.6.13 sleep mode before entering sleep modes the user must make sure the adcifb is idle and that the analog- to-digital converter cell is inactive. to deactivate the analog-to-digital converter cell the sleep bit in the adc configuration register (acr) must be written to one and the adcifb must be idle. to make sure the adcifb is idle, write a zero the trigger mode (trgmod) field in the trigger register (trgr) and wait for the ready bit in the status register (sr) to be set. note that by deactivating the analog-to-digital converter cell, a startup time penalty as defined in the startup field in the adc configur ation register (acr) will apply on the next conversion. 26.6.14 conversion performances for performance and electrical characteristics of the adcifb, refer to the electrical characteris- tics chapter. 26.7 resistive touch screen the adcifb embeds an integrated resistive touch screen sequencer that can be used to cal- culate contact coordinates on a resistive touch screen film. when instructed to start, the integrated resistive touch screen sequencer automatically applies a sequence of voltage pat- terns to the resistive touch screen films and the analog-to-digital conversion cell is used to measure the effects. the resulting measurement s can be used to calculate the horizontal and vertical contact coordinates. it is recommended to use a high resistance touch screen for optimal resolution. the resistive touch screen film is connected to the adcifb using the ad and adp pins. see section 26.7.3 for details. resistive touch screen mode is enabled by writing a one to the touch screen adc mode field in the mode register (mr.tsamod). in this mode, channels tspo+0 though tspo+3 are automatically enabled where tspo refers to the touch screen pin offset field in the mode reg- ister (mr.tspo). for each conversion sequ ence, all enabled channels before tspo+0 and after tspo+3 are converted as ordinary ad c channels, producing 1 conversion result each. when the sequencer enters the tspo+0 channel the resistive touch screen sequencer will take over control and convert the next 4 channels as described in section 26.7.4 .
612 32145c?06/2013 at32uc3l0128/256 26.7.1 resistive touch screen principles a resistive touch screen is based on two resistive films, each one fitted with a pair of electrodes, placed at the top and bottom on one film, and on the right and left on the other. between the two, there is a layer that acts as an insulator, but makes a connection when pressure is applied to the screen. this is illustrated in figure 26-2 on page 612 . figure 26-2. resistive touch screen position measurement 26.7.2 position measurement method as shown in figure 26-2 on page 612 , to detect the position of a contact, voltage is first applied to x p (top) and x m (bottom) leaving y p and y m tristated. due to the linear resistance of the film, there is a voltage gradient from top to bottom on the first film. when a contact is performed on the screen, the voltage at the contact point propagates to the second film. if the input impedance on the y p (right) and y m (left) electrodes are high enough, no current will flow, allowing the volt- age at the contact point to be measured at y p . the value measured represents the vertical position component of the contact point. for the horizontal direction, the same method is used, but by applying voltage from y p (right) to y m (left) and measuring at x p . in an ideal world (linear, with no loss ), the vertical position is equal to: vy p / vdd to compensate for some of the real world imperfections, vx p and vx m can be measured and used to improve accuracy at the cost of two mo re conversions per axes. the new expression for the vertical position then becomes: (vy p - vx m ) / (vx p - vx m ) x m x p y m y p x p x m y p vdd gnd volt horizontal position detection y p y m x p vdd gnd volt vertical position detection pen contact
613 32145c?06/2013 at32uc3l0128/256 26.7.3 resistive touch screen pin connections the resistive touch screen film signals connects to the adcifb using the ad and adp pins. the x p (top) and x m (bottom) film signals are connected to adtspo+0 and adtspo+1 pins, and the y p (right) and y m (left) signals are connected to adts po+2 and adtspo+3 pins. the tspo index is configurable through the touch screen pin offset (tspo) field in the mode register (mr) and allows the user to configure which ad pins to us e for resistive touch scre en applications. writing a zero to the tspo field instructs the adci fb to use ad0 through ad3, where ad0 is con- nected to x p , ad1 is connected to x m and so on. writing a one to the tspo field instructs the adcifb to use ad1 through ad4 for resistiv e touch screen sequencing, where ad1 is con- nected to x p and ad0 is free to be used as an ordinary adc channel. when the analog pin output enable (apoe) bit in the mode register (mr) is zero, the ad pins are used to measure input voltage and drive the gnd sequences, while the adp pins are used to drive the vdd sequences. this arrangement allo ws the user to reduce the voltage seen at the ad input pins by inserting external resistors between adp0 and x p and adp1 and y p signals which are again directly connected to the ad pi ns. it is important that the voltages observed at the ad pins are not higher than the maximum allowed adc input voltage. see figure 26-3 on page 614 for details regarding how to connect the re sistive touch screen films to the ad and adp pins. by adding a resistor between adp0 and x p , and adp1 and y p , the maximum voltage observed at the ad pins can be controlled by th e following voltage divider expressions: the rfilmx parameter is the film resistance observed when measuring between x p and x m . the rresistorx parameter is the resistor size inserted between adp0 and x p . the definition of rfilmy and rresistory is the same but for adp1, y p , and y m instead. table 26-2. resistive touch screen pin connections adcifb pin ts signal, apoe == 0 ts signal, apoe == 1 adp0 xp through a resistor no connect adp1 yp through a resistor no connect adtspo+0 xp xp adtspo+1 xm xm adtspo+2 yp yp adtspo+3 ym ym vad tspo 0 + ?? r filmx r filmx r resistorx + -------------------------------------------- - vdp 0 ?? ? =
614 32145c?06/2013 at32uc3l0128/256 the adp pins are used by default, as the apoe bit is zero after reset. writing a one to the apoe bit instructs the adcifb resistive t ouch screen sequencer to use the already con- nected adtspo+0 and adtspo+2 pins to drive vdd to x p and y p signals directly. in this mode the adp pins can be used as general purpose i/o pins. before writing a one to the apoe bit the user must make sure that the i/o voltage is compatible with the adc input voltage. if the i/o voltage is higher than the maximum input voltage of the adc, permanent damage may occur. refer to the electrical characteristics chapter for details. figure 26-3. resistive touch screen pin connections vad tspo 2 + ?? r filmy r filmy r resistory + -------------------------------------------- - vdp 1 ?? ? = ad tspo+1 x m x p y m y p ad tspo+0 dp 1 dp 0 ad tspo+3 ad tspo+2 analog pin o u tp u t ena b le (mr.apoe) == 0 ad tspo+1 x m x p y m y p ad tspo+0 dp 1 dp 0 ad tspo+3 ad tspo+2 analog pin o u tp u t ena b le (mr.apoe) == 1 n c n c
615 32145c?06/2013 at32uc3l0128/256 26.7.4 resistive touch screen sequencer the resistive touch screen sequencer is responsible for applying voltage to the resistive touch screen films as described in section 26.7.2 . this is done by controlling the output en able and the output value of the adp and ad pins. this allo ws the resistive touch screen sequencer to add a voltage gradient on one film while keeping the other film floating so a touch can be measured. the resistive touch screen sequencer will when measuring the vertical position, apply vdd and gnd to the pins connected to x p and x m . the y p and y m pins are put in tristate mode so the measurement of y p can proceed without interference. to compensate for adc offset errors and non ideal pad drivers, the actual voltage of x p and x m is measured as well, so the real values for vdd and gnd can be used in the contact point calculation to increase accuracy. see second formula in section 26.7.2 . when the vertical values are converted the same setup is applies for the second axes, by setting x p and x m in tristate mode and applying vdd and gnd to y p and y m . refer to section 26.8.3 for details. 26.7.5 pen detect if no contact is applied to the resistive touch screen films, any resistive touch screen conversion result will be undefined as the film being measur ed is floating. this can be avoided by enabling pen detect and only trigger resistive to uch screen conversions when the pen contact (pencnt) status bit in the status register (sr) is one. pen detect is enabled by writing a one to the pen detect (pendet) bit in the mode register (mr). when pen detect is enabled, the adcifb gro unds the vertical panel by applying gnd to x p and x m and polarizes the horizontal panel by enabling pull-up on the pin connected to y p . the y m pin will in this mode be tris tated. since there is no contact, no current is flowin g and there is no related power consumption. as soon as a contact occurs, gnd will propagate to y m by pulling down y p , allowing the contact to be registered by the adcifb. a programmable debouncing filter can be used to filter out false pen detects because of noise. the debouncing filter is programmable from one clk_adc period and up to 2 15 clk_adc peri- ods. the debouncer length is set by writing to the pendbc field in mr.
616 32145c?06/2013 at32uc3l0128/256 figure 26-4. resistive touch screen pen detect the resistive touch screen pen detect can be used to generate an adcifb interrupt request or it can be used to trig a conversion, so that a position can be measured as soon as a contact is detected. the pen detect mode generates two types of status signals, reported in the status register (sr): ? the bit pencnt is set when current flows and remains set until current stops. ? the bit nocnt is set when no current flows and remains set until current flows. before a current change is reflected in the sr, the new status must be stable for the duration of the debouncing time. both status conditions can generate an interrupt request if the corresponding bit in the interrupt mask register (imr) is one. refer to section 26.6.11 on page 610 . x p x m y m y p tr i s t at e gnd pullup to the adc pen int errupt d ebouncer pendbc gnd resistive touch screen sequencer
617 32145c?06/2013 at32uc3l0128/256 26.8 operating modes the adcifb features two operating modes, eac h defining a separate conversion sequence: ? adc mode: at each trigger, all the enabled channels are converted. ? resistive touch screen mode: at each trigger, all enabled channels plus the resistive touch screen channels are converted as described in section 26.8.3 . if channels except the dedicated resistive touch screen channels are enabled, they are converted normally before and after the resistive touch screen channels are converted. the operating mode is selected by the tsamod field in the mode register (mr). 26.8.1 conversion triggers a conversion sequence is started either by a software or by a hardware trigger. when a conver- sion sequence is started, a ll enabled channels will be conver ted and made available in the shared last converted register (lcdr). the software trigger is asserted by writing a one to the start field in th e control register (cr). the hardware trigger can be selected by the trgmod field in the trigger register (trgr). dif- ferent hardware triggers exist: ? external trigger, either rising or falling or any, detected on the external trigger pin trigger ? pen detect trigger, depending the pendet bit in the mode register (mr) ? continuous trigger, meaning the adcifb restarts the next sequence as soon as it finishes the current one ? periodic trigger, which is defined by the trgr.trgper field ? peripheral event trigger, allowing the peripheral event system to synchronize conversion with some configured peripheral event source. enabling a hardware trigger does not disable the software trigger functionality. thus, if a hard- ware trigger is selected, the star t of a conversion can still be in itiated by the so ftware trigger. 26.8.2 adc mode in the adc mode, the active channels are defi ned by the channel status register (chsr). a channel is enabled by writing a one to the corresponding bit in the channel enable register (cher), and disabled by writing a one to the co rresponding bit in the channel disable register (chdr). the conversion results are stored in the last converted data register (lcdr) as they become available, overwriting old conversions. at each trigger, the following sequence is performed: 1. if acr.sleep is one, wake up the adc and wait for the startup time. 2. if channel 0 is enabled, convert channel 0 and store result in lcdr. 3. if channel 1 is enabled, convert channel 1 and store result in lcdr. 4. if channel n is enabled, convert channel n and store result in lcdr. 5. if acr.sleep is one, place the adc cell in a low-power state. if the peripheral dma controller is enabled, all converted values are transferred continuously into the memory buffer. 26.8.3 resistive touch screen mode writing a one to the tsamod field in the mode register (mr) enables resistive touch screen mode. in this mode the channels tspo+0 to ts po+3, corresponding to the resistive touch
618 32145c?06/2013 at32uc3l0128/256 screen inputs, are automatically activated. in addition, if any other channels are enabled, they will be converted before and after th e resistive touch screen conversion. at each trigger, the following sequence is performed: 1. if acr.sleep is one, wake up the a dc cell and wait for the startup time. 2. convert all enabled channels before t spo and store the results in the lcdr. 3. apply supply on the inputs x p and x m during the sample and hold time. 4. convert channel x m and store the result in tmp. 5. apply supply on the inputs x p and x m during the sample and hold time. 6. convert channel x p , subtract tmp from the result a nd store the subtracted result in lcdr. 7. apply supply on the inputs x p and x m during the sample and hold time. 8. convert channel y p , subtract tmp from the result a nd store the subtracted result in lcdr. 9. apply supply on the inputs y p and y m during the sample and hold time. 10. convert channel y m and store the result in tmp. 11. apply supply on the inputs y p and y m during the sample and hold time. 12. convert channel y p , subtract tmp from the result a nd store the subtracted result in lcdr. 13. apply supply on the inputs y p and y m during the sample and hold time. 14. convert channel x p , subtract tmp from the result a nd store the subtracted result in lcdr. 15. convert all enabled channels after tspo + 3 and store results in the lcdr. 16. if acr.sleep is one, place the adc cell in a low-power state. the resulting buffer structure stored in memory is: 1. x p - x m 2. y p - x m 3. y p - y m 4. x p - y m . the vertical position can be easily calculated by dividing the data at offset 1(x p - x m ) by the data at offset 2(y p - x m ). the horizontal position can be easily calculated by dividing the data at offset 3(y p - y m ) by the data at offset 4(x p - y m ).
619 32145c?06/2013 at32uc3l0128/256 26.9 user interface note: 1. the reset values for these registers are device specific. please refer to the module configuration section at the end of this chapter. table 26-3. adcifb register memory map offset register name access reset 0x00 control register cr write-only - 0x04 mode register mr read/write 0x00000000 0x08 adc configuration register acr read/write 0x00000000 0x0c trigger register trgr read/write 0x00000000 0x10 compare value register cvr read/write 0x00000000 0x14 status register sr read-only 0x00000000 0x18 interrupt status register isr read-only 0x00000000 0x1c interrupt clear register icr write-only - 0x20 interrupt enable register ier write-only - 0x24 interrupt disable register idr write-only - 0x28 interrupt mask register imr read-only 0x00000000 0x2c last converted data register lcdr read-only 0x00000000 0x30 parameter register parameter read-only - (1) 0x34 version register version read-only - (1) 0x40 channel enable register cher write-only - 0x44 channel disable register chdr write-only - 0x48 channel status register chsr read-only 0x00000000
620 32145c?06/2013 at32uc3l0128/256 26.9.1 control register register name: cr access type: write-only offset: 0x00 reset value: 0x00000000 ? dis: adcdifb disable writing a zero to this bit has no effect. writing a one to this bit disables the adcifb. note: disabling the adcifb effectively stop s all clocks in the module so the user must make sure the adcifb is idle before disabling the adcifb. ? en: adcifb enable writing a zero to this bit has no effect. writing a one to this bit enables the adcifb. note: the adcifb must be enabled before use. ? start: start conversion writing a zero to this bit has no effect. writing a one to this bit starts an analog-to-digital conversion. ? swrst: software reset writing a zero to this bit has no effect. writing a one to this bit resets the adcifb, simulating a hardware reset. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 ------disen 76543210 ------startswrst
621 32145c?06/2013 at32uc3l0128/256 26.9.2 mode register name: mr access type: read/write offset: 0x04 reset value: 0x00000000 ? pendbc: pen detect debouncing period period = 2 pendbc *t clk_adc ? tspo: touch screen pin offset the touch screen pin offset field is used to indicate which ad pins are connected to the resistive touch screen film edges. onl y an offset is specified and it is assume d that the resistive touch screen films are connected sequentially from the specified of fset pin and up to and including offset + 3 (4 pins). ? apoe: analog pin output enable 0: ad pins are not used to drive v dd in resistive touch screen sequence. 1: ad pins are used to drive vdd in resistive touch screen sequence. note : if the selected i/o voltage configuration is incompatible wit h the analog-to-digital converter cell voltage specification, th is bit must stay cleared to avoid damaging the adc. in this case the adp pins must be used to drive vdd instead, as described in section 26.7.3 . if the i/o and adc voltages are compatible, the ad pins can be used directly by writing a one to this bit. in this case the adp pins can be ignored. ? ace: analog compare enable 0: the analog compare functionality is disabled. 1: the analog compare functionality is enabled. ? pendet: pen detect 0: the pen detect functionality is disabled. 1: the pen detect functionality is enabled. note : touch detection logic can only be enabled when the adc s equencer is idle. for successful pen detection the user must make sure there is enough idle time between consecutive scans for the touch detection logic to settle. ? tsamod: touch screen adc mode 0: touch screen mode is disabled. 1: touch screen mode is enabled. 31 30 29 28 27 26 25 24 pendbc - - - - 23 22 21 20 19 18 17 16 tspo 15 14 13 12 11 10 9 8 -------- 76543210 - apoe ace pendet - - - tsamod
622 32145c?06/2013 at32uc3l0128/256 26.9.3 adc configuration register name: acr access type: read/write offset: 0x08 reset value: 0x00000000 ? shtim: sample & hold time for adc channels ? startup: startup time ? prescal: prescaler rate selection ? res: resolution selection 0: 8-bit resolution. 1: 10-bit resolution. 2: 11-bit resolution, interpolated. 3: 12-bit resolution, interpolated. ? sleep: adc sleep mode 0: adc sleep mode is disabled. 1: adc sleep mode is enabled. 31 30 29 28 27 26 25 24 ---- shtim 23 22 21 20 19 18 17 16 -startup 15 14 13 12 11 10 9 8 - - prescal 76543210 - - res - - - sleep t sample & hold shtim 2 + ?? t clk _ adc ? = tartup startup 1 + ?? 8 ? t clk _ ad ? = t clk _ adc prescal 1 + ?? 2 ? = t clk _ adcifb ?
623 32145c?06/2013 at32uc3l0128/256 26.9.4 trigger register name: trgr access type: read/write offset: 0x0c reset value: 0x00000000 ? trgper: trigger period effective only if trgmod defines a periodic trigger. defines the periodic trigger period, with the following equations: trigger period = trgper *t clk_adc ? trgmod: trigger mode 31 30 29 28 27 26 25 24 trgper[15:8] 23 22 21 20 19 18 17 16 trgper[7:0] 15 14 13 12 11 10 9 8 -------- 76543210 ----- trgmod table 26-4. trigger modes trgmod selected trigger mode 0 0 0 no trigger, only software trigger can start conversions 0 0 1 external trigger rising edge 0 1 0 external trigger falling edge 0 1 1 external trigger any edge 100 pen detect trigger (shall be selected only if pendet is set and tsamod = touch screen mode) 1 0 1 periodic trigger (trgper shall be initiated appropriately) 1 1 0 continuous mode 1 1 1 peripheral event trigger
624 32145c?06/2013 at32uc3l0128/256 26.9.5 compare value register name: cvr access type: read/write offset: 0x10 reset value: 0x00000000 ? hv: high value defines the high value used when comparing analog input. ? lv: low value defines the low value used when comparing analog input. 31 30 29 28 27 26 25 24 - - - - hv[11:8] 23 22 21 20 19 18 17 16 hv[7:0] 15 14 13 12 11 10 9 8 - - - - lv[11:8] 76543210 lv[7:0]
625 32145c?06/2013 at32uc3l0128/256 26.9.6 status register name: sr access type: read-only offset: 0x14 reset value: 0x00000000 ? en: enable status 0: the adcifb is disabled. 1: the adcifb is enabled. this bit is cleared when cr.dis is written to one. this bit is set when cr.en is written to one. ? celse: compare else status this bit is cleared when either clt or cgt are detected or when analog compare is disabled. this bit is set when no clt or cgt are detected on the last converted data and analog compare is enabled. ? cgt: compare greater than status this bit is cleared when no compare great er than cvr.hv is detected on the last converted data or when analog compare is disabled. this bit is set when compare greater than cvr.hv is detect ed on the last converted data and analog compare is enabled. ? clt: compare lesser than status this bit is cleared when no compare lesser than cvr.lv is detected on the last converted data or when analog compare is disabled. this bit is set when compare lesser than cvr.lv is detected on the last converted data and analog compare is enabled. ? busy: busy status this bit is cleared when the adcifb is ready to perform a conversion sequence. this bit is set when the adcifb is busy performing a convention sequence. ? ready: ready status this bit is cleared when the adcifb is busy performing a conversion sequence this bit is set when the adcifb is ready to perform a conversion sequence. ? nocnt: no contact status this bit is cleared when no contact loss is detected or pen detect is disabled this bit is set when contact loss is detected and pen detect is enabled. ? pencnt: pen contact status this bit is cleared when no contact is detected or pen detect is disabled. 31 30 29 28 27 26 25 24 -------en 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 - celse cgt clt - - busy ready 76543210 - - nocnt pencnt - - ovre drdy
626 32145c?06/2013 at32uc3l0128/256 this bit is set when pen contact is detected and pen detect is enabled. ? ovre: overrun error status this bit is cleared when no overrun error has occurred since the start of a conversion sequence. this bit is set when one or more overrun error has occurred since the start of a conversion sequence. ? drdy: data ready status 0: no data has been converted since the last reset. 1: one or more conversions have completed since the last reset and data is available in lcdr. this bit is cleared when cr.swrst is written to one. this bit is set when one or more conversions have completed and data is available in lcdr.
627 32145c?06/2013 at32uc3l0128/256 26.9.7 interrupt status register name: isr access type: read-only offset: 0x18 reset value: 0x00000000 ? celse: compare else status this bit is cleared when the corresponding bit in icr is written to one. this bit is set when the corresponding bit in sr has a zero-to-one transition. ? cgt: compare greater than status this bit is cleared when the corresponding bit in icr is written to one. this bit is set when the corresponding bit in sr has a zero-to-one transition. ? clt: compare lesser than status this bit is cleared when the corresponding bit in icr is written to one. this bit is set when the corresponding bit in sr has a zero-to-one transition. ? busy: busy status this bit is cleared when the corresponding bit in icr is written to one. this bit is set when the corresponding bit in sr has a zero-to-one transition. ? ready: ready status this bit is cleared when the corresponding bit in icr is written to one. this bit is set when the corresponding bit in sr has a zero-to-one transition. ? nocnt: no contact status this bit is cleared when the corresponding bit in icr is written to one. this bit is set when the corresponding bit in sr has a zero-to-one transition. ? pencnt: pen contact status this bit is cleared when the corresponding bit in icr is written to one. this bit is set when the corresponding bit in sr has a zero-to-one transition. ? ovre: overrun error status this bit is cleared when the corresponding bit in icr is written to one. this bit is set when the corresponding bit in sr has a zero-to-one transition. ? drdy: data ready status this bit is cleared when the corresponding bit in icr is written to one. this bit is set when a conversion has completed and new data is available in lcdr. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 - celse cgt clt - - busy ready 76543210 - - nocnt pencnt - - ovre drdy
628 32145c?06/2013 at32uc3l0128/256 26.9.8 interrupt clear register name: icr access type: write-only offset: 0x1c reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in isr and the corresponding interrupt request. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 - celse cgt clt - - busy ready 76543210 - - nocnt pencnt - - ovre drdy
629 32145c?06/2013 at32uc3l0128/256 26.9.9 interrupt enable register name: ier access type: write-only offset: 0x20 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will set the corresponding bit in imr. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 - celse cgt clt - - busy ready 76543210 - - nocnt pencnt - - ovre drdy
630 32145c?06/2013 at32uc3l0128/256 26.9.10 interrupt disable register name : idr access type: write-only offset: 0x24 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in imr. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 - celse cgt clt - - busy ready 76543210 - - nocnt pencnt - - ovre drdy
631 32145c?06/2013 at32uc3l0128/256 26.9.11 interrupt mask register name: imr access type: read-only offset: 0x28 reset value: 0x00000000 0: the corresponding interrupt is disabled. 1: the corresponding interrupt is enabled. a bit in this register is cleared by writing a one to the corresponding bit in interrupt disable register (idr). a bit in this register is set by writing a one to the corresponding bit in interrupt enable register (ier). 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 - celse cgt clt - - busy ready 76543210 - - nocnt pencnt - - ovre drdy
632 32145c?06/2013 at32uc3l0128/256 26.9.12 last converted data register name: lcdr access type: read-only offset: 0x2c reset value: 0x00000000 ? lcch: last converted channel this field indicates what channel was last conv erted, i.e. what channel the ldata represents. ? ldata: last data converted the analog-to-digital conversion data is placed in this register at the end of a conversion on any analog channel and remains until a new conversion on any analog channel is completed. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 lcch 15 14 13 12 11 10 9 8 ---- ldata[11:8] 76543210 ldata[7:0]
633 32145c?06/2013 at32uc3l0128/256 26.9.13 parameter register name: parameter access type: read-only offset: 0x30 reset value: 0x00000000 ? chn: channel n implemented 0: the corresponding channel is not implemented. 1: the corresponding channel is implemented. 31 30 29 28 27 26 25 24 ch31 ch30 ch29 ch28 ch27 ch26 ch25 ch24 23 22 21 20 19 18 17 16 ch23 ch22 ch21 ch20 ch19 ch18 ch17 ch16 15 14 13 12 11 10 9 8 ch15 ch14 ch13 ch12 ch11 ch10 ch9 ch8 76543210 ch7 ch6 ch5 ch4 ch3 ch2 ch1 ch0
634 32145c?06/2013 at32uc3l0128/256 26.9.14 version register name: version access type: read-only offset: 0x34 reset value: 0x00000000 ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
635 32145c?06/2013 at32uc3l0128/256 26.9.15 channel enable register name: cher access type: write-only offset: 0x40 reset value: 0x00000000 ? chn: channel n enable writing a zero to a bit in this register has no effect writing a one to a bit in this register enables the corresponding channel the number of available channels is device dependent. please re fer to the module configuration section at the end of this chapter for information regardin g which channels are implemented. 31 30 29 28 27 26 25 24 ch31 ch30 ch29 ch28 ch27 ch26 ch25 ch24 23 22 21 20 19 18 17 16 ch23 ch22 ch21 ch20 ch19 ch18 ch17 ch16 15 14 13 12 11 10 9 8 ch15 ch14 ch13 ch12 ch11 ch10 ch9 ch8 76543210 ch7 ch6 ch5 ch4 ch3 ch2 ch1 ch0
636 32145c?06/2013 at32uc3l0128/256 26.9.16 channel disable register name: chdr access type: write-only offset: 0x44 reset value : 0x00000000 ? chn: channel n disable writing a zero to a bit in this register has no effect. writing a one to a bit in this register disables the corresponding channel. warning: if the corresponding channel is disabled during a conversion, or if it is disabled and then re-enabled during a conversion, its associated data and its corresponding drdy and ovre bits in sr are unpredictable. the number of available channels is device dependent. please re fer to the module configuration section at the end of this chapter for information regarding how many channels are implemented. 31 30 29 28 27 26 25 24 ch31 ch30 ch29 ch28 ch27 ch26 ch25 ch24 23 22 21 20 19 18 17 16 ch23 ch22 ch21 ch20 ch19 ch18 ch17 ch16 15 14 13 12 11 10 9 8 ch15 ch14 ch13 ch12 ch11 ch10 ch9 ch8 76543210 ch7 ch6 ch5 ch4 ch3 ch2 ch1 ch0
637 32145c?06/2013 at32uc3l0128/256 26.9.17 channel status register name: chsr access type: read-only offset: 0x48 reset value: 0x00000000 ? chn: channel n status 0: the corresponding channel is disabled. 1: the corresponding channel is enabled. a bit in this register is cleared by writing a one to the corresponding bit in channel disable register (chdr). a bit in this register is set by writing a one to the corresponding bit in channel enable register (cher). the number of available channels is device dependent. please re fer to the module configuration section at the end of this chapter for information regarding how many channels are implemented. 31 30 29 28 27 26 25 24 ch31 ch30 ch29 ch28 ch27 ch26 ch25 ch24 23 22 21 20 19 18 17 16 ch23 ch22 ch21 ch20 ch19 ch18 ch17 ch16 15 14 13 12 11 10 9 8 ch15 ch14 ch13 ch12 ch11 ch10 ch9 ch8 76543210 ch7 ch6 ch5 ch4 ch3 ch2 ch1 ch0
638 32145c?06/2013 at32uc3l0128/256 26.10 module configuration the specific configuration for each adcifb inst ance is listed in the following tables.the module bus clocks listed here are connected to the syst em bus clocks. please refer to the power man- ager chapter for details. note: 1. ad3 does not exist table 26-5. module configuration feature adcifb number of adc channels 9 (8 + 1 internal temperature sensor channel) table 26-6. adcifb clocks clock name description clk_adcifb clock for the adcifb bus interface table 26-7. register reset values register reset value version 0x00000110 parameter 0x000003ff table 26-8. adc input channels (1) channel input ch0 ad0 ch1 ad1 ch2 ad2 ch4 ad4 ch5 ad5 ch6 ad6 ch7 ad7 ch8 ad8 ch9 temperature sensor
639 32145c?06/2013 at32uc3l0128/256 27. analog comparator interface (acifb) rev: 2.0.2.2 27.1 features ? controls an array of analog comparators ? low power option ? single shot mode support ? selectable settings for filter option ? filter length and hysteresis ? window mode ? detect inside/outside window ? detect above/below window ? interrupt ? on comparator result rising edge, falling edge, toggle ? inside window, outside window, toggle ? when startup time has passed ? can generate events to the peripheral event system 27.2 overview the analog comparator interface (acifb) is able to control a number of analog comparators (ac) with identical behavior. an analog compar ator compares two voltages and gives a com- pare output depending on this comparison. the acifb can be configured in normal mode using each comparator independently or in win- dow mode using defined comparator pairs to observe a window. the number of channels implemented is device specific. refer to the module configuration sec- tion at the end of this chapter for details.
640 32145c?06/2013 at32uc3l0128/256 27.3 block diagram figure 27-1. acifb block diagram 27.4 i/o lines description there are two groups of analog comparators, a and b, as shown in table 27-1 . in normal mode, this grouping does not have any meaning. in window mode, two analog comparators, one from group a and the corresponding comparator from group b, are paired. ?????. . . trigger events irq gclk peripheral bus acifb analog comparators peripheral event generation - + ac inn inp conf0.inseln - + ac inn inp confn.inseln filter filter interrupt generation clk_acifb ctrl.actest tr.actestn tr.actest0 acout0 acoutn a c p 0 ac n 0 a c r e f n a c p n acnn table 27-1. analog comparator groups for window mode group a group b pair number ac0 ac1 0 ac2 ac3 1 ac4 ac5 2 ac6 ac7 3 table 27-2. i/o line description pin name pin description type acapn positive reference pin for analog comparator a n analog acann negative reference pin for analog comparator a n analog
641 32145c?06/2013 at32uc3l0128/256 the signal names corresponds to the groups a and b of analog comparators. for normal mode, the mapping from input signal names in the bl ock diagram to the signal names is given in table 27-3 . 27.5 product dependencies in order to use this module, other parts of the system must be configured correctly, as described below. 27.5.1 i/o lines the acifb pins are multiplexed with other peripherals. the user must first program the i/o con- troller to give control of the pins to the acifb. 27.5.2 power management if the cpu enters a sleep mode that disables cl ocks used by the acifb, the acifb will stop functioning and resume operation after the system wakes up from sleep mode. 27.5.3 clocks the clock for the acifb bus interface (clk_acifb) is generated by the power manager. this clock is enabled at reset, and can be disabled in the power manager. it is recommended to dis- able the acifb before disabling the clock, to avoid freezing the acifb in an undefined state. the acifb uses a gclk as clock source for the analog comparators. the user must set up this gclk at the right frequency. the clk_acifb cloc k of the interface must be at least 4x the gclk frequency used in the comparators. the gc lk is used both for measuring the startup time of a comparator, and to give a frequency for the comparisons done in continuous measure- ment mode, see section 27.6 . refer to the electrical characteristics chapter for gclk frequency limitations. acbpn positive reference pin for analog comparator b n analog acbnn negative reference pin for analog comparator b n analog acrefn reference voltage for all com parators selectable for inn analog table 27-3. signal name mapping pin name channel number normal mode acap0/acan0 0 acp0/acn0 acbp0/acbn0 1 acp1/acn1 acap1/acan1 2 acp2/acn2 acbp1/acbn1 3 acp3/acn3 acap2/acan2 4 acp4/acn4 acbp2/acbn2 5 acp5/acn5 acap3/acan3 6 acp6/acn6 acbp3/acbn3 7 acp7/acn7 table 27-2. i/o line description pin name pin description type
642 32145c?06/2013 at32uc3l0128/256 27.5.4 interrupts the acifb interrupt request line is connected to the interrupt controller. using the acifb inter- rupt requires the interrupt controller to be programmed first. 27.5.5 peripheral events the acifb peripheral events are connected via the peripheral event system. refer to the peripheral event system chapter for details. 27.5.6 debug operation when an external debugger forces the cpu into debug mode, the acifb continues normal operation. if the acifb is configured in a way that requires it to be periodically serviced by the cpu through interrupts or similar, improper operation or data loss may result during debugging. 27.6 functional description the acifb is enabled by writing a one to the co ntrol register enable bi t (ctrl.en). addition- ally, the comparators must be individually enabled by programming the mode field in the ac configuration register (confn.mode). the results from the individual comparators can either be used directly (normal mode), or the results from two comparators can be grouped to generate a comparison window (window mode). all comparators need not be in the same mode, some comparators may be in normal mode, while others are in window mode. there are restrictions on which ac channels that can be grouped together in a window pair, see section 27.6.5 . 27.6.1 analog comparator operation each ac channel can be in one of four different modes, determined by confn.mode: ?off ? continuous measurement mode (cm) ? user triggered single measurement mode (ut) ? event triggered single measurement mode (et) after being enabled, a startup time defined in ctrl.sut is required before the result of the comparison is ready. the gclk is used for measuring the startup time of a comparator, during the startup time the ac output is not available. when the acn ready bit in the status register (sr.acrdyn) is one, the output of ac n is ready. in window mode the result is avail- able when both the compar ator outputs are ready (sr. acrdyn=1 and sr.acrdyn+1=1). 27.6.1.1 continuous measurement mode in cm, the analog comparator is continuously enabled and performing comparisons. this ensures that the result of the latest comparison is always available in the acn current compari- son status bit in the status register (sr. accsn). comparisons are done on every positive edge of gclk. cm is enabled by writing confn. mode to 1. after the startup time has passed, a comparison is done and sr is updated. appropriate peripheral events and interrupts are also generated. new comparisons are performed continuously until the confn.mode field is written to 0.
643 32145c?06/2013 at32uc3l0128/256 27.6.1.2 user triggered single measurement mode in the ut mode, the user starts a single comparison by writing a one to the user start single comparison bit (ctrl.ustart). th is mode is enabled by writin g confn.mode to 2. after the startup time has passed, a single comparison is done and sr is updated. appropriate peripheral events and interrupts are also generated. no new comparisons will be performed. ctrl.ustart is cleared automatically by hardware when the single comparison has been done. 27.6.1.3 event triggered single measurement mode this mode is enabled by writing confn.mode to 3 and peripheral event trigger enable (ctrl.eventen) to one. the et mode is similar to the ut mode, the difference is that a peripheral event from another hardware module causes the hardware to automatically set the peripheral event start single comparison bit (ctrl.estart). after the startup time has passed, a single comparison is done and sr is updated. appropriate peripheral events and interrupts are also generated. no new compar isons will be perf ormed. ctrl.estart is cleared automatically by hardware when the single comparison has been done. 27.6.1.4 selecting comparator inputs each analog comparator has one positive (i np) and one negative (inn) input. the positive input is fed from an external input pin (acp n). the negative input can either be fed from an external input pin (acnn) or from a reference voltage common to all acs (acrefn). the user selects the in put source as follows: ? in normal mode with the negative input select and positive input select fields (confn.inseln and confn.inselp). ? in window mode with confn.inseln, confn.inselp and confn+1.inseln, confn+1,inselp. the user must configure confn.inseln and confn+1.inselp to the same source. 27.6.2 interrupt generation the interrupt request will be generated if the corresponding bit in the interrupt mask register (imr) is set. bits in imr are set by writing a one to the corresponding bit in the interrupt enable register (ier), and cleared by writing a one to the corresponding bit in the interrupt disable register (idr). the interrupt request remains acti ve until the corresponding bit in isr is cleared by writing a one to the corresponding bit in the interrupt status clear register (icr). 27.6.3 peripheral event generation the acifb can be set up so that certain comparison results notify other parts of the device via the peripheral event system. refer to section 27.6.4.3 and section 27.6.5.3 for information on which comparison results can generate events, and how to configure the acifb to achieve this. zero or one event will be generated pe r comparison. 27.6.4 normal mode in normal mode all analog comparators are operating independently. 27.6.4.1 normal mode output each analog comparator generates one output acout according to the input voltages on inp (ac positive input) and inn (ac negative input):
644 32145c?06/2013 at32uc3l0128/256 ? acout = 1 if v inp > v inn ? acout = 0 if v inp < v inn ? acout = 0 if the ac output is not available (sr.acrdy = 0) the output can optionally be filtered, as described in section 27.6.6 . 27.6.4.2 normal mode interrupt the ac channels can generate interrupts. the inte rrupt settings field in the configuration regis- ter (confn.is) can be configured to select when the ac will generate an interrupt: ? when v inp > v inn ? when v inp < v inn ? on toggle of the ac output (acout) ? when comparison has been done 27.6.4.3 normal mode peripheral events the acifb can generate peripheral events according to the configuration of confn.evenn and confn.evenp. ? when v inp > v inn or ? when v inp < v inn or ? on toggle of the ac output (acout) 27.6.5 window mode in window mode, two acs (an even and the following odd build up a pair) are grouped. the negative input of acn (even) and the positive input of acn+1 (odd) has to be connected together externally to the device and are controlled by the input select fields in the ac configu- ration registers (confn.inseln and confn+1.in selp). the positive input of acn (even) and the negative inpu t of acn+1 (odd) can still be configur ed independently by confn.inselp and confn+1.inseln, respectively.
645 32145c?06/2013 at32uc3l0128/256 figure 27-2. analog comparator interface in window mode 27.6.5.1 window mode output when operating in window mode, each channel generates the same acout outputs as in nor- mal mode, see section 27.6.4.1 . additionally, the acifb generates a window mode signal (acwout) according to the common input voltage to be compared: ? acwout = 1 if the common input voltage is inside the window, v acn(n+1) < v common < v acp(n) ? acwout = 0 if the common input voltage is outside the window, v common < v acn(n+1) or v common > v acp(n) ? acwout = 0 if the window mode output is not available (sr.acrdy n =0 or sr.acrdy n+1=0) 27.6.5.2 window mode interrupts when operating in window mode, each channel can generate the same interrupts as in normal mode, see section 27.6.4.2 . additionally, when channels operate in window mode, programming window mode interrupt set- tings in the window mode configuration register (confwn.wis) can cause interrupts to be generated when: ? as soon as the common inpu t voltage is inside the window. ? as soon as the common input voltage is outside the window. ? on toggle of the window compare output (acwout). ? when the comparison in both channels in the window pair is ready. comparator pair 0 - + ac0 interrupt generator window module acout0 peripheral event generator window window event - + ac1 filter filter sr.accs0 sr.wfcs0 acap0 acan0 acbp0 common acwout acbn0 irq acout1
646 32145c?06/2013 at32uc3l0128/256 27.6.5.3 window mode peripheral events when operating in window mode, each channel ca n generate the same peripheral events as in normal mode, see section 27.6.4.3 . additionally, when channels operate in window mode, programming window mode event selec- tion source (confwn.wevsrc) can cause peri pheral events to be gene rated when: ? as soon as the common inpu t voltage is inside the window. ? as soon as the common input voltage is outside the window. ? on toggle of the window compare output (acwout) ? whenever a comparison is ready and the common input voltage is inside the window. ? whenever a comparison is ready and the common input voltage is outside the window. ? when the comparison in both channels in the window pair is ready. 27.6.6 filtering the output of the comparator can be filtered to reduce noise. the filter length is determined by the filter length field in the confn register (confn.flen). the filter samples the analog comparator output at the gclk frequency for 2 confn.flen samples. a separa te counter (cnt) counts the number of cycles the ac output was one. this filter is deactivated if confn.flen equals 0. if the filter is enabled, the hysteresis value field hys in the confn register (confn.hys) can be used to define a hysteresis value. the hysteresis value should be chosen so that: the filter function is defined by: the filtering algorithm is explained in figure 27-3 . 2 flen measurements are sampled. if the num- ber of measurements that are zero is less than (2 flen /2 - hys), the filtered result is zero. if the number of measurements that are one is more than (2 flen /2 + hys), the filtered result is one. otherwise, the result is unchanged. 2 flen 2 ---------------- hys ? cnt 2 flen 2 --------------- - hys + ?? ?? comp ? ? 1 = 2 flen 2 ---------------- hys + ?? ?? cnt 2 flen 2 --------------- - h ? ys ?? ?? ? ? comp unchanged ? cnt 2 flen 2 ---------------- h ? ys ?? ?? comp ? ? 0 =
647 32145c?06/2013 at32uc3l0128/256 figure 27-3. the filtering algorithm 27.7 peripheral event triggers peripheral events from other modules can trigger comparisons in the acifb. all channels that are set up in event triggered single measurement mode will be started si multaneously when a peripheral event is received. channels that are operating in continuous measurement mode or user triggered single measurement mode will be unaffected by the received event. the soft- ware can still operate these channels indepen dently of channels in event triggered single measurement mode. a peripheral event will trigger one or more comparisons, in normal or window mode. 27.8 ac test mode by writing the analog comparator test mode (cr.actest) bit to one, the outputs from the acs are overridden by the value in the test register (tr), see figure 27-1 . this is useful for software development. 2 flen 2 flen 2 hys hys ?result=0" ?result=1" result = unchanged 0
648 32145c?06/2013 at32uc3l0128/256 27.9 user interface note: 1. the reset values for these registers are device specific. please refer to the module configuration section at the end of this chapter. table 27-4. acifb register memory map offset register register name access reset 0x00 control register ctrl read/write 0x00000000 0x04 status register sr read-only 0x00000000 0x10 interrupt enable register ier write-only 0x00000000 0x14 interrupt disable register idr write-only 0x00000000 0x18 interrupt mask register imr read-only 0x00000000 0x1c interrupt status register isr read-only 0x00000000 0x20 interrupt status clear register icr write-only 0x00000000 0x24 test register tr read/write 0x00000000 0x30 parameter register parameter read-only - (1) 0x34 version register version read-only - (1) 0x80 window0 configuration register confw0 read/write 0x00000000 0x84 window1 configuration register confw1 read/write 0x00000000 0x88 window2 configuration register confw2 read/write 0x00000000 0x8c window3 configuration regi ster confw3 read/write 0x00000000 0xd0 ac0 configuration register conf0 read/write 0x00000000 0xd4 ac1 configuration register conf1 read/write 0x00000000 0xd8 ac2 configuration register conf2 read/write 0x00000000 0xdc ac3 configuration register conf3 read/write 0x00000000 0xe0 ac4 configuration register conf4 read/write 0x00000000 0xe4 ac5 configuration register conf5 read/write 0x00000000 0xe8 ac6 configuration register conf6 read/write 0x00000000 0xec ac7 configuration register conf7 read/write 0x00000000
649 32145c?06/2013 at32uc3l0128/256 27.9.1 control register name: ctrl access type: read/write offset: 0x00 reset value: 0x00000000 ? sut: startup time analog comparator startup time = . each time an ac is enabled, the ac comparison will be enabled after the startup time of the ac. ? actest: analog comparator test mode 0: the analog comparator outputs feeds the channel logic in acifb. 1: the analog comparator outputs ar e bypassed with the ac test register. ? estart: peripheral event start single comparison writing a zero to this bit has no effect. writing a one to this bit starts a comparison and can be used for test purposes. this bit is cleared when comparison is done. this bit is set when an enabled peripheral event is received. ? ustart: user start single comparison writing a zero to this bit has no effect. writing a one to this bit starts a single measurement mode comparison. this bit is cleared when comparison is done. ? eventen: peripheral event trigger enable 0: a peripheral event will not trigger a comparison. 1: enable comparison triggered by a peripheral event. ? en: acifb enable 0: the acifb is disabled. 1: the acifb is enabled. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ------ sut[9:8] 15 14 13 12 11 10 9 8 sut[7:0] 76543210 actest - estart ustart - - -eventen en sut f gclk ---------------- -
650 32145c?06/2013 at32uc3l0128/256 27.9.2 status register name: sr access type: read-only offset: 0x04 reset value: 0x00000000 ? wfcsn: window mode current status this bit is cleared when the common input voltage is outside the window. this bit is set when the common input voltage is inside the window. ? acrdyn: acn ready this bit is cleared when the ac output (acout) is not ready. this bit is set when the ac output (acout) is ready, ac is enabled and its startup time is over. ? accsn: acn current comparison status this bit is cleared when v inp is currently lower than v inn this bit is set when v inp is currently greater than v inn. 31 30 29 28 27 26 25 24 ----wfcs3wfcs2wfcs1wfcs0 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 acrdy7 accs7 acrdy6 accs6 acrdy5 accs5 acrdy4 accs4 76543210 acrdy3 accs3 acrdy2 accs2 acrdy1 accs1 acrdy0 accs0
651 32145c?06/2013 at32uc3l0128/256 27.9.3 interrupt enable register name: ier access type: write-only offset: 0x10 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will set the corresponding bit in imr. 31 30 29 28 27 26 25 24 - - - - wfint3 wfint2 wfint1 wfint0 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 sutint7 acint7 sutint6 acint6 sutint5 acint5 sutint4 acint4 76543210 sutint3 acint3 sutint2 acint2 sutint1 acint1 sutint0 acint0
652 32145c?06/2013 at32uc3l0128/256 27.9.4 interrupt disable register name: idr access type: write-only offset: 0x14 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in imr. 31 30 29 28 27 26 25 24 - - - - wfint3 wfint2 wfint1 wfint0 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 sutint7 acint7 sutint6 acint6 sutint5 acint5 sutint4 acint4 76543210 sutint3 acint3 sutint2 acint2 sutint1 acint1 sutint0 acint0
653 32145c?06/2013 at32uc3l0128/256 27.9.5 interrupt mask register name: imr access type: read-only offset: 0x18 reset value: 0x00000000 ?wfint n: window mode interrupt mask 0: the corresponding interrupt is disabled. 1: the corresponding interrupt is enabled. this bit is cleared when the corresponding bit in idr is written to one. this bit is set when the corresponding bit in ier is written to one. ?sutint n: ac n startup time interrupt mask 0: the corresponding interrupt is disabled. 1: the corresponding interrupt is enabled. this bit is cleared when the corresponding bit in idr is written to one. this bit is set when the corresponding bit in ier is written to one. ? acint n: ac n interrupt mask 0: the corresponding interrupt is disabled. 1: the corresponding interrupt is enabled. this bit is cleared when the corresponding bit in idr is written to one. this bit is set when the corresponding bit in ier is written to one. 31 30 29 28 27 26 25 24 - - - - wfint3 wfint2 wfint1 wfint0 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 sutint7 acint7 sutint6 acint6 sutint5 acint5 sutint4 acint4 76543210 sutint3 acint3 sutint2 acint2 sutint1 acint1 sutint0 acint0
654 32145c?06/2013 at32uc3l0128/256 27.9.6 interrupt status register name: isr access type: read-only offset: 0x1c reset value: 0x00000000 ? wfintn: window mode interrupt status 0: no window mode interrupt is pending. 1: window mode interrupt is pending. this bit is cleared when the corresponding bit in icr is written to one. this bit is set when the corresponding channel pair operating in window mode generated an interrupt. ? sutintn: acn startup time interrupt status 0: no startup time interrupt is pending. 1: startup time interrupt is pending. this bit is cleared when the corresponding bit in icr is written to one. this bit is set when the startup time of the corresponding ac has passed. ? acintn: acn interrupt status 0: no normal mode interrupt is pending. 1: normal mode interrupt is pending. this bit is cleared when the corresponding bit in icr is written to one. this bit is set when the corresponding channel generated an interrupt. 31 30 29 28 27 26 25 24 - - - - wfint3 wfint2 wfint1 wfint0 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 sutint7 acint7 sutint6 acint6 sutint5 acint5 sutint4 acint4 76543210 sutint3 acint3 sutint2 acint2 sutint1 acint1 sutint0 acint0
655 32145c?06/2013 at32uc3l0128/256 27.9.7 interrupt status clear register name: icr access type: write-only offset: 0x20 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in isr and the corresponding interrupt request. 31 30 29 28 27 26 25 24 - - - - wfint3 wfint2 wfint1 wfint0 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 sutint7 acint7 sutint6 acint6 sutint5 acint5 sutint4 acint4 76543210 sutint3 acint3 sutint2 acint2 sutint1 acint1 sutint0 acint0
656 32145c?06/2013 at32uc3l0128/256 27.9.8 test register name: tr access type: read/write offset: 0x24 reset value: 0x00000000 ? actestn: ac output override value if ctrl.actest is set, the acn output is overridden with the value of actestn. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 actest7 actest6 actest5 actest4 a ctest3 actest2 actest1 actest0
657 32145c?06/2013 at32uc3l0128/256 27.9.9 parameter register name: parameter access type: read-only offset: 0x30 reset value: - ? wimpln: window pair n implemented 0: window pair not implemented. 1: window pair implemented. ? acimpln: analog comparator n implemented 0: analog comparator not implemented. 1: analog comparator implemented. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 - - - - wimpl3 wimpl2 wimpl1 wimpl0 15 14 13 12 11 10 9 8 -------- 76543210 acimpl7 acimpl6 acimpl5 acimpl4 acimpl3 acimpl2 acimpl1 acimpl0
658 32145c?06/2013 at32uc3l0128/256 27.9.10 version register name: version access type: read-only offset: 0x34 reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
659 32145c?06/2013 at32uc3l0128/256 27.9.11 window configuration register name: confwn access type: read/write offset: 0x80,0x84,0x88,0x8c reset value: 0x00000000 ? wfen: window mode enable 0: the window mode is disabled. 1: the window mode is enabled. ? weven: window event enable 0: event from awout is disabled. 1: event from awout is enabled. ? wevsrc: event source sel ection for window mode 000: event on acwout rising edge. 001: event on acwout falling edge. 010: event on awout rising or falling edge. 011: inside window. 100: outside window. 101: measure done. 110-111: reserved. ? wis: window mode interrupt settings 00: window interrupt as soon as the input voltage is inside the window. 01: window interrupt as soon as the input voltage is outside the window. 10: window interrupt on toggle of window compare output. 11: window interrupt when evaluation of input voltage is done. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------wfen 15 14 13 12 11 10 9 8 - - - - weven wevsrc 7654321 0 ------ wis
660 32145c?06/2013 at32uc3l0128/256 27.9.12 ac configuration register name: confn access type: read/write offset: 0xd0,0xd4,0xd8,0xdc,0xe0,0xe4,0xe8,0xec reset value: 0x00000000 ? flen: filter length 000: filter off. n: number of samples to be averaged =2 n . ? hys: hysteresis value 0000: no hysteresis. 1111: max hysteresis. ? evenn: event enable negative 0: do not output event when acout is zero. 1: output event when acout is zero. ? evenp: event enable positive 0: do not output event when acout is one. 1: output event when acout is one. ? inselp: positive input select 00: acpn pin selected. 01: reserved. 10: reserved. 11: reserved. ? inseln: negative input select 00: acnn pin selected. 01: acrefn pin selected. 10: reserved. 11: reserved. ?mode: mode 00: off. 01: continuous measurement mode. 10: user triggered single measurement mode. 11: event triggered single measurement mode. 31 30 29 28 27 26 25 24 -flen hys 23 22 21 20 19 18 17 16 - - - - - - evenp evenn 15 14 13 12 11 10 9 8 - - - - inselp inseln 7654321 0 -- mode -- is
661 32145c?06/2013 at32uc3l0128/256 ? is: interrupt settings 00: comparator interrupt when as v inp > v inn . 01: comparator interrupt when as v inp < v inn . 10: comparator interrupt on toggle of analog comparator output. 11: comparator interrupt when comparison of v inp and v inn is done.
662 32145c?06/2013 at32uc3l0128/256 27.10 module configuration the specific configuration for each acifb instan ce is listed in the fo llowing tables.the module bus clocks listed here are conn ected to the system bus clocks. refer to the power manager chapter for details. table 27-5. acifb configuration feature acifb number of channels 8 table 27-6. acifb clocks clock name description clk_acifb clock for the acifb bus interface gclk the generic clock used for the acifb is gclk4 table 27-7. register reset values register reset value version 0x00000202 parameter 0x000f00ff
663 32145c?06/2013 at32uc3l0128/256 28. capacitive touch module (cat) rev: 4.0.0.0 28.1 features ? qtouch? method allows n touch sensors to be implemented using 2n physical pins ? qmatrix method allows x by y matrix of sensor s to be implemented using (x+2y) physical pins ? one autonomous qtouch sensor operates without dma or cpu intervention ? all qtouch sensors can operate in dma- driven mode without cpu intervention ? external synchronization to reduce 50 or 60 hz mains interference ? spread spectrum sensor drive capability 28.2 overview the capacitive touch module (cat) senses touc h on external capacitive touch sensors. capac- itive touch sensors use no external mechan ical components, and therefore demand less maintenance in the user application. the module implements the qtouch method of capturing signals from capacitive touch sensors. the qtouch method is generally suitable for small numbers of sensors since it requires 2 physi- cal pins per sensor. the module also implements the qmatrix method, which is more appropriate for large numbers of sensors since it allows an x by y matrix of sensors to be imple- mented using only (x+2y) physical pins. the module allows methods to function together, so n touch sensors and an x by y matrix of sensors can be implemented us ing (2n+x+2y) physical pins. in addition, the module allows sensors using the qtouch method to be divided into two groups. each qtouch group can be configured with different properties. this eases the implementation of multiple kinds of controls such as push buttons, wheels, and sliders. all of the qtouch sensors can operate in a dma-driven mode, known as dmatouch, that allows detection of touch without cpu intervention. the module also implements one autonomous qtouch sensor that is capable of detecting touch without dma or cpu intervention. this allows proximity or activation detection in low-power sleep modes.
664 32145c?06/2013 at32uc3l0128/256 28.3 block diagram figure 28-1. cat block diagram 28.4 i/o lines description interface registers peripheral bus finite state machine capacitor charge and discharge sequence generator counters csan smp i/o controller pins discharge current sources dis yn analog comparators peripheral event system clk_cat analog comparator interface sync capacitive touch module (cat) csbn gclk_cat vdiven note: italicized signals and blocks are used only for qmatrix operation table 28-1. i/o lines description name description type csan capacitive sense a line n i/o csbn capacitive sense b line n i/o dis discharge current control (only used for qmatrix) analog
665 32145c?06/2013 at32uc3l0128/256 28.5 product dependencies in order to use the cat module, other parts of the system must be configured correctly, as described below. 28.5.1 i/o lines the cat pins may be multiplexed with other peripherals. the user must first program the i/o controller to give control of the pins to the cat module. in qmatrix mode, the y lines must be driven by the cat and analog comparators sense the voltage on the y lines. thus, the cat (not the analog comparator interface) must be the selected function for the y lines in the i/o controller. by writing ones and zeros to bits in the pin mode registers (pinmodex), most of the cat pins can be individually selected to implement the qtouch method or the qmatrix method. each pin has a different name and function depending on whether it is implementing the qtouch method or the qmatrix method. the following table shows the pin names for each method and the bits in the pinmodex registers which control the selection of the qtouch or qmatrix method. smp smp line (only used for qmatrix) output sync synchronize signal input vdiven voltage divider enable (only used for qmatrix) output table 28-1. i/o lines description name description type table 28-2. pin selection guide cat module pin name qtouch method pin name qmatrix method pin name selection bit in pinmodex register csa0 sns0 x0 sp0 csb0 snsk0 x1 sp0 csa1 sns1 y0 sp1 csb1 snsk1 yk0 sp1 csa2 sns2 x2 sp2 csb2 snsk2 x3 sp2 csa3 sns3 y1 sp3 csb3 snsk3 yk1 sp3 csa4 sns4 x4 sp4 csb4 snsk4 x5 sp4 csa5 sns5 y2 sp5 csb5 snsk5 yk2 sp5 csa6 sns6 x6 sp6 csb6 snsk6 x7 sp6 csa7 sns7 y3 sp7 csb7 snsk7 yk3 sp7 csa8 sns8 x8 sp8
666 32145c?06/2013 at32uc3l0128/256 28.5.2 clocks the clock for the cat module, clk_cat, is gen erated by the power manager (pm). this clock is turned on by default, and can be enabled and disabled in the pm. the user must ensure that clk_cat is enabled befo re using the cat module. qmatrix operations also require the cat generic clock, gclk_cat. this generic clock is gener- ated by the system control interface (scif), and is shared between the cat and the analog comparator interface. the user must ensure that the gclk_cat is enabled in the scif before using qmatrix functionality in the cat module. for proper qmatrix operation, the frequency of gclk_cat must be less than half the frequency of clk_cat. if only qtouch functionality is used, then gclk_cat is unnecessary. 28.5.3 interrupts the cat interrupt request line is connected to the interrupt controller. using cat interrupts requires the interrupt controller to be programmed first. 28.5.4 peripheral events the cat peripheral events are connected via the peripheral event system. refer to the periph- eral event system chapter for details. 28.5.5 peripheral direct memory access the cat module provides handshake capability for a peripheral dma controller. one hand- shake controls transfers from the acquired count register (acount) to memory. a second handshake requests burst lengths for each (x,y ) pair to the matrix burst length register csb8 snsk8 x9 sp8 csa9 sns9 y4 sp9 csb9 snsk9 yk4 sp9 csa10 sns10 x10 sp10 csb10 snsk10 x11 sp10 csa11 sns11 y5 sp11 csb11 snsk11 yk5 sp11 csa12 sns12 x12 sp12 csb12 snsk12 x13 sp12 csa13 sns13 y6 sp13 csb13 snsk13 yk6 sp13 csa14 sns14 x14 sp14 csb14 snsk14 x15 sp14 csa15 sns15 y7 sp15 csb15 snsk15 yk7 sp15 csa16 sns16 x16 sp16 csb16 snsk16 x17 sp16 table 28-2. pin selection guide cat module pin name qtouch method pin name qmatrix method pin name selection bit in pinmodex register
667 32145c?06/2013 at32uc3l0128/256 (mblen) when using the qmatrix acquisition me thod. two additional handshakes support dma- touch by regulating transfers from memory to the dmatouch state write register (dmatsw) and from the dmatouch state read register (dmatsr) to memory. the peripheral dma con- troller must be configured properly and enabled in order to perform direct memory access transfers to/from the cat module. 28.5.6 analog comparators when the cat module is performing qmatrix acquisiti on, it requires that on-chip analog compar- ators be used as part of the process. these analog comparators are not controlled directly by the cat module, but by a separate analog comparator (ac) interface. this interface must be configured properly and enabled before the cat module is used. this includes configuring the generic clock input for the analog comparators to the proper sampling frequency. the cat will automatically use the ne gative peripheral events from the ac interface on every y pin in qmatrix mode. when qmatrix acquisition is used the analog comparator corresponding to the selected y pins must be enabled and converting continuously, using the y pin as the positive reference and the acrefn as negative reference. 28.5.7 debug operation when an external debugger forces the cpu into debug mode, the cat continues normal opera- tion. if the cat is configured in a way that requires it to be periodically serviced by the cpu through interrupts or similar, improper operation or data loss may result during debugging. 28.6 functional description 28.6.1 acquisition types the cat module can perform several types of qtouch acquisition from capacitive touch sen- sors: autonomous qtouch (one sensor only), dmatouch, qtouch group a, and qtouch group b. the cat module can also perform qmatrix acquisition. each type of acquisition has an asso- ciated set of pin selection and configuration registers that a llow a large degre e of flexibility. the following schematic diagrams show typica l hardware connections for qtouch and qmatrix sensors, respectively: figure 28-2. cat touch connections avr32 chip qtouch sensor cs (sense capacitor) snskn snsn
668 32145c?06/2013 at32uc3l0128/256 figure 28-3. cat matrix connections in order to use the autonomous qtouch detection capability, the user must first set up the autonomous touch pin select register (atpins) and autonomous/dma touch configuration registers (atcfg0 through 3) with appropriate values. the module can then be enabled using the control register (ctr l). after the module is enabled, th e module will acquire data from the autonomous qtouch sensor and use it to dete rmine whether the sensor is activated. the active/inactive status of the autonomous qtouch sensor is reported in the status register (sr), and it is also possible to configure the cat to generate an interrupt whenever the status changes. the module will continue acquiring autonomous qtouch sensor data and updating autonomous qtouch status until the module is disabled or reset. in order to use the dmatouch capability, it is first necessary to set up the pin mode registers (pinmode0, pinmode1, and pinmode2) so that the desired pins are specified as dma- touch. the autonomous/dma touch configurati on registers (atcfg0 through 3) must also be configured with appropriate values. one channel of the peripheral dma controller must be set up to transfer state words from a block of memory to the dmatsw register, and another chan- nel must be set up to transfer state words from the dmatsr register back to the same block of memory. the module can then be enabled using the ctrl register. after the module is enabled, the module will acquire count values from each dmatouch sensor. once the module has acquired a count value for a sensor, it will use a handshake interface to signal the peripheral dma controller to transfer a stat e word to the dmatsw register. the modu le will use the count value to update the state word , and then the updated state wo rd will be transferred to the dmatsr register. another handshake interface will signal the peripheral dma controller to transfer the contents of the dmatsr register back to memory. the status of the dmatouch sensors can be determined at any time by reading the dmatouch sensor status register (dmatss). avr32 chip cs0 (sense capacitor) x3 yk0 qmatrix sensor array x6 x7 x2 y0 yk1 y1 cs1 (sense capacitor) smp rsmp1 rsmp0 vdiven dis rdis acrefn ra rb note: if the cat internal current sources will be enabled, the smp signal and rsmp resistors should not be included in the design. if the cat internal current sources will not be enabled, the dis signal and rdis resistor should not be included in the design.
669 32145c?06/2013 at32uc3l0128/256 in order to use the qmatrix, qtouch group a, or qtouch group b acquisi tion capabilities, it is first necessary to set up the pin mode regi sters (pinmode0, pinmode1, and pinmode2) and configuration registers (mgcfg0, mgcfg1, tgacfg0, tgacfg1, tgbcfg0, and tgbcfg1). the module must then be enabled using the ctrl register. in order to initiate acquisition, it is necessary to perform a write to the acquisition initiation and selection register (aisr). the specific value written to aisr det ermines which type of acquisition will be per- formed: qmatrix, qtouch group a, or qtouch group b. the cpu can initiate acquisition by writing to the aisr. while qmatrix, qtouch group a, or qtouch group b acquisition is in progress, the module col- lects count values from the sensors and buffers them. availability of acquired count data is indicated by the acquisition read y (acready) bit in the status register (sr). the cpu or the peripheral dma controller can then read the acquired counts from the acount register. because the cat module is config ured with peripheral dma contro ller capability that can trans- fer data from memory to mblen and from acount to memory, the peripheral dma controller can perform long acquisition sequences and stor e results in memory wit hout cpu intervention. 28.6.2 prescaler and charge length each qtouch acquisition type (autonomous qtouch, qtouch group a, and qtouch group b) has its own prescaler. each qtouch prescaler divides down the clk_cat clock to an appropri- ate sampling frequency for its particular acquisition type. typical frequencies are 1mhz for qtouch acquisition and 4mhz for qmatrix burst timing control. each qtouch prescaler is controlled by the div field in the appropriate configuration register 0 (atcfg0, tgacfg0, or tgbcfg0). the qmatrix bur st timing prescaler is controlled by the div field in mgcfg0. each prescaler uses the following formula to generate the sampling clock: sampling clock = clk_cat / (2(div+1)) the capacitive sensor charge length, discharge length, and settle length can be determined for each acquisition type using the ch len, dilen, and selen fields in configuration registers 0 and 1. the lengths are specified in terms of prescaler clocks. in addition, the qmatrix cx dis- charge length can be determined using the cxdilen field in mgcfg2. for qmatrix acquisition, the duration of chlen should not be set to the same value as the period of any periodic signal on any other pin. if the duration of chlen is the same as the period of a signal on another pin, it is likely that the other sig nal will significantly affect measure- ments due to stray capacitive coupling. for exam ple, if a 1 mhz signal is generated on another pin of the chip, then chlen should not be 1 microsecond. for the qmatrix method, burst and capture lengths are set for each (x,y) pair by writing the desired length values to the mblen register. the write must be done before each x line can start its acquisition and is indi cated by the status bit mblreq in the status register (sr). a dma handshake interface is also connected to this status bit to reduce cpu overhead during qmatrix acquisitions. four burst lengths (burst0..3) can be written at one time into the mblen register. if the cur- rent configuration uses y lines larger than y3 the register has to be written a second time. the first write to mblen specifies the burst length for y lines 0 to 3 in the burst0 to burst3 fields, respectively. the seco nd write specifies the burst length fo r y lines 4 to 7 in fields burst0 to burst3, respectively, and so on.
670 32145c?06/2013 at32uc3l0128/256 the y and yk pins remain clamped to ground apart from the specified number of burst pulses, when charge is transferred and captured into the sampling capacitor. 28.6.3 capacitive count acquisition for the qmatrix, qtouch group a, and qtouch group b types of acquisition, the module acquires count values from the sensors, buffers them, and makes them available for reading in the acount register. further processing of the count values must be performed by the cpu. when the module performs qmatrix acquisition us ing multiple y lines, it starts the capture for each y line at the appropriate time in the burst sequence so that all captures finish simultane- ously. for example, suppose that an acquisiti on is performed on y0 and y1 with burst0=53 and burst1=60. the module will first toggle the x line 7 times while capturing on y1 while y0 and yk0 are clamped to ground. th e module will then toggle the x line 53 times while capturing on both y1 and y0. 28.6.4 autonomous qtouch and dmatouch for autonomous qtouch and dmatouch, a complete detection algorithm is implemented within the cat module. the additional parameters needed to control the detection algorithm must be specified by the user in the atcfg2 and atcfg3 registers. autonomous qtouch and dmatouch sensitivity and out-of-touch sensitivity can be adjusted with the sense and outsens fields, respectively, in atcfg2. each field accepts values from one to 255 where 255 is the least sensitive setti ng. the value in the outsens field should be smaller than the value in the sense field. to avoid false positives a detect integration filtering technique can be used. the number of suc- cessive detects required is specified in the filter field of the atcfg2 register. to compensate for changes in capacitance the cat can recalibrate the autonomous qtouch sensor periodically. the timing of this calibration is done with the ndrift and pdrift fields in the configuration register, atcfg3. it is recommended that the pdrift value is smaller than the ndrift value. the autonomous qtouch se nsor and dmatouch sensors will also recalibrate if the count value goes too far positive beyond a threshold. this positive recalibration threshold is specified by the pthr field in the atcfg3 register. the following block diagram shows the sequence of acquisition and processing operations used by the cat module. the aisr written bit is in ternal and not visible in the user interface.
671 32145c?06/2013 at32uc3l0128/256 figure 28-4. cat acquisition and processing sequence 28.6.5 spread spectrum sensor drive to reduce electromagnetic comp atibility issues, the capacitive sensors can be driven with a spread spectrum signal. to enable spread spectr um drive for a specific acquisition type, the user must write a one to the spread bit in the appropriate configuration register 1 (mgcfg1, atcfg1, tgacfg1, or tgbcfg1). during spread spectrum operation, the length of each pulse within a burst is varied in a deter- ministic pattern, so that the exact same burst pattern is used for a specific burst length. the maximum spread is determined by the maxdev field in the spread spectrum configuration register (sscfg) register. the prescaler divisor is varied in a sawtooth pattern from (2(div+1))-maxdev to (2(div+1))+maxdev and then back to (2(div+1))-maxdev. for exam- ple, if div is 2 and maxdev is 3, the prescaler divisor will have the followi ng sequence: 6, 7, 8, idle acquire autonomous touch count acquire counts update autonomous touch detection algorithm wait for all acquired counts to be transferred aisr written flag set? yes no clear aisr written flag no yes autonomous touch enabled (aten)?
672 32145c?06/2013 at32uc3l0128/256 9, 3, 4, 5, 6, 7, 8, 9, 3, 4, etc. maxdev must not exceed the value of (2(div+1)), or undefined behavior will occur. 28.6.6 synchronization to prevent interference from the 50 or 60 hz mains line the cat can trigger acquisition on the sync signal. the sync signal s hould be derived from the mains line. the acquisition will trig- ger on a falling edge of this signal. to enable sy nchronization for a specif ic acquisition type, the user must write a one to the sync bit in the appropriate configuration register 1 (mgcfg1, atcfg1, tgacfg1, or tgbcfg1). for qmatrix acquisition, all x lines must be sampled at a specific phase of the noise signal for the synchronization to be effective. this can be accomplished by the synchronization timer, which is enabled by writing a non-zero value to the synctim field in the mgcfg2 register. this ensures that the start of the acquisition of eac h x line is spaced at regular intervals, defined by the synctim field. 28.6.7 resistive drive by default, the cat pins are driven with normal i/o drive properties. some of the csa and csb pins can optionally drive with a 1k output resi stance for improved emc. the pins that have this capability are listed in the m odule configuration section. 28.6.8 discharge current sources the device integrates discharge current sources, which can be used to discharge the sampling capacitors during the qmatrix measurement pha se. the discharge current sources are enabled by writing the glen bit in the discharge current source (dics) register to one. this enables an internal reference voltage, which can be either the internal 1.1v band gap voltage or vddio/3, as selected by the intvrefsel bit in the dics register. if the dics.intrefsel bit is one, the reference voltage is applied across an internal resistor, r int . otherwise, the voltage is applied to the dis pin, and an external reference resistor must be connected between dis and ground. the nominal discharge current is given by the following formula, where v ref is the reference volt- age, r ref is the value of the reference resistor, trim is the value written to the dics.trim field, and k is a constant of proportionality: i = (v ref /r ref )*(1+(k*trim)) the values for the internal reference resistor, r int , and the constant, k, may be found in the elec- trical characteristics section. the nominal discharge current may be programmed between 2 and 20 a. the reference current can be fine-tuned by adjusting the trim value in the dics.trim field. the reference current is mirrored to each y-pin if the corresponding bit is written to one in the dics.sources field. 28.6.9 voltage divider enable (vdiven) capability in many qmatrix applications, the sense capaci tors will be charged to 50 mv or more and the negative reference pin (acrefn) of the analog comparators can be tied directly to ground. in that case, the relatively small input offset voltage of the comparators will not cause acquisition problems. however, in certain specialized qmat rix applications such as interpolated touch screens, it may be desirable for the sense capacitors to be charged to less than 25 mv. when such small voltages are used on the sense capacitor s, the input offset voltage of the compara- tors becomes an issue and can cause qmatrix acquisition problems.
673 32145c?06/2013 at32uc3l0128/256 problems with qmatrix acquisition of small sens e capacitor voltages can be solved by connect- ing the negative reference pin (acrefn) to a voltage divider that produces a small positive voltage (20 mv, typically) to cancel any negative input offset voltage. with a 3.3v supply, recom- mended values for the voltage divider are ra (res istor from positive supply to acrefn) of 8200 ohm and rb (resistor fr om acrefn to ground) of 50 ohm. these recommende d values will pro- duce 20 mv on the acrefn pin, which should generally be enough to compensate for the worst-case negative input offset of the analog comparators. unfortunately, such a voltage divider constantly draws a small current from the power supply, reducing battery life in portable applications. in or der to prevent this constant power drain, the cat module provides a voltage divider enable pi n (vdiven) that can be used for driving the voltage divider. the vdiven pin provides power to the voltage divider only when the compara- tors are actually performing qmatrix comparisons. when the comparators are inactive, the vdiven output is zero. this minimizes th e power consumed by the voltage divider.
674 32145c?06/2013 at32uc3l0128/256 28.7 user interface table 28-3. cat register memory map offset register register name access reset 0x00 control register ctrl read/write 0x00000000 0x04 autonomous touch pin selection register atpins read/write 0x00000000 0x08 pin mode register 0 p inmode0 read/write 0x00000000 0x0c pin mode register 1 p inmode1 read/write 0x00000000 0x10 autonomous/dma touch configuration register 0 atcfg0 read/write 0x00000000 0x14 autonomous/dma touch configuration register 1 atcfg1 read/write 0x00000000 0x18 autonomous/dma touch configuration register 2 atcfg2 read/write 0x00000000 0x1c autonomous/dma touch configuration register 3 atcfg3 read/write 0x00000000 0x20 touch group a configuration re gister 0 tgacfg0 read/write 0x00000000 0x24 touch group a configuration re gister 1 tgacfg1 read/write 0x00000000 0x28 touch group b configuration re gister 0 tgbcfg0 read/write 0x00000000 0x2c touch group b configuration re gister 1 tgbcfg1 read/write 0x00000000 0x30 matrix group configuration re gister 0 mgcfg0 read/write 0x00000000 0x34 matrix group configuration re gister 1 mgcfg1 read/write 0x00000000 0x38 matrix group configuration re gister 2 mgcfg2 read/write 0x00000000 0x3c status register sr read-only 0x00000000 0x40 status clear register scr write-only - 0x44 interrupt enable register ier write-only - 0x48 interrupt disable register idr write-only - 0x4c interrupt mask register imr read-only 0x00000000 0x50 acquisition initiation and selection register aisr read/write 0x00000000 0x54 acquired count regist er acount read-only 0x00000000 0x58 matrix burst length register mblen write-only - 0x5c discharge current source register dics read/write 0x00000000 0x60 spread spectrum configuration register sscfg read/write 0x00000000 0x64 csa resistor control register csares read/write 0x00000000 0x68 csb resistor control register csbres read/write 0x00000000 0x6c autonomous touch base count register atbase read-only 0x00000000 0x70 autonomous touch current count register atcurr read-only 0x00000000 0x74 pin mode register 2 p inmode2 read/write 0x00000000 0x78 dmatouch state write regi ster dmatsw write-only 0x00000000 0x7c dmatouch state read r egister dmatsr read-only 0x00000000 0x80 analog comparator shift offset register 0 acshi0 read/write 0x00000000 0x84 analog comparator shift offset register 1 acshi1 read/write 0x00000000 0x88 analog comparator shift offset register 2 acshi2 read/write 0x00000000
675 32145c?06/2013 at32uc3l0128/256 note: 1. the reset value for this register is device specific. please refer to the module configuration section at the end of thi s chapter. 0x8c analog comparator shift offset register 3 acshi3 read/write 0x00000000 0x90 analog comparator shift offset register 4 acshi4 read/write 0x00000000 0x94 analog comparator shift offset register 5 acshi5 read/write 0x00000000 0x98 analog comparator shift offset register 6 acshi6 read/write 0x00000000 0x9c analog comparator shift offset register 7 acshi7 read/write 0x00000000 0xa0 dmatouch sensor status register dmatss read-only 0x00000000 0xf8 parameter register parameter read-only - (1) 0xfc version register version read-only - (1) table 28-3. cat register memory map offset register register name access reset
676 32145c?06/2013 at32uc3l0128/256 28.7.1 control register name: ctrl access type: read/write offset: 0x00 reset value: 0x00000000 ? swrst: software reset writing a zero to this bit has no effect. writing a one to this bit resets the module. the module will be disabled after the reset. this bit always reads as zero. ? en: module enable 0: module is disabled. 1: module is enabled. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 swrst------en
677 32145c?06/2013 at32uc3l0128/256 28.7.2 autonomous touch pin selection register name: atpins access type: read/write offset: 0x04 reset value: 0x00000000 ? aten: autonomous touch enable 0: autonomous qtouch acquisitio n and detection is disabled. 1: autonomous qtouch acquisition and detection is enabled using the sense pair specified in atsp. ? atsp: autonomous touch sense pair selects the sense pair that will be used by the autonomous qtouch sensor. a value of n will select sense pair n (csan and csbn pins). 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------aten 76543210 --- atsp
678 32145c?06/2013 at32uc3l0128/256 28.7.3 pin mode registers 0, 1, and 2 name: pinmode0, pinmode1, and pinmode2 access type: read/write offset: 0x08, 0x0c, 0x74 reset value: 0x00000000 ? sp: sense pair mode selection each sp[n] bit determines the operation mode of sens e pair n (csan and csbn pi ns). the (pinmode2.sp[n] pinmode1.sp[n] pinmode0 .sp[n]) bits have the following definitions: 000: sense pair n disabled. 001: sense pair n is assigned to qtouch group a. 010: sense pair n is assigned to qtouch group b. 011: sense pair n is assigned to the qmatrix group. 100: sense pair n is assigned to the dmatouch group. 101: reserved. 110: reserved. 111: reserved. 31 30 29 28 27 26 25 24 - 23 22 21 20 19 18 17 16 - sp[16] 15 14 13 12 11 10 9 8 sp[15:8] 76543210 sp[7:0]
679 32145c?06/2013 at32uc3l0128/256 28.7.4 autonomous/dma touch configuration register 0 name: atcfg0 access type: read/write offset: 0x10 reset value: 0x00000000 ? div: clock divider the prescaler is used to ensure that the clk_cat clock is divided to around 1 mhz to produce the sampling clock.the prescaler uses the following formula to generate the sampling clock: sampling clock = clk_cat / (2(div+1)) ? chlen: charge length for the autonomous qtouch sensor and dmatouch sensors, specifies how many sample clock cycles should be used for transferring charge to the sense capacitor. ? selen: settle length for the autonomous qtouch sensor and dmatouch sensors, specifies how many sample clock cycles should be used for settling after charge transfer. 31 30 29 28 27 26 25 24 div[15:8] 23 22 21 20 19 18 17 16 div[7:0] 15 14 13 12 11 10 9 8 chlen 76543210 selen
680 32145c?06/2013 at32uc3l0128/256 28.7.5 autonomous/dma touch configuration register 1 name: atcfg1 access type: read/write offset: 0x14 reset value: 0x00000000 ? dishift: discharge shift for the autonomous qtouch sensor and dmatouch sensors, specif ies how many bits the dilen field should be shifted before using it to determine the discharge time. ? sync: sync pin for the autonomous qtouch sensor and dmatouch sensors, spec ifies that acquisition shall begin when a falling edge is received on the sync line. ? spread: spread spectrum sensor drive for the autonomous qtouch sensor and dmatouch sensors, spec ifies that spread spectrum sensor drive shall be used. ? dilen: discharge length for the autonomous qtouch sensor and dmatouch sensors, specifies how many sample clock cycles the cat should use to discharge the capacitors before charging them. ? max: maximum count for the autonomous qtouch sensor and dmatouch sensors, spec ifies how many counts the maximum acquisition should be. 31 30 29 28 27 26 25 24 - dishift - sync spread 23 22 21 20 19 18 17 16 dilen 15 14 13 12 11 10 9 8 max[15:8] 76543210 max[7:0]
681 32145c?06/2013 at32uc3l0128/256 28.7.6 autonomous/dma touch configuration register 2 name: atcfg2 access type: read/write offset: 0x18 reset value: 0x00000000 ? filter: autonomous touch filter setting for the autonomous qtouch sensor and dmatouch sensors, specifies how many positive detects in a row the cat needs to have on the sensor before reporting it as a touch. a filter value of 0 is not allowed and will result in undefined behavior. ? outsens: out-of-touch sensitivity for the autonomous qtouch sensor and dmatouch sensors, spec ifies how sensitive the out-of -touch detector should be. ? sense: sensitivity for the autonomous qtouch sensor and dmatouch sensors, specifies how sensitive the touch detector should be. 31 30 29 28 27 26 25 24 - 23 22 21 20 19 18 17 16 -filter 15 14 13 12 11 10 9 8 outsens 76543210 sense
682 32145c?06/2013 at32uc3l0128/256 28.7.7 autonomous/dma touch configuration register 3 name: atcfg3 access type: read/write offset: 0x1c reset value: 0x00000000 ? pthr: positive recalibration threshold for the autonomous qtouch sensor and dmatouch sensors, specif ies how far a sensor?s signal must move in a positive direction from the reference in order to cause a recalibration. ? pdrift: positive drift compensation for the autonomous qtouch sensor and dmatouch sensors, sp ecifies how often a positive drift compensation should be performed. when this field is zero, positive drift compensation will never be performed. when this field is non-zero, the posit ive drift compensation time interval is given by the following formula: tpdrift = pdrift * 65536 * (sample clock period) ? ndrift: negative drift compensation for the autonomous qtouch sensor and dmatouch sensors, spec ifies how often a negative drift compensation should be performed. when this field is zero, negative drift compensatio n will never be performed. when this field is non-zero, the negat ive drift compensation time interval is given by the following formula: tndrift = ndrift * 65536 * (sample clock period) with the typical sample clock frequency of 1 mhz, p drift and ndrift can be set from 0.066 seconds to 16.7 seconds with 0.066 second resolution. 31 30 29 28 27 26 25 24 - 23 22 21 20 19 18 17 16 pthr 15 14 13 12 11 10 9 8 pdrift 76543210 ndrift
683 32145c?06/2013 at32uc3l0128/256 28.7.8 touch group x configuration register 0 name: tgxcfg0 access type : read/write offset: 0x20, 0x28 reset value: 0x00000000 ? div: clock divider the prescaler is used to ensure that the clk_cat clock is divided to around 1 mhz to produce the sampling clock.the prescaler uses the following formula to generate the sampling clock: sampling clock = clk_cat / (2(div+1)) ? chlen: charge length for the qtouch method, specifies how many sample clock cycles sh ould be used for transferring charge to the sense capacitor. ? selen: settle length for the qtouch method, specifies how many sample clock cycles should be used for se ttling after charge transfer. 31 30 29 28 27 26 25 24 div[15:8] 23 22 21 20 19 18 17 16 div[7:0] 15 14 13 12 11 10 9 8 chlen 76543210 selen
684 32145c?06/2013 at32uc3l0128/256 28.7.9 touch group x configuration register 1 name: tgxcfg1 access type: read/write offset: 0x24, 0x2c reset value: 0x00000000 ? dishift: discharge shift for the sensors in qtouch group x, specif ies how many bits the dilen field should be shifted before using it to determine the discharge time. ? sync: sync pin for sensors in qtouch group x, specifies that acquisition shall begin when a falling edge is received on the sync line. ? spread: spread spectrum sensor drive for sensors in qtouch group x, specifies that spread spectrum sensor drive shall be used. ? dilen: discharge length for sensors in qtouch group x, specifies how many clock cycl es the cat should use to discharge the capacitors before charging them. ? max: touch maximum count for sensors in qtouch group x, specifies how many counts the maximum acquisition should be. 31 30 29 28 27 26 25 24 - - dishift - - sync spread 23 22 21 20 19 18 17 16 dilen 15 14 13 12 11 10 9 8 max[15:8] 76543210 max[7:0]
685 32145c?06/2013 at32uc3l0128/256 28.7.10 matrix group configuration register 0 name: mgcfg0 access type: read/write offset: 0x30 reset value: 0x00000000 ? div: clock divider the prescaler is used to ensure that the clk_cat clock is divided to around 4 mhz to produce the burst timing clock.the prescaler uses the following formula to generate the burst timing clock: burst timing clock = clk_cat / (2(div+1)) ? chlen: charge length for qmatrix sensors, specifies how many burst prescaler cloc k cycles should be used for transferring charge to the sense capacitor. ? selen: settle length for qmatrix sensors, specifies how many burst prescaler cloc k cycles should be used for settling after charge transfer. 31 30 29 28 27 26 25 24 div[15:8] 23 22 21 20 19 18 17 16 div[7:0] 15 14 13 12 11 10 9 8 chlen 76543210 selen
686 32145c?06/2013 at32uc3l0128/256 28.7.11 matrix group configuration register 1 name: mgcfg1 access type: read/write offset: 0x34 reset value: 0x00000000 ? dishift: discharge shift for qmatrix sensors, specifies how many bits the dilen field shou ld be shifted before using it to determine the discharge time. ? sync: sync pin for qmatrix sensors, specifies that acquisition shall begin when a falling edge is received on the sync line. ? spread: spread spectrum sensor drive for qmatrix sensors, specifies that spre ad spectrum sensor drive shall be used. ? dilen: discharge length for qmatrix sensors, specifies how many burst prescaler clock cycles the cat should use to discharge the capacitors at the beginning of a burst sequence. ? max: maximum count for qmatrix sensors, specifies how many counts the maximum acquisition should be. 31 30 29 28 27 26 25 24 - dishift - sync spread 23 22 21 20 19 18 17 16 dilen 15 14 13 12 11 10 9 8 max[15:8] 76543210 max[7:0]
687 32145c?06/2013 at32uc3l0128/256 28.7.12 matrix group configuration register 2 name: mgcfg2 access type: read/write offset: 0x38 reset value: 0x00000000 ? acctrl: analog co mparator control when written to one, allows the cat to disable the analog co mparators when they are not needed. when written to zero, the analog comparators are always enabled. ? consen: consensus filter length for qmatrix sensors, specifies that disc harge will be terminated when consen out of the most recent 5 comparator samples are positive. for example, a value of 3 in the consen field will terminate discharge when 3 out of the most recent 5 comparator samples are positive. when consen has the default value of 0, discharge will be terminated immediately when the comparator output goes positive. ? cxdilen: cx capacitor discharge length for qmatrix sensors, specifies how many burst prescaler clock cy cles the cat should use to disc harge the cx capacitor at the end of each burst cycle. ? synctim: sync time interval when non-zero, determines the number of prescaled clock cycles between the start of the acquisition on each x line for qmatrix acquisition. 31 30 29 28 27 26 25 24 acctrl consen - 23 22 21 20 19 18 17 16 cxdilen 15 14 13 12 11 10 9 8 - synctim[11:8] 76543210 synctim[7:0]
688 32145c?06/2013 at32uc3l0128/256 28.7.13 status register name: sr access type: read-only offset: 0x3c reset value: 0x00000000 ? dmatsc: dmatouch sensor state change 0: no change in the dmatss register. 1: one or more bits have changed in the dmatss register. ? dmatsr: dmatouch state read register ready 0: a new state word is not available in the dmatsr register. 1: a new state word is available in the dmatsr register. ? dmatsw: dmatouch state write register request 0: the dmatouch algorithm is not requesting that a state word be written to the dmatsw register. 1: the dmatouch algorithm is requesting that a st ate word be written to the dmatsw register. ? acqdone: acquisition done 0: acquisition is not done (still in progress). 1: acquisition is complete. ? acready: acquired count data is ready 0: acquired count data is not available in the acount register. 1: acquired count data is available in the acount register. ? mblreq: matrix burst length required 0: the qmatrix acquisition does not require any burst lengths. 1: the qmatrix acquisition requires burst lengths for the current x line. ? atstate: autonomous touch sensor state 0: the autonomous qtouch sensor is not active. 1: the autonomous qtouch sensor is active. ? atsc: autonomous touch sensor status interrupt 0: no status change in t he autonomous qtouch sensor. 1: status change in the autonomous qtouch sensor. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 d m at s c - - - - - d m at s r d m at s w 15 14 13 12 11 10 9 8 ------acqdoneacready 76543210 - - - mblreq atstate atsc atcal enabled
689 32145c?06/2013 at32uc3l0128/256 ? atcal: autonomous touch calibration ongoing 0: the autonomous qtouch sensor is not calibrating. 1: the autonomous qtouch sensor is calibrating. ? enabled: module enabled 0: the module is disabled. 1: the module is enabled.
690 32145c?06/2013 at32uc3l0128/256 28.7.14 status clear register name: scr access type: write-only offset: 0x40 reset value: - writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in sr and the corresponding interrupt request. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 dmatsc------- 15 14 13 12 11 10 9 8 ------acqdoneacready 76543210 - - - - - atsc atcal -
691 32145c?06/2013 at32uc3l0128/256 28.7.15 interrupt enable register name: ier access type: write-only offset: 0x44 reset value: - writing a zero to a bit in this register has no effect. writing a one to a bit in this register will set the corresponding bit in imr. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 dmatsc------- 15 14 13 12 11 10 9 8 ------acqdoneacready 76543210 - - - - - atsc atcal -
692 32145c?06/2013 at32uc3l0128/256 28.7.16 interrupt disable register name: idr access type: write-only offset: 0x48 reset value: - writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in imr. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 dmatsc------- 15 14 13 12 11 10 9 8 ------acqdoneacready 76543210 - - - - - atsc atcal -
693 32145c?06/2013 at32uc3l0128/256 28.7.17 interrupt mask register name: imr access type: read-only offset: 0x4c reset value: 0x00000000 0: the corresponding interrupt is disabled. 1: the corresponding interrupt is enabled. a bit in this register is cleared when the corresponding bit in idr is written to one. a bit in this register is set when the corresponding bit in ier is written to one. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 dmatsc------- 15 14 13 12 11 10 9 8 ------acqdoneacready 76543210 - - - - - atsc atcal -
694 32145c?06/2013 at32uc3l0128/256 28.7.18 acquisition initiation and selection register name: aisr access type: read/write offset: 0x50 reset value: 0x00000000 ? acqsel: acquisition type selection a write to this register initiates an acquisition of the following type: 00: qtouch group a. 01: qtouch group b. 10: qmatrix group. 11: undefined behavior. a read of this register will return the value that was previously written. 31 30 29 28 27 26 25 24 - 23 22 21 20 19 18 17 16 - 15 14 13 12 11 10 9 8 - 76543210 -acqsel
695 32145c?06/2013 at32uc3l0128/256 28.7.19 acquired count register name: acount access type: read-only offset: 0x54 reset value: 0x00000000 ?y: y index the y index (for qmatrix method) associated with this count value. ? sporx: sensor pair or x index the sensor pair index (for qtouch method) or x index (for qmatrix method) associated with this count value. ? count: count value the signal (number of counts) acquired on the channel specified in the sporx and y fields. when multiple acquired count values are read from a qtouch acquisition, the y field will always be 0 and the sporx value will increase monotonically. for example, suppose a qtouch acqui sition is performed using sensor pairs sp1, sp4, and sp9. the first count read will have sporx=1, the second read will have sporx=4, and the third read will have sporx=9. when multiple acquired count values are read from a qmatri x acquisition, the sporx value will stay the same while y increases monotonically through all y values in the group. then sporx will increase to the next x value in the group. for example, a qmatrix acquisition with x=2,3 and y=4,7 would provid e count values in the following order: x=2 and y=4, then x=2 and y=7, then x=3 and y=4, and finally x=3 and y=7. 31 30 29 28 27 26 25 24 y 23 22 21 20 19 18 17 16 sporx 15 14 13 12 11 10 9 8 count[15:8] 76543210 count[7:0]
696 32145c?06/2013 at32uc3l0128/256 28.7.20 matrix burst length register name: mblen access type: write-only offset: 0x58 reset value: - ? burstx: burst length x for qmatrix sensors, specifies how many times the switching se quence should be repeated before acquisition begins for each channel. each count in the burstx field sp ecifies 1 repeat of the switching sequence, so the actual burst length will be burst. before doing a qmatrix acquisition on one x line this register has to be written with the burst values for the current xy pairs . for each x line this register needs to be programmed with all the y va lues. if y values larger than 3 are used the register has to be written several times in order to specify all burst lengths. the status register bit mblreq is set to 1 when the cat is waiting for values to be written into this register. 31 30 29 28 27 26 25 24 burst0 23 22 21 20 19 18 17 16 burst1 15 14 13 12 11 10 9 8 burst2 76543210 burst3
697 32145c?06/2013 at32uc3l0128/256 28.7.21 discharge current source register name: dics access type: read/write offset: 0x5c reset value: 0x00000000 ? fsources: force discharge current sources when fsources[n] is 0, the corres ponding discharge current source behavior depends on sources[n]. when fsources[n] is 1, the corresponding di scharge current source is forced to be enabled continuously. this is useful for testing or debugging but should not be done during normal acquisition. ? glen: global enable 0: the current source module is globally disabled. 1: the current source module is globally enabled. ? intvrefsel: internal vo ltage reference select 0: the voltage for the referenc e resistor is generated from the internal band gap circuit. 1: the voltage for the refere nce resistor is vddio/3. ? intrefsel: internal reference select 0: the reference current flows through an external resistor on the dis pin. 1: the reference current flows through the internal reference resistor. ? trim: reference current trimming this field is used to trim the discharge current. 0x00 corresponds to the minimum current value, and 0x1f corresponds to the maximum current value. ? sources: enable discharge current sources when sources[n] is 0, the corresponding discharge current source is disabled. when sources[n] is 1, the corresponding discharge current s ource is enabled at appropriate times during acquisition. 31 30 29 28 27 26 25 24 fsources[7:0] 23 22 21 20 19 18 17 16 glen----- intvrefsel intrefsel 15 14 13 12 11 10 9 8 --- trim 76543210 sources[7:0]
698 32145c?06/2013 at32uc3l0128/256 28.7.22 spread spectrum configuration register name: sscfg access type: read/write offset: 0x60 reset value: 0x00000000 ? maxdev: maximum deviation when spread spectrum burst is enabled, maxdev indicates the maxi mum number of prescaled clock cycles the burst pulse will be extended or shortened. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 maxdev
699 32145c?06/2013 at32uc3l0128/256 28.7.23 csa resistor control register name: csares access type: read/write offset: 0x64 reset value: 0x00000000 ? res: resistive drive enable when res[n] is 0, csa[n] has the same drive properties as normal i/o pads. when res[n] is 1, csa[n] has a nominal outpu t resistance of 1kohm during the burst phase. 31 30 29 28 27 26 25 24 - 23 22 21 20 19 18 17 16 - res[16] 15 14 13 12 11 10 9 8 res[15:8] 76543210 res[7:0]
700 32145c?06/2013 at32uc3l0128/256 28.7.24 csb resistor control register name: csbres access type: read/write offset: 0x68 reset value: 0x00000000 ? res: resistive drive enable when res[n] is 0, csb[n] has the same drive properties as normal i/o pads. when res[n] is 1, csb[n] has a nominal outpu t resistance of 1kohm during the burst phase. 31 30 29 28 27 26 25 24 - 23 22 21 20 19 18 17 16 - res[16] 15 14 13 12 11 10 9 8 res[15:8] 76543210 res[7:0]
701 32145c?06/2013 at32uc3l0128/256 28.7.25 autonomous touch base count register name: atbase access type: read-only offset: 0x6c reset value: 0x00000000 ? count: count value the base count currently stored by the autonomous touch sens or. this is useful for autonomous touch debugging purposes. 31 30 29 28 27 26 25 24 - 23 22 21 20 19 18 17 16 - 15 14 13 12 11 10 9 8 count[15:8] 76543210 count[7:0]
702 32145c?06/2013 at32uc3l0128/256 28.7.26 autonomous touch current count register name: atcurr access type: read-only offset: 0x70 reset value: 0x00000000 ? count: count value the current count acquired by the autonomous touch sensor. this is useful for autonomous touch debugging purposes. 31 30 29 28 27 26 25 24 - 23 22 21 20 19 18 17 16 - 15 14 13 12 11 10 9 8 count[15:8] 76543210 count[7:0]
703 32145c?06/2013 at32uc3l0128/256 28.7.27 dmatouch state write register name: dmatsw access type: write-only offset: 0x78 reset value: 0x00000000 ? notincal: not in calibration mode 0: calibration should be performed on the next iteration of the dmatouch algorithm. 1: calibration should not be performed on the next iteration of the dmatouch algorithm. ? detcnt: detection count this count value is updated and used by the dmatouch algorithm in order to detect when a button has been pushed. ? basecnt: base count this count value represents the average expected acquired count when the sensor/button is not pushed. 31 30 29 28 27 26 25 24 -------notincal 23 22 21 20 19 18 17 16 detcnt[23:16] 15 14 13 12 11 10 9 8 basecnt[15:8] 76543210 basecnt[7:0]
704 32145c?06/2013 at32uc3l0128/256 28.7.28 dma touch state read register name: dmatsr access type: read/write offset: 0x7c reset value: 0x00000000 ? notincal: not in calibration mode 0: calibration should be performed on the next iteration of the dmatouch algorithm. 1: calibration should not be performed on the next iteration of the dmatouch algorithm. ? detcnt: detection count this count value is updated and used by the dmatouch algorithm in order to detect when a button has been pushed. ? basecnt: base count this count value represents the average expected acquired count when the sensor/button is not pushed. 31 30 29 28 27 26 25 24 -------notincal 23 22 21 20 19 18 17 16 detcnt[23:16] 15 14 13 12 11 10 9 8 basecnt[15:8] 76543210 basecnt[7:0]
705 32145c?06/2013 at32uc3l0128/256 28.7.29 analog comparator shift offset register x name: acshix access type: read/write offset: 0x80, 0x84, 0x88, 0x8c, 0x 90, 0x94, 0x98, and 0x9c reset value: 0x00000000 ? shival: shift offset value specifies the amount to shift the count value from each comparator. this al lows the offset of each comparator to be compensated. 31 30 29 28 27 26 25 24 - 23 22 21 20 19 18 17 16 - 15 14 13 12 11 10 9 8 - shival[11:8] 76543210 shival[7:0]
706 32145c?06/2013 at32uc3l0128/256 28.7.30 dmatouch sensor status register name: dmatss access type: read-only offset: 0xa0 reset value: 0x00000000 ? ss: sensor status 0: the dmatouch sensor is not active, i.e. the button is currently not pushed. 1: the dmatouch sensor is active, i.e. the button is currently pushed. 31 30 29 28 27 26 25 24 ss[31:24] 23 22 21 20 19 18 17 16 ss[23:16] 15 14 13 12 11 10 9 8 ss[15:8] 76543210 ss[7:0]
707 32145c?06/2013 at32uc3l0128/256 28.7.31 parameter register name: parameter access type: read-only offset: 0xf8 reset value: - ? sp[n]: sensor pair implemented 0: the corresponding sensor pair is not implemented 1: the corresponding sensor pair is implemented. 31 30 29 28 27 26 25 24 sp[31:24] 23 22 21 20 19 18 17 16 sp[23:16] 15 14 13 12 11 10 9 8 sp[15:8] 76543210 sp[7:0]
708 32145c?06/2013 at32uc3l0128/256 28.7.32 version register name: version access type: read-only offset: 0xfc reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
709 32145c?06/2013 at32uc3l0128/256 28.8 module configuration the specific configuration the cat module is listed in the following tables.the module bus clocks listed here are connecte d to the system bus clocks. pleas e refer to the power manager chapter for details. 28.8.1 resistive drive by default, the cat pins are driven with normal i/o drive properties. some of the csa and csb pins can optionally drive with a 1k output resistance for improved emc. to enable resistive drive on a pin, the user must write a one to the corresponding bit in the csa resistor control register (csares) or csb resistor control register (csbres) register. table 28-4. cat configuration feature cat number of touch sensors/size of matrix allows up to 17 touch sensors, or up to 16 by 8 matrix sensors to be interfaced. table 28-5. cat clocks clock name description clk_cat clock for the cat bus interface gclk the generic clock used for the cat is gclk4 table 28-6. register reset values register reset value version 0x00000400 parameter 0x0001ffff
710 32145c?06/2013 at32uc3l0128/256 29. glue logic controller (gloc) rev: 1.0.0.0 29.1 features ? glue logic for general purpose pcb design ? programmable lookup table ? up to four inputs supported per lookup table ? optional filtering of output 29.2 overview the glue logic controller (gloc) contains programmable logic which can be connected to the device pins. this allows the user to eliminate logic gates for simple glue logic functions on the pcb. the gloc consists of a number of lookup tabl e (lut) units. each lut can generate an output as a user programmable logic expression with fo ur inputs. inputs can be individually masked. the output can be combinatorially generated from the inputs, or filtered to remove spikes. 29.3 block diagram figure 29-1. gloc block diagram peripheral bus truth filter out[0] ... out[n] filten in[3:0] ? in[(4n+3):4n] aen clk_gloc gclk
711 32145c?06/2013 at32uc3l0128/256 29.4 i/o lines description each lut have 4 inputs and one output. the inputs and outputs for the luts are mapped sequentially to the inputs and outputs. this means that lut0 is connected to in0 to in3 and out0. lut1 is connected to in4 to in7 and out1. in general, lutn is connected to in[4n] to in[4n+3] and outn. 29.5 product dependencies in order to use this module, other parts of the system must be configured correctly, as described below. 29.5.1 i/o lines the pins used for interfacing the gloc may be multiplexed with i/o controller lines. the pro- grammer must first program the i/o controller to assign the desired gloc pins to their peripheral function. if i/o lines of the gloc are no t used by the application, they can be used for other purposes by the i/o controller. it is only required to enable the gloc inputs and outputs actually in use. pullups for pins config- ured to be used by the gloc will be disabled. 29.5.2 clocks the clock for the gloc bus interface (clk_gloc) is generated by the power manager. this clock is enabled at reset, and can be disabled in the power manager. it is recommended to dis- able the gloc before disabling the clock, to avoid freezing the module in an undefined state. additionally, the gloc depends on a dedicated generic clock (gclk). the gclk can be set to a wide range of frequencies and clock sources, and must be enabled by the system control interface (scif) before the gloc filter can be used. 29.5.3 debug operation when an external debugger forces the cpu into debug mode, the gloc continues normal operation. 29.6 functional description 29.6.1 enabling the lookup table inputs since the inputs to each lookup table (lut) unit can be multiplexed with other peripherals, each input must be explicitly enabled by writing a one to the corresponding enable bit (aen) in the corresponding contro l register (cr). if no inputs are enabled, the output outn will be the least signif icant bit in the truthn register. table 29-1. i/o lines description pin name pin description type in0-inm inputs to lookup tables input out0-outn output from lookup tables output
712 32145c?06/2013 at32uc3l0128/256 29.6.2 configuring the lookup table the lookup table in each lut unit can generate any logic expression out as a function of up to four inputs, in[3:0]. the truth table for the expression is written to the truth register for the lut. table 29-2 shows the truth table for lut0. the truth table for lutn is written to truthn, and the corresponding input and outputs will be in[4n] to in[4n+3] and outn. 29.6.3 output filter by default, the output outn is a combinatorial func tion of the inputs in[4n] to in[4n+3]. this may cause some short glitches to occur when the inputs change value. it is also possible to clock the output through a fi lter to remove glitches. this requires that the corresponding generic clock (gclk) has been enabled before use. the filter can then be enabled by writing a one to the filter enable (filten) bit in crn. the outn output will be delayed by three to four gclk cycles when the filter is enabled. table 29-2. truth table for the lookup table in lut0 in[3] in[2] in[1] in[0] out[0] 0 0 0 0 truth0[0] 0 0 0 1 truth0[1] 0 0 1 0 truth0[2] 0 0 1 1 truth0[3] 0 1 0 0 truth0[4] 0 1 0 1 truth0[5] 0 1 1 0 truth0[6] 0 1 1 1 truth0[7] 1 0 0 0 truth0[8] 1 0 0 1 truth0[9] 1 0 1 0 truth0[10] 1 0 1 1 truth0[11] 1 1 0 0 truth0[12] 1 1 0 1 truth0[13] 1 1 1 0 truth0[14] 1 1 1 1 truth0[15]
713 32145c?06/2013 at32uc3l0128/256 29.7 user interface note: 1. the reset values are device specific. please refer to the module configuration section at the end of this chapter. table 29-3. gloc register memory map offset register register name access reset 0x00+n*0x08 control register n crn read/write 0x00000000 0x04+n*0x08 truth table register n truthn read/write 0x00000000 0x38 parameter register parameter read-only - (1) 0x3c version register version read-only - (1)
714 32145c?06/2013 at32uc3l0128/256 29.7.1 control register n name: crn access type: read/write offset : 0x00+n*0x08 reset value: 0x00000000 ? filten: filter enable 1: the output is glitch filtered 0: the output is not glitch filtered ? aen: enable in inputs input in[n] is enabled when aen[n] is one. input in[n] is disabled when aen[n] is zero, and will not affect the out value. 31 30 29 28 27 26 25 24 filten------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 ---- aen
715 32145c?06/2013 at32uc3l0128/256 29.7.2 truth table register n name: truthn access type: read/write offset : 0x04+n*0x08 reset value: 0x00000000 ? truth: truth table value this value defines the output ou t as a function of inputs in: out = truth[in] 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 truth[15:8] 76543210 truth[7:0]
716 32145c?06/2013 at32uc3l0128/256 29.7.3 parameter register name: parameter access type: read-only offset :0x38 reset value: - ? luts: lookup table units implemented this field contains the number of lookup table units implemented in this device. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 luts
717 32145c?06/2013 at32uc3l0128/256 29.7.4 version register name: version access type: read-only offset :0x3c reset value: - ? variant: variant number reserved. no functionality associated. ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 ---- variant 15 14 13 12 11 10 9 8 ---- version[11:8] 76543210 version[7:0]
718 32145c?06/2013 at32uc3l0128/256 29.8 module configuration the specific configuration for each gloc instance is listed in the following tables.the gloc bus clocks listed here are connected to the syst em bus clocks. please refer to the power man- ager chapter for details. table 29-4. gloc configuration feature gloc number of lut units 2 table 29-5. gloc clocks clock name description clk_gloc clock for the gloc bus interface gclk the generic clock used for the gloc is gclk5 table 29-6. register reset values register reset value version 0x00000100 parameter 0x00000002
719 32145c?06/2013 at32uc3l0128/256 30. awire uart (aw) rev: 2.3.0.0 30.1 features ? asynchronous receiver or transmitter when the awire system is not used for debugging. ? one- or two-pin operation supported. 30.2 overview if the aw is not used for debugging, the awire ua rt can be used by the user to send or receive data with one start bit, eight data bits, no parit y bits, and one stop bit. this can be controlled through the awire uart user interface. this chapter only describes the awire uart user interface. for a description of the awire debug interface, please see the programming and debugging chapter. 30.3 block diagram figure 30-1. awire debug interface block diagram uart reset filter external reset aw_enable reset_n baudrate detector rw sz addr data crc aw control aw user interface sab interface reset command power manager cpu halt command flash controller chip_erase command awire debug interface pb sab
720 32145c?06/2013 at32uc3l0128/256 30.4 i/o lines description 30.5 product dependencies in order to use this module, other parts of the system must be configured correctly, as described below. 30.5.1 i/o lines the pin used by aw is multiple xed with the reset_n pi n. the reset function ality is the default function of this pin. to enable the awire func tionality on the reset_n pin the user must enable the awire uart user interface. 30.5.2 power management if the cpu enters a sleep mode that disables clocks used by the awire uart user interface, the awire uart user interface will stop functioning and resume operation af ter the system wakes up from sleep mode. 30.5.3 clocks the awire uart uses the internal 120 mhz rc oscillator (rc120m) as clock source for its operation. when using the awire uart user interface rc120m must enabled using the clock request register (see section 30.6.1 ). the clock for the awire uart user interface (clk_aw) is generated by the power manager. this clock is enabled at reset, and can be disabled in the power manager. it is recommended to disable the awire uart user interface before di sabling the clock, to avoid freezing the awire uart user interface in an undefined state. 30.5.4 interrupts the awire uart user interface interrupt request line is connected to the interrupt controller. using the awire uart user interface interrupt requires the interrupt controller to be pro- grammed first. 30.5.5 debug operation if the aw is used for debug ging the awire uart user interface will not be usable. when an external debugger forces the cpu into debug mode, the awire uart user interface continues normal operation. if the awire uart user interface is configured in a way that requires it to be periodically serviced by the cpu through interrupts or similar, improper opera- tion or data loss may result during debugging. 30.5.6 external components the aw needs an external pullup on the reset_n pin to ensure that the pin is pulled up when the bus is not driven. 30.6 functional description the awire uart user interface can be used as a spare asynchronous receiver or transmitter when aw is not used for debugging. table 30-1. i/o lines description name description type data awire data multiplexed wit h the reset_n pin. input/output
721 32145c?06/2013 at32uc3l0128/256 30.6.1 how to initialize the module to initialize the awire uart user interface the user must first enable the clock by writing a one to the clock enable bit in the clock request r egister (clkr.clken) and wait for the clock enable bit in the status register (sr.cenabl ed) to be set. after doing this either receive, transmit or receive with resync must be selected by writing th e corresponding value into the mode field of the control (ctrl.mode) register. due to the rc120m being asynchronous with the system clock values must be allowed to propagate in the syst em. during this time the awire master will set the busy bit in the status register (sr.busy). after the sr.busy bit is cleared the baud rate field in the baud rate register (brr.br) can be written with the wanted baudrate ( ) according to the following formula ( is the rc120m clock frequency): after this operation the user must wait until the sr.busy is clea red. the interface is now ready to be used. 30.6.2 basic asynchronous receiver operation the awire uart user interface must be initialized according to the sequence above, but the ctrl.mode field must be written to one (receive mode). when a data byte arrives the awire uart user interface will indicate this by setting the data ready interrupt bit in the status register (sr.dreadyint). the user must read the data in the receive holding register (rhr.rxdata) and clear the interrupt bit by writing a one to the data ready interrupt clear bit in the status clear register (scr.dreadyint). the interface is now ready to receive another byte. 30.6.3 basic asynchronous transmitter operation the awire uart user interface must be initialized according to the sequence above, but the ctrl.mode field must be written to two (transmit mode). to transmit a data byte the user must write the data to the transmit holding register (the.txdata). before the next byte can be written the sr.busy must be cleared. 30.6.4 basic asynchronous receiver with resynchronization by writing three into ctrl.mode the awire uart user interface will assume that the first byte it receives is a sync byte (0x55) and set brr. br according to this. all subsequent transfers will assume this baudrate, unless brr.br is rewritten by the user. to make the awire uart user interface accept a new sync resynchronization the awire uart user interface must be disabled by writing zero to ctrl.mode and then reenable the interface. 30.6.5 overrun in receive mode an overrun can occur if the user has not read the previous received data from the rhr.rxdata when the newest data should be pl aced there. such a condition is flagged by setting the overrun bit in the status register (sr.overrun). if sr.overrun is set the new- est data received is placed in rhr.rxdata and the data that was there before is overwritten. f br f aw f br 8 f aw br ----------- =
722 32145c?06/2013 at32uc3l0128/256 30.6.6 interrupts to make the cpu able to do other things while wa iting for the awire uart user interface to fin- ish its operations the awire uart user interface supports generating interrupts. all status bits in the status register can be used as interrupt sources, except the sr.busy and sr.cenabled bits. to enable an interrupt the user must write a one to the corresponding bit in the interrupt enable register (ier). upon the next zero to one transition of this sr bit the awire uart user interface will flag this interrupt to the cpu. to clear the interrupt the user must write a one to the corre- sponding bit in the status clear register (scr). interrupts can be disabled by writing a one to th e corresponding bit in the interrupt disable reg- ister (idr). the interrupt mask register (imr) can be read to check if an interrupt is enabled or disabled. 30.6.7 using the peripheral dma controller to relieve the cpu of data transfers the awire uart user interface support using the peripheral dma controller. to transmit using the peripheral dma controller do the following: 1. setup the awire uart user interface in transmit mode. 2. setup the peripheral dma controller with buffer address and length, use byte as trans- fer size. 3. enable the peripheral dma controller. 4. wait until the peripheral dma controller is done. to receive using the peripheral dma controller do the following: 1. setup the awire uart user interface in receive mode 2. setup the peripheral dma controller with buffer address and length, use byte as trans- fer size. 3. enable the peripheral dma controller. 4. wait until the peripheral dma controller is ready.
723 32145c?06/2013 at32uc3l0128/256 30.7 user interface note: 1. the reset values are device specific. please refer to the module configuration section at the end of this chapter. table 30-2. awire uart user interface register memory map offset register register name access reset 0x00 control register ctrl read/write 0x00000000 0x04 status register sr read-only 0x00000000 0x08 status clear register scr write-only - 0x0c interrupt enable register ier write-only - 0x10 interrupt disable register idr write-only - 0x14 interrupt mask register imr read-only 0x00000000 0x18 receive holding register rhr read-only 0x00000000 0x1c transmit holding register thr read/write 0x00000000 0x20 baud rate register brr read/write 0x00000000 0x24 version register version read-only - (1) 0x28 clock request register clkr read/write 0x00000000
724 32145c?06/2013 at32uc3l0128/256 30.7.1 control register name: ctrl access type: read/write offset: 0x00 reset value: 0x00000000 ? mode: awire uart user interface mode 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 ------ mode table 30-3. awire uart user interface modes mode mode description 0 disabled 1receive 2 transmit 3 receive with resync.
725 32145c?06/2013 at32uc3l0128/256 30.7.2 status register name: sr access type: read-only offset: 0x04 reset value: 0x00000000 ? trmis: transmit mismatch 0: no transfers mismatches. 1: the transceiver was active when receiving. this bit is set when the transce iver is active when receiving. this bit is cleared when corresponding bit in scr is written to one. ? overrun: data overrun 0: no data overwritten in rhr. 1: data in rhr has been overwritten before it has been read. this bit is set when data in rhr is overwritten before it has been read. this bit is cleared when corresponding bit in scr is written to one. ? dreadyint: data ready interrupt 0: no new data in the rhr. 1: new data received and placed in the rhr. this bit is set when new data is received and placed in the rhr. this bit is cleared when corresponding bit in scr is written to one. ? readyint: ready interrupt 0: the interface has not generated an ready interrupt. 1: the interface has had a transition from busy to not busy. this bit is set when the interface has transition from busy to not busy. this bit is cleared when corresponding bit in scr is written to one. ? cenabled: clock enabled 0: the awire clock is not enabled. 1: the awire clock is enabled. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 - - trmis - - overrun dreadyint readyint 76543210 - - - - - cenabled - busy
726 32145c?06/2013 at32uc3l0128/256 this bit is set when the clock is disabled. this bit is cleared when the clock is enabled. ? busy: synchronizer busy 0: the asynchronous interface is ready to accept more data. 1: the asynchronous interface is busy and will block writes to ctrl, brr, and thr. this bit is set when the asynchronous interface becomes busy. this bit is cleared when the asynchronous interface becomes ready.
727 32145c?06/2013 at32uc3l0128/256 30.7.3 status clear register name: scr access type: write-only offset: 0x08 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in sr and the corresponding interrupt request. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 - - trmis - - overrun dreadyint readyint 76543210 --------
728 32145c?06/2013 at32uc3l0128/256 30.7.4 interrupt enable register name: ier access type: write-only offset: 0x0c reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will set the corresponding bit in imr. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 - - trmis - - overrun dreadyint readyint 76543210 --------
729 32145c?06/2013 at32uc3l0128/256 30.7.5 interrupt disable register name: idr access type: write-only offset: 0x10 reset value: 0x00000000 writing a zero to a bit in this register has no effect. writing a one to a bit in this register will clear the corresponding bit in imr. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 - - trmis - - overrun dreadyint readyint 76543210 --------
730 32145c?06/2013 at32uc3l0128/256 30.7.6 interrupt mask register name: imr access type: read-only offset: 0x14 reset value: 0x00000000 0: the corresponding interrupt is disabled. 1: the corresponding interrupt is enabled. a bit in this register is cleared when the corresponding bit in idr is written to one. a bit in this register is set when the corresponding bit in ier is written to one. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 - - trmis - - overrun dreadyint readyint 76543210 --------
731 32145c?06/2013 at32uc3l0128/256 30.7.7 receive holding register name: rhr access type: read-only offset: 0x18 reset value: 0x00000000 ? rxdata: received data the last byte received. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 rxdata
732 32145c?06/2013 at32uc3l0128/256 30.7.8 transmit holding register name: thr access type: read/write offset: 0x1c reset value: 0x00000000 ? txdata: transmit data the data to send. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 txdata
733 32145c?06/2013 at32uc3l0128/256 30.7.9 baud rate register name: brr access type: read/write offset: 0x20 reset value: 0x00000000 ? br: baud rate the baud rate ( ) of the transmission, calculated us ing the following formula ( is the rc120m frequency): br should not be set to a value smaller than 32. writing a value to this field will update the baud rate of the transmission. reading this field will give the current baud rate of the transmission. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 br[15:8] 76543210 br[7:0] f br f aw f br 8 f aw br ----------- =
734 32145c?06/2013 at32uc3l0128/256 30.7.10 version register name: version access type: read-only offset: 0x24 reset value: 0x00000200 ? version: version number version number of the module. no functionality associated. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 - - - - version[11:8] 76543210 version[7:0]
735 32145c?06/2013 at32uc3l0128/256 30.7.11 clock request register name: clkr access type: read/write offset: 0x28 reset value: 0x00000000 ? clken: clock enable 0: the awire clock is disabled. 1: the awire clock is enabled. writing a zero to this bit will disable the awire clock. writing a one to this bit will enable the awire clock. 31 30 29 28 27 26 25 24 -------- 23 22 21 20 19 18 17 16 -------- 15 14 13 12 11 10 9 8 -------- 76543210 -------clken
736 32145c?06/2013 at32uc3l0128/256 30.8 module configuration the specific configuration for each awire instan ce is listed in the following tables.the module bus clocks listed here are connected to the syst em bus clocks. please refer to the power man- ager chapter for details. table 30-4. aw clocks clock name description clk_aw clock for the aw bus interface table 30-5. register reset values register reset value version 0x00000230
737 32145c?06/2013 at32uc3l0128/256 31. programming and debugging 31.1 overview the at32uc3l0128/256 supports programming and debugging through two interfaces, jtag or awire. jtag is an industry standard interface and allows boundary scan for pcb testing, as well as daisy-chaining of multiple devices on the pcb. awire is an atmel proprietary protocol which offers higher throughput and robust communication, and does not require application pins to be reserved. either interface provides access to the intern al service access bus (sab), which offers a bridge to the high speed bus, giving access to memories and peripherals in the device. by using this bridge to the bus system, the flas h and fuses can thus be programmed by access- ing the flash controller in the same manner as the cpu. the sab also provides access to the nexus-compliant on-chip debug (ocd) system in the device, which gives the user non-intrusive run-time control of the program execution. addition- ally, trace information can be output on the auxiliary (aux) deb ug port or buff ered in internal ram for later retrieval by jtag or awire. 31.2 service access bus the avr32 architecture offers a common interface for access to on-chip debug, programming, and test functions. these are mapped on a common bus called t he service acce ss bus (sab), which is linked to the jtag and awire port th rough a bus master module, which also handles synchronization between th e debugger and sab clocks. when accessing the sab through th e debugger ther e are no limitations on debugger frequency compared to chip frequency, although there must be an active system clock in order for the sab accesses to complete. if the system clock is swit ched off in sleep mode, activity on the debugger will restart the system clock automatically, without waking the device from sleep. debuggers may optimize the transfer rate by adjusting the frequency in relation to the system clock. this ratio can be measured with debug protocol specific instructions. the service access bus uses 36 address bits to address memory or registers in any of the slaves on the bus. the bus supports sized accesses of bytes (8 bits), halfwords (16 bits), or words (32 bits). all accesses must be aligned to the size of th e access, i.e. halfword accesses must have the lowest address bit cleared, and word accesses must have the two lowest address bits cleared. 31.2.1 sab address map the sab gives the user access to the internal address space and other features through a 36 bits address space. the 4 msbs identify the slave number, while the 32 lsbs are decoded within the slave?s addre ss space. the sab slaves are shown in table 31-1 . table 31-1. sab slaves, addresses and descriptions slave address [35:32] description unallocated 0x0 intentionally unallocated ocd 0x1 ocd registers hsb 0x4 hsb memory space, as seen by the cpu
738 32145c?06/2013 at32uc3l0128/256 31.2.2 sab security restrictions the service access bus can be restricted by internal security measures. a short description of the security measures are found in the table below. 31.2.2.1 security measure and control location a security measure is a mechanism to either block or a llow sab access to a certain address or address range. a security measure is enabled or disabled by one or several control signals. this is called the control location for the security measure. these security measures can be used to prevent an end user from reading out the code pro- grammed in the flash, for instance. below follows a more in depth description of what locations are accessible when the security measures are active. note: 1. second word of the user page, refer to the fuses settings section for details. hsb 0x5 alternative mapping for hsb space, for compatibility with other 32-bit avr devices. memory service unit 0x6 memory service unit registers reserved other unused table 31-1. sab slaves, addresses and descriptions slave address [35:32] description table 31-2. sab security measures security measure control location description secure mode flashcdw secure bits set allocates a portion of the flash for secure code. this code cannot be read or debugged. the user page is also locked. security bit flashcdw security bit set programming and debugging not possible, very restricted access. user code programming flashcdw uprot + security bit set restricts all access except parts of the flash and the flash controller for programming user code. debugging is not possible unless an os running from the secure part of the flash supports it. table 31-3. secure mode sab restrictions name address start address end access secure flash area 0x580000000 0x580000000 + (userpage[15:0] << 10) blocked secure ram area 0x500000000 0x500000000 + (userpage[31:16] << 10) blocked user page 0x580800000 0x581000000 read other accesses - - as normal
739 32145c?06/2013 at32uc3l0128/256 table 31-4. security bit sab restrictions name address start address end access ocd dccpu, ocd dcemu, ocd dcsr 0x100000110 0x100000118 read/write user page 0x580800000 0x581000000 read other accesses - - blocked table 31-5. user code programming sab restrictions name address start address end access ocd dccpu, ocd dcemu, ocd dcsr 0x100000110 0x100000118 read/write user page 0x580800000 0x581000000 read flashcdw pb interface 0x5fffe0000 0x5fffe0400 read/write flash pages outside bootprot 0x580000000 + bootprot size 0x580000000 + flash size read/write other accesses - - blocked
740 32145c?06/2013 at32uc3l0128/256 31.3 on-chip debug rev: 2.1.2.0 31.3.1 features ? debug interface in compli ance with ieee-isto 5001- 2003 (nexus 2.0) class 2+ ? jtag or awire access to all on-chip debug functions ? advanced program, data, ownership, and watchpoint trace supported ? nanotrace awire- or jtag-based trace access ? auxiliary port for high-speed trace information ? hardware support for 6 prog ram and 2 data breakpoints ? unlimited number of softw are breakpoints supported ? automatic crc check of memory regions 31.3.2 overview debugging on the at32uc 3l0128/256 is facilitated by a powerful on-chip debug (ocd) sys- tem. the user accesses this through an exte rnal debug tool which connects to the jtag or awire port and the auxiliary (aux) port if implemented. the aux port is primarily used for trace functions, and an awire- or jtag-based deb ugger is sufficient for basic debugging. the debug system is based on the nexus 2.0 standard, class 2+, which includes: ? basic run-time control ? program breakpoints ? data breakpoints ?program trace ? ownership trace ? data trace in addition to the mandatory nexus debug features, the at32uc3l0128/256 implements sev- eral useful ocd features, such as: ? debug communication channel between cpu and debugger ? run-time pc monitoring ? crc checking ? nanotrace ? software quality assurance (sqa) support the ocd features are controlled by ocd registers, which can be accessed by the debugger, for instance when the nexus_access jtag inst ruction is loaded. the cpu can also access ocd registers directly using mtdr/mfdr instructions in any privileged mode. the ocd registers are implemented based on the recommendations in the nexus 2.0 standard, and are detailed in the avr32uc technical reference manual. 31.3.3 i/o lines description the ocd aux trace port contains a number of pins, as shown in table 31-6 on page 741 . these are multiplexed with i/o co ntroller lines, and must explicit ly be enabled by writing ocd registers before the debug session starts. the aux port is mapped to two different locations,
741 32145c?06/2013 at32uc3l0128/256 selectable by ocd registers, minimizing the chance that the aux port will need to be shared with an application. 31.3.4 product dependencies in order to use this module, other parts of the system must be configured correctly, as described below. 31.3.4.1 power management the ocd clock operates independently of the cpu clock. if enabled in the power manager, the ocd clock (clk_ocd) will continue running even if the cpu enters a sleep mode that disables the cpu clock. 31.3.4.2 clocks the ocd has a clock (clk_ocd) running synchronou sly with the cpu clock. this clock is gen- erated by the power manager. the clock is enabled at reset, and can be disabled by writing to the power manager. 31.3.4.3 interrupt the ocd system interrupt request lines are connected to the interrupt controller. using the ocd interrupts requires the interrupt controller to be programmed first. table 31-6. auxiliary port signals pin name pin description direction active level type mcko trace data output clock output digital mdo[5:0] trace data output output digital mseo[1:0] trace frame control output digital evti_n event in input low digital evto_n event out output low digital
742 32145c?06/2013 at32uc3l0128/256 31.3.5 block diagram figure 31-1. on-chip debug block diagram 31.3.6 sab-based debug features a debugger can control all ocd fe atures by writing ocd register s over the sab interface. many of these do not depend on output on the aux port, allowing an awire- or jtag-based debugger to be used. a jtag-based debugger should connect to the device through a standard 10-pin idc connector as described in the avr32uc technical reference manual. an awire-based deb ugger should connect to the device through the reset_n pin. on-chip debug jtag debug pc debug instruction cpu breakpoints program trace data trace ownership trace watchpoints transmit queue aux jtag internal sram s e r v i c e a c c e s s b u s memory service unit hsb bus matrix memories and peripherals awire awire
743 32145c?06/2013 at32uc3l0128/256 figure 31-2. jtag-based debugger figure 31-3. awire-based debugger 31.3.6.1 debug communication channel the debug communication channel (d cc) consists of a pair ocd registers with associated handshake logic, accessible to both cpu and debugger. the registers can be used to exchange data between the cpu and the debugmaster, both runtime as well as in debug mode. 32-bit avr jtag-based debug tool pc jtag 10-pin idc 32-bit avr awire-based debug tool pc awire
744 32145c?06/2013 at32uc3l0128/256 the ocd system can generate an interrupt to the cpu when dccpu is read and when dcemu is written. this enables the user to build a cust um debug protocol using only these registers. the dccpu and dcemu registers are available even when the security bit in the flash is active. for more information refer to the avr32uc technical reference manual. 31.3.6.2 breakpoints one of the most fundamental debug features is the abilit y to halt the cpu, to examine registers and the state of the system. this is accomplish ed by breakpoints, of which many types are available: ? unconditional breakpoints are set by writing ocd registers by the debugger, halting the cpu immediately. ? program breakpoints halt the cpu when a specific address in the program is executed. ? data breakpoints halt the cpu when a specific memory address is read or written, allowing variables to be watched. ? software breakpoints halt the cpu when the breakpoint instruction is executed. when a breakpoint triggers, the cpu enters debug mode, and the d bit in the status register is set. this is a privileged mode with dedicated return address and return status registers. all privi- leged instructions are permitted. debug mode can be entered as either ocd mode, running instructions from the debugger, or monitor mode, running instructions from program memory. 31.3.6.3 ocd mode when a breakpoint triggers, the cpu enters ocd mode, and instructions are fetched from the debug instruction ocd register. each time this register is written by the debugger, the instruc- tion is executed, allowing the debugger to execut e cpu instructions directly. the debug master can e.g. read out the register file by issuing mtdr instructions to the cpu, writing each register to the debug communication channel ocd registers. 31.3.6.4 monitor mode since the ocd registers are directly accessible by the cpu, it is possible to build a software- based debugger that runs on the cpu itself. setting the monitor mode bit in the development control register causes the cpu to enter monitor mode instead of ocd mode when a breakpoint triggers. monitor mode is similar to ocd mode, except that instructions are fetched from the debug exception vector in regular program memory, instead of issued by the debug master. 31.3.6.5 program counter monitoring normally, the cpu would need to be halted for a debugger to examine the current pc value. however, the at32uc3l0128/256 also proves a debug program counter ocd register, where the debugger can continuously read the current pc without affecting the cpu. this allows the debugger to generate a simple statistic of the time spent in various areas of the code, easing code optimization. 31.3.7 memory service unit the memory service unit (msu) is a block dedicated to test and debug functionality. it is con- trolled through a dedicated set of registers addressed through the service access bus.
745 32145c?06/2013 at32uc3l0128/256 31.3.7.1 cyclic redundancy check (crc) the msu can be used to automatically calculate the crc of a block of data in memory. the msu will then read out each word in the specif ied memory block and re port the crc32-value in an msu register. 31.3.7.2 nanotrace the msu additionally supports nanotrace. this is a 32-bit avr-specific feature, in which trace data is output to memory instead of the aux port . this allows the trace data to be extracted by the debugger through the sab, enabling trace features for awire- or jtag-based debuggers. the user must write msu registers to configure the address and size of the memory block to be used for nanotrace. the nanotrace buffer can be anywhere in the physical address range, including internal and external ram, through an ebi, if present. this area may not be used by the application running on the cpu. 31.3.8 aux-based debug features utilizing the auxiliary (aux) port gives a ccess to a wide range of advanced debug features. of prime importance are the trace features, which allow an external debugger to receive continuous information on the program execution in the cpu. additionally, event in and event out pins allow external events to be correlated with the program flow. debug tools utilizing the aux port should connec t to the device throug h a nexus-compliant mic- tor-38 connector, as described in the avr32uc technical reference manual. this connector includes the jtag signals and the reset_n pin, giving full access to the programming and debug features in the device.
746 32145c?06/2013 at32uc3l0128/256 figure 31-4. aux+jtag based debugger 31.3.8.1 trace operation trace features are enabled by writing ocd re gisters by the debugger. the ocd extracts the trace information from the cpu, compresses this information and formats it into variable-length messages according to the nexus standard. the messages are buffered in a 16-frame transmit queue, and are output on the aux port one frame at a time. the trace features can be configured to be very selective, to reduce the bandwidth on the aux port. in case the transmit queue overflows, er ror messages are produced to indicate loss of data. the transmit queue module can optionally be configured to halt the cpu when an overflow occurs, to prevent the loss of messages, at the expense of longer run-time for the program. 31.3.8.2 program trace program trace allows the debugger to continuously monitor the program execution in the cpu. program trace messages are generated for every branch in the program, and contains com- pressed information, which allows the debugger to correlate the message with the source code to identify the branch instruction and target address. 31.3.8.3 data trace data trace outputs a message every time a specific location is read or written. the message contains information about the type (read/write) and size of the access, as well as the address and data of the accessed location. the at32uc3l0128/256 contains two data trace channels, avr32 aux+jtag debug tool jtag aux high speed m ictor38 trace buffer pc
747 32145c?06/2013 at32uc3l0128/256 each of which are controlled by a pair of ocd registers which determine the range of addresses (or single address) which should produce data trace messages. 31.3.8.4 ownership trace program and data trace operate on virtual addresses. in cases where an operating system runs several processes in overlapping virtual memory segments, the ownership trace feature can be used to identify th e process switch. when the o/s activates a process, it will write the process id number to an ocd register, which produces an ownership trace message, allowing the debug- ger to switch context for the subsequent progra m and data trace messages. as the use of this feature depends on the software running on the cpu, it can also be used to extract other types of information from the system. 31.3.8.5 watchpoint messages the breakpoint modules normally used to generate program and data breakpoints can also be used to generate watchpoint messages, allowing a debugger to monitor program and data events without halting the cpu. watchpoints can be enabled independently of breakpoints, so a breakpoint module can optionally halt the cpu w hen the trigger condition occurs. data trace modules can also be configured to produce watc hpoint messages instead of regular data trace messages. 31.3.8.6 event in and event out pins the aux port also contains an event in pin (evti_n) and an event out pin (evto_n). evti_n can be used to trigger a breakpoint when an external event occurs. it can also be used to trigger specific program and data trace synchronization messages, allowing an external event to be correlated to the program flow. when the cpu enters debug mode, a debug status message is transmitted on the trace port. all trace messages can be timestamped when they are received by the debug tool. however, due to the latency of the trans mit queue buffering, the timest amp will not be 100% accurate. to improve this, evto_n can toggle every time a message is inserted into the transmit queue, allowing trace messages to be timestamped prec isely. evto_n can also toggle when a break- point module triggers, or when the cpu enters debug mode, for any reason. this can be used to measure precisely when the respective internal event occurs. 31.3.8.7 software quality analysis (sqa) software quality analysis (sqa) deals with two important issues regarding embedded software development. code coverage involves identifying untested parts of the embedded code, to improve test procedures and thus the quality of the released software. performance analysis allows the developer to precisely quantify the time spent in various parts of the code, allowing bottlenecks to be identified and optimized. program trace must be used to accomplish these tasks without instrumenting (altering) the code to be examined. however, traditional program trace cannot reconstruct the current pc value without correlating the trace information with the source code, which cannot be done on-the-fly. this limits program trace to a relatively short time segment, determined by the size of the trace buffer in the debug tool. the ocd system in at32uc3l0128/256 extends program trace with sqa capabilities, allowing the debug tool to reconstruct the pc value on -the-fly. code coverage and performance analysis can thus be reported for an unlimited execution sequence.
748 32145c?06/2013 at32uc3l0128/256 31.3.9 module configuration the bit mapping of the peripheral debug register (pdbg) is described in table 31-7 . please refer to the on-chip debug chapter in the avr32uc technical reference manual for details. table 31-7. bit mapping of the peripheral debug register (pdbg) bit peripheral 0 ast 1wdt
749 32145c?06/2013 at32uc3l0128/256 31.4 jtag and boundary-scan (jtag) rev: 2.2.2.4 31.4.1 features ? ieee1149.1 compli ant jtag interface ? boundary-scan chain fo r board-level testing ? direct memory access and programming capabilities through jtag interface 31.4.2 overview the jtag interface offers a four pin programming and debug solution, including boundary-scan support for board-level testing. figure 31-5 on page 750 shows how the jtag is connected in an 32-bit avr device. the tap controller is a state machine controlled by the tck and tms signals. the tap controller selects either the jtag instruction register or one of several data registers as the scan chain (shift register) between the tdi-input and tdo-output. the instruction register holds jt ag instructions controlling the be havior of a data register. the device identification register, bypass register , and the boundary-scan chain are the data reg- isters used for board-level testing. the reset register can be used to keep the device reset during test or programming. the service access bus (sab) interface cont ains address and data registers for the service access bus, which gives access to on-chip debug, programming, and other functions in the device. the sab offers se veral modes of access to the addr ess and data regist ers, as described in section 31.4.11 . section 31.5 lists the supported jtag instructions, wit h references to the description in this document.
750 32145c?06/2013 at32uc3l0128/256 31.4.3 block diagram figure 31-5. jtag and boundary-scan access 31.4.4 i/o lines description 31.4.5 product dependencies in order to use this module, other parts of the system must be configured correctly, as described below. table 31-8. i/o line description pin name pin description type active level reset_n external reset pin. used when enabling and disabling the jtag. input low tck test clock input. fully asynchronous to system clock frequency. input tms test mode select, sampled on rising tck. input tdi test data in, sampled on rising tck. input tdo test data out, driven on falling tck. output 32-bit avr device jtag data registers tap controller instruction register device identification register by-pass register reset register service access bus interface boundary scan chain pins and analog blocks data register scan enable jtag pins boundary scan enable 2nd jtag device jtag master tdi tdo part specific registers ... tdo tdi tms tms tck tck instruction register scan enable sab internal i/o lines jtag tms tdi tdo tck
751 32145c?06/2013 at32uc3l0128/256 31.4.5.1 i/o lines the tms, tdi, tdo, and tck pins are multiplexed with i/o lines. when the jtag is used the associated pins must be enabled. to enable the jtag pins, refer to section 31.4.7 . while using the multiplexed jtag lines all norma l peripheral activity on these lines is disabled. the user must make sure that no external peripheral is blocking the jtag lines while debugging. 31.4.5.2 power management when an instruction that accesses the sab is loaded in the instruction register, before entering a sleep mode, the system clocks are not switched off to a llow debugging in sleep modes. this can lead to a program behaving differently when debugging. 31.4.5.3 clocks the jtag interface uses the external tck pin as clock source. this clock must be provided by the jtag master. instructions that use the sab bus requires the internal main clock to be running. 31.4.6 jtag interface the jtag interface is accessed through the dedicated jtag pins shown in table 31-8 on page 750 . the tms control line navigates the tap controller, as shown in figure 31-6 on page 752 . the tap controller manages the serial access to the jtag instruction and data registers. data is scanned into the selected instruction or data register on tdi, and out of the register on tdo, in the shift-ir and shift-dr states, respectively. th e lsb is shifted in and out first. tdo is high- z in other states than shift-ir and shift-dr. the device implements a 5-bit instruction regist er (ir). a number of public jtag instructions defined by the jtag standard are supported, as described in section 31.5.2 , as well as a num- ber of 32-bit avr-specific private jtag instructions described in section 31.5.3 . each instruction selects a specific data register for the shift-dr path, as described for each instruction.
752 32145c?06/2013 at32uc3l0128/256 figure 31-6. tap controller state diagram test-logic- reset run-test/ idle select-dr scan select-ir scan capture-dr capture-ir shift-dr shift-ir exit1-dr exit1-ir pause-dr pause-ir exit2-dr exit2-ir update-dr update-ir 0 1 1 1 0 0 1 0 1 1 0 0 1 0 1 1 1 0 1 1 0 0 1 1 0 1 0 0 0 0 0 1
753 32145c?06/2013 at32uc3l0128/256 31.4.7 how to initialize the module to enable the jtag pins the tck pin must be held low while the reset_n pin is released. after enabling the jtag interface the halt bit is set automatically to prevent the system from run- ning code after the interface is enabled. to make the cpu run again set halt to zero using the halt command.. jtag operation when reset_n is pulled low is not possible. independent of the initial state of the tap controller, the test-logic-reset state can always be entered by holding tms high for 5 tck clock periods. this sequence should always be applied at the start of a jtag session and after enabling the jtag pins to bring the tap controller into a defined state before applying jtag commands. applying a 0 on tms for 1 tck period brings the tap controller to the run-test/idle state, which is the starting point for jtag operations. 31.4.8 how to disable the module to disable the jtag pins the tck pin must be held high while reset _n pin is released. 31.4.9 typical sequence assuming run-test/idle is the present state, a typical scenario for using the jtag interface follows. 31.4.9.1 scanning in jtag instruction at the tms input, apply the sequence 1, 1, 0, 0 at the rising edges of tck to enter the shift instruction register (shift-ir) state. while in this state, shift the 5 bits of the jtag instructions into the jtag instruction register from the tdi input at the rising edge of tck. during shifting, the jtag outputs status bits on tdo, refer to section 31.5 for a description of these. the tms input must be held low during input of the 4 lsbs in order to remain in the shift-ir state. the jtag instruction selects a particular data register as path between tdi and tdo and controls the circuitry surrounding th e selected data register. apply the tms sequence 1, 1, 0 to re-enter the run-test/idle state. the instruction is latched onto the parallel output from the shift register path in the update-ir state. the exit-ir, pause-ir, and exit2-ir states are only used for navigating the state machine. figure 31-7. scanning in jtag instruction 31.4.9.2 scanning in/out data at the tms input, apply the sequence 1, 0, 0 at the rising edges of tck to enter the shift data register (shift-dr) state. while in this state, upload the selected data register (selected by the present jtag instruction in the jtag instruction register) from the tdi input at the rising edge tck tap state tlr rti seldr selir capir shir ex1ir updir rti tms tdi instruction tdo impldefined
754 32145c?06/2013 at32uc3l0128/256 of tck. in order to remain in the shift-dr state, the tms input must be held low. while the data register is shifted in from the tdi pin, the parallel inputs to the data register captured in the capture-dr state is shifted out on the tdo pin. apply the tms sequence 1, 1, 0 to re-enter the run-test/idle state. if the selected data register has a latched parallel-output, the latching takes place in the update-dr state. the exit-dr, pause-dr, and exit2-dr states are only used for navigating the state machine. as shown in the state diagram, the run-test/idle state need not be entered between selecting jtag instruction and using data registers. 31.4.10 boundary-scan the boundary-scan chain has the c apability of driving an d observing the logic levels on the digi- tal i/o pins, as well as the boundary between digi tal and analog logic for analog circuitry having off-chip connections. at system level, all ics having jtag capabilities ar e connected serially by the tdi/tdo signals to form a long shift register. an external controller sets up the devices to drive values at their output pins, and observe the input values received from other devices. the controller compares the received data with the expected result. in this way, boundary-scan pro- vides a mechanism for testing interconnections and integrity of components on printed circuits boards by using the 4 tap signals only. the four ieee 1149.1 defined mandatory jtag in structions idcode, bypass, sample/pre- load, and extest can be used for testing the printed circuit board. initial scanning of the data register path will show th e id-code of the device, sinc e idcode is the default jtag instruction. it may be desirable to have the 32-bit avr device in reset during test mode. if not reset, inputs to the device may be determined by the scan operations, and the internal software may be in an undetermi ned state when exiting the test mode. if nee ded, the bypass instruction can be issued to make the shortest possible scan chain through the device. the device can be set in the reset state either by pulling the external resetn pi n low, or issuing the avr_reset instruction with appropriate setting of the reset data register. the extest instruction is used for sampling external pins and loading output pins with data. the data from the output latch will be driven out on the pins as soon as the extest instruction is loaded into the jtag ir-register. therefore, the sample/preload should also be used for setting initial values to the scan ring, to avoid damaging the board when issuing the extest instruction for the first time. sample/preload c an also be used for taking a snapshot of the external pins during normal operation of the part. when using the jtag interface for boundary-scan, the jtag tck clock is independent of the internal chip clock. the internal chip clock is not required to run during boundary-scan operations. note: for pins connected to 5v lines care should be taken to not drive the pins to a logic one using boundary-scan, as this will create a current fl owing from the 3,3v driver to the 5v pull-up on the line. optionally a series resistor can be added between the line and the pin to reduce the current. details about the boundary-scan chain can be found in the bsdl file for the device. this can be found on the atmel website. 31.4.11 service access bus the avr32 architecture offers a common interface for access to on-chip debug, programming, and test functions. these are mapped on a common bus called t he service acce ss bus (sab),
755 32145c?06/2013 at32uc3l0128/256 which is linked to the jtag through a bus master module, which also handles synchronization between the tck and sab clocks. for more information about the sab and a list of sab slaves see the service access bus chapter. 31.4.11.1 sab address mode the memory_sized_access instruction allows a sized read or write to any 36-bit address on the bus. memory_word_access is a shorthand instruction for 32-bit accesses to any 36-bit address, while the nexus_access instruct ion is a nexus-compliant shorthand instruc- tion for accessing the 32-bit ocd registers in the 7-bit address space reserved for these. these instructions require two passes through the shift-dr tap state: one for the address and control information, and one for data. 31.4.11.2 block transfer to increase the transfer rate, consecutive me mory accesses can be accomplished by the memory_block_access instruction, which only r equires a single pass through shift-dr for data transfer only. the address is automatically incremented according to the size of the last sab transfer. 31.4.11.3 canceling a sab access it is possible to abort an ongoing sab access by the cancel_access instruction, to avoid hanging the bus due to an extremely slow slave. 31.4.11.4 busy reporting as the time taken to perform an access may vary depending on system activity and current chip frequency, all the sab access jtag instructions can return a busy indicator. this indicates whether a delay needs to be inserted, or an operation needs to be repeated in order to be suc- cessful. if a new acce ss is requested while the sab is busy, the request is ignored. the sab becomes busy when: ? entering update-dr in the address phase of any read operation, e.g., after scanning in a nexus_access address wit h the read bit set. ? entering update-dr in the data phase of any write operation, e.g., after scanning in data for a nexus_access write. ? entering update-dr during a memory_block_access. ? entering update-dr after scanning in a counter value for sync. ? entering update-ir af ter scanning in a memory_block _access if the previous access was a read and data was scanned after scanning the address. the sab becomes ready again when: ? a read or write operation completes. ? a sync countdown completed. ? a operation is cancelled by the cancel_access instruction. what to do if the busy bit is set: ? during shift-ir: the new instruction is selected, but the previous operation has not yet completed and will continue (u nless the new instruction is cancel_access). you may
756 32145c?06/2013 at32uc3l0128/256 continue shifting the same instruction until the busy bit clears, or start shifting data. if shifting data, you must be prepared that the data shift may also report busy. ? during shift-dr of an address: the new address is ignored. the sab stays in address mode, so no data must be shifted. repeat the address until the busy bit clears. ? during shift-dr of read data: the read data is invalid. the sab stays in data mode. repeat scanning until the busy bit clears. ? during shift-dr of write data: the write data is ignored. the sab stays in data mode. repeat scanning until the busy bit clears. 31.4.11.5 error reporting the service access bus may not be able to comp lete all accesses as requested. this may be because the address is invalid, the addressed area is read-only or cannot handle byte/halfword accesses, or because the chip is set in a protected mode where only limited accesses are allowed. the error bit is updated when an access complete s, and is cleared when a new access starts. what to do if the error bit is set: ? during shift-ir: the new instruction is selected. the last operation performed using the old instruction did not complete successfully. ? during shift-dr of an address: the previous operation failed. the new address is accepted. if the read bit is set, a read operation is started. ? during shift-dr of read data: the read operation failed, and the read data is invalid. ? during shift-dr of write data: the previous write operation failed. the new data is accepted and a write operation started. this should only occur during block writes or stream writes. no error can occur between scanning a write address and the following write data. ? while polling with cancel_access: the previous access was cancelled. it may or may not have actually completed. ? after power-up: the error bit is set after power up, but there has been no previous sab instruction so this error can be discarded. 31.4.11.6 protected reporting a protected status may be reported during shift-ir or shift-dr. this indicates that the security bit in the flash controller is set and that the chip is locked for access, according to section 31.5.1 . the protected state is reported when: ? the flash controller is under reset. this can be due to the avr_reset command or the reset_n line. ? the flash controller has not read the security bit from the flash yet (this will take a a few ms). happens after the flash controller reset has been released. ? the security bit in the flash controller is set. what to do if the protected bit is set: ? release all active avr_ reset domains, if any. ? release the reset_n line. ? wait a few ms for the security bit to clear. it can be set temporarily due to a reset.
757 32145c?06/2013 at32uc3l0128/256 ? perform a chip_erase to clear the security bit. note : this will erase all the contents of the non-volatile memory. 31.5 jtag instruction summary the implemented jtag instructions in the 32-bit avr are shown in the table below. 31.5.1 security restrictions when the security fuse in the flash is programmed, the following jtag instructions are restricted: ? nexus_access ? memory_word_access ? memory_block_access ? memory_sized_access for description of what memory locations remain accessible, please refer to the sab address map. full access to these instructions is re-enabled when the security fuse is erased by the chip_erase jtag instruction. table 31-9. jtag instruction summary instruction opcode instruction description 0x01 idcode select the 32-bit device ident ification register as data register. 0x02 sample_preload take a snapshot of external pin values without affe cting system operation. 0x03 extest select boundary-scan chain as data register for testing circuitry external to the device. 0x04 intest select boundary-scan chain for internal testing of the device. 0x06 clamp bypass device through bypass register, while driving outputs from boundary- scan register. 0x0c avr_reset apply or remove a static reset to the device 0x0f chip_erase erase the device 0x10 nexus_access select the sab address and data regist ers as data register for the tap. the registers are accessed in nexus mode. 0x11 memory_word_access select the sab address an d data registers as data register for the tap. 0x12 memory_block_access select the sab data register as data register for the tap. the address is auto-incremented. 0x13 cancel_access cancel an ongoing nexus or memory access. 0x14 memory_service select the sab address and data regist ers as data register for the tap. the registers are accessed in memory service mode. 0x15 memory_sized_access select the sab address an d data registers as data register for the tap. 0x17 sync synchronization counter 0x1c halt halt the cpu for safe programming. 0x1f bypass bypass this device through the bypass register. others n/a acts as bypass
758 32145c?06/2013 at32uc3l0128/256 note that the security bit will read as programmed and block these instructions also if the flash controller is statically reset. other security mechanisms can also restrict these functions. if such mechanisms are present they are listed in the sab address map section. 31.5.1.1 notation table 31-11 on page 758 shows bit patterns to be shifted in a format like " peb01 ". each charac- ter corresponds to one bit, and eight bits are grouped together for readability. the least significantbit is always shifted first, and the mo st significant bit shifted last. the symbols used are shown in table 31-10 . in many cases, it is not required to shift all bits through the data register. bit patterns are shown using the full width of the shift register, but the suggested or required bits are emphasized using bold text. i.e. given the pattern " aaaaaaar xxxxxxxx xxxxxxxx xxxxxxxx xx", the shift register is 34 bits, but the test or debug unit may choose to shift only 8 bits " aaaaaaar ". the following describes how to interpret the fields in the instruction description tables: table 31-10. symbol description symbol description 0 constant low value - always reads as zero. 1 constant high value - always reads as one. a an address bit - always scanned with the least significant bit first b a busy bit. reads as one if the sab was busy, or zero if it was not. see section 31.4.11.4 for details on how the busy reporting works. d a data bit - always scanned with the least significant bit first. e an error bit. reads as one if an error occurred, or zero if not. see section 31.4.11.5 for details on how the error reporting works. p the chip protected bit. some devices may be se t in a protected state where access to chip internals are severely restricted. see the docum entation for the specific device for details. on devices without this possibility, this bit always reads as zero. r a direction bit. set to one to request a read, set to zero to request a write. s a size bit. the size encoding is described where used. x a don?t care bit. any value can be shifted in, and output data should be ignored. table 31-11. instruction description instruction description ir input value shows the bit pattern to shift into ir in t he shift-ir state in order to select this instruction. the pattern is show both in binary and in hexadecimal form for convenience. example: 10000 (0x10) ir output value shows the bit pattern shifted out of ir in t he shift-ir state when this instruction is active. example: peb01
759 32145c?06/2013 at32uc3l0128/256 31.5.2 public jtag instructions the jtag standard defines a number of public jtag instructions. these instructions are described in the sections below. 31.5.2.1 idcode this instruction selects the 32 bit device identification register (did) as data register. the did register consists of a version number, a device number, and the manufacturer code chosen by jedec. this is the default instruction after a jtag reset. details about the did register can be found in the module configuration section at the end of this chapter. starting in run-test/idle, the device identification register is accessed in the following way: 1. select the ir scan path. 2. in capture-ir: the ir output value is latched into the shift register. 3. in shift-ir: the instruction register is shifted by the tck input. 4. return to run-test/idle. 5. select the dr scan path. 6. in capture-dr: the idcode value is latched into the shift register. 7. in shift-dr: the idcode scan ch ain is shifted by the tck input. 8. return to run-test/idle. 31.5.2.2 sample_preload this instruction takes a snap-shot of the input/ output pins without affect ing the system operation, and pre-loading the scan chain without updating the dr-latch. the boundary-scan chain is selected as data register. starting in run-test/idle, the device identification register is accessed in the following way: dr size shows the number of bits in the data register chain when this instruction is active. example: 34 bits dr input value shows which bit pattern to shift into the data register in the shift-dr state when this instruction is active. multiple such line s may exist, e.g., to distinguish between reads and writes. example: aaaaaaar xxxxxxxx xxxxxxxx xxxxxxxx xx dr output value shows the bit pattern shifted out of the dat a register in the sh ift-dr state when this instruction is active. multiple such line s may exist, e.g., to distinguish between reads and writes. example: xx xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxeb table 31-11. instruction description (continued) instruction description table 31-12. idcode details instructions details ir input value 00001 (0x01) ir output value p0001 dr size 32 dr input value xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx dr output value device identification register
760 32145c?06/2013 at32uc3l0128/256 1. select the ir scan path. 2. in capture-ir: the ir output value is latched into the shift register. 3. in shift-ir: the instruction register is shifted by the tck input. 4. return to run-test/idle. 5. select the dr scan path. 6. in capture-dr: the data on the external pins are sampled into the boundary-scan chain. 7. in shift-dr: the boundary-scan chain is shifted by the tck input. 8. return to run-test/idle. 31.5.2.3 extest this instruction selects the boundary-scan chain as data register for testing circuitry external to the 32-bit avr package. the contents of the latched outputs of the boundary-scan chain is driven out as soon as the jtag ir-register is loaded with the extest instruction. starting in run-test/idle, the extest instruction is accessed the following way: 1. select the ir scan path. 2. in capture-ir: the ir output value is latched into the shift register. 3. in shift-ir: the instruction register is shifted by the tck input. 4. in update-ir: the data from the boundary-scan chain is applied to the output pins. 5. return to run-test/idle. 6. select the dr scan path. 7. in capture-dr: the data on the external pins is sampled into the boundary-scan chain. 8. in shift-dr: the boundary-scan chain is shifted by the tck input. 9. in update-dr: the data from the scan chain is applied to the output pins. 10. return to run-test/idle. table 31-13. sample_preload details instructions details ir input value 00010 (0x02) ir output value p0001 dr size depending on boundary-scan chain, see bsdl-file. dr input value depending on boundary-scan chain, see bsdl-file. dr output value depending on boundary-scan chain, see bsdl-file. table 31-14. extest details instructions details ir input value 00011 (0x03) ir output value p0001 dr size depending on boundary-scan chain, see bsdl-file. dr input value depending on boundary-scan chain, see bsdl-file. dr output value depending on boundary-scan chain, see bsdl-file.
761 32145c?06/2013 at32uc3l0128/256 31.5.2.4 intest this instruction selects the boundary-scan chain as data register for testing internal logic in the device. the logic inputs are determined by the boundary-scan chain, and the logic outputs are captured by the boundary-scan chain. the device output pins are driven from the boundary-scan chain. starting in run-test/idle, the intest instruction is access ed the following way: 1. select the ir scan path. 2. in capture-ir: the ir output value is latched into the shift register. 3. in shift-ir: the instruction register is shifted by the tck input. 4. in update-ir: the data from the boundary-scan chain is applied to the internal logic inputs. 5. return to run-test/idle. 6. select the dr scan path. 7. in capture-dr: the data on the internal logic is sampled into the boundary-scan chain. 8. in shift-dr: the boundary-scan chain is shifted by the tck input. 9. in update-dr: the data from the boundary-scan chain is applied to internal logic inputs. 10. return to run-test/idle. 31.5.2.5 clamp this instruction selects the bypass register as data register. the device output pins are driven from the boundary-scan chain. starting in run-test/idle, the clamp instruction is acce ssed the following way: 1. select the ir scan path. 2. in capture-ir: the ir output value is latched into the shift register. 3. in shift-ir: the instruction register is shifted by the tck input. 4. in update-ir: the data from the boundary-scan chain is applied to the output pins. 5. return to run-test/idle. 6. select the dr scan path. 7. in capture-dr: a logic ?0? is loaded into the bypass register. 8. in shift-dr: data is scanned from tdi to tdo through the bypass register. table 31-15. intest details instructions details ir input value 00100 (0x04) ir output value p0001 dr size depending on boundary-scan chain, see bsdl-file. dr input value depending on boundary-scan chain, see bsdl-file. dr output value depending on boundary-scan chain, see bsdl-file.
762 32145c?06/2013 at32uc3l0128/256 9. return to run-test/idle. 31.5.2.6 bypass this instruction selects the 1-bit bypass register as data register. starting in run-test/idle, the clamp instruction is acce ssed the following way: 1. select the ir scan path. 2. in capture-ir: the ir output value is latched into the shift register. 3. in shift-ir: the instruction register is shifted by the tck input. 4. return to run-test/idle. 5. select the dr scan path. 6. in capture-dr: a logic ?0? is loaded into the bypass register. 7. in shift-dr: data is scanned from tdi to tdo through the bypass register. 8. return to run-test/idle. 31.5.3 private jtag instructions the 32-bit avr defines a number of private jt ag instructions, not defined by the jtag stan- dard. each instruction is br iefly described in text, with det ails following in table form. 31.5.3.1 nexus_access this instruction allows nexus-compliant access to the on-chip debug registers through the sab. the 7-bit register index, a read/write co ntrol bit, and the 32-bit data is accessed through the jtag port. the data register is alternately interpreted by the sab as an address register and a data regis- ter. the sab starts in addre ss mode after the nexus_access instruction is selected, and toggles between address and data mode each time a data scan completes with the busy bit cleared. note : the polarity of the direction bit is inverse of the nexus standard. table 31-16. clamp details instructions details ir input value 00110 (0x06) ir output value p0001 dr size 1 dr input value x dr output value x table 31-17. bypass details instructions details ir input value 11111 (0x1f) ir output value p0001 dr size 1 dr input value x dr output value x
763 32145c?06/2013 at32uc3l0128/256 starting in run-test/idle, ocd registers are accessed in the following way: 1. select the ir scan path. 2. in capture-ir: the ir output value is latched into the shift register. 3. in shift-ir: the instruction register is shifted by the tck input. 4. return to run-test/idle. 5. select the dr scan path. 6. in shift-dr: scan in the direction bit (1=read, 0=write) and the 7-bit address for the ocd register. 7. go to update-dr and re-enter select-dr scan. 8. in shift-dr: for a read operation, scan out the contents of the addressed register. for a write operation, scan in the new contents of the register. 9. return to run-test/idle. for any operation, the full 7 bits of the address must be provided. for write operations, 32 data bits must be provided, or the result will be undefined. for read ope rations, shifting may be termi- nated once the required number of bits have been acquired. 31.5.3.2 memory_service this instruction allows access to registers in an optional memory service unit. the 7-bit register index, a read/write control bit, and the 32-bit data is accessed through the jtag port. the data register is alternately interpreted by the sab as an address register and a data regis- ter. the sab starts in addres s mode after the memory_service instruction is selected, and toggles between address and data mode each time a data scan completes with the busy bit cleared. starting in run-test/idle, memory service registers are accessed in the following way: 1. select the ir scan path. 2. in capture-ir: the ir output value is latched into the shift register. 3. in shift-ir: the instruction register is shifted by the tck input. 4. return to run-test/idle. 5. select the dr scan path. 6. in shift-dr: scan in the direction bit (1=read, 0=write) and the 7-bit address for the memory service register. table 31-18. nexus_access details instructions details ir input value 10000 (0x10) ir output value peb01 dr size 34 bits dr input value (address phase) aaaaaaar xxxxxxxx xxxxxxxx xxxxxxxx xx dr input value (data read phase) xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx xx dr input value (data write phase) dddddddd dddddddd dddddddd dddddddd xx dr output value (address phase) xx xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxeb dr output value (data read phase) eb dddddddd dddddddd dddddddd dddddddd dr output value (data write phase) xx xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxeb
764 32145c?06/2013 at32uc3l0128/256 7. go to update-dr and re-enter select-dr scan. 8. in shift-dr: for a read operation, scan out the contents of the addressed register. for a write operation, scan in the new contents of the register. 9. return to run-test/idle. for any operation, the full 7 bits of the address must be provided. for write operations, 32 data bits must be provided, or the result will be undefined. for read ope rations, shifting may be termi- nated once the required number of bits have been acquired. 31.5.3.3 memory_sized_access this instruction allows access to the entire service access bus data area. data is accessed through a 36-bit byte index, a 2-bit size, a direction bit, and 8, 16, or 32 bits of data. not all units mapped on the sab bus may support all sizes of accesses, e.g., some may only support word accesses. the data register is alternately interpreted by the sab as an address register and a data regis- ter. the sab starts in address mode after the memory_sized_access instruction is selected, and toggles between address and data mode each time a data scan completes with the busy bit cleared. table 31-19. memory_service details instructions details ir input value 10100 (0x14) ir output value peb01 dr size 34 bits dr input value (address phase) aaaaaaar xxxxxxxx xxxxxxxx xxxxxxxx xx dr input value (data read phase) xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx xx dr input value (data write phase) dddddddd dddddddd dddddddd dddddddd xx dr output value (address phase) xx xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxeb dr output value (data read phase) eb dddddddd dddddddd dddddddd dddddddd dr output value (data write phase) xx xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxeb
765 32145c?06/2013 at32uc3l0128/256 the size field is encoded as i table 31-20 . starting in run-test/idle, sab data is accessed in the following way: 1. select the ir scan path. 2. in capture-ir: the ir output value is latched into the shift register. 3. in shift-ir: the instruction register is shifted by the tck input. 4. return to run-test/idle. 5. select the dr scan path. 6. in shift-dr: scan in the direction bit (1=read, 0=write), 2-bit access size, and the 36-bit address of the data to access. 7. go to update-dr and re-enter select-dr scan. 8. in shift-dr: for a read operation, scan out the contents of the addressed area. for a write operation, scan in the new contents of the area. 9. return to run-test/idle. for any operation, the full 36 bits of the address must be provided. for write operations, 32 data bits must be provided, or the result will be undefined. for read ope rations, shifting may be termi- nated once the required number of bits have been acquired. table 31-20. size field semantics size field value access size data alignment 00 byte (8 bits) address modulo 4 : data alignment 0: dddddddd xxxxxxxx xxxxxxxx xxxxxxxx 1: xxxxxxxx dddddddd xxxxxxxx xxxxxxxx 2: xxxxxxxx xxxxxxxx dddddddd xxxxxxxx 3: xxxxxxxx xxxxxxxx xxxxxxxx dddddddd 01 halfword (16 bits) address modulo 4 : data alignment 0: dddddddd dddddddd xxxxxxxx xxxxxxxx 1: not allowed 2: xxxxxxxx xxxxxxxx dddddddd dddddddd 3: not allowed 10 word (32 bits) address modulo 4 : data alignment 0: dddddddd dddddddd dddddddd dddddddd 1: not allowed 2: not allowed 3: not allowed 11 reserved n/a table 31-21. memory_sized_access details instructions details ir input value 10101 (0x15) ir output value peb01 dr size 39 bits dr input value (address phase) aaaaaaaa aaaaaaaa aaaaaaaa aaaaaaaa aaaassr dr input value (data read phase) xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxx dr input value (data write phase) dddddddd dddddddd dddddddd dddddddd xxxxxxx
766 32145c?06/2013 at32uc3l0128/256 31.5.3.4 memory_word_access this instruction allows access to the entire service access bus data area. data is accessed through the 34 msb of the sab addr ess, a direction bit, and 32 bits of data. this instruction is identical to memory_sized_access except that it always does word sized accesses. the size field is implied, and the two lowest address bits are removed and not scanned in . note: this instruction was previously known as memory_access, and is provided for back- wards compatibility. the data register is alternately interpreted by the sab as an address register and a data regis- ter. the sab starts in address mode after the memory_word_access instruction is selected, and toggles between address and data mode each time a data scan completes with the busy bit cleared. starting in run-test/idle, sab data is accessed in the following way: 1. select the ir scan path. 2. in capture-ir: the ir output value is latched into the shift register. 3. in shift-ir: the instruction register is shifted by the tck input. 4. return to run-test/idle. 5. select the dr scan path. 6. in shift-dr: scan in the direction bit (1=read, 0=write) and the 34-bit address of the data to access. 7. go to update-dr and re-enter select-dr scan. 8. in shift-dr: for a read operation, scan out the contents of the addressed area. for a write operation, scan in the new contents of the area. 9. return to run-test/idle. for any operation, the full 34 bits of the address must be provided. for write operations, 32 data bits must be provided, or the result will be undefined. for read ope rations, shifting may be termi- nated once the required number of bits have been acquired. dr output value (address phase) xxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxeb dr output value (dat a read phase) xxxxxeb dddddddd dddddddd dddddddd dddddddd dr output value (d ata write phase) xxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxeb table 31-21. memory_sized_access details (continued) instructions details table 31-22. memory_word_access details instructions details ir input value 10001 (0x11) ir output value peb01 dr size 35 bits dr input value (address phase) aaaaaaaa aaaaaaaa aaaaaaaa aaaaaaaa aar dr input value (data read phase) xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx xxx dr input value (data write phase) dddddddd dddddddd dddddddd dddddddd xxx
767 32145c?06/2013 at32uc3l0128/256 31.5.3.5 memory_block_access this instruction allows ac cess to the entire sab data area. up to 32 bits of data is accessed at a time, while the address is sequentially incr emented from the previously used address. in this mode, the sab a ddress, size, and access direction is not provided with each access. instead, the previous address is auto-incremente d depending on the specified size and the pre- vious operation repeated. the address must be set up in advance with memory_size_access or memory_word_acce ss. it is allowed, but not required, to shift data after shifting the address. this instruction is primarily intended to speed up large quantities of sequential word accesses. it is possible to use it also for byte and halfword accesses, but the overhead in this is case much larger as 32 bits must still be shifted for each access. the following sequence should be used: 1. use the memory_size_access or memory_word_access to read or write the first location. 2. return to run-test/idle. 3. select the ir scan path. 4. in capture-ir: the ir output value is latched into the shift register. 5. in shift-ir: the instruction register is shifted by the tck input. 6. return to run-test/idle. 7. select the dr scan path. t he address will now have increment ed by 1, 2, or 4 (corre- sponding to the next byte, halfword, or word location). 8. in shift-dr: for a read operation, scan out the contents of the next addressed location. for a write operation, scan in the new contents of the next addressed location. 9. go to update-dr. 10. if the block access is not complete, return to select-dr scan and repeat the access. 11. if the block access is complete, return to run-test/idle. for write operations, 32 data bits must be pr ovided, or the result will be undefined. for read operations, shifting may be terminated once the required number of bits have been acquired. dr output value (address phase) xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx xeb dr output value (data read phase) xeb dddddddd dddddddd dddddddd dddddddd dr output value (d ata write phase) xxx xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxeb table 31-22. memory_word_access details (continued) instructions details table 31-23. memory_block_access details instructions details ir input value 10010 (0x12) ir output value peb01 dr size 34 bits dr input value (data read phase) xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx xx
768 32145c?06/2013 at32uc3l0128/256 the overhead using block word access is 4 cycles per 32 bits of data, resulting in an 88% trans- fer efficiency, or 2.1 mbytes per second with a 20 mhz tck frequency. 31.5.3.6 cancel_access if a very slow memory location is accessed du ring a sab memory access, it could take a very long time until the busy bit is cleared, and the sab becomes ready for the next operation. the cancel_access instruction pr ovides a possibility to abort an ongoing transfer and report a timeout to the jtag master. when the cancel_access instruction is select ed, the current access will be terminated as soon as possible. there are no guarantees about how long this will take, as the hardware may not always be able to cancel the access immedi ately. the sab is ready to respond to a new command when the busy bit clears. starting in run-test/idl e, cancel_access is acce ssed in the following way: 1. select the ir scan path. 2. in capture-ir: the ir output value is latched into the shift register. 3. in shift-ir: the instruction register is shifted by the tck input. 4. return to run-test/idle. 31.5.3.7 sync this instruction allows external debuggers and testers to measure the ratio between the external jtag clock and the internal system clock. the sync data register is a 16-bit counter that counts down to zero using the internal system clock. the busy bit stays high until the counter reaches zero. starting in run-test/idle, sync inst ruction is used in the following way: 1. select the ir scan path. 2. in capture-ir: the ir output value is latched into the shift register. 3. in shift-ir: the instruction register is shifted by the tck input. 4. return to run-test/idle. 5. select the dr scan path. dr input value (data write phase) dddddddd dddddddd dddddddd dddddddd xx dr output value (data read phase) eb dddddddd dddddddd dddddddd dddddddd dr output value (data write phase) xx xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxeb table 31-23. memory_block_access details (continued) instructions details table 31-24. cancel_access details instructions details ir input value 10011 (0x13) ir output value peb01 dr size 1 dr input value x dr output value 0
769 32145c?06/2013 at32uc3l0128/256 6. scan in an 16-bit counter value. 7. go to update-dr and re-enter select-dr scan. 8. in shift-dr: scan out the busy bit, and until the busy bit clears goto 7. 9. calculate an approximation to the internal clock speed using the elapsed time and the counter value. 10. return to run-test/idle. the full 16-bit counter value must be provided when starting the synch operation, or the result will be undefined. when r eading status, shifting may be term inated once the required number of bits have been acquired. 31.5.3.8 avr_reset this instruction allows a debugger or tester to directly control separate reset domains inside the chip. the shift register contains one bit for each controllable reset domain. setting a bit to one resets that domain and holds it in reset. setting a bit to zero releases the reset for that domain. the avr_reset instruction can be used in the following way: 1. select the ir scan path. 2. in capture-ir: the ir output value is latched into the shift register. 3. in shift-ir: the instruction register is shifted by the tck input. 4. return to run-test/idle. 5. select the dr scan path. 6. in shift-dr: scan in the value corresponding to the reset domains the jtag master wants to reset into the data register. 7. return to run-test/idle. 8. stay in run test idle for at least 10 tck clock cycles to let the reset propagate to the system. see the device specific document ation for the number of reset domains, and what these domains are. for any operation, all bits must be provided or th e result will be undefined. table 31-25. sync_access details instructions details ir input value 10111 (0x17) ir output value peb01 dr size 16 bits dr input value dddddddd dddddddd dr output value xxxxxxxx xxxxxxeb table 31-26. avr_reset details instructions details ir input value 01100 (0x0c) ir output value p0001
770 32145c?06/2013 at32uc3l0128/256 31.5.3.9 chip_erase this instruction allows a programmer to completely erase all nonvolatile memories in a chip. this will also clear any security bits that are set, so the device can be accessed normally. in devices without non-volatile memories this instruction does nothing, and appears to complete immediately. the erasing of non-volatile memo ries starts as soon as the chi p_erase instruction is selected. the chip_erase instruction selects a 1 bit bypass data register. a chip erase operation should be performed as: 1. reset the system and stop the cpu from executing. 2. select the ir scan path. 3. in capture-ir: the ir output value is latched into the shift register. 4. in shift-ir: the instruction register is shifted by the tck input. 5. check the busy bit that was scanned out during shift-ir. if the busy bit was set goto 2. 6. return to run-test/idle. 31.5.3.10 halt this instruction allows a programmer to easily stop the cpu to ensure that it does not execute invalid code during programming. this instruction selects a 1-bit halt register. setting this bit to one halts the cpu. setting this bit to zero releases the cpu to run normally. the value shifted out from the data register is one if the cpu is halted. before releasing the halt command the cpu needs to be reset to ensure that it will start at the re set startup address. the halt instruction can be used in the following way: 1. select the ir scan path. 2. in capture-ir: the ir output value is latched into the shift register. 3. in shift-ir: the instruction register is shifted by the tck input. 4. return to run-test/idle. 5. select the dr scan path. dr size device specific. dr input value device specific. dr output value device specific. table 31-26. avr_reset details (continued) instructions details table 31-27. chip_erase details instructions details ir input value 01111 (0x0f) ir output value p0b01 where b is the busy bit. dr size 1 bit dr input value x dr output value 0
771 32145c?06/2013 at32uc3l0128/256 6. in shift-dr: scan in the value 1 to halt the cpu, 0 to start cpu execution. 7. return to run-test/idle. table 31-28. halt details instructions details ir input value 11100 (0x1c) ir output value p0001 dr size 1 bit dr input value d dr output value d
772 32145c?06/2013 at32uc3l0128/256 31.5.4 jtag data registers the following device specific registers can be selected as jtag scan chain depending on the instruction loaded in the jtag instruction register. additional registers exist, but are implicitly described in the functional description of the relevant instructions. 31.5.4.1 device identification register the device identification register contains a unique identifier for each product. the register is selected by the idcode instruction, which is the default instruction after a jtag reset. device specific id codes the different device configurations have different jtag id codes, as shown in table 31-29 . note that if the flash co ntroller is statically rese t, the id code will be undefined . 31.5.4.2 reset register the reset register is selected by the avr_reset instruction and co ntains one bit for each reset domain in the device. se tting each bit to one will keep that domain reset until the bit is cleared. msb lsb bit 31 28 27 12 11 1 0 device id revision part number manufacturer id 1 4 bits 16 bits 11 bits 1 bit revision this is a 4 bit number identifying the revision of the component. rev a = 0x0, b = 0x1, etc. part number the part number is a 16 bit code identifying the component. manufacturer id the manufacturer id is a 11 bit code identifying the manufacturer. the jtag manufacturer id for atmel is 0x01f. table 31-29. device and jtag id device name jtag id code (r is the revision number) at32uc3l0256 0xr21c903f at32uc3l0128 0xr21ca03f bit 0 reset domain system system resets the whole chip, except the jtag itself.
773 32145c?06/2013 at32uc3l0128/256 31.5.4.3 boundary-scan chain the boundary-scan chain has the c apability of driving an d observing the logic levels on the digi- tal i/o pins, as well as driving and observing the lo gic levels between the digital i/o pins and the internal logic. typically, output value, output enable, and input data are all available in the boundary scan chain. the boundary scan chain is described in the bsdl (boundary scan description language) file available at the atmel web site.
774 32145c?06/2013 at32uc3l0128/256 31.6 awire debug interface (aw) rev.: 2.3.0.1 31.6.1 features ? single pin debug system. ? half duplex asynchronous communication (uart compatible). ? full duplex mode for direct uart connection. ? compatible with jtag function ality, except boundary scan. ? failsafe packet-oriented protocol. ? read and write on-chip memory and program on -chip flash and fuses through sab interface. ? on-chip debug access through sab interface. ? asynchronous receiver or transmitter when the awire system is not used for debugging. 31.6.2 overview the awire debug interface (aw) offers a single pi n debug solution that is fully compatible with the functionality offered by the jtag interface, except boundary scan. this functionality includes memory access, programming capab ilities, and on-chip debug access. figure 31-8 on page 775 shows how the aw is connected in a 32-bit avr device. the reset_n pin is used both as reset and debug pin. a specia l sequence on reset_n is needed to block the normal reset functionality and enable the aw. the service access bus (sab) interface cont ains address and data registers for the service access bus, which gives access to on-chip debug, programming, and other functions in the device. the sab offers several modes of access to the addre ss and data registers, as dis- cussed in section 31.6.6.8 . section 31.6.7 lists the supported awire commands and responses, with references to the description in this document. if the aw is not used for debugging, the awire ua rt can be used by the user to send or receive data with one stop bit, eight data bits, no parity bits, and one stop bit. this can be controlled through the awire user interface.
775 32145c?06/2013 at32uc3l0128/256 31.6.3 block diagram figure 31-8. awire debug interface block diagram 31.6.4 i/o lines description 31.6.5 product dependencies in order to use this module, other parts of the system must be configured correctly, as described below. table 31-30. i/o lines description name description type data awire data multiplexed wit h the reset_n pin. input/output dataout awire data output in 2-pin mode. output uart reset filter external reset aw_enable reset_n baudrate detector rw sz addr data crc aw control aw user interface sab interface reset command power manager cpu halt command flash controller chip_erase command awire debug interface pb sab
776 32145c?06/2013 at32uc3l0128/256 31.6.5.1 i/o lines the pin used by aw is multiple xed with the reset_n pi n. the reset function ality is the default function of this pin. to enable the awire func tionality on the reset_n pin the user must enable the aw either by sending the enable sequence over the reset_n pin from an external awire master or by enabling the awire user interface. in 2-pin mode data is received on the rese t_n line, but transmitted on the dataout line. after sending the 2_pin_mode command the dataout line is automatically enabled. all other peripheral functions on this pin is disabled. 31.6.5.2 power management when debugging through aw the system clocks are automatically turned on to allow debugging in sleep modes. 31.6.5.3 clocks the awire uart uses the internal 120 mhz rc oscillator (rc120m) as clock source for its operation. when enabling the aw the rc120m is automatically started. 31.6.5.4 external components the aw needs an external pullup on the reset_n pin to ensure that the pin is pulled up when the bus is not driven. 31.6.6 functional description 31.6.6.1 awire communication protocol the aw is accessed through the reset_n pin shown in table 31-30 on page 775 . the aw communicates through a uart operating at vari able baud rate (depending on a sync pattern) with one start bit, 8 data bits (lsb first), one stop bit, and no parity bits. the awire protocol is based upon command packets from an externalmaster and response packets from the slave (aw). the master always initiates communication and decides the baud rate. the packet contains a sync byte (0x55), a comm and/response byte, two length bytes (optional), a number of data bytes as defined in the length field (optional), and two crc bytes. if the com- mand/response has the most significant bit set, the command/response also carries the optional length and data fields. the crc field is not checked if the crc value transmitted is 0x0000. table 31-31. awire packet format field number of bytes description comment optional sync 1 sync pattern (0x55). used by the receiver to set the baud rate clock. no command/ response 1 command from the master or response from the slave. when the most significant bit is set the command/response has a length field. a response has the next most significant bit set. a command does not have this bit set. no
777 32145c?06/2013 at32uc3l0128/256 crc calculation the crc is calculated from th e command/response, length, and data fields. the polynomial used is the fcs16 (or crc-16-ccit) in reve rse mode (0x8408) and the starting value is 0x0000. example command below is an example command from the master with additional data. figure 31-9. example command example response below is an example response from the slave with additional data. figure 31-10. example response length 2 the number of bytes in the data field. ye s data length data according to command/ response. ye s crc 2 crc calculated with the fcs16 polynomial. crc value of 0x0000 makes the awire disregard the crc if the master does not support it. no table 31-31. awire packet format field number of bytes description comment optional baud_rate_clk data_pin ... field sync(0x55) command(0x81) length(msb) length(lsb) ... data(msb) data(lsb) crc(msb) crc(lsb) baud_rate_clk data_pin ... field sync(0x55) response(0xc1) length(msb) length(lsb) ... data(msb) data(lsb) crc(msb) crc(lsb)
778 32145c?06/2013 at32uc3l0128/256 avoiding drive contention when changing direction the awire debug protocol uses one dataline in both directions. to avoid both the master and the slave to drive this line when changing direction the aw has a built in guard time before it starts to drive the line. at reset this guard time is set to maximum (128 bit cycles), but can be lowered by the master upon command. the aw will release the line im mediately after the stop char acter has been transmitted. during the direction change there can be a period when the line is not driven. an external pullup has to be added to reset_n to keep the signal stable when neither master or slave is actively driving the line. 31.6.6.2 the reset_n pin normal reset functionality on the reset_n pin is disabled when using awire. however, the user can reset the system through the reset awire command. during awire operation the reset_n pin should not be conne cted to an external reset ci rcuitry, but disconnected via a switch or a jumper to avoid drive contention and speed problems. figure 31-11. reset circuitry and awire. 31.6.6.3 initializing the aw to enable aw, the user has to send a 0x55 patte rn with a baudrate of 1 khz on the reset_n pin. the aw is enabled after tr ansmitting this pattern and the us er can start transmitting com- mands. this pattern is not the sync pattern for the first command. after enabling the awire debug interface the halt bit is set automatically to prevent the system from running code after the interface is enabled. to make the cpu run again set halt to zero using the halt command. 31.6.6.4 disabling the aw to disable aw, the user can keep the reset_n pin low for 100 ms. this will disable the aw, return reset_n to its normal fu nction, and reset the device. an awire master can also disable awire by sending the disable command. after acking the command the aw will be disabled and rese t_n returns to its normal function. reset_n aw debug interface jumper mcu power manager awire master connector board reset circuitry
779 32145c?06/2013 at32uc3l0128/256 31.6.6.5 resetting the aw the awire master can reset the aw slave by pulling the reset_n pin low for 20 ms. this is equivalent to disabling and then enabling aw. 31.6.6.6 2-pin mode to avoid using special hardware when using a normal uart device as awire master, the awire slave has a 2-pin mode where one pin is used as input and on pin is used as output. to enable this mode the 2_pin_mode command must be sent. after sending the command, all responses will be sent on the dataout pin instead of th e reset_n pin. commands are still received on the reset_n pin. 31.6.6.7 baud rate clock the communication speed is set by the master in the sync field of the command. the aw will use this to resynchronize its baud rate cloc k and reply on this frequency. the minimum fre- quency of the communication is 1 khz. the maximum frequency depends on the internal clock source for the aw (rc120m). the baud rate cl ock is generated by aw with the following formula: where is the baud rate frequency and is the frequency of the internal rc120m. tune is the value returned by the baud_rate response. to find the max frequency the user can issue the tune command to the aw to make it return the tune value. this value can be used to compute the . the maximum operational fre- quency ( ) is then: 31.6.6.8 service access bus the avr32 architecture offers a common interface for access to on-chip debug, programming, and test functions. these are mapped on a common bus called t he service acce ss bus (sab), which is linked to the awire through a bus mast er module, which also handles synchronization between the awire and sab clocks. for more information about the sab and a list of sab slaves see the service access bus chapter. sab clock when accessing the sab through the awire there are no limitations on baud rate frequency compared to chip frequency, although there must be an active system clock in order for the sab accesses to complete. if the system clock (clk_sys) is switched off in sleep mode, activity on the awire pin will restart the clk_sys automatica lly, without waking t he device from sleep. awire masters may optimize the transfer rate by adjusting the baud rate frequency in relation to the clk_sys. this ratio can be measur ed with the memory_speed_request command. when issuing the memory_speed_request comm and a counter value cv is returned. cv can be used to calculate the sab speed ( ) using this formula: f aw tune f ? br 8 ---------------------------- - = f br f aw f aw f brmax f brmax f aw 4 ------- = f sab
780 32145c?06/2013 at32uc3l0128/256 sab address mode the service access bus uses 36 address bits to address memory or registers in any of the slaves on the bus. the bus supports sized accesses of bytes (8 bits), halfwords (16 bits), or words (32 bits). all accesses must be aligned to the size of th e access, i.e. halfword accesses must have the lowest address bit cleared, and word accesses must have the two lowest address bits cleared. two instructions exist to access the sab: memory_write and memory_read. these two instructions write and read words, halfwords, and bytes from the sab. busy reporting if the awire master, during a memory_wri te or a memory_read command, transmit another byte when the awire is still busy send ing the previous byte to the sab, the aw will respond with a memory_read_write_status error. see chapter section 31.6.8.5 for more details. the awire master should adjust its baudrate or delay between bytes when doing sab accesses to ensure that the sab is not overwhelmed with data. error reporting if a write is performed on a non-existing memory location the sab interface will respond with an error. if this happens, all further writes in this command will not be performed and the error and number of bytes written is reported in the memory_readwrite_status message from the aw after the write. if a read is performed on a non-existing memo ry location, the sab in terface will respond with an error. if this happens, the data bytes read after this event are not valid. the aw will include three extra bytes at the end of the transfer to indicate if the transfer was successful, or in the case of an error, how many valid bytes were received. 31.6.6.9 crc errors/nack response the aw will calculate a crc value when receiving the command, lengt h, and data fields of the command packets. if this value differs from the value from the crc field of the packet, the aw will reply with a nack response. otherwis e the command is carried out normally. an unknown command will be re plied with a nack response. in worst case a transmission error can happen in the length or command field of the packet. this can lead to the awire slave trying to receive a command with or without length (opposite of what the master intended) or receive an incorrect number of byte s. the awire slave will then either wait for more data when the master has finished or already have transmitted the nack response in congestion with the master. the master can implement a timeout on every com- mand and reset the slave if no response is returned after the timeout period has ended. f sab 3 f aw cv 3 ? ---------------- - =
781 32145c?06/2013 at32uc3l0128/256 31.6.7 awire command summary the implemented awire commands are shown in the table below. the responses from the aw are listed in section 31.6.8 . all awire commands are described below, with a summary in table form. 31.6.7.1 aya this command asks the aw: ?are you aliv e?, where the aw should respond with an acknowledge. table 31-32. awire command summary command instruction description 0x01 aya ?are you alive?. 0x02 jtag_id asks aw to return the jtag idcode. 0x03 status_request request a status message from the aw. 0x04 tune tell the aw to report the current baud rate. 0x05 memory_speed_request reports the speed difference between the awire control and the sab clock domains. 0x06 chip_erase erases the flash and all volatile memories. 0x07 disable disables the aw. 0x08 2_pin_mode enables the dataout pin and puts the awire in 2-pin mode, where all responses are sent on the dataout pin. 0x80 memory_write writes words, halfwords, or bytes to the sab. 0x81 memory_read reads words, halfwords, or bytes from the sab. 0x82 halt issues a halt command to the device. 0x83 reset issues a reset to the reset controller. 0x84 set_guard_time sets the guard time for the aw. table 31-33. command/response description notation command/response description command/response value shows the command/response value to put into the command/response field of the packet. additional data shows the format of the optional data field if applicable. possible responses shows the possi ble responses for this command. table 31-34. aya details command details command value 0x01 additional data n/a possible responses 0x40: ack ( section 31.6.8.1 ) 0x41: nack ( section 31.6.8.2 )
782 32145c?06/2013 at32uc3l0128/256 31.6.7.2 jtag_id this command instructs the aw to output the jtag idcode in the following response. 31.6.7.3 status_request asks the aw for a status message. 31.6.7.4 tune asks the aw for the current baud rate counter value. 31.6.7.5 memory_speed_request asks the aw for the relative speed between the awire clo ck (rc120m) and t he sab interface. 31.6.7.6 chip_erase this instruction allows a programmer to completely erase all nonvolatile memories in the chip. this will also clear any security bits that are set, so the device can be accessed normally. the command is acked immediately, but the status of the command can be monitored by checking table 31-35. jtag_id details command details command value 0x02 additional data n/a possible responses 0xc0: idcode ( section 31.6.8.3 ) 0x41: nack ( section 31.6.8.2 ) table 31-36. status_request details command details command value 0x03 additional data n/a possible responses 0xc4: status_info ( section 31.6.8.7 ) 0x41: nack ( section 31.6.8.2 ) table 31-37. tune details command details command value 0x04 additional data n/a possible responses 0xc3: baud_rate ( section 31.6.8.6 ) 0x41: nack ( section 31.6.8.2 ) table 31-38. memory_speed_request details command details command value 0x05 additional data n/a possible responses 0xc5: memory_speed ( section 31.6.8.8 ) 0x41: nack ( section 31.6.8.2 )
783 32145c?06/2013 at32uc3l0128/256 the chip erase ongoing bit in the status bytes received after the status_request command. 31.6.7.7 disable disables the aw. the aw will re spond with an ack response and then disable itself. 31.6.7.8 2_pin_mode enables the dataout pin as an ou tput pin. all respons es sent from the awire slave will be sent on this pin, instead of the reset_n pin, starting with the ack for the 2_pin_mode command. 31.6.7.9 memory_write this command enables programming of memo ry/writing to registers on the sab. the memory_write command allows words, halfwords, and bytes to be programmed to a contin- uous sequence of addresses in one operation. before transferring the data, the user must supply: 1. the number of data bytes to write + 5 (size and starting address) in the length field. 2. the size of the transfer: words, halfwords, or bytes. 3. the starting address of the transfer. table 31-39. chip_erase details command details command value 0x06 additional data n/a possible responses 0x40: ack ( section 31.6.8.1 ) 0x41: nack ( section 31.6.8.2 ) table 31-40. disable details command details command value 0x07 additional data n/a possible responses 0x40: ack ( section 31.6.8.1 ) 0x41: nack ( section 31.6.8.2 ) table 31-41. disable details command details command value 0x07 additional data n/a possible responses 0x40: ack ( section 31.6.8.1 ) 0x41: nack ( section 31.6.8.2 )
784 32145c?06/2013 at32uc3l0128/256 the 4 msb of the 36 bit sab address are submitted to gether with the size fi eld (2 bits). then fol- lows the 4 remaining address bytes and finally the data bytes. the size of the transfer is specified using the values from the following table: below is an example write command: 1. 0x55 (sync) 2. 0x80 (command) 3. 0x00 (length msb) 4. 0x09 (length lsb) 5. 0x25 (size and address msb, the two msb of this byte are unused and set to zero) 6. 0x00 7. 0x00 8. 0x00 9. 0x04 (address lsb) 10. 0xca 11. 0xfe 12. 0xba 13. 0xbe 14. 0xxx (crc msb) 15. 0xxx (crc lsb) the length field is set to 0x0009 because there are 9 bytes of additional data: 5 address and size bytes and 4 bytes of data. the address and size field indicates that words should be written to address 0x5000 00004. the data written to 0x500000004 is 0xcafebabe. 31.6.7.10 memory_read this command enables reading of memory/registers on the service access bus (sab). the memory_read command allows words, halfwords, and bytes to be read from a continuous sequence of addresses in one operation. the user must supply: table 31-42. size field decoding size field description 00 byte transfer 01 halfword transfer 10 word transfer 11 reserved table 31-43. memory_write details command details command value 0x80 additional data size, address and data possible responses 0xc2: memory_readwrite_status ( section 31.6.8.5 ) 0x41: nack ( section 31.6.8.2 )
785 32145c?06/2013 at32uc3l0128/256 1. the size of the data field: 7 (size and starting address + read length indicator) in the length field. 2. the size of the transfer: words, halfwords, or bytes. 3. the starting address of the transfer. 4. the number of bytes to read (max 65532). the 4 msb of the 36 bit sab addre ss are submitted together with th e size field (2 bits). the 4 remaining address bytes are submitted before the number of bytes to read. the size of the transfer is specified using the values from the following table: below is an example read command: 1. 0x55 (sync) 2. 0x81 (command) 3. 0x00 (length msb) 4. 0x07 (length lsb) 5. 0x25 (size and address msb, the two msb of this byte are unused and set to zero) 6. 0x00 7. 0x00 8. 0x00 9. 0x04 (address lsb) 10. 0x00 11. 0x04 12. 0xxx (crc msb) 13. 0xxx (crc lsb) the length field is set to 0x0007 because there ar e 7 bytes of additional data: 5 bytes of address and size and 2 bytes with the number of bytes to read. the address and size field indicates one word (four bytes) should be read from address 0x500000004. table 31-44. size field decoding size field description 00 byte transfer 01 halfword transfer 10 word transfer 11 reserved table 31-45. memory_read details command details command value 0x81 additional data size, address and length possible responses 0xc1: memdata ( section 31.6.8.4 ) 0xc2: memory_readwrite_status ( section 31.6.8.5 ) 0x41: nack ( section 31.6.8.2 )
786 32145c?06/2013 at32uc3l0128/256 31.6.7.11 halt this command tells the cpu to halt code execution for safe programming. if the cpu is not halted during programming it can start executing partially loaded programs. to halt the proces- sor, the awire master should send 0x01 in the data field of the command. after programming the halting can be released by sending 0x00 in the data field of the command. 31.6.7.12 reset this command resets different domains in the part. the awire master sends a byte with the reset value. each bit in the reset value byte corresponds to a reset domain in the chip. if a bit is set the reset is activated and if a bit is not set the reset is released. the number of reset domains and their destinations are identical to the resets described in the jtag data registers chapter under reset register. 31.6.7.13 set_guard_time sets the guard time value in the aw, i.e. how long the aw will wait before starting its transfer after the master has finished. the guard time can be either 0x00 (128 bit lengths), 0x01 (16 bit lengths), 0x2 (4 bit lengths) or 0x3 (1 bit length). table 31-46. halt details command details command value 0x82 additional data 0x01 to halt the cpu 0x00 to release the halt and reset the device. possible responses 0x40: ack ( section 31.6.8.1 ) 0x41: nack ( section 31.6.8.2 ) table 31-47. reset details command details command value 0x83 additional data reset value for each reset do main. the number of reset domains is part specific. possible responses 0x40: ack ( section 31.6.8.1 ) 0x41: nack ( section 31.6.8.2 ) table 31-48. set_guard_time details command details command value 0x84 additional data guard time possible responses 0x40: ack ( section 31.6.8.1 ) 0x41: nack ( section 31.6.8.2 )
787 32145c?06/2013 at32uc3l0128/256 31.6.8 awire response summary the implemented awire responses are shown in the table below. 31.6.8.1 ack the aw has received the command successfully and performed the operation. 31.6.8.2 nack the aw has received the command, but got a crc mismatch. 31.6.8.3 idcode the jtag idcode for this device. 31.6.8.4 memdata the data read from the address specified by the memory_read command. the last 3 bytes are status bytes from the read. the first status byte is the status of the command described in the table below. the last 2 bytes are the number of remaining data bytes to be sent in the data field of the packet when the error occurred. if the read was not successful all data bytes after the failure are undefined. a su ccessful word read (4 bytes) will look like this: table 31-49. awire response summary response instruction description 0x40 ack acknowledge. 0x41 nack not acknowledge. sent after crc errors and after unknown commands. 0xc0 idcode the jtag idcode. 0xc1 memdata values read from memory. 0xc2 memory_readwrite_status status after a memory_write or a memory_read command. ok, busy, error. 0xc3 baud_rate the current baudrate. 0xc4 status_info status information. 0xc5 memory_speed sab to awire speed information. table 31-50. ack details response details response value 0x40 additional data n/a table 31-51. nack details response details response value 0x41 additional data n/a table 31-52. idcode details response details response value 0xc0 additional data jtag idcode
788 32145c?06/2013 at32uc3l0128/256 1. 0x55 (sync) 2. 0xc1 (command) 3. 0x00 (length msb) 4. 0x07 (length lsb) 5. 0xca (data msb) 6. 0xfe 7. 0xba 8. 0xbe (data lsb) 9. 0x00 (status byte) 10. 0x00 (bytes remaining msb) 11. 0x00 (bytes remaining lsb) 12. 0xxx (crc msb) 13. 0xxx (crc lsb) the status is 0x00 and all data read are valid. an unsuccessful four byte read can look like this: 1. 0x55 (sync) 2. 0xc1 (command) 3. 0x00 (length msb) 4. 0x07 (length lsb) 5. 0xca (data msb) 6. 0xfe 7. 0xxx (an error has occurred. data read is undefined. 5 bytes remaining of the data field) 8. 0xxx (more undefined data) 9. 0x02 (status byte) 10. 0x00 (bytes remaining msb) 11. 0x05 (bytes remaining lsb) 12. 0xxx (crc msb) 13. 0xxx (crc lsb) the error occurred after reading 2 bytes on the sab. the rest of the bytes read are undefined. the status byte indicates the error and the bytes remaining indicates how many bytes were remaining to be sent of the data field of the packet when the error occurred. table 31-53. memdata status byte status byte description 0x00 read successful 0x01 sab busy 0x02 bus error (wrong address) other reserved table 31-54. memdata details response details response value 0xc1 additional data data read, status byte, and byte count (2 bytes)
789 32145c?06/2013 at32uc3l0128/256 31.6.8.5 memory_readwrite_status after a memory_write command this response is sent by aw. the response can also be sent after a memory_read command if aw encountered an error when receiving the address. the response contains 3 bytes, where t he first is the status of the command and the 2 next contains the byte count when the first erro r occurred. the first byte is encoded this way: 31.6.8.6 baud_rate the current baud rate in the aw. see section 31.6.6.7 for more details. 31.6.8.7 status_info a status message from aw. table 31-55. memory_readwrite_status status byte status byte description 0x00 write successful 0x01 sab busy 0x02 bus error (wrong address) other reserved table 31-56. memory_readwrite_status details response details response value 0xc2 additional data status byte and byte count (2 bytes) table 31-57. baud_rate details response details response value 0xc3 additional data baud rate table 31-58. status_info contents bit number name description 15-9 reserved 8protected the protection bit in the in ternal flash is set. sab acce ss is restricted. this bit will read as one during reset. 7 sab busy the sab bus is busy with a previous transf er. this could indi cate that the cpu is running on a very slow clock, the cpu clock has stopped for some reason or that the part is in constant reset. 6 chip erase ongoing the chip erase operation has not finished. 5 cpu halted this bit will be set if the cpu is halted. this bit will read as zero during reset. 4-1 reserved 0 reset status this bit will be set if aw has reset the cpu using the reset command.
790 32145c?06/2013 at32uc3l0128/256 31.6.8.8 memory_speed counts the number of rc120m clock cycles it takes to sync one message to the sab interface and back again. the sab clock speed ( ) can be calculated using the following formula: 31.6.9 security restrictions when the security fuse in the flash is programmed, the following awire commands are limited: ? memory_write ? memory_read unlimited access to these instructions is restored when the security fuse is erased by the chip_erase awire command. note that the security bit will read as programmed and block these instructions also if the flash controller is statically reset. table 31-59. status_info details response details response value 0xc4 additional data 2 status bytes table 31-60. memory_speed details response details response value 0xc5 additional data clock cycle count (ms) f sab f sab 3 f aw cv 3 ? ---------------- - =
791 32145c?06/2013 at32uc3l0128/256 32. electrical characteristics 32.1 absolute maximum ratings* notes: 1. 5v tolerant pins, see section 3.2 ?peripheral multiplexing on i/o lines? on page 8 2. v vdd corresponds to either v vddin or v vddio , depending on the supply for the pin. refer to section 3.2 on page 8 for details. 32.2 supply characteristics the following characteristics are applicable to the operating temperature range: t a =-40c to 85c, unless otherwise specified and are valid for a junction temperature up to t j =100c. please refer to section 6. ?supply and startup considerations? on page 36 table 32-1. absolute maximum ratings operating temperature..................................... -40 ? c to +85 ? c *notice: stresses beyond those listed under ?absolute maximum ratings? may cause permanent damage to the device. this is a stress rating only and functional opera- tion of the device at these or other condi- tions beyond those indicated in the operational sections of this specification is not implied. exposure to absolute maxi- mum rating conditions for extended peri- ods may affect device reliability. storage temperature...................................... -60c to +150c voltage on input pins (except for 5v pins) with respect to ground .................................................................-0.3v to v vdd (2) +0.3v voltage on 5v tolerant (1) pins with respect to ground ............... .............................................................................-0.3v to 5.5v total dc output current on all i/ o pins - vddio ........... 120ma total dc output current on all i/o pins - vddin ............. 36ma maximum operating voltage v ddcore.............. ........... 1.98v maximum operating voltage vddio, vddin .................... 3.6v table 32-2. supply characteristics symbol parameter voltag e min max unit v vddio dc supply peripheral i/os 1.62 3.6 v v vddin dc supply peripheral i/os, 1.8v single supply mode 1.62 1.98 v dc supply peripheral i/os and internal regulator, 3.3v supply mode 1.98 3.6 v v vddcore dc supply core 1.62 1.98 v v vddana analog supply voltage 1.62 1.98 v
792 32145c?06/2013 at32uc3l0128/256 note: 1. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same process technology. t hese values are not covered by test limits in production. 32.3 maximum clock frequencies these parameters are given in the following conditions: ?v vddcore = 1.62v to 1.98v ? temperature = -40c to 85c 32.4 power consumption the values in table 32-5 are measured values of power consumption under the following condi- tions, except where noted: ? operating conditions, internal core supply ( figure 32-1 ) - this is the default configuration table 32-3. supply rise rates and order (1) symbol parameter rise rate min max unit comment v vddio dc supply peripheral i/os 0 2.5 v/s v vddin dc supply peripheral i/os and internal regulator 0.002 2.5 v/s slower rise time requires external power-on reset circuit. v vddcore dc supply core 0 2.5 v/s rise before or at the same time as vddio v vddana analog supply voltage 0 2.5 v/s rise together with vddcore table 32-4. clock frequencies symbol parameter description min max units f cpu cpu clock frequency 50 mhz f pba pba clock frequency 50 f pbb pbb clock frequency 50 f gclk0 gclk0 clock frequency dfllif main reference, gclk0 pin 50 f gclk1 gclk1 clock frequency dfllif dithering and ssg reference, gclk1 pin 50 f gclk2 gclk2 clock frequency ast, gclk2 pin 20 f gclk3 gclk3 clock frequency pwma, gclk3 pin 140 f gclk4 gclk4 clock frequency cat, acifb, gclk4 pin 50 f gclk5 gclk5 clock frequency gloc 80 f gclk6 gclk6 clock frequency 50 f gclk7 gclk7 clock frequency 50 f gclk8 gclk8 clock frequency pll source clock 50 f gclk9 gclk9 clock frequency freqm, gclk0-8 150
793 32145c?06/2013 at32uc3l0128/256 ?v vddin = 3.0v ?v vddcore = 1.62v, supplied by the internal regulator ? corresponds to the 3.3v supply mode with 1.8v regulated i/o lines, please refer to the supply and startup considerations section for more details ? equivalent to the 3.3v single supply mode ? consumption in 1.8v single supply mode can be estimated by subtracting the regula- tor static current ? operating conditions, external core supply ( figure 32-2 ) - used only when noted ?v vddin = v vddcore = 1.8v ? corresponds to the 1.8v single supply mode, please refer to the supply and startup considerations section for more details ?t a = 25 ? c ? oscillators ? osc0 (crystal o scillator) stopped ? osc32k (32khz crystal oscillator) running with external 32khz crystal ? dfll running at 50mhz with osc32k as reference ? clocks ? dfll used as main clock source ? cpu, hsb, and pbb clocks undivided ? pba clock divided by 4 ? the following peripheral clocks running ? pm, scif, ast, flashcdw, pba bridge ? all other peripheral clocks stopped ? i/os are inactive with internal pull-up ? flash enabled in high speed mode ? por18 enabled ? por33 disabled
794 32145c?06/2013 at32uc3l0128/256 note: 1. these numbers are valid for the measured condition only and must not be extrapolated to other frequencies. figure 32-1. measurement schematic, internal core supply table 32-5. power consumption for different operating modes mode conditions measured on consumption typ unit active (1) cpu running a recursive fibonacci algorithm amp0 300 a/mhz cpu running a division algorithm 174 idle (1) 96 frozen (1) 57 standby (1) 46 stop 38 a deepstop 25 static -osc32k and ast stopped -internal core supply 14 -osc32k running -ast running at 1khz -external core supply ( figure 32-2 ) 7.3 -osc32k and ast stopped -external core supply ( figure 32-2 ) 6.7 shutdown -osc32k running -ast running at 1khz 800 na ast and osc32k stopped 220 amp0 vddin vddcore vddana vddio
795 32145c?06/2013 at32uc3l0128/256 figure 32-2. measurement schematic, external core supply amp0 vddin vddcore vddana vddio
796 32145c?06/2013 at32uc3l0128/256 32.5 i/o pin c haracteristics notes: 1. v vdd corresponds to either v vddin or v vddio , depending on the supply for the pin. refer to section 3.2.1 on page 8 for details. 2. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. table 32-6. normal i/o pin characteristics (1) symbol parameter condition min typ max units r pullup pull-up resistance 75 100 145 kohm v il input low-level voltage v vdd = 3.0v -0.3 0.3*v vdd v v vdd = 1.62v -0.3 0.3*v vdd v ih input high-level voltage v vdd = 3.6v 0.7*v vdd v vdd + 0.3 v v vdd = 1.98v 0.7*v vdd v vdd + 0.3 v ol output low-level voltage v vdd = 3.0v, i ol = 3ma 0.4 v v vdd = 1.62v, i ol = 2ma 0.4 v oh output high-level voltage v vdd = 3.0v, i oh = 3ma v vdd - 0.4 v v vdd = 1.62v, i oh = 2ma v vdd - 0.4 f max output frequency (2) v vdd = 3.0v, load = 10pf 45 mhz v vdd = 3.0v, load = 30pf 23 t rise rise time (2) v vdd = 3.0v, load = 10pf 4.7 ns v vdd = 3.0v, load = 30pf 11.5 t fall fall time (2) v vdd = 3.0v, load = 10pf 4.8 v vdd = 3.0v, load = 30pf 12 i leak input leakage current pull-up resistors disabled 1 a c in input capacitance, all normal i/o pins except pa 0 5 , pa 0 7 , pa 1 7 , pa 2 0 , pa21, pb04, pb05 tqfp48 package 1.4 pf qfn48 package 1.1 tllga48 package 1.1 c in input capacitance, pa20 tqfp48 package 2.7 qfn48 package 2.4 tllga48 package 2.4 c in input capacitance, pa05, pa07, pa17, pa21, pb04, pb05 tqfp48 package 3.8 qfn48 package 3.5 tllga48 package 3.5 table 32-7. high-drive i/o pin characteristics (1) symbol parameter condition min typ max units r pullup pull-up resistance pa06 30 50 110 kohm pa02, pb01, reset 75 100 145 pa08, pa09 10 20 45
797 32145c?06/2013 at32uc3l0128/256 notes: 1. v vdd corresponds to either v vddin or v vddio , depending on the supply for the pin. refer to section 3.2.1 on page 8 for details. 2. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. v il input low-level voltage v vdd = 3.0v -0.3 0.3*v vdd v v vdd = 1.62v -0.3 0.3*v vdd v ih input high-level voltage v vdd = 3.6v 0.7*v vdd v vdd + 0.3 v v vdd = 1.98v 0.7*v vdd v vdd + 0.3 v ol output low-level voltage v vdd = 3.0v, i ol = 6ma 0.4 v v vdd = 1.62v, i ol = 4ma 0.4 v oh output high-level voltage v vdd = 3.0v, i oh = 6ma v vdd -0.4 v v vdd = 1.62v, i oh = 4ma v vdd -0.4 f max output frequency, all high-drive i/o pins, except pa08 and pa09 (2) v vdd = 3.0v, load = 10pf 45 mhz v vdd = 3.0v, load = 30pf 23 t rise rise time, all high-drive i/o pins, except pa08 and pa09 (2) v vdd = 3.0v, load = 10pf 4.7 ns v vdd = 3.0v, load = 30pf 11.5 t fall fall time, all high-drive i/o pins, except pa08 and pa09 (2) v vdd = 3.0v, load = 10pf 4.8 v vdd = 3.0v, load = 30pf 12 f max output frequency, pa08 and pa 0 9 (2) v vdd = 3.0v, load = 10pf 54 mhz v vdd = 3.0v, load = 30pf 40 t rise rise time, pa08 and pa09 (2) v vdd = 3.0v, load = 10pf 2.8 ns v vdd = 3.0v, load = 30pf 4.9 t fall fall time, pa08 and pa09 (2) v vdd = 3.0v, load = 10pf 2.4 v vdd = 3.0v, load = 30pf 4.6 i leak input leakage current pull-up resistors disabled 1 a c in input capacitance, all high-drive i/o pins, except pa08 and pa09 tqfp48 package 2.2 pf qfn48 package 2.0 tllga48 package 2.0 c in input capacitance, pa08 and pa09 tqfp48 package 7.0 qfn48 package 6.7 tllga48 package 6.7 table 32-7. high-drive i/o pin characteristics (1) symbol parameter condition min typ max units table 32-8. high-drive i/o, 5v toler ant, pin characteristics (1) symbol parameter condition min typ max units r pullup pull-up resistance 30 50 110 kohm v il input low-level voltage v vdd = 3.0v -0.3 0.3*v vdd v v vdd = 1.62v -0.3 0.3*v vdd
798 32145c?06/2013 at32uc3l0128/256 notes: 1. v vdd corresponds to either v vddin or v vddio , depending on the supply for the pin. refer to section 3.2.1 on page 8 for details. 2. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. v ih input high-level voltage v vdd = 3.6v 0.7*v vdd 5.5 v v vdd = 1.98v 0.7*v vdd 5.5 v ol output low-level voltage v vdd = 3.0v, i ol = 6ma 0.4 v v vdd = 1.62v, i ol = 4ma 0.4 v oh output high-level voltage v vdd = 3.0v, i oh = 6ma v vdd -0.4 v v vdd = 1.62v, i oh = 4ma v vdd -0.4 f max output frequency (2) v vdd = 3.0v, load = 10pf 87 mhz v vdd = 3.0v, load = 30pf 58 t rise rise time (2) v vdd = 3.0v, load = 10pf 2.3 ns v vdd = 3.0v, load = 30pf 4.3 t fall fall time (2) v vdd = 3.0v, load = 10pf 1.9 v vdd = 3.0v, load = 30pf 3.7 i leak input leakage current 5.5v, pull-up resistors disabled 10 a c in input capacitance tqfp48 package 4.5 pf qfn48 package 4.2 tllga48 package 4.2 table 32-8. high-drive i/o, 5v toler ant, pin characteristics (1) symbol parameter condition min typ max units table 32-9. twi pin characteristics (1) symbol parameter condition min typ max units r pullup pull-up resistance 25 35 60 kohm v il input low-level voltage v vdd = 3.0v -0.3 0.3*v vdd v v vdd = 1.62v -0.3 0.3*v vdd v ih input high-level voltage v vdd = 3.6v 0.7*v vdd v vdd + 0.3 v v vdd = 1.98v 0.7*v vdd v vdd + 0.3 input high-level voltage, 5v tolerant smbus compliant pins v vdd = 3.6v 0.7*v vdd 5.5 v v vdd = 1.98v 0.7*v vdd 5.5 v ol output low-level voltage i ol = 3ma 0.4 v i leak input leakage current pull-up resistors disabled 1 a i il input low leakage 1 i ih input high leakage 1 c in input capacitance tqfp48 package 3.8 pf qfn48 package 3.5 tllga48 package 3.5
799 32145c?06/2013 at32uc3l0128/256 note: 1. v vdd corresponds to either v vddin or v vddio , depending on the supply for the pin. refer to section 3.2.1 on page 8 for details. 32.6 oscillator characteristics 32.6.1 oscillator 0 (osc0) characteristics 32.6.1.1 digital clock characteristics the following table describes the characteristics for the oscillator when a digital clock is applied on xin. note: 1. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. 32.6.1.2 crystal oscillator characteristics the following table describes the characteristics for the oscillator when a crystal is connected between xin and xout as shown in figure 32-3 . the user must choose a crystal oscillator where the crystal load capacitance c l is within the range given in the table. the exact value of c l can be found in the crystal datasheet. the capacitance of the external capacitors (c lext ) can then be computed as follows: where c pcb is the capacitance of the pcb and c i is the internal equivalent load capacitance. t fall fall time cbus = 400pf, v vdd > 2.0v 250 ns cbus = 400pf, v vdd > 1.62v 470 f max max frequency cbus = 400pf, v vdd > 2.0v 400 khz table 32-9. twi pin characteristics (1) symbol parameter condition min typ max units table 32-10. digital clock ch aracteristics symbol parameter conditions min typ max units f cpxin xin clock frequency 50 mhz t cpxin xin clock duty cycle (1) 40 60 % t startup startup time 0 cycles c in xin input capacitance tqfp48 package 7.0 pf qfn48 package 6.7 tllga48 package 6.7 c lext 2c l c i ? ?? c pcb ? =
800 32145c?06/2013 at32uc3l0128/256 notes: 1. please refer to th e scif chapter for details. 2. nominal crystal cycles. 3. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. figure 32-3. oscillator connection 32.6.2 32khz crystal oscillator (osc32k) characteristics figure 32-3 and the equation above also applies to the 32khz oscillator connection. the user must choose a crystal oscillator wh ere the crystal load capacitance c l is within the range given in the table. the exact value of c l can then be found in the crystal datasheet. table 32-11. crystal oscillator characteristics symbol parameter conditions min typ max unit f out crystal oscillator frequency (3) 0.45 10 16 mhz c l crystal load capacitance (3) 618 pf c i internal equivalent load capacitance 2 t startup startup time scif.oscctrl.gain = 2 (1) 30 000 (2) cycles i osc current consumption active mode, f = 0.45mhz, scif.oscctrl.gain = 0 30 a active mode, f = 10mhz, scif.oscctrl.gain = 2 220 xin xout c lext c lext c l c i uc3l
801 32145c?06/2013 at32uc3l0128/256 notes: 1. nominal crystal cycles. 2. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. 32.6.3 phase locked loop (pll) characteristics note: 1. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. table 32-12. 32 khz crystal oscillator characteristics symbol parameter conditions min typ max unit f out crystal oscillator frequency 32 768 hz t startup startup time r s = 60kohm, c l = 9pf 30 000 (1) cycles c l crystal load capacitance (2) 612.5 pf c i internal equivalent load capacitance 2 i osc32 current consumption 0.6 a r s equivalent series resistance (2) 32 768hz 35 85 kohm table 32-13. phase locked loop characteristics symbol parameter conditions min typ max unit f out output frequency (1) 40 240 mhz f in input frequency (1) 416 i pll current consumption 8 a/mhz t startup startup time, from enabling the pll until the pll is locked f in = 4mhz 200 s f in = 16mhz 155
802 32145c?06/2013 at32uc3l0128/256 32.6.4 digital frequency locked loop (dfll) characteristics notes: 1. spread spectrum generator (ssg) is disabled by wr iting a zero to the en bit in the dfll0ssg register. 2. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. 3. the fine and coarse values are selected by wrirting to the dfll0val.fine and dfll0val.coarse field respectively. table 32-14. digital frequency locked loop characteristics symbol parameter conditions min typ max unit f out output frequency (2) 20 150 mhz f ref reference frequency (2) 8 150 khz fine resolution step fine > 100, all coarse values (3) 0.38 % frequency drift over voltage and temperature open loop mode see figure 32- 4 accuracy (2) fine lock, f ref = 32khz, ssg disabled 0.1 0.5 % accurate lock, f ref = 32khz, dither clk rcsys/2, ssg disabled 0.06 0.5 fine lock, f ref = 8-150khz, ssg disabled 0.2 1 accurate lock, f ref = 8-150khz, dither clk rcsys/ 2, ssg disabled 0.1 1 i dfll power consumption 25 a/mhz t startup startup time (2) within 90% of final values 100 s t lock lock time f ref = 32khz, fine lock, ssg disabled 8 ms f ref = 32khz, accurate lock, dithering clock = rcsys/2, ssg disabled 28
803 32145c?06/2013 at32uc3l0128/256 figure 32-4. dfll open loop frequency variation (1)(2) notes: 1. the plot shows a typical open loop mode behavior with coarse= 99 and fine= 255 2. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. 32.6.5 120mhz rc oscillator (rc120m) characteristics note: 1. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. table 32-15. internal 120mhz rc oscillator characteristics symbol parameter conditions min typ max unit f out output frequency (1) 88 120 152 mhz i rc120m current consumption 1.2 ma t startup startup time (1) v vddcore = 1.8v 3 s dfll open loop frequency variation 80 90 100 110 120 130 140 150 160 -40-20 0 204060 80 temperature frequencies (mhz) 1,98v 1,8v 1.62v
804 32145c?06/2013 at32uc3l0128/256 32.6.6 32khz rc oscillator (rc32k) characteristics note: 1. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. 32.6.7 system rc oscillator (rcsys) characteristics 32.7 flash characteristics table 32-18 gives the device maximum operating frequency depending on the number of flash wait states and the flash read mode. the fsw bit in the flashcdw fsr register controls the number of wait states used wh en accessing the flash memory. table 32-16. 32khz rc oscillator characteristics symbol parameter conditions min typ max unit f out output frequency (1) 20 32 44 khz i rc32k current consumption 0.7 a t startup startup time (1) 100 s table 32-17. system rc oscillator characteristics symbol parameter conditions min typ max unit f out output frequency calibrated at 85 ? c 111.6 115 118.4 khz table 32-18. maximum operating frequency flash wait states read mode maximum operating frequency 1 high speed read mode 50mhz 0 25mhz 1 normal read mode 30mhz 0 15mhz table 32-19. flash characteristics symbol parameter conditions min typ max unit t fpp page programming time f clk_hsb = 50mhz 5 ms t fpe page erase time 5 t ffp fuse programming time 1 t fea full chip erase time (ea) 6 t fce jtag chip erase time (chip_erase) f clk_hsb = 115khz 310
805 32145c?06/2013 at32uc3l0128/256 32.8 analog characteristics 32.8.1 voltage regulator characteristics note: 1. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. note: 1. refer to section 6.1.2 on page 36 . table 32-20. flash endurance and data retention symbol parameter condit ions min typ max unit n farray array endurance (write/page) 100k cycles n ffuse general purpose fuses endurance (write/bit) 10k t ret data retention 15 years table 32-21. vreg electrical characteristics symbol parameter condition min typ max units v vddin input voltage range 1.98 3.3 3.6 v v vddcore output voltage, calibrated value v vddin >= 1.98v 1.8 output voltage accuracy (1) i out = 0.1ma to 60ma, v vddin > 1.98v 2 % i out = 0.1ma to 60ma, v vddin < 1.98v 4 i out dc output current (1) normal mode 60 ma low power mode 1 i vreg static current of internal regulator normal mode 13 a low power mode 4 table 32-22. decoupling requirements symbol parameter condition typ techno. units c in1 input regulator capacitor 1 33 nf c in2 input regulator capacitor 2 100 c in3 input regulator capacitor 3 10 f c out1 output regulator capacitor 1 100 nf c out2 output regulator capacitor 2 2.2 tantalum 0.5 806 32145c?06/2013 at32uc3l0128/256 32.8.2 power-on reset 18 characteristics note: 1. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. figure 32-5. por18 operating principle table 32-23. por18 characteristics symbol parameter condition min typ max units v pot+ voltage threshold on v vddcore rising 1.45 1.58 v v pot- voltage threshold on v vddcore falling 1.2 1.32 t det detection time (1) time with vddcore < v pot- necessary to generate a reset signal 460 s i por18 current consumption 4 a t startup startup time (1) 6s reset v vddcore v pot+ v pot- time
807 32145c?06/2013 at32uc3l0128/256 32.8.3 power-on reset 33 characteristics note: 1. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. figure 32-6. por33 operating principle table 32-24. por33 characteristics symbol parameter condition min typ max units v pot+ voltage threshold on v vddin rising 1.49 1.58 v v pot- voltage threshold on v vddin falling 1.3 1.45 t det detection time (1) time with vddin < v pot- necessary to generate a reset signal 460 s i por33 current consumption 20 a t startup startup time (1) 400 s reset v vddin v pot+ v pot- time
808 32145c?06/2013 at32uc3l0128/256 32.8.4 brown out detector characteristics the values in table 32-25 describe the values of the bodlevel in the flash general purpose fuse register. 32.8.5 supply monitor 33 characteristics notes: 1. calibration value can be read from the sm33.calib field. this field is updated by the flash fuses after a reset. refer to scif chapter for details. 2. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. table 32-25. bodlevel values bodlevel value min typ max units 011111 binary (31) 0x1f 1.60 v 100111 binary (39) 0x27 1.69 table 32-26. bod characteristics symbol parameter condition min typ max units v hyst bod hysteresis t = 25 ? c10mv t det detection time time with vddcore < bodlevel necessary to generate a reset signal 1s i bod current consumption 7 a t startup startup time 5s table 32-27. sm33 characteristics symbol parameter condition min typ max units v th voltage threshold calibrated (1) , t = 25 ? c 1.675 1.75 1.825 v step size, between adjacent values in scif.sm33.calib (2) 11 mv v hyst hysteresis (2) 30 t det detection time time with vddin < v th necessary to generate a reset signal 280 s i sm33 current consumption normal mode 17 a t startup startup time normal mode 140 s
809 32145c?06/2013 at32uc3l0128/256 32.8.6 analog to digital converter characteristics note: these values are based on simulation and characterization of other avr microcontrollers ma nufactured in the same process technology. these values are not covered by test limits in production. 32.8.6.1 inputs and sample and hold acquisition times note: 1. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. the analog voltage source must be able to charge the sample and hold (s/h) capacitor in the adc in order to achieve maximum accuracy. seen externally the adc input consists of a resis- tor ( ) and a capacitor ( ). in addition, the resistance ( ) and capacitance ( ) of the pcb and source must be taken into account when calculating the required sample and hold time. figure 32-7 shows the adc input channel equivalent circuit. table 32-28. adc characteristics symbol parameter conditions min typ max units f adc adc clock frequency 12-bit resolution mode 6 mhz f adc adc clock frequency 10-bit resolution mode 6 mhz 8-bit resolution mode 6 t startup startup time return from idle mode 15 s t conv conversion time (latency) f adc = 6mhz 11 26 cycles throughput rate v vdd > 3.0v, f adc = 6mhz, 12-bit resolution mode, low impedance source 28 ksps throughput rate v vdd > 3.0v, f adc = 6mhz, 10-bit resolution mode, low impedance source 460 ksps v vdd > 3.0v, f adc = 6mhz, 8-bit resolution mode, low impedance source 460 v advrefp reference voltage range v advrefp = v vddana 1.62 1.98 v i adc current consumption on v vddana adc clock = 6mhz 350 a i advrefp current consumption on advrefp pin f adc = 6mhz 150 table 32-29. analog inputs symbol parameter conditions min typ max units v adn input voltage range 12-bit mode 0v advrefp v 10-bit mode 8-bit mode c onchip internal capacitance (1) 22.5 pf r onchip internal resistance (1) v vddio = 3.0v to 3.6v, v vddcore = 1.8v 3.15 kohm v vddio = v vddcore = 1.62v to 1.98v 55.9 r onchip c onchip r source c source
810 32145c?06/2013 at32uc3l0128/256 figure 32-7. adc input the minimum sample and hold time (in ns) can be found using this formula: where n is the number of bits in the conversion. is defined by the shtim field in the adcifb acr register. please refer to the adcifb chapter for more information. 32.8.6.2 applicable cond itions and derating data note: 1. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. adcvrefp/2 c onchip r onchip positive input r source c source v in t samplehold r onchip r source + ?? c onchip c source + ?? ? 2 n 1 + ?? ln ? ? t samplehold table 32-30. transfer characteristics 12-bit resolution mode (1) parameter conditions min typ max units resolution 12 bit integral non-linearity adc clock frequency = 6mhz, input voltage range = 0 - v advrefp +/-4 lsb adc clock frequency = 6mhz, input voltage range = (10% v advrefp ) - (90% v advrefp ) +/-2 differential non-linearity adc clock frequency = 6mhz -1.5 1.5 offset error +/-3 gain error +/-5 table 32-31. transfer characteristics, 10-bit resolution mode (1) parameter conditions min typ max units resolution 10 bit integral non-linearity adc clock frequency = 6mhz +/-1 lsb differential non-linearity -1 1 offset error +/-1 gain error +/-2
811 32145c?06/2013 at32uc3l0128/256 note: 1. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. note: 1. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. 32.8.7 temperature sensor characteristics note: 1. the temperature sensor is not calib rated. the accuracy of the temperature s ensor is governed by the adc accuracy. table 32-32. transfer characteristics, 8-bit resolution mode (1) parameter conditions min typ max units resolution 8bit integral non-linearity adc clock frequency = 6mhz +/-0.5 lsb differential non-linearity -0.3 0.3 offset error +/-1 gain error +/-1 table 32-33. temperature sensor characteristics (1) symbol parameter condition min typ max units gradient 1mv/ c i ts current consumption 1 a t startup startup time 0s
812 32145c?06/2013 at32uc3l0128/256 32.8.8 analog comparator characteristics notes: 1. ac.confn.flen and ac.confn.hys fields, refe r to the analog comparator interface chapter. 2. referring to f ac . 3. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. 32.8.9 capacitive touch characteristics 32.8.9.1 discharge current source note: 1. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. table 32-34. analog comparator characteristics symbol parameter condition min typ max units positive input voltage range (3) -0.2 v vddio + 0.3 v negative input voltage range (3) -0.2 v vddio - 0.6 statistical offset (3) v acrefn = 1.0v, f ac = 12mhz, filter length = 2, hysteresis = 0 (1) 20 mv f ac clock frequency for gclk4 (3) 12 mhz throughput rate (3) f ac = 12mhz 12 000 000 comparisons per second propagation delay delay from input change to interrupt status register changes ns i ac current consumption (3) all channels, vddio = 3.3v, f a = 3mhz 420 a t startup startup time 3 cycles input current per pin (3) 0.2 a/mhz (2) table 32-35. dics characteristics symbol parameter min typ max unit r ref internal resistor 170 kohm k trim step size (1) 0.7 % 1 t clkacifb f ac ? ---------------------------------------- 3 + ?? ?? t clkacifb ?
813 32145c?06/2013 at32uc3l0128/256 32.8.9.2 strong pull-up pull-down table 32-36. strong pull-up pull-down parameter min typ max unit pull-down resistor 1 kohm pull-up resistor 1
814 32145c?06/2013 at32uc3l0128/256 32.9 timing characteristics 32.9.1 startup, reset, and wake-up timing the startup, reset, and wake-up timings are calculated using the following formula: where and are found in table 32-37 . is the period of the cpu clock. if a clock source other than rcsys is selected as the cpu clock, the oscillator startup time, , must be added to the wake-up time from the stop, deepstop, and static sleep modes. please refer to the source for the cpu clock in the ?oscillator characteristics? on page 799 for more details about oscillator startup times. note: 1. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. 32.9.2 reset_n timing note: 1. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. tt const n cpu t cpu ? + = t const n cpu t cpu t oscstart table 32-37. maximum reset and wake-up timing (1) parameter measuring max (in s) max startup time from power-up, using regulator time from vddin crossing the v pot+ threshold of por33 to the first instruction entering the decode stage of cpu. vddcore is supplied by the internal regulator. 2210 0 startup time from power-up, no regulator time from vddin crossing the v pot+ threshold of por33 to the first instruction entering the decode stage of cpu. vddcore is connected to vddin. 1810 0 startup time from reset release time from releasing a reset source (except por18, por33, and sm33) to the first instruction entering the decode stage of cpu. 170 0 wake-up idle from wake-up event to the first instruction of an interrupt routine entering the decode stage of the cpu. 019 frozen 0110 standby 0110 stop 27 + 116 deepstop 27 + 116 static 97 + 116 wake-up from shutdown from wake-up event to the first instruction entering the decode stage of the cpu. 1180 0 t const n cpu t oscstart t oscstart t oscstart table 32-38. reset_n waveform parameters (1) symbol parameter conditions min max units t reset reset_n minimum pulse length 10 ns
815 32145c?06/2013 at32uc3l0128/256 32.9.3 usart in spi mode timing 32.9.3.1 master mode figure 32-8. usart in spi master mode with (cpo l= cpha= 0) or (cpol= cpha= 1) figure 32-9. usart in spi master mode with (cpol= 0 and cpha= 1) or (cpol= 1 and cpha= 0) notes: 1. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. 2. where: uspi0 uspi1 miso spck mosi uspi2 uspi3 uspi4 miso spck mosi uspi5 table 32-39. usart in spi mode timing, master mode (1) symbol parameter conditions min max units uspi0 miso setup time before spck rises v vddio from 3.0v to 3.6v, maximum external capacitor = 40pf 28.7 + t sample (2) ns uspi1 miso hold time after spck rises 0 uspi2 spck rising to mosi delay 16.5 uspi3 miso setup time be fore spck falls 25.8 + t sample (2) uspi4 miso hold time after spck falls 0 uspi5 spck falling to mosi delay 21.19 t sample t spck t spck 2 t clkusart ? ------------------------------------ 1 2 -- - ?? ?? t clkusart ? ? =
816 32145c?06/2013 at32uc3l0128/256 maximum spi frequency, master output the maximum spi master output frequenc y is given by the following formula: where is the mosi delay, uspi2 or uspi5 depending on cpol and ncpha. is the maximum frequency of the spi pins. please refe r to the i/o pin characteristics section for the maximum frequency of the pins. is the maximum frequency of the clk_spi. refer to the spi chapter for a description of this clock. maximum spi frequency, master input the maximum spi master input frequenc y is given by the following formula: where is the miso setup and hold time, uspi0 + uspi1 or uspi3 + uspi4 depending on cpol and ncpha. is the spi slave response time. please refer to the spi slave datasheet for . is the maximum frequency of the clk_spi. refer to the spi chap- ter for a description of this clock. 32.9.3.2 slave mode figure 32-10. usart in spi slave mode with (cpol= 0 and cpha= 1) or (cpol= 1 and cpha= 0) f spckmax min f pinmax 1 spin ------------ f clkspi 2 ? 9 ---------------------------- - ? (, ) = spin f pinmax f clkspi f spckmax min 1 spin t valid + ----------------------------------- - f clkspi 2 ? 9 ---------------------------- - (,) = spin t valid t valid f clkspi uspi7 uspi8 miso spck mosi uspi6
817 32145c?06/2013 at32uc3l0128/256 figure 32-11. usart in spi slave mode with (cpol= cpha= 0) or (cpol= cpha= 1) figure 32-12. usart in spi slave mode, npcs timing notes: 1. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. 2. where: uspi10 uspi11 miso spck mosi uspi9 uspi14 uspi12 uspi15 uspi13 nss spck, cpol=0 spck, cpol=1 table 32-40. usart in spi mode timing, slave mode (1) symbol parameter conditions min max units uspi6 spck falling to miso delay v vddio from 3.0v to 3.6v, maximum external capacitor = 40pf 37.3 ns uspi7 mosi setup time before spck rises 2.6 + t sample (2) + t clk_usart uspi8 mosi hold time after spck rises 0 uspi9 spck rising to miso delay 37.0 uspi10 mosi setup time before spck falls 2.6 + t sample (2) + t clk_usart uspi11 mosi hold time after spck falls 0 uspi12 nss setup time before spck rises 27.2 uspi13 nss hold time after spck falls 0 uspi14 nss setup time before spck falls 27.2 uspi15 nss hold time after spck rises 0 t sample t spck t spck 2 t clkusart ? ------------------------------------ 1 2 -- - + ?? ?? t clkusart ? ? =
818 32145c?06/2013 at32uc3l0128/256 maximum spi frequency, slave input mode the maximum spi slave input frequency is given by the following formula: where is the mosi setup and hold time, uspi7 + uspi8 or uspi10 + uspi11 depending on cpol and ncpha. is the maximum frequency of the clk_spi. refer to the spi chapter for a description of this clock. maximum spi frequency, slave output mode the maximum spi slave output frequency is given by the following formula: where is the miso delay, uspi6 or uspi9 depending on cpol and ncpha. is the spi master setup time. please refer to the spi master datasheet for . is the maximum frequency of the clk_spi. refer to the spi chapter for a description of this clock. is the maximum frequency of the spi pins. please refer to the i/o pin characteris- tics section for the maximum frequency of the pins. 32.9.4 spi timing 32.9.4.1 master mode figure 32-13. spi master mode with (cpol= nc pha= 0) or (cpol= ncpha= 1) f spckmax min f clkspi 2 ? 9 ---------------------------- - 1 spin ------------ (,) = spin f clkspi f spckmax min f clkspi 2 ? 9 ---------------------------- - f pinmax ? 1 spin t setup + ------------------------------------ (,) = spin t setup t setup f clkspi f pinmax spi0 spi1 miso spck mosi spi2
819 32145c?06/2013 at32uc3l0128/256 figure 32-14. spi master mode with (cpol= 0 and ncpha= 1) or (cpol= 1 and ncpha= 0) note: 1. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. maximum spi frequency, master output the maximum spi master output frequenc y is given by the following formula: where is the mosi delay, spi2 or spi5 depending on cpol and ncpha. is the maximum frequency of the spi pins. please refer to the i/o pin characteristics section for the maximum frequency of the pins. maximum spi frequency, master input the maximum spi master input frequenc y is given by the following formula: where is the miso setup and hold time, spi0 + spi1 or spi3 + spi4 depending on cpol and ncpha. is the spi slave response time. please refer to the spi slave datasheet for . spi3 spi4 miso spck mosi spi5 table 32-41. spi timing, master mode (1) symbol parameter conditions min max units spi0 miso setup time before spck rises v vddio from 3.0v to 3.6v, maximum external capacitor = 40pf 33.4 + (t clk_spi )/2 ns spi1 miso hold time after spck rises 0 spi2 spck rising to mosi delay 7.1 spi3 miso setup time befo re spck falls 29.2 + (t clk_spi )/2 spi4 miso hold time after spck falls 0 spi5 spck falling to mosi delay 8.63 f spckmax min f pinmax 1 spin ------------ (,) = spin f pinmax f spckmax 1 spin t valid + ----------------------------------- - = spin t valid t valid
820 32145c?06/2013 at32uc3l0128/256 32.9.4.2 slave mode figure 32-15. spi slave mode with (cpol= 0 and ncpha= 1) or (cpol= 1 and ncpha= 0) figure 32-16. spi slave mode with (cpol= ncp ha= 0) or (cpol= ncpha= 1) figure 32-17. spi slave mode, npcs timing spi7 spi8 miso spck mosi spi6 spi10 spi11 miso spck mosi spi9 spi14 spi12 spi15 spi13 npcs spck, cpol=0 spck, cpol=1
821 32145c?06/2013 at32uc3l0128/256 note: 1. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. maximum spi frequency, slave input mode the maximum spi slave input frequency is given by the following formula: where is the mosi setup and hold time, spi7 + spi8 or spi10 + spi11 depending on cpol and ncpha. is the maximum frequency of the clk_spi. refer to the spi chap- ter for a description of this clock. maximum spi frequency, slave output mode the maximum spi slave output frequency is given by the following formula: where is the miso delay, spi6 or spi9 depending on cpol and ncpha. is the spi master setup time. please refer to the spi master datasheet for . is the max- imum frequency of the spi pins. please refer to the i/o pin characteristics section for the maximum frequency of the pins. 32.9.5 twim/twis timing figure 32-43 shows the twi-bus timing requirements and the compliance of the device with them. some of these requirements (t r and t f ) are met by the device without requiring user inter- vention. compliance with the other requirements (t hd-sta , t su-sta , t su-sto , t hd-dat , t su-dat-twi , t low- twi , t high , and f twck ) requires user intervention through appropriate programming of the relevant table 32-42. spi timing, slave mode (1) symbol parameter conditions min max units spi6 spck falling to miso delay v vddio from 3.0v to 3.6v, maximum external capacitor = 40pf 29.4 ns spi7 mosi setup time before spck rises 0 spi8 mosi hold time after spck rises 6.0 spi9 spck rising to miso delay 29.0 spi10 mosi setup time before spck falls 0 spi11 mosi hold time after spck falls 5.5 spi12 npcs setup time before spck rises 3.4 spi13 npcs hold time after spck falls 1.1 spi14 npcs setup time before spck falls 3.3 spi15 npcs hold time after spck rises 0.7 f spckmax min f clkspi 1 spin ------------ (,) = spin f clkspi f spckmax min f pinmax 1 spin t setup + ------------------------------------ (, ) = spin t setup t setup f pinmax
822 32145c?06/2013 at32uc3l0128/256 twim and twis user interface registers. please refer to the twim and twis sections for more information. notes: 1. standard mode: ; fast mode: . 2. a device must internally provide a hold time of at least 300 ns for twd with reference to the falling edge of twck. notations: c b = total capacitance of one bus line in pf t clkpb = period of twi peripheral bus clock t prescaled = period of twi internal prescaled clock (see chapters on twim and twis) the maximum t hd;dat has only to be met if the device does not stretch the low period (t low-twi ) of twck. table 32-43. twi-bus timing requirements symbol parameter mode minimum maximum unit requirement device requirement device t r twck and twd rise time standard (1) - 1000 ns fast (1) 20 + 0.1c b 300 t f twck and twd fall time standard - 300 ns fast 20 + 0.1c b 300 t hd-sta (repeated) start hold time standard 4 t clkpb - ? s fast 0.6 t su-sta (repeated) start set-up time standard 4.7 t clkpb - ? s fast 0.6 t su-sto stop set-up time standard 4.0 4t clkpb - ? s fast 0.6 t hd-dat data hold time standard 0.3 (2) 2t clkpb 3.45 () 15t prescaled + t clkpb ? s fast 0.9 () t su-dat-twi data set-up time standard 250 2t clkpb -ns fast 100 t su-dat --t clkpb -- t low-twi twck low period standard 4.7 4t clkpb - ? s fast 1.3 t low --t clkpb -- t high twck high period standard 4.0 8t clkpb - ? s fast 0.6 f twck twck frequency standard - 100 khz fast 400 1 12t clkpb ----------------------- - f twck 100 khz ? f twck 100 khz ?
823 32145c?06/2013 at32uc3l0128/256 32.9.6 jtag timing figure 32-18. jtag interface signals note: 1. these values are based on simulation and characterization of other avr microcontrollers manufactured in the same pro- cess technology. these values are not covered by test limits in production. jtag2 jtag3 jtag1 jtag4 jtag0 tms/tdi tck tdo jtag5 jtag6 jtag7 jtag8 jtag9 jtag10 boundary scan inputs boundary scan outputs table 32-44. jtag timings (1) symbol parameter conditions min max units jtag0 tck low half-period v vddio from 3.0v to 3.6v, maximum external capacitor = 40pf 21.8 ns jtag1 tck high half-period 8.6 jtag2 tck period 30.3 jtag3 tdi, tms setup before tck high 2.0 jtag4 tdi, tms hold after tck high 2.3 jtag5 tdo hold time 9.5 jtag6 tck low to tdo valid 21.8 jtag7 boundary scan inputs setup time 0.6 jtag8 boundary scan inputs hold time 6.9 jtag9 boundary scan outputs hold time 9.3 jtag10 tck to boundary scan outputs valid 32.2
824 32145c?06/2013 at32uc3l0128/256 33. mechanical characteristics 33.1 thermal considerations 33.1.1 thermal data table 33-1 summarizes the thermal resistance data depending on the package. 33.1.2 junction temperature the average chip-junction temperature, t j , in c can be obtained from the following: 1. 2. where: ? ? ja = package thermal resistance, junction-to-ambient (c/w), provided in table 33-1 . ? ? jc = package thermal resistance, junction-to-ca se thermal resistance (c/w), provided in table 33-1 . ? ? heat sink = cooling device thermal resistance (c/w), provided in the device datasheet. ?p d = device power consumption (w) estimated from data provided in section 32.4 on page 792 . ?t a = ambient temperature (c). from the first equation, the user can derive the estimated lifetime of the chip and decide if a cooling device is necessary or not. if a coolin g device is to be fitted on the chip, the second equation should be used to compute the resulting average chip-junction temperature t j in c. table 33-1. thermal resistance data symbol parameter condition package typ unit ? ja junction-to-ambient thermal resistance still air tqfp48 54.4 ? c/w ? jc junction-to-case thermal resistance tqfp48 15.7 ? ja junction-to-ambient thermal resistance still air qfn48 26.0 ? c/w ? jc junction-to-case thermal resistance qfn48 1.6 ? ja junction-to-ambient thermal resistance still air tllga48 25.4 ? c/w ? jc junction-to-case thermal resistance tllga48 12.7 t j t a p d ? ja ? ?? + = t j t a p ? d ? ? heatsink ?? jc ?? ++ =
825 32145c?06/2013 at32uc3l0128/256 33.2 package drawings figure 33-1. tqfp-48 package drawing table 33-2. device and package maximum weight 140 mg table 33-3. package characteristics moisture sensitivity level msl3 table 33-4. package reference jedec drawing reference ms-026 jesd97 classification e3
826 32145c?06/2013 at32uc3l0128/256 figure 33-2. qfn-48 package drawing note: the exposed pad is not connected to anything internally, but should be soldered to ground to increase board level reliabil ity. table 33-5. device and package maximum weight 140 mg table 33-6. package characteristics moisture sensitivity level msl3 table 33-7. package reference jedec drawing reference m0-220 jesd97 classification e3
827 32145c?06/2013 at32uc3l0128/256 figure 33-3. tllga-48 package drawing table 33-8. device and package maximum weight 39.3 mg table 33-9. package characteristics moisture sensitivity level msl3 table 33-10. package reference jedec drawing reference n/a jesd97 classification e4
828 32145c?06/2013 at32uc3l0128/256 33.3 soldering profile table 33-11 gives the recommended soldering profile from j-std-20. a maximum of three reflow passes is allowed per component. table 33-11. soldering profile profile feature green package average ramp-up rate (217c to peak) 3c/s max preheat temperature 175c 25c 150-200c time maintained above 217c 60-150 s time within 5 ? c of actual peak temperature 30 s peak temperature range 260c ramp-down rate 6c/s max time 25 ? c to peak temperature 8 minutes max
829 32145c?06/2013 at32uc3l0128/256 34. ordering information table 34-1. ordering information device ordering code carrier type package package type temperature operating range at32uc3l0256 at32uc3l0256-autes es tqfp 48 jesd97 classification e3 industrial (-40 ? c to 85 ? c) at32uc3l0256-aut tray at32uc3l0256-aur tape & reel at32uc3l0256-zautes es qfn 48 at32uc3l0256-zaut tray at32uc3l0256-zaur tape & reel at32uc3l0256-d3hes es tllga 48 jesd97 classification e4 at32uc3l0256-d3ht tray at32uc3l0256-d3hr tape & reel at32uc3l0128 at32uc3l0128-aut tray tqfp 48 jesd97 classification e3 at32uc3l0128-aur tape & reel at32uc3l0128-zaut tray qfn 48 at32uc3l0128-zaur tape & reel at32uc3l0128-d3ht tray tllga 48 jesd97 classification e4 at32uc3l0128-d3hr tape & reel
830 32145c?06/2013 at32uc3l0128/256 35. errata 35.1 rev. c 35.1.1 scif 1. the rc32k output on pa20 is not always permanently disabled the rc32k output on pa20 may sometimes re-appear. fix/workaround before using rc32k for other purposes, the following procedure has to be followed in order to properly disable it: - run the cpu on rcsys - disable the output to pa20 by writing a zero to pm.ppcr.rc32out - enable rc32k by writing a one to scif.rc 32kcr.en, and wait for this bit to be read as one - disable rc32k by writing a ze ro to scif.rc32kcr.en, and wait for this bit to be read as zero. 2. pllcount value larger than zero can cause pllen glitch initializing the pllcount with a value greater than zero creates a glitch on the pllen sig- nal during asynchronous wake up. fix/workaround the lock-masking mechanism for the pll should not be used. the pllcount field of the pll control register should always be written to zero. 3. writing 0x5a5a5a5a to the scif memory range will enable the scif unlock feature the scif unlock feature will be enabled if the val ue 0x5a5a5a5a is written to any loca- tion in the scif memory range. fix/workaround none. 35.1.2 spi 1. spi data transfer hangs with csr0.csaat==1 and mr.modfdis==0 when csr0.csaat==1 and mode fault detection is enabled (mr.modfdis==0), the spi module will not start a data transfer. fix/workaround disable mode fault detection by writing a one to mr.modfdis. 2. disabling spi has no effect on the sr.tdre bit disabling spi has no effect on the sr.tdre bit whereas the write data command is filtered when spi is disabled. writing to tdr when spi is disabled will not clear sr.tdre. if spi is disabled during a pdca transfer, the pdca will continue to write data to tdr until its buffer is empty, and this data will be lost. fix/workaround disable the pdca, add two nops, and disable the spi. to continue the transfer, enable the spi and pdca. 3. spi disable does not work in slave mode spi disable does not work in slave mode. fix/workaround read the last received data, then perform a software reset by writing a one to the software reset bit in the control register (cr.swrst).
831 32145c?06/2013 at32uc3l0128/256 4. spi bad serial clock generation on 2nd chip_select when scbr=1, cpol=1, and ncpha=0 when multiple chip selects (cs) are in use, if one of the baudrates equal 1 while one (csrn.scbr=1) of the others do not equal 1, and csrn.cpol=1 and csrn.ncpha=0, then an additional pulse will be genera ted on sck. fix/workaround when multiple cs are in use, if one of the baudrates equals 1, the others must also equal 1 if csrn.cpol=1 and csrn.ncpha=0. 5. spi mode fault detection enable causes incorrect behavior when mode fault detection is enabled (mr. modfdis==0), the spi module may not operate properly. fix/workaround always disable mode fault detection before using the spi by writing a one to mr.modfdis. 6. spi rdr.pcs is not correct the pcs (peripheral chip select) field in th e spi rdr (receive data register) does not correctly indicate the value on the npcs pins at the end of a transfer. fix/workaround do not use the pcs field of the spi rdr. 35.1.3 twi 1. smbalert bit may be set after reset the smbus alert (smbalert) bit in the status register (sr) might be erroneously set after system reset. fix/workaround after system reset, clear the sr.smbalert bit before commencing any twi transfer. 2. clearing the nak bit before the btf bit is set locks up the twi bus when the twis is in transmit mode, clearing the nak received (nak) bit of the status reg- ister (sr) before the end of the acknowl edge/not acknowledge cycle will cause the twis to attempt to continue transmitting data, thus locking up the bus. fix/workaround clear sr.nak only after the byte transfer finished (btf) bit of the same register has been set. 35.1.4 tc 1. channel chaining skips first pulse for upper channel when chaining two channels using the block mode register, the first pulse of the clock between the channels is skipped. fix/workaround configure the lower channel with ra = 0x1 and rc = 0x2 to produce a dummy clock cycle for the upper channel. after the dummy cycle has been generated, indicated by the sr.cpcs bit, reconfigure the ra and rc registers for the lower channel with the real values. 35.1.5 cat 1. cat qmatrix sense capacitors discharged prematurely at the end of a qmatrix burst charging sequence that uses different burst count values for different y lines, the y lines may be incorrectly grounded for up to n-1 periods of the periph-
832 32145c?06/2013 at32uc3l0128/256 eral bus clock, where n is the ratio of t he pb clock frequency to the gclk_cat frequency. this results in premature loss of charge from the sense capacitors and thus increased vari- ability of the acquir ed count values. fix/workaround enable the 1kohm drive resistors on all implemented qmatrix y lines (csa 1, 3, 5, 7, 9, 11, 13, and/or 15) by writing ones to the corresponding odd bits of the csares register. 2. autonomous cat acquisition must be longer than ast source clock period when using the ast to trigger cat autonomous touch acquisition in sleep modes where the cat bus clock is turned off, the cat will start several acquisitions if the period of the ast source clock is larger than one cat acquisition. one ast clock period after the ast trigger, the cat clock will automatically stop and t he cat acquisition can be stopped prematurely, ruining the result. fix/workaround always ensure that the atcfg1.max field is set so that the duration of the autonomous touch acquisition is greater than one clock period of the ast source clock. 35.1.6 awire 1. awire memory_speed_request comm and does not return correct cv the awire memory_speed_request command does not retu rn a cv co rresponding to the formula in the awire debug interface chapter. fix/workaround issue a dummy read to address 0x100000000 before issuing the memory_speed_request command and use this formula instead: 35.1.7 flash 1. corrupted data in flash may happen after flash page write operations after a flash page write operation from an external in situ programmer, reading (data read or code fetch) in flash may fail. this may lead to an exception or to others errors derived from this corrupted read access. fix/workaround before any flash page write operation, each write in the page buffer must preceded by a write in the page buffer with 0xffff_ffff content at any address in the page. 35.2 rev. b 35.2.1 scif 1. the rc32k output on pa20 is not always permanently disabled the rc32k output on pa20 may sometimes re-appear. fix/workaround before using rc32k for other purposes, the following procedure has to be followed in order to properly disable it: - run the cpu on rcsys - disable the output to pa20 by writing a zero to pm.ppcr.rc32out - enable rc32k by writing a one to scif.rc 32kcr.en, and wait for this bit to be read as one f sab 7 f aw cv 3 ? ---------------- - =
833 32145c?06/2013 at32uc3l0128/256 - disable rc32k by writing a ze ro to scif.rc32kcr.en, and wait for this bit to be read as zero. 2. pllcount value larger than zero can cause pllen glitch initializing the pllcount with a value greater than zero creates a glitch on the pllen sig- nal during asynchronous wake up. fix/workaround the lock-masking mechanism for the pll should not be used. the pllcount field of the pll control register should always be written to zero. 3. writing 0x5a5a5a5a to the scif memory range will enable the scif unlock feature the scif unlock feature will be enabled if the val ue 0x5a5a5a5a is written to any loca- tion in the scif memory range. fix/workaround none. 35.2.2 wdt 1. wdt control register does not have synchronization feedback when writing to the timeout pr escale select (psel), time ban prescale select (tban), enable (en), or wdt mode (mode) fieldss of the wdt control register (ctrl), a synchro- nizer is started to propagate the values to the wdt clcok domain. this synchronization takes a finite amount of time, but only the status of the synchronization of the en bit is reflected back to the user. writing to the synch ronized fields during synchronization can lead to undefined behavior. fix/workaround -when writing to the affected fields, the user must ensure a wait corresponding to 2 clock cycles of both the wdt peripheral bus clock and the selected wdt clock source. -when doing writes that changes the en bit, the en bit can be read back until it reflects the written value. 35.2.3 spi 1. spi data transfer hangs with csr0.csaat==1 and mr.modfdis==0 when csr0.csaat==1 and mode fault detection is enabled (mr.modfdis==0), the spi module will not start a data transfer. fix/workaround disable mode fault detection by writing a one to mr.modfdis. 2. disabling spi has no effect on the sr.tdre bit disabling spi has no effect on the sr.tdre bit whereas the write data command is filtered when spi is disabled. writing to tdr when spi is disabled will not clear sr.tdre. if spi is disabled during a pdca transfer, the pdca will continue to write data to tdr until its buffer is empty, and this data will be lost. fix/workaround disable the pdca, add two nops, and disable the spi. to continue the transfer, enable the spi and pdca. 3. spi disable does not work in slave mode spi disable does not work in slave mode. fix/workaround read the last received data, then perform a software reset by writing a one to the software reset bit in the control register (cr.swrst).
834 32145c?06/2013 at32uc3l0128/256 4. spi bad serial clock generation on 2nd chip_select when scbr=1, cpol=1, and ncpha=0 when multiple chip selects (cs) are in use, if one of the baudrates equal 1 while one (csrn.scbr=1) of the others do not equal 1, and csrn.cpol=1 and csrn.ncpha=0, then an additional pulse will be genera ted on sck. fix/workaround when multiple cs are in use, if one of the baudrates equals 1, the others must also equal 1 if csrn.cpol=1 and csrn.ncpha=0. 5. spi mode fault detection enable causes incorrect behavior when mode fault detection is enabled (mr. modfdis==0), the spi module may not operate properly. fix/workaround always disable mode fault detection before using the spi by writing a one to mr.modfdis. 6. spi rdr.pcs is not correct the pcs (peripheral chip select) field in th e spi rdr (receive data register) does not correctly indicate the value on the npcs pins at the end of a transfer. fix/workaround do not use the pcs field of the spi rdr. 35.2.4 twi 1. twis may not wake the device from sleep mode if the cpu is put to a sleep mode (except idle and frozen) directly after a twi start condi- tion, the cpu may not wake upon a twis address match. the request is nacked. fix/workaround when using the twi address match to wake t he device from sleep, do not switch to sleep modes deeper than frozen. another solution is to enable asynchronous eic wake on the twis clock (twck) or twis data (twd) pi ns, in order to wake the system up on bus events. 2. smbalert bit may be set after reset the smbus alert (smbalert) bit in the status register (sr) might be erroneously set after system reset. fix/workaround after system reset, clear the sr.smbalert bit before commencing any twi transfer. 3. clearing the nak bit before the btf bit is set locks up the twi bus when the twis is in transmit mode, clearing the nak received (nak) bit of the status reg- ister (sr) before the end of the acknowl edge/not acknowledge cycle will cause the twis to attempt to continue transmitting data, thus locking up the bus. fix/workaround clear sr.nak only after the byte transfer finished (btf) bit of the same register has been set. 35.2.5 pwma 1. the sr.ready bit cannot be cleared by writing to scr.ready the ready bit in the status register will not be cleared when writing a one to the corre- sponding bit in the status clear register. th e ready bit will be cleared when the busy bit is set. fix/workaround
835 32145c?06/2013 at32uc3l0128/256 disable the ready interrupt in the interrupt handler when receiving the interrupt. when an operation that triggers the busy/ready bit is started, wait until the ready bit is low in the sta- tus register before enabling the interrupt. 35.2.6 tc 1. channel chaining skips first pulse for upper channel when chaining two channels using the block mode register, the first pulse of the clock between the channels is skipped. fix/workaround configure the lower channel with ra = 0x1 and rc = 0x2 to produce a dummy clock cycle for the upper channel. after the dummy cycle has been generated, indicated by the sr.cpcs bit, reconfigure the ra and rc registers for the lower channel with the real values. 35.2.7 cat 1. cat qmatrix sense capacitors discharged prematurely at the end of a qmatrix burst charging sequence that uses different burst count values for different y lines, the y lines may be incorrectly grounded for up to n-1 periods of the periph- eral bus clock, where n is the ratio of t he pb clock frequency to the gclk_cat frequency. this results in premature loss of charge from the sense capacitors and thus increased vari- ability of the acquir ed count values. fix/workaround enable the 1kohm drive resistors on all implemented qmatrix y lines (csa 1, 3, 5, 7, 9, 11, 13, and/or 15) by writing ones to the corresponding odd bits of the csares register. 2. autonomous cat acquisition must be longer than ast source clock period when using the ast to trigger cat autonomous touch acquisition in sleep modes where the cat bus clock is turned off, the cat will start several acquisitions if the period of the ast source clock is larger than one cat acquisition. one ast clock period after the ast trigger, the cat clock will automatically stop and t he cat acquisition can be stopped prematurely, ruining the result. fix/workaround always ensure that the atcfg1.max field is set so that the duration of the autonomous touch acquisition is greater than one clock period of the ast source clock. 3. cat consumes unnecessary power when disabled or when autonomous touch not used a cat prescaler controlled by the atcfg0.div field will be active even when the cat mod- ule is disabled or when the autonomous touch feature is not used, thereby causing unnecessary power consumption. fix/workaround if the cat module is not used, disable the clk_cat clock in the pm module. if the cat module is used but the autonomous touch feature is not used, the power consumption of the cat module may be reduced by writing 0xffff to the atcfg0.div field. 35.2.8 awire 1. awire memory_speed_request comm and does not return correct cv the awire memory_speed_request command does not retu rn a cv co rresponding to the formula in the awire debug interface chapter. fix/workaround
836 32145c?06/2013 at32uc3l0128/256 issue a dummy read to address 0x100000000 before issuing the memory_speed_request command and use this formula instead: 35.2.9 flash 1. corrupted data in flash may happen after flash page write operations after a flash page write operation from an external in situ programmer, reading (data read or code fetch) in flash may fail. this may lead to an exception or to others errors derived from this corrupted read access. fix/workaround before any flash page write operation, each write in the page buffer must preceded by a write in the page buffer with 0xffff_ffff content at any address in the page. 35.3 rev. a 35.3.1 device 1. jtagid is wrong the jtagid is 0x021df03f. fix/workaround none. 35.3.2 flashcdw 1. general-purpose fuse programming does not work the general-purpose fuses cannot be programmed and are stuck at 1. please refer to the fuse settings chapter in the flashcdw for more information about what functions are affected. fix/workaround none. 2. set security bit command does not work the set security bit (ssb) command of the flas hcdw does not work . the device cannot be locked from external jtag, awire, or other debug accesses. fix/workaround none. 3. flash programming time is longer than specified f sab 7 f aw cv 3 ? ---------------- - =
837 32145c?06/2013 at32uc3l0128/256 the flash programming time is now : fix/workaround none. 4. power manager 5. clock failure detector (cfd) can be issued while turning off the cfd while turning off the cfd, the cfd bit in the status register (sr) can be set. this will change the main cl ock source to rcsys. fix/workaround solution 1: enable cfd in terrupt. if cfd interrupt is issues after turning off the cfd, switch back to original main clock source. solution 2: only turn off the cfd while running the main clock on rcsys. 6. sleepwalking in idle and frozen sleep mode will mask all other pb clocks if the cpu is in idle or frozen sleep mode and a module is in a state that triggers sleep walk- ing, all pb clocks will be masked except the pb clock to the sleepwalking module. fix/workaround mask all clock requests in the pm.ppcr register before going into idle or frozen mode. 2. unused pb clocks are running three unused pba clocks are en abled by default and will cause increased active power consumption. fix/workaround disable the clocks by writing zeroes to bits [27: 25] in the pba clock mask register. 35.3.3 scif 1. the rc32k output on pa20 is not always permanently disabled the rc32k output on pa20 may sometimes re-appear. fix/workaround before using rc32k for other purposes, the following procedure has to be followed in order to properly disable it: - run the cpu on rcsys - disable the output to pa20 by writing a zero to pm.ppcr.rc32out - enable rc32k by writing a one to scif.rc 32kcr.en, and wait for this bit to be read as one - disable rc32k by writing a ze ro to scif.rc32kcr.en, and wait for this bit to be read as zero. 2. pll lock might not clear after disable table 35-1. flash characteristics symbol parameter conditions min typ max unit t fpp page programming time f clk_hsb = 50mhz 7.5 ms t fpe page erase time 7.5 t ffp fuse programming time 1 t fea full chip erase time (ea) 9 t fce jtag chip erase time (chip_erase) f clk_hsb = 115khz 250
838 32145c?06/2013 at32uc3l0128/256 under certain circumstances, the lock signal from the phase locked loop (pll) oscillator may not go back to zero after th e pll oscillator has been disabl ed. this can cause the prop- agation of clock signals with the wrong frequency to parts of the system that use the pll clock. fix/workaround pll must be turned off befor e entering stop, deepstop or static sleep modes. if pll has been turned off, a delay of 30us must be observed after the pll has been enabled again before the scif.pll0lock bit can be used as a valid indication that the pll is locked. 3. pllcount value larger than zero can cause pllen glitch initializing the pllcount with a value greater than zero creates a glitch on the pllen sig- nal during asynchronous wake up. fix/workaround the lock-masking mechanism for the pll should not be used. the pllcount field of the pll control register should always be written to zero. 4. rcsys is not calibrated the rcsys is not calibrated and will run faster than 115.2khz. fre quencies around 150khz can be expected. fix/workaround if a known clock source is available the rc sys can be runtime calibrated by using the fre- quency meter (freqm) and tuning the rcsys by writing to the rccr register in scif. 5. writing 0x5a5a5a5a to the scif memory range will enable the scif unlock feature the scif unlock feature will be enabled if the val ue 0x5a5a5a5a is written to any loca- tion in the scif memory range. fix/workaround none. 35.3.4 wdt 1. clearing the watchdog timer (wdt) counter in second half of timeout period will issue a watchdog reset if the wdt counter is cleared in the second half of the ti meout period, the wdt will immedi- ately issue a watchdog reset. fix/workaround use twice as long timeout period as needed and clear the wdt counter within the first half of the timeout period. if the wdt counter is cl eared after the first half of the timeout period, you will get a watchdog reset immediately. if the wdt counter is not clea red at all, the time before the reset will be tw ice as long as needed. 2. wdt control register does not have synchronization feedback when writing to the timeout pr escale select (psel), time ban prescale select (tban), enable (en), or wdt mode (mode) fieldss of the wdt control register (ctrl), a synchro- nizer is started to propagate the values to the wdt clcok domain. this synchronization takes a finite amount of time, but only the status of the synchronization of the en bit is reflected back to the user. writing to the synch ronized fields during synchronization can lead to undefined behavior. fix/workaround -when writing to the affected fields, the user must ensure a wait corresponding to 2 clock cycles of both the wdt peripheral bus clock and the selected wdt clock source. -when doing writes that changes the en bit, the en bit can be read back until it reflects the written value.
839 32145c?06/2013 at32uc3l0128/256 35.3.5 gpio 1. clearing interrupt flags can mask other interrupts when clearing interrupt flags in a gpio port, interrupts on other pins of that port, happening in the same clock cycle will not be registered. fix/workaround read the pvr register of the port before and af ter clearing the interrupt to see if any pin change has happened while clearing the interr upt. if any change occurred in the pvr between the reads, they must be treated as an interrupt. 35.3.6 spi 1. spi data transfer hangs with csr0.csaat==1 and mr.modfdis==0 when csr0.csaat==1 and mode fault detection is enabled (mr.modfdis==0), the spi module will not start a data transfer. fix/workaround disable mode fault detection by writing a one to mr.modfdis. 2. disabling spi has no effect on the sr.tdre bit disabling spi has no effect on the sr.tdre bit whereas the write data command is filtered when spi is disabled. writing to tdr when spi is disabled will not clear sr.tdre. if spi is disabled during a pdca transfer, the pdca will continue to write data to tdr until its buffer is empty, and this data will be lost. fix/workaround disable the pdca, add two nops, and disable the spi. to continue the transfer, enable the spi and pdca. 3. spi disable does not work in slave mode spi disable does not work in slave mode. fix/workaround read the last received data, then perform a software reset by writing a one to the software reset bit in the control register (cr.swrst). 4. spi bad serial clock generation on 2nd chip_select when scbr=1, cpol=1, and ncpha=0 when multiple chip selects (cs) are in use, if one of the baudrates equal 1 while one (csrn.scbr=1) of the others do not equal 1, and csrn.cpol=1 and csrn.ncpha=0, then an additional pulse will be genera ted on sck. fix/workaround when multiple cs are in use, if one of the baudrates equals 1, the others must also equal 1 if csrn.cpol=1 and csrn.ncpha=0. 5. spi mode fault detection enable causes incorrect behavior when mode fault detection is enabled (mr. modfdis==0), the spi module may not operate properly. fix/workaround always disable mode fault detection before using the spi by writing a one to mr.modfdis. 6. spi rdr.pcs is not correct the pcs (peripheral chip select) field in th e spi rdr (receive data register) does not correctly indicate the value on the npcs pins at the end of a transfer. fix/workaround do not use the pcs field of the spi rdr.
840 32145c?06/2013 at32uc3l0128/256 35.3.7 twi 1. twis may not wake the device from sleep mode if the cpu is put to a sleep mode (except idle and frozen) directly after a twi start condi- tion, the cpu may not wake upon a twis address match. the request is nacked. fix/workaround when using the twi address match to wake t he device from sleep, do not switch to sleep modes deeper than frozen. another solution is to enable asynchronous eic wake on the twis clock (twck) or twis data (twd) pi ns, in order to wake the system up on bus events. 2. smbalert bit may be set after reset the smbus alert (smbalert) bit in the status register (sr) might be erroneously set after system reset. fix/workaround after system reset, clear the sr.smbalert bit before commencing any twi transfer. 3. clearing the nak bit before the btf bit is set locks up the twi bus when the twis is in transmit mode, clearing the nak received (nak) bit of the status reg- ister (sr) before the end of the acknowl edge/not acknowledge cycle will cause the twis to attempt to continue transmitting data, thus locking up the bus. fix/workaround clear sr.nak only after the byte transfer finished (btf) bit of the same register has been set. 4. twis stretch on address match error when the twis stretches twck due to a slave address match, it also holds twd low for the same duration if it is to be receiving data. when twis releases twck, it releases twd at the same time. this can cause a twi timing violation. fix/workaround none. 5. twim twalm polarity is wrong the twalm signal in the twim is active high instead of active low. fix/workaround use an external inverter to invert the signal going into the twim. when using both twim and twis on the same pins, the twalm cannot be used. 35.3.8 pwma 1. the sr.ready bit cannot be cleared by writing to scr.ready the ready bit in the status register will not be cleared when writing a one to the corre- sponding bit in the status clear register. th e ready bit will be cleared when the busy bit is set. fix/workaround disable the ready interrupt in the interrupt handler when receiving the interrupt. when an operation that triggers the busy/ready bit is started, wait until the ready bit is low in the sta- tus register before enabling the interrupt. 35.3.9 tc 1. channel chaining skips first pulse for upper channel when chaining two channels using the block mode register, the first pulse of the clock between the channels is skipped.
841 32145c?06/2013 at32uc3l0128/256 fix/workaround configure the lower channel with ra = 0x1 and rc = 0x2 to produce a dummy clock cycle for the upper channel. after the dummy cycle has been generated, indicated by the sr.cpcs bit, reconfigure the ra and rc registers for the lower channel with the real values. 35.3.10 adcifb 1. adcifb dma transfer does not work with divided pba clock dma requests from the adcifb will not be performed when the pba clock is slower than the hsb clock. fix/workaround do not use divided pba clock when th e pdca transfers from the adcifb. 35.3.11 cat 1. cat qmatrix sense capacitors discharged prematurely at the end of a qmatrix burst charging sequence that uses different burst count values for different y lines, the y lines may be incorrectly grounded for up to n-1 periods of the periph- eral bus clock, where n is the ratio of t he pb clock frequency to the gclk_cat frequency. this results in premature loss of charge from the sense capacitors and thus increased vari- ability of the acquir ed count values. fix/workaround enable the 1kohm drive resistors on all implemented qmatrix y lines (csa 1, 3, 5, 7, 9, 11, 13, and/or 15) by writing ones to the corresponding odd bits of the csares register. 2. autonomous cat acquisition must be longer than ast source clock period when using the ast to trigger cat autonomous touch acquisition in sleep modes where the cat bus clock is turned off, the cat will start several acquisitions if the period of the ast source clock is larger than one cat acquisition. one ast clock period after the ast trigger, the cat clock will automatically stop and t he cat acquisition can be stopped prematurely, ruining the result. fix/workaround always ensure that the atcfg1.max field is set so that the duration of the autonomous touch acquisition is greater than one clock period of the ast source clock. 3. cat consumes unnecessary power when disabled or when autonomous touch not used a cat prescaler controlled by the atcfg0.div field will be active even when the cat mod- ule is disabled or when the autonomous touch feature is not used, thereby causing unnecessary power consumption. fix/workaround if the cat module is not used, disable the clk_cat clock in the pm module. if the cat module is used but the autonomous touch feature is not used, the power consumption of the cat module may be reduced by writing 0xffff to the atcfg0.div field. 4. cat module does not terminate qtouch burst on detect the cat module does not terminate a qtouch burst when the detection voltage is reached on the sense capacitor. this can ca use the sense capacitor to be charged more than necessary. depending on the dielectric abso rption characteristics of the capacitor, this can lead to unstable measurements. fix/workaround use the minimum possible value for the max field in the atcfg1, tg0cfg1, and tg1cfg1 registers.
842 32145c?06/2013 at32uc3l0128/256 35.3.12 awire 1. awire memory_speed_request comm and does not return correct cv the awire memory_speed_request command does not retu rn a cv co rresponding to the formula in the awire debug interface chapter. fix/workaround issue a dummy read to address 0x100000000 before issuing the memory_speed_request command and use this formula instead: 35.3.13 flash 1. corrupted data in flash may happen after flash page write operations after a flash page write operation from an external in situ programmer, reading (data read or code fetch) in flash may fail. this may lead to an exception or to others errors derived from this corrupted read access. fix/workaround before any flash page write operation, each write in the page buffer must preceded by a write in the page buffer with 0xffff_ffff content at any address in the page. 35.3.14 i/o pins 1. pa05 is not 3.3v tolerant. pa05 should be grounded on the pcb and left unused if vddio is above 1.8v. fix/workaround none. 2. no pull-up on pins that are not bonded pb13 to pb27 are not bonded on uc3l0256/128, but has no pull-up and can cause current consumption on vddio/vddin if left undriven. fix/workaround enable pull-ups on pb13 to pb27 by writing 0x0fffe000 to the puers1 register in the gpio. 3. pa17 has low esd tolerance pa17 only tolerates 500v esd pulses (human body model). fix/workaround care must be taken during manufacturing and pcb design. f sab 7 f aw cv 3 ? ---------------- - =
843 32145c?06/2013 at32uc3l0128/256 36. datasheet revision history please note that the referring page numbers in th is section are referred to this document. the referring revision in this section are referring to the document revision. 36.1 rev. c ? 06/2013 36.2 rev. b ? 01/2012 36.3 rev. a ? 12/2011 1. updated the datasheet with new atmel blue logo and the last page. 2. added flash errata. 1. description: dfll frequency is 20 to 150mhz, not 40 to 150mhz. 2. description: ?one touch sensor can be configured to operate autonomously...? replaced by ?all touch sensors can be configured to operate autonomously...?. 3. block diagram: gclk_in is input, not output, and is 2 bits wide (gclk_in[1..0]). cat smp corrected from i/o to output. spi np cs corrected from output to i/o. 4. package and pinout: prnd signal removed from signal descriptions list table and gpio controller function multiplexing table. 5. supply and startup considerations: in 1.8v si ngle supply mode figure, the input voltage is 1.62-1.98v, not 1.98-3.6v . ?on system start-up, the dfll is disabled? is replaced by ?on system start-up, all high-speed clocks are disabled?. 6. adcifb: prnd signal removed from block diagram. 7. electrical characteristics: added pll source clock in the clock frequencies table in the maximum clock frequencies section. removed 64-pin package information from i/o pin characteristics tables and digital clock characteristics table. 8. electrical characteristics: removed usb transce iver characteristics, as the device contains no usb. 9. mechanical characteristics: added notes to package drawings. 10. summary: removed programming and debugging chapter, added processor and architecture chapter. 11. datasheet revision history: corrected release dat e for datasheet rev. a; the correct date is 12/2011. 1. initial revision.
i 32145c?06/2013 at32uc3l0128/256 table of contents features ................ ................ .............. ............... .............. .............. ............ 1 1 description ............ .............. .............. ............... .............. .............. ............ 3 2 overview ............ ................ ................ ............... .............. .............. ............ 5 2.1 block diagram ...................................................................................................5 2.2 configuration summary .....................................................................................6 3 package and pinout ................. ................ ................. ................ ............... 7 3.1 package .............................................................................................................7 3.2 peripheral multiplexing on i/o lines ..................................................................8 3.3 signal descriptions ..........................................................................................13 3.4 i/o line considerations ...................................................................................16 4 processor and architecture .... ................ ................. ................ ............. 18 4.1 features ..........................................................................................................18 4.2 avr32 architecture .........................................................................................18 4.3 the avr32uc cpu ........................................................................................19 4.4 programming model ........................................................................................23 4.5 exceptions and interrupts ................................................................................27 5 memories ............... .............. .............. ............... .............. .............. .......... 32 5.1 embedded memories ......................................................................................32 5.2 physical memory map .....................................................................................32 5.3 peripheral address map ..................................................................................33 5.4 cpu local bus mapping .................................................................................34 6 supply and startup c onsiderations ............ ................. .............. .......... 36 6.1 supply considerations .....................................................................................36 6.2 startup considerations ....................................................................................40 7 peripheral dma controller (pdca) ................ .............. .............. .......... 41 7.1 features ..........................................................................................................41 7.2 overview ..........................................................................................................41 7.3 block diagram .................................................................................................42 7.4 product dependencies ....................................................................................42 7.5 functional description .....................................................................................43 7.6 performance monitors .....................................................................................45 7.7 user interface ..................................................................................................47
ii 32145c?06/2013 at32uc3l0128/256 7.8 module configuration ......................................................................................75 8 flash controller (fla shcdw) .................. ................ ................. .......... 77 8.1 features ..........................................................................................................77 8.2 overview ..........................................................................................................77 8.3 product dependencies ....................................................................................77 8.4 functional description .....................................................................................78 8.5 flash commands ............................................................................................83 8.6 general-purpose fuse bits ..............................................................................85 8.7 security bit ......................................................................................................88 8.8 user interface ..................................................................................................89 8.9 fuse settings ...................................................................................................99 8.10 serial number ................................................................................................102 8.11 module configuration ....................................................................................102 9 secure access unit (sau) .... ................ ................. ................ ............. 104 9.1 features ........................................................................................................104 9.2 overview ........................................................................................................104 9.3 block diagram ...............................................................................................105 9.4 product dependencies ..................................................................................106 9.5 functional description ...................................................................................106 9.6 user interface ................................................................................................110 9.7 module configuration ....................................................................................125 10 hsb bus matrix (hmatrixb) ................ ................. ................ ............. 126 10.1 features ........................................................................................................126 10.2 overview ........................................................................................................126 10.3 product dependencies ..................................................................................126 10.4 functional description ...................................................................................126 10.5 user interface ................................................................................................130 10.6 module configuration ....................................................................................138 11 interrupt controller (intc) ............. .............. .............. .............. ........... 140 11.1 features ........................................................................................................140 11.2 overview ........................................................................................................140 11.3 block diagram ...............................................................................................140 11.4 product dependencies ..................................................................................141 11.5 functional description ...................................................................................141 11.6 user interface ................................................................................................144
iii 32145c?06/2013 at32uc3l0128/256 11.7 module configuration ....................................................................................148 12 power manager (pm) ......... .............. .............. .............. .............. ........... 151 12.1 features ........................................................................................................151 12.2 overview ........................................................................................................151 12.3 block diagram ...............................................................................................152 12.4 i/o lines description .....................................................................................152 12.5 product dependencies ..................................................................................152 12.6 functional description ...................................................................................153 12.7 user interface ................................................................................................162 12.8 module configuration ....................................................................................185 13 system control interface (sci f) ........... ................. ................ ............. 186 13.1 features ........................................................................................................186 13.2 overview ........................................................................................................186 13.3 i/o lines description .....................................................................................186 13.4 product dependencies ..................................................................................186 13.5 functional description ...................................................................................187 13.6 user interface ................................................................................................207 13.7 module configuration ....................................................................................260 14 asynchronous timer (ast) ...... ................. ................ .............. ........... 264 14.1 features ........................................................................................................264 14.2 overview ........................................................................................................264 14.3 block diagram ...............................................................................................265 14.4 product dependencies ..................................................................................265 14.5 functional description ...................................................................................266 14.6 user interface ................................................................................................272 14.7 module configuration ....................................................................................293 15 watchdog timer (wdt) ......... ................ ................. ................ ............. 294 15.1 features ........................................................................................................294 15.2 overview ........................................................................................................294 15.3 block diagram ...............................................................................................294 15.4 product dependencies ..................................................................................294 15.5 functional description ...................................................................................295 15.6 user interface ................................................................................................300 15.7 module configuration ....................................................................................306
iv 32145c?06/2013 at32uc3l0128/256 16 external interrupt controller (eic) ..... .............. .............. ............ ........ 307 16.1 features ........................................................................................................307 16.2 overview ........................................................................................................307 16.3 block diagram ...............................................................................................307 16.4 i/o lines description .....................................................................................308 16.5 product dependencies ..................................................................................308 16.6 functional description ...................................................................................308 16.7 user interface ................................................................................................312 16.8 module configuration ....................................................................................328 17 frequency meter (freqm) ...... ................ ................. ................ ........... 329 17.1 features ........................................................................................................329 17.2 overview ........................................................................................................329 17.3 block diagram ...............................................................................................329 17.4 product dependencies ..................................................................................329 17.5 functional description ...................................................................................330 17.6 user interface ................................................................................................332 17.7 module configuration ....................................................................................343 18 general-purpose input/output controller (gpio) .... .............. ........... 345 18.1 features ........................................................................................................345 18.2 overview ........................................................................................................345 18.3 block diagram ...............................................................................................345 18.4 i/o lines description .....................................................................................346 18.5 product dependencies ..................................................................................346 18.6 functional description ...................................................................................347 18.7 user interface ................................................................................................352 18.8 module configuration ....................................................................................375 19 universal synchro nous asynchronous receiver tr ansmitter (usart) 376 19.1 features ........................................................................................................376 19.2 overview ........................................................................................................376 19.3 block diagram ...............................................................................................377 19.4 i/o lines description ....................................................................................378 19.5 product dependencies ..................................................................................378 19.6 functional description ...................................................................................379 19.7 user interface ................................................................................................405
v 32145c?06/2013 at32uc3l0128/256 19.8 module configuration ....................................................................................427 20 serial peripheral interface (spi) ................ ................ .............. ........... 428 20.1 features ........................................................................................................428 20.2 overview ........................................................................................................428 20.3 block diagram ...............................................................................................429 20.4 application block diagram .............................................................................429 20.5 i/o lines description .....................................................................................430 20.6 product dependencies ..................................................................................430 20.7 functional description ...................................................................................430 20.8 user interface ................................................................................................441 20.9 module configuration ....................................................................................468 21 two-wire master interface (twim) ........ ................. ................ ............. 469 21.1 features ........................................................................................................469 21.2 overview ........................................................................................................469 21.3 list of abbreviations ......................................................................................470 21.4 block diagram ...............................................................................................470 21.5 application block diagram .............................................................................471 21.6 i/o lines description .....................................................................................471 21.7 product dependencies ..................................................................................471 21.8 functional description ...................................................................................473 21.9 user interface ................................................................................................485 21.10 module configuration ....................................................................................502 22 two-wire slave interface (twis) ........... ................. ................ ............. 503 22.1 features ........................................................................................................503 22.2 overview ........................................................................................................503 22.3 list of abbreviations ......................................................................................504 22.4 block diagram ...............................................................................................504 22.5 application block diagram .............................................................................505 22.6 i/o lines description .....................................................................................505 22.7 product dependencies ..................................................................................505 22.8 functional description ...................................................................................506 22.9 user interface ................................................................................................516 22.10 module configuration ....................................................................................532 23 pulse width modulation c ontroller (pwma) .......... ................ ........... 533 23.1 features ........................................................................................................533
vi 32145c?06/2013 at32uc3l0128/256 23.2 overview ........................................................................................................533 23.3 block diagram ...............................................................................................534 23.4 i/o lines description .....................................................................................534 23.5 product dependencies ..................................................................................534 23.6 functional description ...................................................................................535 23.7 user interface ................................................................................................541 23.8 module configuration ....................................................................................558 24 timer/counter (tc) ........... .............. .............. .............. .............. ........... 559 24.1 features ........................................................................................................559 24.2 overview ........................................................................................................559 24.3 block diagram ...............................................................................................560 24.4 i/o lines description .....................................................................................560 24.5 product dependencies ..................................................................................560 24.6 functional description ...................................................................................561 24.7 user interface ................................................................................................576 24.8 module configuration ....................................................................................599 25 peripheral event system .... .............. ............... .............. .............. ........ 600 25.1 features ........................................................................................................600 25.2 overview ........................................................................................................600 25.3 peripheral event system block diagram .......................................................600 25.4 functional description ...................................................................................600 25.5 application example ......................................................................................603 26 adc interface (adcifb) .... ............. .............. .............. .............. ........... 604 26.1 features ........................................................................................................604 26.2 overview ........................................................................................................604 26.3 block diagram ...............................................................................................605 26.4 i/o lines description .....................................................................................606 26.5 product dependencies ..................................................................................606 26.6 functional description ...................................................................................607 26.7 resistive touch screen .................................................................................611 26.8 operating modes ...........................................................................................617 26.9 user interface ................................................................................................619 26.10 module configuration ....................................................................................638 27 analog comparator interface (acifb) ........ .............. .............. ........... 639 27.1 features ........................................................................................................639
vii 32145c?06/2013 at32uc3l0128/256 27.2 overview ........................................................................................................639 27.3 block diagram ...............................................................................................640 27.4 i/o lines description .....................................................................................640 27.5 product dependencies ..................................................................................641 27.6 functional description ...................................................................................642 27.7 peripheral event triggers ..............................................................................647 27.8 ac test mode ................................................................................................647 27.9 user interface ................................................................................................648 27.10 module configuration ....................................................................................662 28 capacitive touch module (cat ) ............. ................. ................ ........... 663 28.1 features ........................................................................................................663 28.2 overview ........................................................................................................663 28.3 block diagram ...............................................................................................664 28.4 i/o lines description .....................................................................................664 28.5 product dependencies ..................................................................................665 28.6 functional description ...................................................................................667 28.7 user interface ................................................................................................674 28.8 module configuration ....................................................................................709 29 glue logic controller (gloc) .............. ................. ................ ............. 710 29.1 features ........................................................................................................710 29.2 overview ........................................................................................................710 29.3 block diagram ...............................................................................................710 29.4 i/o lines description .....................................................................................711 29.5 product dependencies ..................................................................................711 29.6 functional description ...................................................................................711 29.7 user interface ................................................................................................713 29.8 module configuration ....................................................................................718 30 awire uart (aw) .............. .............. .............. .............. .............. ........... 719 30.1 features ........................................................................................................719 30.2 overview ........................................................................................................719 30.3 block diagram ...............................................................................................719 30.4 i/o lines description .....................................................................................720 30.5 product dependencies ..................................................................................720 30.6 functional description ...................................................................................720 30.7 user interface ................................................................................................723
viii 32145c?06/2013 at32uc3l0128/256 30.8 module configuration ....................................................................................736 31 programming and debugging .. ................. ................ .............. ........... 737 31.1 overview ........................................................................................................737 31.2 service access bus .......................................................................................737 31.3 on-chip debug ..............................................................................................740 31.4 jtag and boundary-scan (jtag) .................................................................749 31.5 jtag instruction summary ...........................................................................757 31.6 awire debug interface (aw) .........................................................................774 32 electrical characteristics ... .............. ............... .............. .............. ........ 791 32.1 absolute maximum ratings* .........................................................................791 32.2 supply characteristics ...................................................................................791 32.3 maximum clock frequencies ........................................................................792 32.4 power consumption ......................................................................................792 32.5 i/o pin characteristics ...................................................................................796 32.6 oscillator characteristics ...............................................................................799 32.7 flash characteristics .....................................................................................804 32.8 analog characteristics ...................................................................................805 32.9 timing characteristics ...................................................................................814 33 mechanical characteristics ..... ................ ................. ................ ........... 824 33.1 thermal considerations ................................................................................824 33.2 package drawings .........................................................................................825 33.3 soldering profile ............................................................................................828 34 ordering information .......... .............. ............... .............. .............. ........ 829 35 errata ........... ................ ................ ................. ................ .............. ........... 830 35.1 rev. c ............................................................................................................830 35.2 rev. b ............................................................................................................832 35.3 rev. a ............................................................................................................836 36 datasheet revision history .. ................ ................. ................ ............. 843 36.1 rev. c ? 06/2013 ...........................................................................................843 36.2 rev. b ? 01/2012 ...........................................................................................843 36.3 rev. a ? 12/2011 ...........................................................................................843 table of contents.......... ................. ................ ................. ................ ........... i
atmel corporation 1600 technology drive san jose, ca 95110 usa tel: (+1) (408) 441-0311 fax: (+1) (408) 487-2600 www.atmel.com atmel asia limited unit 01-5 & 16, 19f bea tower, millennium city 5 418 kwun tong roa kwun tong, kowloon hong kong tel: (+852) 2245-6100 fax: (+852) 2722-1369 atmel munich gmbh business campus parkring 4 d-85748 garching b. munich germany tel: (+49) 89-31970-0 fax: (+49) 89-3194621 atmel japan g.k. 16f shin-osaki kangyo bldg 1-6-4 osaki, shinagawa-ku tokyo 141-0032 japan tel: (+81) (3) 6417-0300 fax: (+81) (3) 6417-0370 ? 2013 atmel corporation. all rights reserved. / rev.: 32145c?avr32?06/2013 atmel ? , logo and combinations thereof, avr ? , picopower ? , qtouch ? , aks ? and others are registered trademarks or trademarks of atmel corpo- ration or its subsidiaries. other terms and product names may be trademarks of others. disclaimer: the information in this document is provided in conn ection with atmel products. no license, express or implied, by estoppel or otherwise, to any intellectual property right is granted by this document or in connection with the sale of atmel products. except as set forth in the atmel term s and conditions of sales locat ed on the atmel website, atmel assumes no liability whatsoever and disclaims any express, implied or st atutory warranty relating to its products including, but not li mited to, the implied warranty of merchantability, fitness for a particular purpose, or non-infringement. in no event shall atmel be liable for any d irect, indirect, consequential, punitive, special or incidental damages (including, without limi tation, damages for loss and profits, business i nterruption, or loss of information) arising out of the use or inability to use this document, even if atmel has been advised of the possibility of suc h damages. atmel makes no representations or warranties with respect to the accuracy or co mpleteness of the contents of this document and reserves the ri ght to make changes to specifications and products descriptions at any time without notice. atmel does not make any commitment to update the information contained herein. unless specifically provided oth erwise, atmel products are not suitable for, and shall not be used in, automotive applications. atmel products are not intended, authorized, or warranted for use as components in applications intend ed to support or sustain life.

▲Up To Search▲

Price & Availability of TS80C51RA2-LCBB

	To Download TS80C51RA2-LCBB Datasheet File
If you can't view the Datasheet, Please click here to try to view without PDF Reader .