View MT90812_8362031.PDF datasheet online --- IC-ON-LINE

Datasheet File OCR Text:

1 features ? 192 channel x 192 channel non-blocking switching ? 2 local bus streams @ 2mb/s supports up to 64 channels ? in tdm mode, the expansion bus supports up to 128 channels at 8.192 mb/s ? rate conversion capability between local and expansion bus streams ? integrated conference bridge, supporting 15 parties over 5 bridges ? integrated pll ? frequency shift keying (fsk) 1200 baud transmitter, meeting bell 202 or ccitt v.23 standards ? 32 channel dual tone generator, including 16 standard dtmf tones and tone ringer ? expansion bus in idx link mode, allows the interconnection of up to 4 idx devices ? programmable per channel gain control from +3 to -27db, increments of 1db for output channels ? supervisory signalling cadence detection capability ? hdlc resource allocator ? d-channel buffering of message information ? c-channel access for control and status registers ? provides both variable and constant delay modes ? parallel microprocessor port, compatible to intel and motorola and national cpus ? supports both a-law or u-law operation ? supports both st-bus, gci and hmvip framing formats applications ? computer telephony integration (cti) ? key telephone systems ? private branch exchange (pbx) systems description by integrating key functions needed in voice telecom application, the integrated digital switch (idx) provides a solution-on-a-chip for key telephone systems, pbx applications or cti designs. figure 2 shows a typical con?guration. the MT90812 provides non-blocking timeslot interchange capability for b, c and d channels, up to a maximum of 192 channels. it offers conference call capability for 15 parties over a maximum of 5 conference bridges. with its integrated pll, the MT90812 provides the necessary clocks to support peripheral devices, such as codecs or interconnected idx devices. integrated into the idx is the capability to detect supervisory signalling and to generate fsk 1200-baud signals. in addition, an integrated digital tone generator produces continuous dual tones, including standard dtmf. with its programmable gain control, the idx allows users to use codecs without gain control and also centrally manage conference calls. to support both small and large switching platforms, a built-in expansion bus allows the interconnection of up to 4 idx devices or external components such as digital switches. when 4 idx devices are interconnected, the array is capable of switching 256 channels (64x4), handling 60 conference parties (15x4) and generating additional tones including programmable ones. other functions are also increased in this con?guration. the functional block diagram is shown in figure 1. an evaluation board, meb90812, is available complete with software and a user manual, which demonstrates the layout of a typical application board and facilitates the use of the MT90812, and peripheral devices such as mitels dnic products. ds5219 issue 2 december 1999 ordering information MT90812ap 68 pin plcc MT90812al 64 pin mqfp -40 to +85 c MT90812 integrated digital switch (idx) advance information
MT90812 advance information 2 figure 1 - functional block diagram figure 2 - system blocks - typical con?guration sti0 sti1 sto0 sto1 serial to parallel converter output mux parallel to serial converter microprocessor connect memory interface est1 m u x est0 fsk and tone conference d-channel tx/rx hdlc resource allocator hdlc controller pll timing & control frame pulses and clocks cpu gain control energy detect ring r+ r- data memory frequency generation local c.o. trunks 2b+d ports analog integrated system expansion bus (tdm) mpu interface - d-channel - chip control ports digital switch - digital switch - conferencing - tone generation - energy detection 2b+d 2b+d 2b+d 2b+d 2b+d 2b+d 2b+d 2b+d tdm bus 2b+d ports codec . . . 2b+d 2b+d 2b+d codec
advance information MT90812 3 figure 3 - pin connections 60 27 nc ic 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 9 8 7 6 5 4 3 2 1 68 67 66 65 64 63 62 61 59 58 57 56 55 54 53 52 51 50 49 48 47 46 45 44 nc rxcen txcen reop teop vss1 ad7 ad6 ad5 ad4 ad3 ad2 ad1 ad0 vss5 nc vbuf vssa vdda f4o c2o c4o c10o vss3 r+ r - reset ic vdd vss4 nc nc nc c4i fpi osco c8p_c16i f8 c8 ode est1 est0 dper sto1 sto0 sti1 sti0 vss2 nc a0 a1 a2 a3 a4 a5 a6 a7 a8 a9 ds/ rd cs as/ale irq 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 68 pin plcc im dta r/ w \ wr 64 pin mqfp 17 19 15 13 11 9 7 5 3 12 nc nc vss2 sti0 sti1 sto0 sto1 dper est0 est1 ode c8 f8 c8p_c16i osco fpi c4i ic reop vss1 ad7 ad6 ad5 ad4 ad3 ad2 ad1 ad0 rxcen teop txcen vss5 a2 a1 as/ale cs r/ w \ wr ds/ rd a9 a8 a6 irq dta a0 a3 a4 nc a5 a7 vssa vdda f4o c2o c10o reset vdd c4o vss3 r+ r - ic vss4 46 81012141618 vbuf 32 31 30 29 28 27 26 25 24 23 22 21 20 52 53 54 55 56 57 58 59 60 60 62 63 64 im 51 49 48 50 47 46 45 44 43 42 41 40 39 38 37 36 35 34 33
MT90812 advance information 4 pin description pin # name description 64 pin mqfp 68 pin plcc 1 25-26 nc no connect . ground 2-11 27-36 a0 - a9 address 0 - 9(input) . when non-multiplexed cpu bus is selected, these lines provide the a0 - a9 address lines to idx internal memories. 12 37 ds/ rd data strobe/ read (input) . for motorola multiplexed bus operation, this active high ds input works with cs to enable the read and write operations. for motorola non-multiplexed cpu bus operation, this input is ds. this active low input works in conjunction with cs to enable the read and write operations. for intel/national multiplexed bus operations, this input is rd. this active low input sets the data bus lines (ad0-ad7) as outputs. 13 38 r/ w \ wr read/ write \ write (input) . in case of non-multiplexed and motorola multiplexed buses, this input is read/ wr ite. this input controls the direction of the data bus lines (ad0 - ad7) during a microprocessor access. for intel/national multiplexed bus, this input is wr. this active low signal con?gures the data bus lines (ad0-ad7) as inputs. 14 39 cs chip select (input) . active low input enabling a microprocessor read or write of internal memories. 15 40 as/ale address strobe or latch enable (input) . this input is only used if multiplexed bus is selected via im input pin. 16 41 im cpu interface mode (input) . if high, this input sets the device in the multiplexed microprocessor mode. if this input is grounded, the device resumes non- multiplexed cpu interface. 17 42 dt a data acknowledgment (open drain output) . this active low output indicates that a data bus transfer is complete. a 10kohm pull-up resistor is required at this output. 18 43 irq interrupt request output (open drain output) . this active low output noti?es the controlling microprocessor of an interrupt request. it goes low only when the bits in the interrupt enable register are programmed to acknowledge the source of the interrupt as de?ned in the interrupt status register. -44nc no connect . ground 19 45 vss5 ground . 20-27 46-53 ad0 - ad7 data bus (bidirectional) . these pins provide microprocessor access to the internal memories. in the multiplexed bus mode, these pins also provide the input address to the internal address latch circuit. 28 54 vss1 ground . 29 55 teop transmit end of packet (input) . this is a strobe that is generated by the hdlc controller chip for one bit period during the last bit of the closing ?ag of the transmit packet. 30 56 reop receive end of packet (input) . a receive packet will normally be terminated when the hdlc controller asserts the reop strobe for one bit period, one bit time after the closing ?ag is received. 31 57 txcen transmit clock enable (output) . the hdlc transmitter is controlled by the idx- generated transmit clock enable signal, txcen. 32 58 rxcen receive clock enable (output) . the hdlc receiver is controlled by the idx- generated receive clock enable signal, rxcen.
advance information MT90812 5 33-34 59-61 nc no connect . ground 35 62 vss2 ground . 36-37 63-64 sti0-1 serial tdm input streams 0 and 1 (input) . serial data input streams which have data rates of 2.048 mb/s with 32 channels. 38-39 65-66 sto0-1 serial tdm output streams 0 and 1 (three-state output) . serial data output streams which have data rates of 2.048 mb/s with 32 channels. 40 67 dper d-channel input in st-bus format (input) . the mt8952b cdsto stream containing formatted d-channel data. 41 68 est0 expansion bus serial data stream 0 (three-state output/input) . this is a bi- directional pin at 8.192 mb/s in idx link mode. in tdm link mode this is a 2.048, 4.096 or 8.192 mb/s output stream. 42 1 est1 expansion bus serial data stream 1 (three-state output/input) . this is a bi- directional pin at 8.192 mb/s in idx link mode. in tdm link mode this is a 2.048, 4.096 or 8.192 mb/s input stream. 43 2 ode output device enable (input) . this is the output enable input for the serial outputs. if this input is low, sto0, sto1, est0, est1 are high impedance. if this input is high, each channel may still be put into high impedance state by using per channel control bit in the connection memory. 44 3 c8 clock 8.192 (bidirectional) . as an input this signal is used for expansion bus and/or internal clock source at 8.192 mhz depending on the timing mode selected. as an output this signal is an 8.192 mhz output clock locked to the reference input signal. 45 4 f8 frame pulse for 8.192 mhz (bidirectional) . as an input accepts and automatically identi?es frame synchronization signals formatted according to st- bus and gci interface speci?cations. as an output is an 8 khz frame pulse that indicates the start of the active frame. either f8 or fpi are used for frame synchronization depending on the timing mode selected 46 5 c8p_c16i oscillator master clock (cmos input) . for crystal operation, a 8.192mhz crystal is connected to this pin from osco. for clock oscillator operation, this pin is connected to a clock source. the clock source of either 8.192 mhz or 16.384 mhz can be used as selected in the timing control register (tc). 47 6 osco oscillator master clock (cmos output) . for crystal operation, a 8.192mhz crystal is connected from this pin to c8p_c16i. for clock oscillator operation, this pin is left unconnected. 48 7 fpi frame pulse (input). this input accepts and automatically identi?es frame synchronization signals formatted according to st-bus and gci interface speci?cations. either f8 or fpi are used for frame synchronization depending on the timing mode selected. 49 8 c4i clock 4.096 mhz (input) . this input is the 4.096 mhz clock input. -9nc no connect . ground 50-51 10-11 ic internal connect . open. 52 12 vssa analog ground . 53 13 vdda +5 volt power supply (analog) . pin description (continued) pin # name description 64 pin mqfp 68 pin plcc
MT90812 advance information 6 overview the MT90812 integrated digital switch (idx) provides the integration of several functions required in a telecom application. the idx includes a digital switch for switching up to 192 x 192 channels, ?ve conference bridges, a dtmf/supervisory-tone bus and two digital energy detector circuits for trunk call progress tone detection. there are two 2.048 mbit/s serial links and an expansion bus that can operate at 2.048, 4.096 or 8.192 mb/s. a digital frequency shift keying (fsk) transmitter, compatible to bell 202 or ccitt v.23 1200 baud is provided. d-channel control is realized by microprocessor access to the MT90812. d-channel messages are relayed to and from the 2b+d line transceivers via the local tdm link. in d-channel basic receive/transmit mode a 32 byte buffer is provided for each of the transmit and receive directions, and these can be independently assigned to speci?c d-channels. alternatively, the hdlc controller mode can be selected, providing an interface to the mt8952 hdlc. the hdlc controller mode provides the necessary control signals to operate the mt8952 hdlc in external timing mode, allowing multiplexing of the mt8952 over the local tdm links. the MT90812 also provides an interface to the expansion bus capable of linking a number of MT90812 devices together directly or in a larger matrix through other digital switches. each MT90812 can route any of the 64 channels associated with the local tdm streams onto the expansion bus. thus, system growth is easily achieved via the addition of a MT90812 device onto the expansion bus. very little hardware overhead is required to enable cost effective system growth. in a multi-idx system functions including conferencing, tone generation, supervisory signal detection, d- channel receiver and transmitter, tone ringer and fsk transmitter can be shared across the system. for example, in a system consisting of four MT90812 devices, 20 three party conferences can be supported, independent of which MT90812 the party originated. for tone generation, 6 programmable tones per MT90812 translates to 24 programmable tones in a four MT90812 system, all of which can be routed to any channel in the system. 54 14 f4o frame pulse for 4.096 mhz (output) . this is an 8 khz output frame pulse that indicates the start of the active st-bus/gci frame. the pulse width is based upon the period of the c4o clock. 55 15 c2o clock 2.048 mhz (output) . this output is an 2.048 mhz output clock locked to the reference input signal. 56 16 c4o clock 4.096 mhz (output) . this output is an 4.096 mhz output clock locked to the reference input signal. 57 17 c10o clock 10.24mhz (output) . this output is a 10.24 mhz clock locked to the reference input signal. 58 18 vss3 ground . 59 19 r+ ringing generator +ve output . this output is a 16, 20, 25 or 50 hz square wave. 60 20 r- ringing generator -ve output . square wave output 180 degrees out of phase with r+. 61 21 reset device reset (input). when 0, reset the device internal counters, registers and tri-states sto0, sto1, est0, est1 and data outputs from the microport. 62 22 ic internal connection . tie to vss for normal operation. 63 23 vdd +5 volt power supply . 64 24 vss4 ground . pin description (continued) pin # name description 64 pin mqfp 68 pin plcc
MT90812 vii 1.0 functional description.................................................................................................... 7 2.0 local tdm streams ......................................................................................................... 8 3.0 expansion bus............................................................................................................... .. 8 3.1 tdm link mode ............................................................................................................... .........................8 3.2 idx link mode............................................................................................................... ...........................9 4.0 switching................................................................................................................... ..... 10 4.1 switching functions ......................................................................................................... ......................10 4.1.1 data transfer from local tdm streams to data memory .............................................................. 10 4.1.2 data transfer from the expansion bus to data memory................................................................ 10 4.1.3 connect memory and data memory structure ............................................................................... 10 4.1.4 switching output data from either data memory or connect memory .......................................... 10 4.2 connection memory ........................................................................................................... ....................10 4.2.1 connection memory usage for switch connection or message mode .......................................... 11 4.2.2 connection memory select .................................................................................................. .......... 11 4.2.3 connect memory configurations for expansion bus modes .......................................................... 11 5.0 address memory map ................................................................................................... 12 5.1 memory page select .......................................................................................................... ....................12 5.1.1 addressing memory pages in multiplexed microport mode ........................................................... 12 5.1.2 addressing memory pages in non-multiplexed microport mode.................................................... 12 5.2 data memory and connect memory .............................................................................................. ........13 5.2.1 local data memory ......................................................................................................... ............... 13 5.3 connection memory use in conferencing, gain control and specifying incoming sources for .............14 5.4 use of data memory reserved for expansion bus streams .................................................................15 6.0 conferencing ................................................................................................................ .18 6.1 channel attenuation......................................................................................................... ......................20 6.2 noise suppression and channel inversion ..................................................................................... .......20 6.3 tone insertion .............................................................................................................. ..........................20 6.4 conference overflow......................................................................................................... .....................21 6.5 starting a new conference ................................................................................................... .................21 6.6 removing a channel from a conference........................................................................................ ........21 7.0 gain control ................................................................................................................ ... 22 8.0 delays through the MT90812 ....................................................................................... 22 8.1 minimum delay mode (cst bit=0) .............................................................................................. ...........22 8.2 constant delay mode (cst bit=1) ............................................................................................. ............24 8.3 delays in conferencing ...................................................................................................... ....................24
MT90812 viii 9.0 timing and clock control............................................................................................. 24 9.1 input timing reference ...................................................................................................... ....................25 9.2 serial data interface timing................................................................................................ ...................26 9.2.1 local streams, sti/o0 and sti/o1.......................................................................................... ........ 26 9.2.2 expansion bus, est0/1..................................................................................................... ............. 26 9.2.3 hmvip frame alignment..................................................................................................... ........... 26 9.2.4 output clock and frame pulse signals...................................................................................... .... 26 9.2.5 selecting timing from the input clock reference or from the pll ................................................ 27 9.3 phase lock loop (pll)....................................................................................................... ...................27 9.3.1 master and slave pll modes ................................................................................................ ........ 28 9.4 watchdog timer .............................................................................................................. .......................28 9.5 c8p pin timing source ....................................................................................................... ...................29 9.5.1 clock oscillator.......................................................................................................... ..................... 29 9.5.2 crystal oscillator ........................................................................................................ .................... 29 10.0 d-channel signalling support ..................................................................................... 30 11.0 d-channel basic receive transmit block................................................................... 30 11.1 receiver operation ......................................................................................................... .......................31 11.1.1 receiver interrupt handling.............................................................................................. .............. 32 12.0 transmitter operation ................................................................................................... 33 12.1 transmitter interrupt handling............................................................................................. ...................34 13.0 hdlc resource allocator module ............................................................................... 34 13.1 general description of MT90812 and shared hdlc configuration.......................................................35 13.2 connection to mt8952 hdlc controller and mt9171/72b dnic .........................................................35 13.2.1 connection to mt8952b hdlc controller .................................................................................... . 35 13.2.2 connection to mt9171/72b dnic ............................................................................................ ...... 36 13.2.3 data stream flow......................................................................................................... .................. 36 13.3 tx control ................................................................................................................. .............................37 13.3.1 generation of txcen ...................................................................................................... ............... 37 13.3.2 end of the transmission of a packet...................................................................................... ........ 38 13.3.3 tx and rx handshaking .................................................................................................... ............ 38 13.3.4 merging of d and c-channels.............................................................................................. ........... 39 14.0 rx control................................................................................................................. ..... 39 14.1 rx circuit functions....................................................................................................... ........................39 14.1.1 generation of rxcen...................................................................................................... ............... 40 14.1.2 dedicated receive mode ................................................................................................... ............ 40 14.1.3 multiplexed receive mode ................................................................................................. ............ 40 14.1.4 auto-hunt monitoring ..................................................................................................... ................. 40 14.1.5 cts generation........................................................................................................... ................... 41 14.1.6 circumstances when monitoring a channel is stopped ................................................................ 41
MT90812 ix 14.1.7 receive packet termination............................................................................................... ............ 42 15.0 c-channel data............................................................................................................. .42 16.0 tone generation ............................................................................................................ 42 16.1 tone ringer................................................................................................................ ............................44 17.0 frequency shift keying (fsk) transmitter.................................................................. 45 18.0 ringing generator ......................................................................................................... 4 7 19.0 supervisory signal detection and cadence measurement ....................................... 47 20.0 microprocessor port...................................................................................................... 49 21.0 connection memory bits .............................................................................................. 50 21.1 connection memory high..................................................................................................... ..................50 21.2 connection memory low...................................................................................................... ..................51 22.0 detailed register descriptions .................................................................................... 52 22.1 address memory select register (ams)....................................................................................... .........53 22.2 control register (ctl) ..................................................................................................... ......................54 22.3 timing control register (tc) ............................................................................................... ..................55 22.4 output clocking control register (occ) ..................................................................................... ..........56 22.5 interrupt status register (ints)........................................................................................... ..................56 22.6 interrupt enable register (inte)........................................................................................... .................57 22.7 ringer and fsk control register (rfc) ...................................................................................... ..........58 22.8 fsk transmit memory (fskm) ................................................................................................. .............59 22.9 conference overflow status register (confo)................................................................................ ....59 22.10 conference control register (cc) .......................................................................................... ...............60 22.11 tone generation and energy detect control register (tedc) .............................................................61 22.12 energy detect a - low threshold (edalt) .................................................................................. ........62 22.13 energy detect a - high threshold (edaht) ................................................................................. ........62 22.14 supervisory signal cadence register a (ssca) .............................................................................. .....62 22.15 energy detect b - low threshold register (edblt) .......................................................................... ...63 22.16 energy detect b - high threshold register (edbht)........................................................................ ...63 22.17 supervisory signal cadence register b (sscb) ............................................................................. .....63 22.18 low tone coefficient registers 1-7 (ltc1-7).............................................................................. .........64 22.19 high tone coefficient registers 1-7 (htc1-7) .............................................................................. ........64 22.20 conference party control register (cpc1-15) .............................................................................. .......65 22.21 d-channel receive interrupt threshold (drxit)............................................................................. ......66 22.21.1 message length interrupt mode.......................................................................................... .......... 66 22.21.2 fifo level interrupt mode .............................................................................................. .............. 66 22.22 d-channel rx control (drxc) ............................................................................................... ...............67 22.23 d-channel br status (drxs)................................................................................................ ................68
MT90812 x 22.24 d-channel rx fifo output (drxout)......................................................................................... ........68 22.25 d-channel tx fifo input (dtxin) ........................................................................................... .............68 22.26 d-channel tx control (dtxc)............................................................................................... ................69 22.27 hra ctrl register 1(hc1) .................................................................................................. ................70 22.28 hra ctrl register 2 (hc2) ................................................................................................. ................71 22.29 hra ctrl register 3 (hc3) ................................................................................................. ................72 22.30 hra lock out register 1 (hlo1) ............................................................................................ ..............72 22.31 hra lock out register 2 (hlo2) ............................................................................................ ..............72 22.32 hra status 1 (hs1) ........................................................................................................ .......................73 22.33 hra status 2 (hs2) ........................................................................................................ .......................74 22.34 hra status 3 (hs3) ........................................................................................................ .......................74 22.35 hra status 4 (hs4) ........................................................................................................ .......................75 23.0 applications............................................................................................................... .... 76 23.1 local tdm channel assignment............................................................................................... .............76 23.2 dnic channel assignment .................................................................................................... ................76 24.0 ac/dc electrical characteristics ................................................................................. 78 24.1 timing references for tdm streams.......................................................................................... ...........85 24.2 ac parameters referenced to incoming clock signals........................................................................ 89 24.3 ac parameters referenced to outgoing c4 or c8 clock signals ........................................................91 24.4 hra timing ................................................................................................................. ...........................92
advance information MT90812 7 1.0 functional description the functional block diagram of fig. 1 depicts the main operations performed by the MT90812. the integrated digital switch has three tdm streams. the two local tdm serial streams, sti/o0 and sti/o1, operate at 2048kbit/s and are arranged in 125us wide frames each containing 32 8-bit channels. the third tdm stream, comprised of est0 and est1 can be used as an additional serial stream at 2.048, 4.096 or 8.192 mb/s, supporting 32, 64 or 128 channels, respectively. the expansion bus, est0/1 operates in two modes, tdm link and idx link modes. idx link mode allows multiple MT90812 devices to be linked together very ef?ciently. in idx link mode, the incoming data on the local tdm streams of each MT90812 is transferred onto the expansion bus to enable switching channels between a maximum of four MT90812 devices. for tdm link mode, the expansion bus is con?gured as a tdm serial stream which can run at 2.048, 4.096 or 8.192 mb/s. in tdm link mode, the MT90812 can be connected to more peripheral devices or to other digital switches (i.e. mt8980/1/2 or mt8985/6) to support larger matrices of MT90812 devices. the MT90812 can switch data from channels on the local and expansion input streams to channels on the local and expansion output streams. the controlling microprocessor can simultaneously read channels on tdm inputs or write to channels on tdm outputs (message mode). to the microprocessor, the MT90812 looks like a memory peripheral. the microprocessor can write to the MT90812 to establish switched connections between input tdm channels and output tdm channels, or to transmit messages on output tdm channels. by reading from the MT90812, the microprocessor can receive messages from tdm input channels or check which switched connections have already been established. the MT90812 provides conference call capability and supports a total of 15 parties, distributed over a maximum of 5 conferences. (i.e. 1x15 parties, 3x5 parties, 5x3 parties etc.). conference parties can be from any of the incoming channels on the local or expansion tdm streams. gain control is provided on the outgoing channels, with a range of +3 to -27 db in steps of 1db, as well as - db. if a channel is in a conference, incoming and/or outgoing gain control is provided. conference incoming gain also ranges from +3 to -27 db in steps of 1db, as well as - db. conference outgoing gain can range from 0 to -9 db in steps of 3 db. a tone source of 32 dual tones is generated from the tone generator block and stored in data memory. outgoing gain control of +3 to -27 db in steps of 1db, as well as - db, is provided for each tone. seven of the thirty-two tones are programmable in frequency. the 32 locations can be switched to outgoing channels or accessed by the microprocessor. a phase coherent fsk transmitter generates two output frequencies, representing the marks and spaces, selectable to bell 202 or ccitt v.23 standards at 1200 baud. the fsk transmitter output is a pcm coded signal that can be directed to any outgoing local tdm channel. two energy detect blocks provide monitoring capability of supervisory signalling for any of the tdm channels. d-channel access is provided to link the microprocessor to the transceivers. there are two modes of message formats that can be selected. the d-channel basic receive transmit (dbrt) mode provides basic formatting of the data, which includes start and stop signalling and parity checking. in dbrt mode there are rx and tx buffers, 32 bytes in length, which can be allocated to any of the incoming/outgoing channels of the tdm streams. the hdlc mode provides a control interface to facilitate the multiplexing of an external hdlc controller (mt8952) over any of the local tdm streams d-channels. c-channel access for control of st-bus family of devices (e.g. mt9160b, mt9171/72, mt8930, mt8910) is provided through message mode. each of the programmable parameters within the functional blocks are accessed through a parallel microprocessor port compatible with cpu non-multiplexed bus and intel ? , motorola ? and national ?
MT90812 advance information 8 multiplexed bus speci?cations. the MT90812 can operate in either a -law or m- law as de?ned in control register as speci?ed in section control register (ctl) on page 54. 2.0 local tdm streams there are two local serial time division multiplexed (tdm) streams. these streams at sti/o0 and sti/o1 provide a link between the MT90812 and other peripheral devices, including those in the st-bus family (e.g. mt9160b, mt9171/72, mt8930, mt8910). the two serial streams operate at 2.048 mbit/s and are arranged in 125us wide frames, each comprising 32 8-bit channels. refer to section 9.2, serial data interface timing. the MT90812 can support primary rate or basic rate devices. using the basic rate devices (e.g. mt9171/72 dnic) in dual port mode the d-channel should be assigned to sti/o1 streams. d-channel signalling support is provided for any timeslot on the sti/o1 streams for the hdlc controller mode. the dbrt can access any timeslot and stream. refer to d-channel signalling support on page 30 and local tdm channel assignment on page 76. 3.0 expansion bus the expansion bus operates in 2 modes, idx link and tdm link modes. the modes will be described in the following sections. section 24.1 describes the timing references used for both expansion bus modes. 3.1 tdm link mode in this mode, the expansion bus at est0 and est1, are regular output and input serial streams, respectively. they operate at either 2.048, 4.096 or 8.192 mbit/s. refer to fig. 4. figure 4 - expansion bus (tdm link mode) at 2.048mb/s, the ?rst block of 32 locations of data memory reserved for the expansion bus are used. the 32 incoming channels on est1 are placed in expansion block 1 of data memory. at 4.096 mb/s the 64 bytes are utilized in expansion block 1 and 2. at 8.192mb/s all 128 locations of data memory reserved for the expansion bus are used. refer to the description of the memory allocation in section 5, address memory map. this mode allows linking larger matrices of MT90812 devices together. for example, at 2.048 mb/s, eight MT90812 devices can be connected together via a mt8980d (dx) switch, as shown in fig. 5. figure 5 - eight idx con?guration using expansion bus tdm link mode f0i channel i+32 frame n i sti/o1 . . . est0/1 . . . a1 a2 e1 e2 e3 e4 e5 e6 e7 e8 e128 trunks and ports idxa est1 sti/o0-1 trunk s and ports idxb est1 sti/o0-1 est0 uport trunks and ports idxh est1 sti/o0-1 est0 uport uport est0 dx mpu uport
advance information MT90812 9 each of the 32 channels of the 8 streams connecting the dx and idx devices can be switched to any outgoing channel and stream. this provides switching across all eight of the MT90812 devices. 3.2 idx link mode in idx link mode the expansion bus allows up to four idx devices to be connected together as shown in fig. 6. the data ?ow between each MT90812 is supported on est0 and est1, two serial streams which operate at 8.192mbit/s and each consist of 128 8-bit channels. the four MT90812 devices are labelled a, b, c and d. tdm link mode can be used to link more than four MT90812 devices, refer to section 3.1. the expansion bus channel assignment for est0 and est1 in idx link mode is shown in fig. 7. for the est0 stream, each of the four MT90812 devices place data into 32 timeslots and read in data from the other 96 timeslots. the ebus position, as de?ned in section control register (ctl) on page 54, designates which timeslots the MT90812 writes to and which timeslots it reads from. for example, idx a would output channel 1 at the timeslot shown as a1 in fig. 7 and input data during timeslots b1, c1 and d1. for the est1 stream, each MT90812 reads in 32 channels and can output data to the other 96 channels. which channels are read are also determined from the ebus position bits. for example, idx a will read data during timeslots ea1, ea2,... ea32. each MT90812 will receive a total of 128 channels from the two streams est0 and est1. for idx a, the 96 channels, b1-b32, c1-c32, d1-d32 will be taken from est0 and 32 channels ea1-ea32 will be taken from est1. a description of programming the switch in idx link mode is given in connection memory on page 10. the data memory allocation is described in section 5, address memory map. on power up, the ebus is set to high impedance and the ebus position bits must be programmed to avoid any contention on the two streams. figure 6 - four idx con?guration using expansion bus idx link mode figure 7 - expansion bus - idx link mode idx a system expansion bus (tdm) mpu st/o1 sti/o0 uport trunks and port s idx b sti/o1 sti/o0 uport idx c sti/o1 sti/o0 uport idx d sti/o1 sti/o0 uport trunks and ports trunks and ports trunks and ports f0i frame n est0 est1 a1 b1 c1 d1 a2 b2 c2 d2 a32 b32 c32 d32 ea32 eb32 ec32 ed32 ea1 eb1 ec1 ed1
MT90812 advance information 10 4.0 switching the switching function of the MT90812 is described in four parts: ? how incoming data from the local tdm streams are transferred to data memory. ? how incoming data from the expansion bus is transferred to data memory for both idx link and tdm link modes. ? connect memory and data memory structure. ? how output data can be switched from either data memory or connect memory. each will be described below with reference to figure 1 - functional block diagram. this is followed by a more detailed description of connect memory in section 4.2. 4.1 switching functions 4.1.1 data transfer from local tdm streams to data memory the serial data incoming to the MT90812 is converted into parallel format (8 bits per channel) with the parallel to serial converters for both the local and expansion bus streams. this data is written to consecutive locations in data memory. the two local tdm streams, sti/o0 and sti/o1, operate at 2.048 mb/s for a total of 64 channels per frame. the 64 bytes are stored in the local data memory page in locations 00 h to 3f h . refer to the memory map in ta b l e 2 . 4.1.2 data transfer from the expansion bus to data memory in tdm link mode, the expansion bus rate can be set to 2.048, 4.096 or 8.192 mb/s, where the number of channels used for the expansion bus stream are 32, 64 and 128, respectively. in data memory, 128 bytes are reserved in the expansion data memory page. if there are only 32 or 64 channels (i.e. at 2.048 or 4.096 mb/s) then the ?rst 32 or 64 locations of the 128 reserved for the expansion bus are used. in idx link mode, the expansion bus rate is set to 8.192 mb/s. there are 96 channels incoming from est0 and 32 channels from esti, for a total of 128 incoming channels read into expansion data memory. 4.1.3 connect memory and data memory structure there are 64 locations in local data memory reserved for the incoming channels of the local tdm streams and 128 locations in expansion data memory, for the incoming channels of the expansion bus. in addition, there are another 32 locations of local data memory reserved for the tone generator. refer to the memory map in table 2. for each output channel there is an associated connect memory location. there are 128 locations for the outgoing expansion bus channels and 64 for the local tdm streams. 4.1.4 switching output data from either data memory or connect memory when the MT90812 switches data from input channels to outgoing channels, the address for data memory is read from the connection memory location corresponding to the desired output channel. in message mode, output data is read directly from the connection memory location corresponding to the output channel. the details for setting up the connections are given in the following section. 4.2 connection memory the use of connection memory in the MT90812 is described in three parts: ? connection memory usage for switch connection or message mode
advance information MT90812 11 ? control register and the use of connection memory ? connect memory con?gurations for expansion bus modes each will be described below. a full description of addressing memory in the MT90812 is given in address memory map starting on page 12. refer to section 21.0 for a de?nition of the connection memory high and low bits. 4.2.1 connection memory usage for switch connection or message mode locations in the connection memory, which is split into high and low parts, are associated with particular tdm output streams. when a channel is due to be transmitted on an tdm output stream, the data for the channel can either be switched from an tdm input stream or it can originate from the microprocessor. if the data is switched from an input, then the contents of the connection memory low location associated with the output channel is used to address the data memory. the data memory address corresponds to the channel on the input stream on which the data for switching arrived. if the data for the output channel originates from the microprocessor (message mode), then the contents of the connection memory low location associated with the output channel are output directly, and this data is output repetitively on the channel once every frame until the microprocessor intervenes. the connection memory high determines whether individual output channels are in message mode, controls individual output channels to go into a high-impedance state and speci?es the gain for each outgoing channel. 4.2.2 connection memory select if the microport is operating in multiplexed mode, addressing the high and low sections of connection memory is done by setting the memory select bits in control register. if the microport is operating in non-multiplexed mode, addressing the high and low sections of connection memory is done by setting the external address bits a9,a8,a7. refer to address memory select register (ams) on page 53, and microprocessor port on page 49. the control register also consists of mode control bits that allows the chip to broadcast messages on all tdm output channels (i.e., to put every channel into message mode). mode control bit 5, ct2:msg bit, puts every output channel on every output stream into active message mode; i.e., the contents of the connection memory low are output on the tdm output streams once every frame unless the ode pin is low. in this mode the chip behaves as if bits 2 and 0 of every connection memory high location were 1, regardless of the actual values. if car:msg bit is 0, then bits 2 and 0 of each connection memory high location function as follows. if cmh:bit 2 is set to 1, the associated tdm output channel is in message mode; i.e., the byte in the corresponding connection memory low location is transmitted on the stream at that channel. otherwise, the serial input is transmitted and the connection memory low de?nes the associated input stream and channel where the byte is to be found. if the ode pin is low, then all serial outputs are high-impedance. if the ode pin is high and car:msg bit is 1, then all outputs are active. if the ode pin is high and car:msg bit is 0, then the bit 0 in the connection memory high location enables the output driver for the corresponding individual output stream and channel. cmh:bit 0=1 enables the driver and cmh:bit 0=0 disables it. 4.2.3 connect memory con?gurations for expansion bus modes in tdm link mode, the 128 connect memory locations reserved for the expansion bus are associated with the outgoing channels of est0. in idx link mode, there are 32 outgoing channels for the est0 stream. for the est1 stream there are 96 outgoing channels. in this mode the connection memory is con?gured such that the ?rst 32 locations are used for the est0 stream. the next 96 locations are for the est1 stream as selected by the expansion bus position bits as described in address memory map on page 12.
MT90812 advance information 12 5.0 address memory map the MT90812 memory is accessed via the microport. the microport can operate in multiplexed or non- multiplexed mode as described in microprocessor port on page 49 the access to the MT90812 memory for multiplexed and non-multiplexed mode is described below. 5.1 memory page select the MT90812 memory is divided into 7 pages, as listed in table 1. in multiplexed mode, the memory select bits in the address memory select register (ams) determine the page that is addressed. in non-multiplexed mode, the external address bits a9,a8,a7, determine the page that is addressed, eliminating the need to access the ams register for memory page select. the control registers, described in detailed register descriptions on page 52, consist of one page of 128 locations. in multiplexed mode the control registers are accessed independent of the setting of the memory select bits in the ams register, by setting the external address bit a7 to low. in non-multiplexed model the control registers are accessed by setting address bits a9, a8, and a7 to high. the control register at location 61 h ( 3e1 h in motorola non-muxed, 061 h in in multiplexed mode) must be initialized to 080 h . the addressing of the other blocks and memory pages are described below. each data and connect memory page consists of 128 locations, as shown in table 2. 5.1.1 addressing memory pages in multiplexed microport mode an MT90812 memory address, in multiplexed microport mode, consists of two portions. the higher order bits(3) originate from the control address memory select (ams) register, which may be written to or read from via the control interface. the control interface receives address information at a7 to a0, data information at d7 to d0 and handles the microprocessor control signals cs, dt a, r/ w and ds. the lower order bits(8) originate from the address lines directly. the address lines a6-a0, on the control interface, give access to the MT90812 registers directly if a7 is zero, or depending on the contents of ams register, to the high or low sections of the connection memory, or to the data memory. 5.1.2 addressing memory pages in non-multiplexed microport mode. a MT90812 memory address, in non-multiplexed microport mode, consists of a9 to a0. the higher order bits(3) originating from the external address bits a9,a8,a7, control which page is accessed. the control interface receives address information at a9 to a0, data information at d7 to d0 and handles the microprocessor control signals cs, dt a, r/ w and ds. the lower order bits(7) originating from the external non-multiplexed mode multiplexed mode memory pages external address a9,a8,a7 memory select bits external address a7 111 xxx 0 control registers (section 22.0) 000 000 1 local data memory 001 001 1 expansion data memory 010 010 1 local connect memory low 011 011 1 expansion connect memory low 100 100 1 local connect memory high 101 101 1 expansion connect memory high table 1 - MT90812 memory page select
advance information MT90812 13 address bits a6-a0, give access to the control registers if a9,a8,a7=111, or depending on the high order bits a9,a8,a7, to the high or low sections of the connection memory, or to the data memory, as shown in table 1. 5.2 data memory and connect memory the data memory and connect memory map, as shown in table 2, illustrates the direct relationship between dm and cm for each of the channels. as described earlier in switching on page 10, the data from the incoming tdm streams are written to data memory and the data is switched to the outgoing streams as programmed in cm. table 2 - data memory and connect memory pages in addition to holding the incoming data from the tdm streams, data memory holds the output of the other MT90812 blocks. this is described below in the following section. expansion memory page is used to hold the incoming data for the expansion bus tdm streams. refer to use of data memory reserved for expansion bus streams on page 15 for further description. connection memory is used to specify the source for the outgoing channels and to connect the conference, energy detect and dbrt blocks to incoming channels. in addition cm is used to specify the gain for the outgoing streams and the gain for the tones. 5.2.1 local data memory table 3 details the mapping for local data memory. the ?rst block of 32 locations in data memory is used to store the 32 bytes of data from the sti0 stream. locations 20-3f are used for the 32 bytes of data from the sti1 stream. the third block of 32 locations in data memory is used for output of the tone generator block as described in tone generation on page 42 the next 32-bytes consist of conference outputs(15), dchout(1), and some unused locations(14). address a6-a0 memory pages data memory pages connect memory high pages connect memory low pages local memory page 00-1f sti0 32 channels sto0 outgoing gain sto0 32 channels 20-3f sti1 32 channels sto1 outgoing gain sto1 32 channels 40-5f tones(32), fsk tone gain (not used) 60-7f confout(15), unused(1) dchout(1) unused(15) conf - incoming gain(15) it - incoming gain(1) dch - incoming gain(1) eda -incoming gain(1) edb -incoming gain(1) unused(13) conf incoming channel(15) insertion tone - channel(1) dch incoming channel(1) eda -incoming channel(1) edb -incoming channel(1) unused(13) expansion memory page 00-1f ei1 32 channels ei1 outgoing gain ei1 32 channels 20-3f ei2 32 channels ei2 outgoing gain ei2 32 channels 40-5f ei3 32 channels ei3 outgoing gain ei3 32 channels 60-7f ei4 32 channels ei4 outgoing gain ei4 32 channels address a6-a0 local data memory description 00-1f sti0 32 channels sti0 32 channels 20-3f sti1 32 channels sti1 32 channels
MT90812 advance information 14 table 3 - local data memory 5.3 connection memory use in conferencing, gain control and specifying incoming sources for energy detect and dbr connection memory is used to specify the source and gain for the 64 outgoing channels of sto0 and sto1 and the 128 outgoing channels of the expansion bus. message mode, minimum or constant delay and output enable are speci?ed in cmh for each of these channels. the conference circuit incoming channels are speci?ed in cml 60-6e. the incoming conference gain, inversion bit and noise suppression are speci?ed in cmh. message mode, minimum or constant delay and output enable are not used for locations 60-6e reserved for conference control. see the description of connection memory high on page 50. channels are transferred from data memory with constant delay to the conference circuit. channels that can be included in a conference include; the 64 channels of sti0 and sti1, and the channels of the four expansion bus blocks. in fact any location in local or expansion data memory can be speci?ed as in incoming conference source, including the 32 tones or the dbt output. the conference party control registers specify which conference the channel is participating in, output attenuation levels, tone insertion and conference initialization. see the description of conference party control register (cpc1-15) on page 65. cm is also used to connect the energy detect and dbrt blocks to incoming channels. the incoming channels are speci?ed in cml. the locations are listed below in table 4. channels can be transferred from data memory in either minimum or constant delay to the energy detect and dbrt blocks as speci?ed in cmh. cmh message mode and output enable bits are ignored for locations 70-72 of cm. in addition cmh is used to specify the gain for the outgoing streams and the gain for the tones. table 4 - connect memory 40-59 dtmf tones(26) dtmf tone generator output 59-5a tone ringer or dtmf tone ringer or tone generator output 5b-5e tones(4) tone generator output 5f fsk or dtmf fsk transmitter output or tone generator output 60-6e confout(15) conference output 6f unused 70 dchout(1) output from the d-channel tx fifo buffer. allows d-channel tx buffer to be directed to any outgoing channel. 71-7f unused(14) unused(14) hex address a6-a0 connect memory low description 60-6e confin conference incoming 6f insertion tone insertion tone incoming channel 70 dch in(1) incoming channel to be transferred to the dbr fifo. 71 eda incoming channel to be transferred to the energy detect a block 72 edb incoming channel to be transferred to the energy detect b block 72-7f unused(14) unused(14) address a6-a0 local data memory description
advance information MT90812 15 channels can be transferred from data memory in either minimum or constant delay to the energy detect and dbrt blocks as speci?ed in cmh. cmh message mode and output enable bits are ignored for locations 70-72 of cm. in addition, cmh is used to specify the gain for the outgoing streams and the gain for the tones. 5.4 use of data memory reserved for expansion bus streams the use of the four blocks reserved for the expansion bus is dependent on the expansion bus mode set for the device. the two expansion bus modes, tdm link and idx link, are described on page 8. in tdm link mode, the four blocks are used according to the data rate set for the expansion bus. at 2.048 mb/s the ?rst 32 bytes, 00-1f, are used to store the incoming data. at 4.096 mb/s the ?rst 64 bytes, 00-3f are used. at 8.192 mb/s all 128 locations, 00-7f, are used. in idx link mode, 128 channels are read, 32 from est1 and 96 from est0. the positions that the MT90812 will read and write to the expansion bus are controlled by the ep1 and ep0 bits in control register b. for example, a group of four MT90812 devices are labelled a, b, c, and d. the 128 channels on the expansion bus streams are identi?ed as a1, b1, c1, d1, a2, b2, c2, d2,...., a128, b128, c128, d128. the MT90812 with ep1,ep0 set to 0,0 will read and write to est0 and est1 as listed in table 5. table 5 - expansion bus read/write timeslots for idx a the MT90812 with ep1,ep0 set to 0,0 will output on est0 during channel ai and will read the next three channels bi, ci and di. channels bi, ci and di go into data memory at expansion block 2, 3 and 4 respectively, as shown in table 2. expansion block 1 will contain incoming channels on est1 sent to idx a in timeslots labelled ea1,...,ea32 as shown in figure 7 on page 9. the MT90812 with ep1,ep0 set to 0,1 will output on est0 and read est1 during channel bi and will read est0 and output on est1 for channels ai, ci and di. the memory map for the expansion bus timeslots are shown in the figures 8 - 12 for each of the four settings of ep1 and ep0. expansion bus channel (i=1,32) ai bi ci di est0 write read read read est1 read write write write
MT90812 advance information 16 figure 8 - data memory assignment for expansion bus timeslots for ep1,ep0 = 00 figure 9 - data memory assignment for expansion bus timeslots for ep1,ep0 = 01 est0 a1 b1 c1 d1 a2 b2 c2 d2 est1 a1 b1 c1 d1 a2 b2 c2 d2 fp expansion bus data memory hex a6-a0 dm 00 ea1 01 ea2 -- 1f ea32 20 eb1 21 eb2 -- 3f eb32 40 ec1 41 ec2 -- 5f ec32 60 ed1 61 ed2 -- 7f ed32 . expansion bus data memory hex a6-a0 dm 00 ea1 01 ea2 -- 1f ea32 20 eb1 21 eb2 -- 3f eb32 40 ec1 41 ec2 -- 5f ec32 60 ed1 61 ed2 -- 7f ed32 est0 a1 b1 c1 d1 a2 b2 c2 d2 est1 a1 b1 c1 d1 a2 b2 c2 d2 fp
advance information MT90812 17 figure 10 - data memory assignment for expansion bus timeslots for ep1,ep0 = 10 figure 11 - data memory assignment for expansion bus timeslots for ep1,ep0 = 11 the previous diagrams illustrate the data memory allocation for the timeslots on est0 and est1. fig. 12 illustrates connect memory allocation for the timeslots on est0 and est1. for the idx a, which has ep0,ep1 = 00, the circled timeslots are read as incoming data to data memory. the other timeslots are outgoing . expansion bus data memory hex 00 a6-a0 dm 00 ea1 01 ea2 -- 1f ea32 20 eb1 21 eb2 -- 3f eb32 40 ec1 41 ec2 -- 5f ec32 60 ed1 61 ed2 -- 7f ed32 est0 a1 b1 c1 d1 a2 b2 c2 d2 est1 a1 b1 c1 d1 a2 b2 c2 d2 fp . expansion bus data memory hex a6-a0 dm 00 ea1 01 ea2 -- 1f ea32 20 eb1 21 eb2 -- 3f eb32 40 ec1 41 ec2 -- 5f ec32 60 ed1 61 ed2 -- 7f ed32 est0 a1 b1 c1 d1 a2 b2 c2 d2 est1 a1 b1 c1 d1 a2 b2 c2 d2 fp
MT90812 advance information 18 timeslots, where idx a can either write to est0 and est1 during these channels or place est0 or est1 in high impedance. figure 12 - data memory assignment for expansion bus timeslots for ep1,ep0 = 00 6.0 conferencing the conference block provides conference call capability in the MT90812 and supports a total of 15 parties, distributed over a maximum of 5 conferences. (i.e. 1x15 parties, 3x5 parties, 5x3 parties etc.). a/m-law companded data from an incoming channel is converted to linear format, applied incoming gain, processed by a dedicated arithmetic unit, applied outgoing conference gain and stored in data memory in linear format. the output signal contains all the information of each channel connected in conference except its own. for each of the 15 conference parties there is a conference party control register. the conference party control register contains the conference id, start bit, insertion tone enable and outgoing channel attenuation. refer to conference party control register (cpc1-15) on page 65. the output for each conference is stored in 1 of 15 data memory locations which can be switched to any outgoing channel. for each of the 15 dm locations the corresponding connect memory low byte is used to specify the incoming source channel. the conference circuit incoming channels are speci?ed in cml 60-6e. the incoming conference gain, inversion bit and noise suppression are speci?ed in cmh. message mode, minimum or constant delay and output enable are not used for locations 60-6e reserved for conference control. see the description of connection memory high on page 50. channels are transferred from data memory with constant delay to the conference circuit. . expansion bus data memory hex a6-a0 dm 00 ea1 01 ea2 -- 1f ea32 20 eb1 21 eb2 -- 3f eb32 40 ec1 41 ec2 -- 5f ec32 60 ed1 61 ed2 -- 7f ed32 est0 a1 b1 c1 d1 a2 b2 c2 d2 est1 a1 b1 c1 d1 a2 b2 c2 d2 fp . expansion bus connect memory hex a6-a0 cm 00 ea1 01 ea2 -- 1f ea32 20 eb1 21 eb2 -- 3f eb32 40 ec1 41 ec2 -- 5f ec32 60 ed1 61 ed2 -- 7f ed32
advance information MT90812 19 channels that can be included in a conference include; the 64 channels of sti0 and sti1, and the channels of the four expansion bus blocks. in fact any location in local or expansion data memory can be speci?ed as in incoming conference source, including the 32 tones or the dbt output. figure 13 - conference circuit block diagram figure 14 - four party conference example cpc1-15: data memory connect memory 60-6e h * control registers gcin,inv,ns 30-3e h low incoming channel high conference output 60-6e h * incoming data incoming channel gain pad accumulator 1 accumulator 5 add / subtract gain pad noise suppression insertion tone gain channel inversion confo: cid 08 h cc: td2-0, cfen 09 h ints: cfs 04 h inte: cfe 05 h 6f h *refer to fig. 15 for address information. cid,st,it,gcout 1 frame (32 timeslots) slot 8 slot 9 slot 10 slot 11 sti sto (input) (output) abcd b+c+f a+c+f a+b+f h slot 12 slot 13 slot 14 slot 15 efgh g a+b+c e d conf #1
MT90812 advance information 20 figure 15 - conference control with conference party control registers and connect memory 6.1 channel attenuation channel attenuation is provided on incoming and outgoing channels that are in a conference. the gain can range from +3 to -27 db in steps of 1db, as well as - db for the incoming pcm data and +0 to -9 db in steps of 3db for outgoing pcm data. if an over?ow condition occurs, then the input from each channel in a conference can be independently attenuated, by setting the incoming channel attenuation bits in cmh for the speci?c conference party. the outgoing gain bits are in the conference party control register. 6.2 noise suppression and channel inversion channel inversion and noise suppression bits are speci?ed in connect memory high for locations 60-6e h . when noise suppression is enabled for a speci?c input channel, then the pcm bytes for this channel, when below the selected threshold level, are converted to pcm bytes corresponding to the minimum pcm code level before being added to the conference sum. the four threshold levels available correspond to the ?rst, ?fth, ninth, and sixteenth step of the ?rst segment. these are 1/4096, 9/4096, 16/4096, and 32/4096 with respect to full scale a-law, and 1/8159, 9/8159, 16/8159, and 32/8159 with respect to full scale ulaw. the threshold level is set using the threshold bits ns1, ns0. the inversion bit allows for every other channel in a conference to be inverted. this reduces noise due to re?ections and line impedance mismatch. 6.3 tone insertion as a party is added to a conference, if the insertion tone bit (it) is set, all channels connected in a conference will have the tone added to the conference output. this allows for conference users to be informed of a new party being added to the conference, or to be reminded that they are in a conference. the dm address of the desired tone must be programmed at location 6f h of cml, (16f h non-mux mode, ef h for mux-mode addressing). the tone source may be from any location in dm, including any of the 32 tones from the tone generation block. the pcm data from the speci?ed data memory location will be added to the conference output for a speci?ed tone duration. the tone duration is speci?ed in the conference control register (cc). refer to section 22.10 for a description of the conference control register (cc). the tone duration can be set from 0.125 to 1.0 seconds in steps of 0.125 seconds. the tone duration is from the time the party is added to the conference by writing the conference party control register. cid,st,it,gcout data memory connect memory 60-6e h* control registers parallel to serial output stream gcin,inv,ns 30-3e hex low incoming channel high conference output 60-6e h* *in non-mux micro-port mode the addresses for dm,cml and cmh are 060-06e, 160-16e and 260-26e respectively and the control registers addresses are 3b0-3be. for mux-mode the addresses dm,cml and cmh are e0-ee and the control registers addresses are 30-3e with the appropriate page selected in the ams register. incoming data incoming channel 6f h insertion tone gain conf party control
advance information MT90812 21 6.4 conference over?ow a peak clipping indicator identi?es the conference causing conference bridge over?ow whenever a 14-bit two complement over?ow occurs 1 . once a conference bridge over?ow occurs an interrupt is asserted, the conference over?ow bit in the interrupt status register (ints) is set and the conference id is placed in the conference over?ow status register (confo). refer to the confo register description on page 59. reading the interrupt status register (ints) will clear the conference over?ow bit. a conference over?ow will not trigger an interrupt until the conference over?ow bit is cleared. the conference over?ow interrupt is maskable using the interrupt enable register (inte). the conference interrupt mask does not disable updates of the confo register. the conference id in this register will not be updated again until it is reset. the register is reset following a read of the register or resetting the conference block or MT90812 device. note 1 - the over?ow limit is the same whether ulaw or alaw companding is used. following gain adjustment companding will then implement clipping to the ulaw and alaw max values of 8031 and 4032 respectively. 6.5 starting a new conference in order to initiate a conference, the conference party control register as well as connection memory low/ high must be programmed. setting the st bit for the ?rst party programmed for a conference will remove any other parties that may have been previously programmed for that conference. the following steps outline initialization of a conference. conference block initialization 1) perform MT90812 reset or conference reset. the conference party control registers are reset and all conference id numbers are set to null. 2) disable the conference over?ow interrupt until after a conference is set up. set cfe bit low in the interrupt enable register (inte). 3) enable the conference block by writing to the conference control register, setting cfen=1 and setting the tone duration. 4) set up tone insertion: program the tone by writing the tone coef?cient registers if a different programmable tone is required. refer to tone generation on page 42 identify the tone by writing cml location 6f with the dm address of the tone. write cmh location 6f with the gain setting for the insertion tone. conference party initialization 5) write cmh with the incoming gain, inversion bit and noise suppression for the ?rst party. 6) write cml with the incoming channel dm address for the ?rst party. 7) write the conference party control register with a conference id from 1-5. set the st bit for the ?rst party programmed for a conference. setting the st bit will remove any other parties that may have been previously programmed for that conference. set the insertion tone bit and outgoing conference gain control required. 8) write cml of the incoming channel with dm address of the conference output location (60-6e h ). write cmh of the location with the required outgoing gain if it has not been previously set. also set the oe bit. 9) repeat steps 5 to 8 for each additional party in the conference. 10) enable the conference over?ow interrupt if required, by setting cfe, bit 1 in the interrupt enable register (inte). read the confo register to ensure it is reset allowing the next over?ow to update the conference id value. 6.6 removing a channel from a conference setting the conference id number, in conference party control register, to 0 will disconnect the selected channel from the conference. once the selected channel is removed from the conference, the output enable
MT90812 advance information 22 (oe) bit of local (or expansion depending on the output stream number) connect memory high must be set to 0 in order to put the output driver of the corresponding stream into high-impedance state during the selected timeslot. 7.0 gain control gain control is provided on the outgoing channels, with a range of +3 to -27 db in steps of 1db, as well as - db. if a channel is in a conference, incoming and/or outgoing gain control is provided, with a range of 0 to -9 db in steps of 3 db. refer to section 6.0 for a description of gain control for a conference. outgoing gain control of +3 to -27 db in steps of 1db, as well as - db is also provided on the 32 tones from the tone generator. the gain for each outgoing channel is speci?ed in connect memory high. refer to connection memory high on page 50. there are ?ve bits g4-g0 which are used to set the gain. if 0 db value is selected then the gain control circuitry is bypassed. otherwise the 8 bit pcm value for the outgoing channel is read from data memory, expanded to a 14 bit linear pcm value, multiplied by the appropriate gain factor, and compressed to an 8 bit pcm value, before being output on the outgoing serial stream. the output of the tone generator and conference blocks are multiplied by the gain factor speci?ed for the tone or conference party and stored as a 14 bit linear pcm value in data memory. when the outgoing channel is connected to a tone or conference output location, the 14 bit linear pcm value is then read from dm, multiplied by the gain factor speci?ed for the outgoing channel, and compressed to an 8 bit pcm value, before being output on the outgoing serial stream. 8.0 delays through the MT90812 a delay results when transferring channel information from a MT90812 local input stream to an output stream. this delay varies according to the switch mode programmed in the cst bit of connect memory high; i.e. minimum or constant throughput delay mode. 8.1 minimum delay mode (cst bit=0) in minimum delay mode the delay is dependent on the combination of source and destination channels, the input and output streams and the data rate of the expansion bus. data transfers between streams operating at the same data rate (i.e. sti1 to sto1, est1 to est0, etc.) can be described as follows. channel information for a particular timeslot n from the input stream is sent to data memory in timeslot n+1 . channel information is queued for an output channel n in timeslot n-1 . thus, information entering the MT90812 from timeslot n , cannot be transmitted in the same timeslot n or timeslot n+1 , without a frame delay. information switched to a timeslot of m=n+2 or later will be switched within the same frame. the relationship that is required between incoming and outgoing timeslots are shown in table 6. for all but four cases, if the outgoing timeslot, m, is greater than or equal to n+2, the data is switched within the same frame. the throughput delay is m-n timeslots. there are four cases where there are data transfers between streams operating at different data rates. this occurs when the expansion bus is running at 4.096mb/s or 8.192 mb/s and the data is transferred between the expansion bus and either sto0 or sto1. the channel numbers range from 0 to 31 for a stream operating at 2.048mb/s, and from 0 to 63 and 0 to 127 for streams operating at 4.096mb/s and 8.192 mb/s, respectively.
advance information MT90812 23 table 6 - output channels for minimum delay the output channel number m, speci?ed for minimum delay in these four cases account for there being two 4.096mb/s channels and four 8.192mb/s channels for every one 2.048mb/s channel. table 7 lists the condition required for a throughput delay of less than one frame period, the throughput delay if this condition is met and the throughput delay expressed in timeslots when switching is made in the following frame. for cases where there are different data rates the delay is expressed in timeslots associated with the fastest data rate. i.e. with the source channel from est1 (@8mb/s) and destination channel on sto1 (@2mb/s) the delay is expressed in 8mb/s timeslots. if the incoming 8mb/s channel, n = 119, outgoing 2mb/s channel m =31, then the delay = 4m-n = 4(31)-119= ?ve 8mb/s timeslots. if m=1 the delay=128-(n-4m)=128-(119- 4)=thirteen 8mb/s timeslots. table 7 - throughput delay for minimum delay mode notes: t.s. = time-slot. t.s 2 . =2mb/s t.s. = 3.9 us. t.s 4 . =4mb/s t.s.=1.95 us. t.s 8 .=8mb/s t.s.=0.975 us. delays are measured in timeslots and at the point in time from when the input channel is completely shifted in and when the out put channel is completely shifted out. source and destination streams expansion bus data rate 2.048 mb/s 4.096 mb/s 8.192 mb/s sti0/1 -> sto0/1 m>=n+2 n/a n/a est0/1 -> est0/1 m>=n+2 sti0/1 -> est0/1 m>=2n+3 m>=4n+5 est0/1 -> sto0/1 m>=(n+3)/2 m>=(n+5)/4 source and destination streams input channel, n, range output channel, m, range condition for switching within same frame throughput delay within same frame throughput delay if condition not met sti0/1 -> sto0/1 0-31 0-31 m>=n+2 m-n t.s 2 . 32-(n-m) t.s 2 . est0/1 -> est0/1 2.048 mb/s 0-31 0-31 m>=n+2 m-n t.s 2 . 32-(n-m) t.s 2 . est0/1 -> est0/1 4.096 mb/s 0-64 0-64 m-n t.s 4 . 64-(n-m) t.s 4 . est0/1 -> est0/1 8.192 mb/s 0-127 0-127 m-n t.s 8 . 128-(n-m) t.s 8 . sti0/1 -> est0/1 2.048 mb/ s 0-31 0-31 m>=n+2 m-n t.s 2 . 32-(n-m) t.s 2 . sti0/1 -> est0/1 4.096 mb/ s 0-31 0-64 m>=2n+3 m-2n t.s 4 . 64-(2n-m) t.s 4 . sti0/1 -> est0/1 8.192 mb/ s 0-31 0-127 m>=4n+5 m-4n t.s 8 . 128-(4n-m) t.s 8 . est0/1 2.048 mb/s -> sti0/ 1 0-31 0-31 m>=n+2 m-n t.s 2 . 32-(n-m) t.s 2 . est0/1 4.096 mb/s -> sti0/1 0-64 0-31 m>=(n+3)/2 2m-n t.s 4 . 64-(n-2m) t.s 4 . est0/1 8.192 mb/s -> sti0/1 0-127 0-31 m>=(n+5)/4 4m-n t.s 8 . 128-(n-4m) t.s 8 .
MT90812 advance information 24 8.2 constant delay mode (cst bit=1) in constant delay mode, channel integrity is maintained by making use of a multiple data memory buffer technique. the input channels written in any of the buffers during frame n will be read out during frame n+2. table 8 lists the throughput delay for constant delay mode for all combinations of source and destination streams. table 8 - throughput delay for constant delay mode notes: t.s. = time-slot. t.s 2 . =2mb/s t.s. = 3.9 us. t.s 4 . =4mb/s t.s.=1.95 us. t.s 8 .=8mb/s t.s.=0.975 us. delays are measured in timeslots and at the point in time from when the input channel is completely shifted in and when the out put channel is completely shifted out. 8.3 delays in conferencing in a conference the data is read from data memory and transferred to the conference block as in constant delay mode, with a 2 frame delay. if the incoming data is in frame n, then within the ?rst half of frame n+2 the conference output is calculated and stored in the conference output locations in data memory. the conference output data is then switched to the outgoing data channel in minimum delay mode. the minimum delay possible in a conference is one frame + two 2mb/s-timeslots = 34 2mb/s-timeslots. the maximum delay possible is approximately 2 frames + 1.5 frames + two 2mb/s-timeslots = 82 2mb/s-timeslots. 9.0 timing and clock control the MT90812 clock control circuitry selects one of ?ve possible input clock and frame pulse references. the input clock can be either 4.092, 8.192, or 16.384 mhz as described in section 9.1, input timing reference. fig. 16 shows the clock control functional diagram. the clock control circuitry provides an internal master clock of 8.192 mhz, generates 2.048, 4.096, 8.192, and 10.24 mhz output clocks, f4 and f8 frame pulse signals, as well as the serial interface timing for sti/o0, sti/o1 and est0/1 serial streams. these signals are either generated directly from the input clock source or from an on-chip analog pll. the on-chip analog pll may be used to generate 2.048, 4.096, 8.192, and 10.24 mhz clocks. the pll operates in master and slave modes. master mode provides more jitter attenuation while slave mode minimizes phase delay. the pll can provide the required 4.096 and 10.24 mhz clocks (c4 and c10) to be supplied to the mt9171/72 dnic devices. the c4 and c10 clocks meet the requirement that they be frequency locked and maintain a jitter of less than or equal to 15ns with respect to each other, while maintaining at least 40/60 duty cycle for c10o. refer to section 9.3.1, master and slave pll modes. source and destination streams input channel, n, range output channel, m, range throughput delay sti0/1 -> sto0/1 0-31 0-31 2x32-(n-m) t.s 2 . est0/1 -> est0/1 2.048 mb/s 0-31 0-31 2x32-(n-m) t.s 2 . est0/1 -> est0/1 4.096 mb/s 0-64 0-64 2x64-(n-m) t.s 4 . est0/1 -> est0/1 8.192 mb/s 0-127 0-127 2x128-(n-m) t.s 8 . sti0/1 -> est0/1 2.048 mb/s 0-31 0-31 2x32-(n-m) t.s 2 . sti0/1 -> est0/1 4.096 mb/s 0-31 0-64 2x64-(2n-m) t.s 4 . sti0/1 -> est0/1 8.192 mb/s 0-31 0-127 2x128-(4n-m) t.s 8 . est0/1 2.048 mb/s -> sti0/1 0-31 0-31 2x32-(n-m) t.s 2 . est0/1 4.096 mb/s -> sti0/1 0-64 0-31 2x64-(n-2m) t.s 4 . est0/1 8.192 mb/s -> sti0/1 0-127 0-31 2x128-(n-4m) t.s 8 .
advance information MT90812 25 multiple idx systems are supported by allowing the idx to either drive or receive an 8.192 mhz clock. the master idx in the system may supply c8 while the slave idx derive their timing from the master. in a multi- chassis application a slave idx may be required to generate its own c10. c8 is distributed between master and slave idx devices and the pll is then used to phase lock c10, c4 and c2 to the c8 input. the MT90812 expansion bus can be supported with either an 8.192 mhz or 16.384 mhz clock when operating at 8.192 mb/s. when the input clock source is selected as 16.384 mhz either hmvip and non-hmvip mode may be used. in a multiple idx system the slave idx devices are supplied c8 from a master idx. a watchdog timer on the idx allows the slave idx to monitor the c8 input. in the event of the loss of the c8 clock the slave idx can be switched to be master idx and supply c8 to the system. this provides redundancy for the clock source allowing idx operation to remain independent of the other idx devices if necessary. figure 16 - clock control functional diagram *non-mux mode addresses for tcr, occr, inte and ints are 382 h , 383 h , 384 h and 385 h , respectively. **c8 and f8 are bi-directional pads. they are used as inputs in c8 timing mode, otherwise as output pads. 9.1 input timing reference the input timing reference is selected setting cr1-0 and hmvip bits in the timing control register (tc) described on page 55. one of ?ve possible clock and frame pulse references, c4/f4, c8/f8, c8p, or c16/f8, c16/hmvip, can be selected, as listed in table 9. table 9 - clock modes the MT90812 requires at least an 8.192 mhz clock internally. when the c4 input clock is selected the 8.192 mhz clock is derived from the pll. c4 is not a valid clock reference when the pll is disabled. cr1,0 hmvip clock reference frame pulse reference 00 x c4 fpi(f4) 01 x c8 f8i 10 x c8p (default) no frame pulse 11 0 c16 non-hmvip fpi (f8) 11 1 c16 hmvip fpi(f4) c8p_c16 pll c8 internal clk wd c8f external 8.192 mhz crystal c8** f8** c4o wde,fpo,cr1-0,hmvip,pe,pms,pcs *02 h timing control register pcos,-,c10e,c8e,f8e,c4e,f4e,c2e *03 h output clock control register c8 c8fe *04 h interrupt enable register c8f *05 h interrupt status register timing interface c10o f4o c2o f8** fpi c8** c4i osco st_clk est_clk_in est_clk_out
MT90812 advance information 26 the MT90812 defaults to c8p input clock reference when reset. when c8p is selected as the input clock reference the clock oscillator pins c8p_c16 and osc can be used with an external 8.192 mhz crystal or pin c8p_c16 can be used directly as a clock input with osc left unconnected. refer to section 9.5, c8p pin timing source. when c8p is selected, no frame pulse is used and the MT90812 generates f4o and f8o. f4o and f8o can be disabled by setting f4e and f8e bits low in the output clocking control register (occ). with c16 as the clock reference the hmvip frame alignment interface can be selected with the hmvip bit in the timing control register (tc). 9.2 serial data interface timing st-bus, gci or hmvip serial data interface timing modes are supported on the serial streams of the MT90812. the two local streams, sti/o0, sti/o1 operate at 2.048 mb/s. the expansion bus can operate in two modes, tdm link and idx link, as described in section 3.0. in tdm link, est0/1 can operate at 2, 4 or 8 mb/s. in idx link, est0/1 operates at 8 mb/s. the incoming 8khz frame pulse used for frame synchronization for both local and expansion bus streams can be either st-bus, gci or hmvip format. in all timing modes except c16-hmvip mode, the MT90812 automatically detects the presence of an input frame pulse on f8 or fpi pins and identi?es it as either st-bus or gci. for c16-hmvip mode, the frame pulse must be in st-bus format. 9.2.1 local streams, sti/o0 and sti/o1 for sti/o0, sti/o1 2.048 mb/s streams a 4.096 mhz clock is used for the serial interface timing and is generated in the clock control block. the pcs bit in the tc register selects the source of this clock as either derived from the input clock or from the pll and is described below. in st-bus format, every second edge of the 4.096 mhz clock marks the bit boundary and the data is clocked in on the rising edge the 4.096 mhz clock, three quarters of the way into the bit cell, see figure 40 on page 86. in gci format, every second rising edge of the 4.096 mhz clock marks the bit boundary and data is clocked in on the falling edge of the 4.096 mhz clock at three quarters of the way into the bit cell, see figure 41 on page 87. 9.2.2 expansion bus, est0/1 the expansion bus can run at 2, 4 or 8mb/s. in tdmlink, bits ep0 and ep1 of the tc register de?ne the data rate. in idx link the date rate is always 8mb/s. in both modes the timing is similar to that used for st0/1 streams when double rate clock is used. for example at 2, 4 and 8 mb/s rates, clk is 4.096, 8.192 or 16.384 mhz, respectively. refer to figure 40 on page 86 for st-bus timing and figure 41 on page 87 for gci. when c8 timing mode is selected, the incoming data, on the expansion bus running at 8mb/s, is clocked at either the 1/2 bit time or 3/4 bit time. fig. 42 and fig. 43 shows the expansion bus timing with the data clocked at the 1/2 bit time for st-bus and gci, respectively. without the presence of a 16.384 mhz input clock reference, the pll must be used to clock data at the 3/4 bit time. the pll must be enabled and the pcs bit set to 1. for further description see section 9.2.5. 9.2.3 hmvip frame alignment when c16 timing mode is selected, the hmvip bit in the timing control register (tc) enables the hmvip frame alignment interface. the c8p_c16i input must be at 16.384 mhz, c4i must be 4.096mhz. c4i is used to sample the 8khz st-bus frame pulse. the timing relationship between the two clocks and the frame pulse is de?ned in figure 44 on page 88. in c16 non-hmvip mode frame synchronization is made using f8. c16/f8 timing is shown in figure 37 on page 84 for st-bus and figure 38 on page 84 for gci. 9.2.4 output clock and frame pulse signals the MT90812 generates c2o, f4o, c4o, f8, c8, and c10o signals. these outputs are enabled by setting the corresponding bits in the output clocking control register (occ) described on page 56. f8 and c8 signals are bi-directional pins. when the c8/f8 input clock reference mode is selected they are used as inputs and the c8e and f8e output enables of the occ register are ignored.
advance information MT90812 27 the MT90812 generates c4o, f4o, c8, and f8 signals in either st-bus or gci formats as selected by the fpo bit in the timing control register (tc). this selection is independent of the incoming frame synchronization used. 9.2.5 selecting timing from the input clock reference or from the pll as mention above, the clocks used for the local and expansion bus streams can be derived directly from the input clock reference or from the pll. two bits are used to control this, the pll clock select ( pcs ) bit in the tc register and the pll clock output select ( pcos ) bit in the output clocking control register (occ). referring to figure 16, clock control functional diagram, on page 25, the clock used for the sti/o0 and sti/ o1 serial streams, is stclk. the clocks used for clocking in data and clocking out data on the expansion bus are est_clk_in and est_clk_out, respectively. with pcs set high, est_clk_in, stclk and c4o are generated from the pll. otherwise they are derived from the input clock. with the pcs set high, c4o and c10 are both supplied from the pll for clock supply to the mt9171/72/73 dnic devices. the other advantage with pcs high allows for clocking in data at three quarters into the bit cell on the 8mb/s expansion bus. with pcos set high, est_clk_out and c8 (output) are generated from the pll. otherwise, they are derived from the input clock. for example, with pcos=0 in c8p timing mode, c8p is used to generate c8 (output) and in c16 timing modes, c8 (output) is c16 divided by 2. when c4f4 is used as the timing reference, est0/1, 4 and 8 mb/s timing, as well as f8o and c8o, are generated from the pll independent of the pcs and pcos settings. 9.3 phase lock loop (pll) as shown in fig. 17, the pll of the MT90812 consists of a phase and frequency detector, loop filter, voltage controlled oscillator and divider circuit. the phase and frequency detector compares the reference signal selected by cr0-1 bits in the tc register, with the feedback signal from the divider circuit, and provides an error signal corresponding to the phase difference between the two. this error signal is passed to the loop filter. the loop filter is a second order low pass ?lter which operates in two modes, master and slave. these modes are described in section 9.3.1. the voltage controlled oscillator receives the ?ltered signal from the loop filter and based on its value, generates a corresponding digital output signal. the divider circuit uses the vco output to generate ?ve clock outputs, c10, c8, c8_75, c4 and c4_75. c10, c8 and c4 are 10.29 mhz, 8.192 mhz and 4.096 mhz signals respectively. c8_75 and c4_75 are 8.192 mhz and 4.096 mhz signals with a 75/25% duty cycle. these two clock signals are used in the serial interface circuit for the expansion bus. refer to section 9.2. the pll is enabled with pe bit set in the timing control register (tc). with the pll off, c10 is disabled and c4 as an input clock reference is not valid. figure 17 - pll block diagram clock reference phase & frequency detector charge pump loop filter v to i current controlled oscillator divider c8 c8_75 c4 c4_75 c10 wde , fpo , cr1-0 , hmvip , pe , pms , pcs *02 h timing control register
MT90812 advance information 28 9.3.1 master and slave pll modes the pll master/slave (pms) bit in the tc register selects the pll mode. in master mode the pll loop ?lter is selected to minimize the magnitude of any one clock correction. this preserves the c10o 40/60 duty cycle given the input jitter as speci?ed on page 95. for example this provides the 10.24 mhz clock required for dnic operation with up to 32ns of clock correction on the input clock once per 125us frame. refer to intrinsic jitter for master and slave modes on page 93 and typical input to output jitter transfer for master mode on page 93. in slave mode the pll loop ?lter is selected to minimize the phase delay on the output clock with respect to the input clock reference. the typical input to output jitter transfer for slave mode is shown on page 94. in slave mode the loop filter ensures a phase difference of less than 15 ns. this is assuming an input clock reference from the master idx, where the master idx has up to 32 ns of clock correction on the input clock once per 125 us. in an application where the clock reference does not require jitter attenuation the pll can be used in slave mode. for example in a multi-idx application the master idx could have its pll in master mode, and generate clocks for the other idx devices. setting the plls of these slave idx devices in slave mode lessens phase delay while taking advantage of the clock source from the master idx. note: the master idx pll maybe placed in slave mode as well if jitter attenuation is not required. 9.4 watchdog timer a watchdog timer monitors c8 input. this requires the presence of a 8.192 mhz clock at c8p_c16. in the event of the loss of the c8 clock an interrupt is generated and the c8f bit in the interrupt status register (ints) is set. the system can service the interrupt and maintain operation of the MT90812 by switching clock input reference from c8 (cr0-1=01) to c8p (cr0-1=10) and enable the MT90812 to supply c8 output. figure 18 - watchdog con?guration this provides redundancy for the clock source in a multiple idx system. the watchdog timer is enabled by setting wde bit in timing control register (tc).the interrupt is enabled by setting c8fe bit in interrupt enable register (inte). sti0 sti1 sto0 sto1 idx a c4i f4i c8 f8 f4o c4o c10o c8p osco est0 est1 idx b c8p osco crystal 8.192 mhz crystal 8.192 mhz sti0 sti1 sto0 sto1 c4i f4i c8 f8 f4o c4o c10o est0 est1
advance information MT90812 29 9.5 c8p pin timing source the MT90812 can use either a clock or crystal, connecting to pins c8p_c16i and osco, as a reference timing source. 9.5.1 clock oscillator fig. 19 shows a 8.192mhz clock oscillator, with 32 ppm tolerance, directly connected to c8p_c16i pin of the MT90812. the output clock should be connected directly (not ac coupled) to the c8p_c16i pin, and the osco output should be left open. figure 19 - clock oscillator circuit 9.5.2 crystal oscillator alternatively, a crystal oscillator may be used. a complete oscillator circuit made up of a crystal, resistor and capacitors is shown in fig. 20. the accuracy of a crystal oscillator depends on the crystal tolerance as well as the load capacitance tolerance. typically, for a 8.192mhz crystal speci?ed with a 32pf load capacitance, each 1pf change in load capacitance contributes approximately 9ppm to the frequency deviation. consequently, capacitor tolerances, and stray capacitances have a major effect on the accuracy of the oscillator frequency. the trimmer capacitor shown in fig. 20 may be used to compensate for capacitive effects. if accuracy is not a concern, then the trimmer may be removed, the 39pf capacitor may be increased to 56pf, and a wider tolerance crystal may be substituted. the crystal should be a fundamental mode type. the fundamental mode crystal permits a simpler oscillator circuit with no additional ?lter components and is less likely to generate spurious responses. the crystal speci?cation is as follows: frequency: 8.192mhz tolerance: as required oscillation mode: fundamental resonance mode: parallel load capacitance: 20pf maximum series resistance: 35 w approximate drive level: 1mw e.g. cts r1b23b32-8.192mhz (20ppm absolute 6 ppm 0c to 50c, 32pf, 25 +5v 8.192mhz gnd 0.1uf +5v c8p_c16i no connection osco MT90812
MT90812 advance information 30 figure 20 - crystal oscillator circuit 10.0 d-channel signalling support the MT90812 can support communications over the d-channel in one of the following methods: ? basic receive transmit method ? shared hdlc resource method the ?rst method is supported by the d-channel basic receive transmit (dbrt) block. the dbrt supports the communication over the d-channel with the use of start and stop signalling and buffering of messages for both receive and transmit directions. the second method supports the use of the mt8952 hdlc protocol controller for communication over the d- channel. the hdlc resource allocator (hra) block in the MT90812 provides an interface to the mt8952 hdlc protocol controller. refer to the msan-122 note for a description of how voice/data channels and signalling information channels on a digital communications link are supported. msan-178 note provides a programming example for the hra. each of the blocks are described in the following sections. 11.0 d-channel basic receive transmit block the MT90812 can support communications over the d-channel with the use of the d-channel basic receiver/ transmitter (dbrt). the d-channel basic receiver and transmitter are used to transfer data between a channel on a serial stream and the parallel micro-port with the use of two 32 byte fifos. there are two modes which the receiver or transmitter may be placed in: message length interrupt mode (mlim) and fifo level interrupt mode (flim). mli is suited to smaller messages, where ef?cient use of bandwidth is important and interrupts are generated on the completion of a message. fli mode is suited to the transfer of large amount of data, where due to the size of the message, start, parity and stop bits are required more often and interrupt generation can support a large data transfer through the fifo. the bit formatting for these two modes are summarized in table 10 and fig. 21. in mli mode a start bit is sent, followed by the message bits and parity (if enabled) and stop bit. in fli mode start, parity and stop bits are added to every 8 bits. it is also possible to send unframed data in fli mode. the bit rate can be set to 1, 2, or 8 bits per frame for any mode. osco 56pf 1m w 39pf 3-50pf 8mhz MT90812 c8p_c16i 100 w 1uh 1uh inductor: may improve stability and is optional
advance information MT90812 31 table 10 - dbrt modes of operation *x = dont care figure 21 - dbrt modes 11.1 receiver operation the receiver transfers incoming data for a speci?ed channel, identi?ed in cm location 70 h , to the rx fifo. the system read of d-channel rx fifo output (drxout) register at 43 h accesses the next data byte in the rx fifo buffer. the following diagram illustrates the data ?ow for the d-channel data. the rx control register d-channel rx fifo control bits (drxc) at 41 h is used to specify the receiver bit order, data rate at 1, 2 or 8 bits per frame, message length or fifo level interrupt mode, select start and stop bits, enable parity, and activate the receiver. refer to page 66 for more description. the receiver bit order de?nes whether the ?rst bit received on the tdm channel is the lsb or msb read on the microport data bus (d0 or d7, respectively). in mli mode, when the data is transferred to the rx fifo the start and stop bits are automatically stripped off. the start and stop enable (se) bit in drxc register is not used and the received message (1 to 256 bits) is always assumed to be framed by the start and stop bits. the received data is only transferred to the rx fifo following the reception of the start bit (the ?rst 0). the status of the parity enable (pe) bit in the drxc register mode name interrupt mode mode bit (m) start-stop bits parity bit bit rate message oriented mlim 1 x 0 1, 2, or 8 bits/frame message oriented with parity mlim 1 x 1 1, 2, or 8 bits/frame unframed flim 0 0 x 1, 2, or 8 bits/frame byte oriented flim 0 1 0 1, 2, or 8 bits/frame byte oriented with parity flim 0 1 1 1, 2, or 8 bits/frame st stp st stp b7 b6 b5 b4 b3 b2 b1 b0 b7 b6 b5 b4 b3 b2 b1 b0 b7 b6 b5 b4 b3 b2 b1 b0 b7 b6 b5 b4 b3 b2 b1 b0 st p st stp stp st stp b7 b6 b5 b4 b3 b2 b1 b0 p b7 b6 b5 b4 b3 b2 b1 b0 st stp p mlim* mlim* flim* flim* flim* message oriented message oriented with parity unframed data byte oriented byte oriented with parity * the bit order can be reversed using rxbo and txbo in drxc register
MT90812 advance information 32 will specify whether the received data will have a parity bit and consequently the receiver will perform a parity check on the received data. in fli mode, the start and stop bits and the parity bit can be enabled or disabled with the se and pe bits in the drxc control register, respectively. with the start and stop bits enabled, the start of the message is identi?ed by the ?rst 0 received after the dbr is enabled. when the start and stop bits are disabled, no parity check will be performed (regardless of the status of pe bit) and the data will be transferred from the incoming tdm stream to the rx fifo following the rx being enabled. figure 22 - data flow for d-channel receiver 11.1.1 receiver interrupt handling there are four interrupts associated with the d-channel receiver. they are listed in table 11. table 11 - d-channel receive interrupts the dree and drxe bits in the interrupt enable register (inte) on page 57 enable/disable the above interrupts. the main difference in mli and fli modes is in determining when the interrupt occurs. in mli mode the interrupt occurs when the full message has been received. in fli mode the interrupt occurs when the number of bytes in the fifo equals the trigger level. the d-channel receive interrupt threshold (drxit) register is used to program when an interrupt occurs for either mli or fli mode. in the latter mode, the interrupt is to indicate that the fifo level is attained and not necessarily the end of the message. in the former mode, the interrupt solely indicates the end of the message. as listed in table 11 an interrupt is also triggered when one of the following error conditions occurs: ? the rx fifo is full and the next byte of data has been received and is to be transferred to the fifo then the overrun status bit is set and an interrupt occurs (rx overrun error). interrupts register reference page description drx ints page 56 d-channel receive message length or fifo level interrupt dre ints page 56 d-channel receive fifo error. status of error in dchs register. oe drxs page 68 receive overrun error pe drxs page 68 receive parity error se drxs page 68 receive stop bit error channel m data memory rx fifo channel m dchin cm 70h dsti drxout = 43 hex control registers uport read serial to parallel rx dbrx
advance information MT90812 33 ? the stop bit was not detected (rx stop bit error). ? the status of the received parity bit did not equate to the calculated parity (rx parity bit error). 12.0 transmitter operation fig. 23 illustrates the data ?ow for the d-channel data in the transmit direction. a system write to the tx fifo buffer is performed by addressing the d-channel tx fifo input (dtxin) register at location 44 hex of the control register page. up to 32 bytes can be written to the fifo. the output of the dbtx is memory mapped to data memory location 70 hex . the output of the transmitter can then be directed to the speci?ed output channel by programming connect memory low as 70 hex for the intended output channel. as with the receiver, the transmitter also operates in two modes: mlim and flim. the tx control register d- channel tx control (dtxc) at location 45 hex of the control register page is used to program the transmitter for interrupt select, transmission rate at 1, 2, or 8 bits per frame, mli or fli mode, start and stop bit enable (se), parity enable (pe), and start transmission (st). the transmission sequence starts when bit st is set followed by writing the ?rst byte to the fifo. the transmission bit order is determined from txb0 bit in the d- channel receive interrupt threshold (drxit) register. refer to page 66 for a description on the transmission bit order. in mli mode, the transmitter automatically appends the start and stop bits to an n byte message with an optional parity bit. the status of se bit has no effect on this mode but the pe bit speci?es whether the message to be transmitted requires a parity bit. in fli mode, the status of se and pe bits can provide three transmission methods. if the se bit is disabled, the transmitter does not include start, parity and stop bits to the message regardless of the status of pe bit. if se is enabled, the user has the option of enabling or disabling the pe bit. the start, parity, and stop bits are applied on a per byte basis. the transmitted messages are always padded with stop bits for any remaining bits. for example in the case where a byte long message is to be sent with start and stop bits and a data rate at 8 bits per frame, the transmission of the message will take two frames and the stop bit will be in the second bit of the second frame and the 3rd to 8th bits will be padded with stop bits. the transmitter operation also provides a message to be broadcast to several channels simultaneously. this is accomplished by programming the connect memory of the intended outgoing channels all to 70 hex . one application for broadcasting messages might be to update displays on all phone sets. figure 23 - data flow for d-channel transmitter dtxin data memory tx fifo uport write connect memory 70 hex control registers parallel to serial output stream channel m 70 hex 44 hex tx dbtx
MT90812 advance information 34 12.1 transmitter interrupt handling in either mli or fli modes interrupts are generated on tx fifo empty or 3/4 empty, or end of transmission. the tx fifo interrupt select (is) and interrupt level (il) bits in the dtxc register specify the condition for an interrupt to occur. if the is bit is set high, then the interrupt occurs when transmission is complete. table 12 - number of available frames to continue the message following tx fifo empty interrupt if the is bit is zero, then an interrupt occurs when either the tx fifo is empty or 3/4 empty depending on the status of the il bit. this type of interrupt can be used to continue the message. therefore, to continue a message after the tx fifo empty interrupt occurs, the user can write to the tx fifo within the number of frames as shown in table 12. the d-channel tx enable interrupt (dtxe) bit in the interrupt enable register (inte) on page 57 enables or disables the transmitter interrupts. 13.0 hdlc resource allocator module the hdlc resource allocator (hra) block in the MT90812 provides an interface to the mt8952 hdlc protocol controller. this interface supports the sharing of the hdlc resource across several mt9171/72 dnic devices for communication over the d-channel. the msan-122 application note describes how voice/data channels and signalling information channels on a digital communications link are supported. refer to the msan-122 note for a general description of: ? mt8952 hdlc protocol controller ? mt9171/72 digital network interface circuit ? shared hdlc resource method the hra block is described in the following sections. ? general description of MT90812 and shared hdlc con?guration ? connection to mt8952 hdlc controller and mt9171/72 dnic ? connection to mt8952b hdlc controller ? connection to mt9171/72 dnic ? data stream flow ? tx control ? generation of txcen ? end of the transmission of a packet ? tx and rx handshaking ? merging of d and c-channels ? rx control ? generation of rxcen mode name interrupt mode bit rate 1 bit/frame 2 bits/frame 8 bits/frame message oriented mlim 8 frames 4 frames 1 frame message oriented with parity mlim 8 frames 4 frames 1 frame unframed flim 8 frames 4 frames 1 frame byte oriented flim 10 frames 5 frames 1 frame byte oriented with parity flim 11 frames 6 frames 1 frame
advance information MT90812 35 ? dedicated receive mode ? multiplexed receive mode ? cts generation ? receive packet termination ? rx channel auto-hunt ? auto-hunt monitoring ? circumstances when monitoring a channel is stopped refer to the hdlc control and status registers starting on page 70. refer to msan-178 note for a programming example for the hra. 13.1 general description of MT90812 and shared hdlc con?guration the hra provides the capability to multiplex the hdlc controller over a maximum of 16 channels of the sto1 link. signalling information can be passed to and from the far end through the 8, 16, or 64 kb/s d-channels. the microprocessor uses the mt8952 to send and receive the signalling information in the hdlc protocol. to send information, the microprocessor writes to the hdlc transmit buffer. the message will then be sent out 1,2, or 8 bits per channel to the far end. messages from the peripherals can be received through the rx and the auto- hunt blocks of the hra. the hra is used to specify which channel, and hence which of the dnics, the message is intended. the hra allows the transmit and receive sections of the mt8952 to act independently. when the microprocessor is transmitting to one peripheral it may receive from another simultaneously. to specify the receive and transmit channels the channel number is written to the drx4-1 bits in hra ctrl register 2 (hc2) and ntx4-1 bits in hra ctrl register 3 (hc3). the channel number for the rx circuit may also be supplied from the auto-hunt circuit of the hra. the auto-hunt circuit supplies the channel number when the rx circuit is in multiplexed mode. when in dedicated mode the channel number is supplied by a system write to the drx4-1 bits in hra ctrl register 2 (hc2). when in multiplexed mode the hra supports polling of the peripherals. in this mode, sharing of the single hdlc amongst several peripherals (i.e. mt9171/72 dnic devices) is simpli?ed through the use of the auto- hunt circuit. when it is necessary for a peripheral to send a message to the central processor, the peripheral continually sends a request-to-send (rts) message. the auto-hunt block in the hra monitors each incoming channel in turn and checks for a rts message. upon detection of a rts from the peripheral the hra latches the channel number to be used as the next receive channel by the rx circuit of the hra. the auto-hunt circuit functions independently of the tx and rx circuits, thereby reducing the workload demanded of the mt8952 hdlc controller. there is some handshaking that is required between the tx, rx and auto-hunt blocks and will be described further in the next sections. 13.2 connection to mt8952 hdlc controller and mt9171/72b dnic fig. 24 shows a typical application of the MT90812 connected to a mt8952 hdlc protocol controller and several mt9171/72 dnics. placing the mt8952b in the external timing mode and the dnics in the dual-port digital-network mode allows for easy interface through the hra block. 13.2.1 connection to mt8952b hdlc controller the multiplex timing for the hdlc-channel is performed by the hra block. the mt8952b transmit and receive sections, in external timing mode, are independently controlled by the hardware pins txcen and rxcen, respectively. to assist in multiplexing, the mt8952b outputs teop and reop indicate when it has ?nished transmitting or receiving a packet. the txcen and rxcen clock enable strobes are generated by the hra during the speci?ed-channel and bit times to allow the hdlc controller to transmit to, and receive from, any channel.
MT90812 advance information 36 the MT90812 output c2o is a 2.048 mhz clock provided for the mt8952 hdlc controller bit rate clock input. 13.2.2 connection to mt9171/72b dnic the dnic, as mentioned earlier, is used in dual-port mode. the b1 and b2 channels are input/output at dnic port dsti/dsto in timeslots 0 (b1) and 16 (b2) relative to f0i. the d and c information is input/output at port cdsti/cdsto on channels 0 (d) and 16 (c). the dnics are 'daisy-chained' together using the delayed frame pulse output f0o. in dual-port mode the signal f0o comes at the end of channel 0. supplying it to the next dnic in the chain skews its active channels by one channel. because two tdm links are used to support these 'daisy-chained' dnics, up to 16 line circuits may be served by one hdlc protocol controller with this con?guration. 13.2.3 data stream flow fig. 24 shows ?ve numbered streams which connect a MT90812 to the dnics and the mt8952 hdlc protocol controller. they are: 1 . the sto0 stream from the MT90812 to the dnics dsti containing b channels. 2 . the sto1 stream from the MT90812 to the dnics cdsti containing d/c-channels. 3 . the sti0 stream of the MT90812 from the dnics dsto containing b channels. 4 . the dnics cdsto containing d/c-channels to the MT90812 sti1 and the mt8952b cdsti. 5 . the mt8952b cdsto stream to the MT90812 dper containing formatted d-channel data. streams 1 and 3 contain only b channel information. stream 1 originates from the sto0 of the MT90812 and is input to dsti of the dnic devices. stream 3 is the opposite direction of stream 1, transferring b-channel information from dsto of the dnics to sti0 of the MT90812. streams 2, 4 and 5 are used to pass d- and c-channel information from the microprocessor to the dnics through the MT90812 and hdlc protocol controller. d-channel information starting at the microprocessor is written to the mt8952 transmit buffer. it is formatted and sent out stream 5. stream 5 connects cdsto of the mt8952b to the dper input of the MT90812. in the MT90812 the d-channel and c-channel information are merged to form stream 2 . the c-channel information is written by the microprocessor to the connect memory of the MT90812. the c-channel information from connect memory is sent out sto1 of the MT90812 along with the d-channel information from dper. stream 2 connects sto1 from the MT90812 to cdsti of the dnics. the merging of the d- and c- channels is described in section 13.3.4. the d- and c-channel information is transferred from the dnics cdsto on stream 4 , to both the MT90812 and the hdlc. the c-channel information is transferred to the data memory of the MT90812, where it may be read directly by the microprocessor. the d-channel information is transferred to the microprocessor through the mt8952 hdlc protocol controller. the mt8952 stores the d-channel information in the 19-byte buffer to be read by the microprocessor.
advance information MT90812 37 figure 24 - typical application using the hdlc resource allocator 13.3 tx control the tx circuit performs the following functions: ? generate txcen to enable the hdlc transmitter. ? handle transmit packet termination. ? allow rx and tx handshaking ? merge d-channel data from the hdlc controller into the local tdm stream. these functions will be described in the sections below. 13.3.1 generation of txcen the hdlc transmitter is controlled by the MT90812 generated transmit clock enable signal, txcen. the txcen output signal enables the hdlc transmitter in the appropriate channel as speci?ed by the system, via c4i f0i lout lin cdsti dsti c10i f0o dnic interface t r dsto cdsto sti0 sto1 dper reop rxcen teop txcen idx mt8952b micro d0 d7 a0 a9 e r/ w mrdy irq a0 a3 d0 d7 cs r/ w e irq txcen teop rxcen reop cdsto cdsti f0i c2i processor c d sti1 sto0 c10o a0-a9 d0-d7 cs r/ w irq c4i f0i lout lin cdsti dsti c10i f0o dnic interface t r dsto cdsto c4i f0i lout lin cdsti dsti c10i f0o dnic interface t r dsto cdsto 3 1 2 4 5 mclko mclki crystal clock 20.48 mhz c2o c4o f4o
MT90812 advance information 38 the MT90812 microport. txcen output signal is enabled for one to eight bits per channel per frame, depending upon the selected baud rate. the desired active channel is selected by the system via a write to the MT90812 devices next transmit channel (ntx) bits de?ned in hra ctrl register 3 (hc3). a write to this register will start txcen to be enabled for the channel speci?ed as soon as the transmitter becomes inactive. the ntx bits are double buffered which allows the system to specify the next transmit packet at any time, without concern for whether the present transmit packet is ?nished. at the start of the transmission of a packet the following three actions are taken: ? the ntx bits are latched into the ptx (present transmit channel) bits in hra status 4 (hs4) register. ? status bit txchnl is set, ? a transmitter-active ?ag (txact) is set latching the system's ntx request into the ptx (present transmit channel) register rede?nes the active transmit channel time. when the status bit txchnl is set, this informs the system that its transmit channel request has been satis?ed. when the transmitter-active ?ag (txact) is set, this allows txcen to be enabled during the appropriate bit times. 13.3.2 end of the transmission of a packet a transmit packet will be terminated by a transmit end-of-packet (teop) strobe from the hdlc controller chip or by setting the steop ?ag through the hra ctrl register 2 (hc2). the hdlc controller chip asserts teop for one bit period during the last bit of the closing ?ag of the transmit packet. the system may, at any time, raise its own steop ?ag through the hra ctrl register 2 (hc2). when the steop ?ag has been read by the system, it will be automatically cleared. in either case, the end-of-packet signal will cause the transmitter to go inactive, thus allowing another packet transmission to be initiated (if desired). whenever txact is low, the hdlc controller's transmit clock enable txcen is disabled, and idles are transmitted for the unused remainder (if any) of the d-channel time. 13.3.3 tx and rx handshaking transmit and receive functions are generally independent of each other. but there is a coupling of the two functions caused by the requirement for a go-ahead handshake. this handshake must be sent by the MT90812 when a peripheral requests to transmit to the system. while the MT90812 is generating and transmitting this go-ahead on the channel reserved for the speci?c peripheral, it pre-empts the system's d-channel transmit time. this interference occurs for that particular peripheral only, for the duration of the clear-to-send (cts) nine-bit go-ahead pattern. if the system transmits a packet to a peripheral which is also currently being sent a go-ahead this would result in the loss of data from the system transmit packet. to prevent data loss a number of approaches can be taken. after specifying the ntx channel, and waiting until txchnl goes high, the system can read the present receive channel (prx). if prx is the same as ntx, then the system may either: ? send the packet anyway and retransmit on request, ? assert a software transmit end-of-packet (steop) to terminate the system's request for that particular transmit channel, ? wait for a maximum of ?ve frames (at a 16k baud rate) for the cts to complete, ? or monitor ctsact until it goes low. in dedicated receive mode, where the generation and transmission of cts by the MT90812 is inhibited, no possibility of contention exists between the system and the MT90812. so such restrictions need not apply.
advance information MT90812 39 13.3.4 merging of d and c-channels. the hra block multiplexes the d-channel, originating at the hdlc protocol controller, and the c-channels into a common output stream. c-channel and d-channel information destined for the line circuit are also fed through a multiplexer controlled by txcen and a half-frame count. this does the merging of the outgoing d- and c-channel information and produces a high (all-ones) pattern on the unused d-channels. figure 25 - composite st-bus frame for hra application fig. 25 shows the composition of the st-bus in the transmit direction. the streams are also labelled according to the numbers assigned as in fig. 24. the b1 and b2 channels for up to 16 line circuits arrive at sti0 stream #3 and depart at sto0 stream #1 . all dnic line circuits are connected in parallel to this bus. the channel assignments and line destination addresses are shown in fig. 24. the c-channel information, which have been written to the MT90812 connect memory, are assigned to the ?rst 16 channels of sto1, stream #2 . the d-channel from the mt8952b is input to the hra block of the MT90812 at dper, stream #5 , during the times enabled by txcen. the c and d-channels are combined to produce sto1, stream #2 , which is then routed to the dnics. the composition of cdsto is shown, in fig. 24, with the d-channel enabled for line circuit 3 (channel 18). in the opposite direction, c-channel information from the line circuits arrives at sti1, stream #4 , in the ?rst 16 channels. the mt8952b, enabled by rxcen, receives the active incoming d-channel directly from the same bus. 14.0 rx control 14.1 rx circuit functions the rx circuit performs the following functions: ? generate rxcen to enable the hdlc receiver ? operate in dedicated mode or multiplexed modes ? handle receive packet termination ? generate the cts pattern to be transmitted 0123456 15 16 17 18 19 20 21 22 23 24 25 31 channel f0i f0o csti dper txcen sto1 sti/o0 c0 c1 c2 c3 c4 c5 c6 c7 c8 c9 c10 c15 d d 0 d 1 c0 c1 c2 c3 c4 c5 c6 c7 c8 c9 c10 c15 d b2 b2 b2 b2 b2 b2 b2 b2 b2 b2 b2 b2 b1 b1 b1 b1 b1 b1 b1 b1 1234567891011 1234567 16 #3/#1 #2 #5 16 26
MT90812 advance information 40 14.1.1 generation of rxcen the rx circuit performs two functions. as with the tx circuit, it must generate the proper hdlc receive clock enable signal ( rxcen). this signal has the same characteristics as txcen. the rxcen signal enables the hdlc receiver in the appropriate channel as speci?ed by the system, via the MT90812 microport. rxcen is enabled for one to eight bits per channel per frame, depending upon the selected baud rate, speci?ed in hra ctrl register 1(hc1). the desired active channel corresponding to a speci?c d-channel is determined from either the dedicated rx channel (drx) bits in hra ctrl register 2 (hc2) or the present receive channel (prx) bits in hra status registers in dedicated and multiplexed modes, respectively. 14.1.2 dedicated receive mode when the receiver is operated in dedicated receive mode, the auto-hunt circuitry is ignored and the system provides the channel number. dedicated receive mode is selected by setting ddrx bit in hra ctrl register 2 (hc2). when ddrx bit is set high it enables dedicated reception from the channel selected by the drxi bits in hra ctrl register 2 (hc2). in dedicated receive mode rxchnl and rxact bits will be held high. the ctsact bit will be held low, disabling the cts generator. these three status bits are in hra status 1 (hs1). a write to the drx bits will start rxcen to be enabled for the channel speci?ed as soon as the receiver becomes inactive.the drx bits are double buffered which allows the next receive channel to be speci?ed at any time, without a concern for whether the present receive packet is ?nished. 14.1.3 multiplexed receive mode the primary function of the rx channel auto-hunt circuit is to provide the rx circuit with the next receive channel number. the auto-hunt circuit has to interface with both the rx and tx circuits as described in section 13.3.3. the main function of the auto-hunt circuit is described in the following sections. 14.1.4 auto-hunt monitoring the auto-hunt circuit scans each of the 16 incoming d-channels on stream sti1 to determine if any one of them is requesting to send a packet to the system. the auto-hunt circuit will scan the channels which are enabled for receive activity. on reset, all 16 channels are locked out, and the system must specify which channels are to be scanned for request-to-send (rts) ?ags. the two hra lockout registers described in section 22.30 and section 22.31 are used to specify the channels. the auto-hunt circuit will stop monitoring a channel for ?ags in three other circumstances which are described in section 14.1.6. the hdlc ?ag control character '01111110', (7e hex ), is used to indicate an rts to the system. for each d- channel which is not disabled, the auto-hunt circuit monitors during the appropriate bit times for ?ags. if a ?ag is undetected after 15 consecutive bit times on the currently scanned d-channel, the auto-hunt circuit moves to the next channel. the amount of time required to accumulate 15 consecutive bits is dependent on the selected baud rate; for a 16k baud rate, each non-disabled-channel is monitored for up to eight frames before being rejected. this hunt sequence continues until a ?ag is detected, at which point the scanning stops and the flag bit is set high in the hra status register. scanning for ?ags, at least until an rts is actually detected, is independent of rx circuit activity or inactivity. the auto-hunt circuitry cycles endlessly through all available (non-disabled) channels, listening for a peripheral's request to send. the more channels there are with rts, the faster the auto-hunt circuit ?nds the next receive channel number, thereby keeping the receiver inactive time to a minimum. at the start of receiving a packet the rx circuit will latch the nrx channel number, clearing the flag bit, and moving the auto-hunt circuit on to begin scanning the next potential receive channel.
advance information MT90812 41 in multiplexed operation, the receiver obtains the next rx channel (nrx) number from the rx channel auto- hunt circuit. three conditions must be met before the rx circuitry will initiate receiving a packet: ? the auto-hunt circuit must have detected a request-to-send (rts) ?ag. ? the system must have read the channel number for the most recently active receive channel. this condition is not required if the received rts ?ag is the ?rst detected ?ag after reset. ? the receiver must currently be inactive. the order in which these three conditions are satis?ed is not important. for the ?rst condition to be true the auto-hunt circuit must detect the rts pattern '01111110' on the current d-channel . the second condition is met by setting the status bit (rxchnl) high in hra status 1 (hs1) register when a new receive channel number is latched. this disables subsequent receive channel number latching until the system clears that bit by reading prx in hra status 3 (hs3) register. if the system wants to stop the hdlc receiver, it may do so by not reading prx. the third condition, that the hdlc receiver be inactive, is achieved by the detection of a receive end-of-packet strobe (reop) from the hdlc controller chip. at the start of the reception of a packet the following ?ve actions are taken: ? the nrx channel number will be latched into the prx (present receive channel) bits, where it will de?ne the active receive channel time; ? an internal receive-active ?ag (rxact) will be set, allowing rxcen to be enabled during the appropriate channel time; ? the rxchnl status bit will be set, to inform the system that a new prx channel number is available; ? the auto-hunt circuitry will be cleared, and forced to move on to the next channel to be scanned; ? and a cts go-ahead pattern will be initiated (ctsact high). 14.1.5 cts generation clear-to-send active (ctsact) is in hra status 1 (hs1) register. ctsact set high indicates that the receiver is currently transmitting a clear-to-send go-ahead pattern to the peripheral on the transmit channel denoted by prxi. the cts pattern generated is the 9-bit hdlc go-ahead character '011111110' (7f +0). for a 16k baud rate, this bit pattern is sent out 2 bits/frame in the appropriate d-channel. when the go-ahead is complete, ctsact will be forced low, and control of the particular d-channel used for the go-ahead transmission will be returned to the system. the cts generator can be re-triggered by the system, if desired, by setting the re-initiate clear to send (rects) bit in hra ctrl register 2 (hc2) on page 71. once a new cts pattern has been initiated, the rects control bit will automatically be cleared by the rx circuitry. rects is acted upon with the same timing as the normal source of cts initiation. 14.1.6 circumstances when monitoring a channel is stopped there are four conditions which will cause the auto-hunt circuit to stop monitoring a particular channel for ?ags. they are: ? rts is not detected ? the channel is the current active receive channel ? the channel is locked out ? the channel is the current active transmit channel the ?rst, which was described above, is when a ?ag has not been detected within 15 valid bit times.
MT90812 advance information 42 the second condition occurs if the auto-hunt circuit manages to cycle back to monitor a peripheral which is already the currently active receive channel. because the peripheral continues to send request-to-sends during the time when it is being acknowledged with a go-ahead from the rx circuitry, it is important that ?ags not be detected twice for the same request. this condition is prevented by disabling scanning on a currently active receive channel. the third condition that causes the auto-hunt circuit to skip a channel, is if it has been locked out by the system. this allows the system to independently disable reception from any channel. on reset, all 16 d- channels are locked out, and the system must specify which channels are to be enabled for receive activity. the two hdlc ctrl lockout registers are described in section 22.30 and section 22.31. the fourth condition that will cause the auto-hunt circuit to skip a channel occurs when the scanned-channel number is the same as an active tx channel. the channel is skipped in this case to reduce the contention between a system transmit packet and the transmission of a cts. if a ?ag has been detected before the transmitter goes active on the same channel, then the contention, described in tx and rx handshaking on page 38, will occur. however, if the transmitter goes active at any point before a ?ag is detected, then contention will be avoided. 14.1.7 receive packet termination a receive packet will be terminated when the hdlc controller asserts the reop strobe for one bit period, one bit time after the closing ?ag is received. if desired, the system may assert its own sreop ?ag via the micro interface. the sreop bit is in hra ctrl register 2 (hc2) on page 71. this ?ag has the same effect as the normal reop strobe, but may be written asynchronously and will be cleared by the rx circuit after it has been acted upon. in either case, a receive end-of-packet will cause rxact to go low, disabling rxcen. 15.0 c-channel data c-channel access for codec or dnic control is provided through message mode (refer to connection memory on page 10). the c-channel information can be read from data memory or written to connect memory. c-channel data status information is very static. therefore, the interface as provided via message mode is single buffered. hence, the c-channel information may change every frame. 16.0 tone generation the MT90812 generates the standard 16 dtmf frequencies within +/- 0.6% of the nominal standard frequencies. table 13 shows the standard dtmf frequencies, the coef?cient used to generate the closest frequency, the actual frequency generated and the percent deviation of the generated tone from the nominal. table 13 - dtmf frequencies tone group frequency hz coef?cient actual frequency % deviation low 697 59h 695.31 -0.24% 770 62h 765.63 -0.56% 852 6dh 851.56 -0.05% 941 78h 937.50 -0.37% high 1209 8dh 1203.12 -0.49% 1336 96h 1343.75 +0.58% 1477 9fh 1484.38 +0.50% 1633 a9h 1640.62 +0.47%
advance information MT90812 43 a further 9 other standard tones are available for use as call progress and supervisory tones. also available are 7 programmable tones, which may be single or dual frequency. this totals 32 available tone outputs. the composite signal output level in the transmit direction is -4 dbm0 ( m -law) and -10 dbm0 (a-law), when the programmable gains are set at zero db. pre-twist of 2.0 db is incorporated into the composite signal resulting in a low tone output level of -8.12 dbm0 and a high group level of -6.12dbm0 (for m -law, 6 db lower for a-law). note that the digitally generated signals will be gain adjusted as programmed in the connect memory high for each outgoing channel. refer to section 7.0 for a description of gain control. each of the seven programmable locations has two 8-bit registers accessible via the parallel microprocessor interface used to program the two tone frequencies. if a single tone is desired then one of the registers is programmed to zero. a tone output is disabled if both low and high coef?cient registers are programmed to zero. the register descriptions of the coef?cient registers, low tone coef?cient 1-7, and high tone coef?cient 1-7 (ltc1-7, htc1-7) are listed on page 64. of the 7 programmable tone generators the ?rst two can also be used to provide dual frequency squarewave ringing signals. tone ringer enable bits, tre1 and tre2, in the tedc register, enable the generation of the squarewave ringing signals by the two tone ringer circuits. the tone ringer function is described in section 16.1. the tone generation circuit can be enabled with tge bit set to 1 in the tone generation and energy detect control register (tedc) listed on page 61. with the dtmf circuit reset there is no output generated for all the 32 tones including the tone ringers and fsk transmitter output. frequencies in the ?xed and programmable locations tone locations are generated according to one of three formulas, depending on what range the coef?cient value is in, as shown in table 14. the coef?cients of the ?xed tones cannot be changed. the coef?cient is an integer value from 0 to 255. these single and dual frequency tones are written to the local data memory page at the tone block addresses. the tones will be selected by writing the corresponding tones address to the connect memory low of the outgoing channel. the address and description of each tone is listed in table 15. outgoing gain control of +3 to -27 db in steps of 1db, as well as - db, is provided for outgoing channels connected to the tone generator output locations. refer to section 7.0 for further description of gain control. all in-band (i.e. 500hz to 3.5khz) harmonics and noise components are at least 35db below the fundamental frequencies. total signal to distortion ratio over this band is at least 30db. table 14 - programmable frequencies available coef?cient value formula hz frequency range hz resolution hz 0 disabled - - 1-63 250+(coef x 3.90625) 253.91-496.09 3.90625 64-127 coef x 7.8125 500.0 - 992.19 7.8125 128-255 (coef x 15.625) - 1000 1000.0 - 2984.38 15.625
MT90812 advance information 44 table 15 - tone block address note 1: 1st and 2nd programmable tones can be programmed as dual frequency squarewave tone ringer signals note 2: 7th programmable tone is replaced with fsk signal when fsk transmitter is enabled. 16.1 tone ringer of the 7 programmable tone generators the ?rst two can be used to provide dual frequency squarewave ringing signals. to enable this mode and generate the squarewave ringing signals, the tone ringer enable bits, tre1 and tre2, in the tone generation and energy detect control register (tedc) must be set. tre1 and tre2 enable the ?rst and second tone ringer circuits, respectively. the digital tone generator uses the values programmed into the low and high tone coef?cient registers to generate two different squarewave frequencies. both coef?cients are determined by the following equation: coef = [8000/frequency(hz)] - 1 where coef is an integer between 1 and 255. this produces frequencies between 31.25 - 4000 hz with a non- linear resolution as shown in table 16. the ringer program switches between these two frequencies at a 5 hz or 10 hz rate as selected by the wr bit in the tedc register. addr frequency(hz) application addr frequency(hz) application (type of cadenced tones generated) 00 697+1209 dtmf digits 1 10 350+440 dial tone, recall dial tone, con?rmation tone 01 697+1336 dtmf digit 2 11 440 call waiting, busy veri?cation, executive override 02 697+1477 dtmf digit 3 12 440+480 audible ringback, special audible ringback 03 697+1633 dtmf digit a 13 440+620 intercept tone 04 770+1209 dtmf digit 4 14 480+620 reorder tone, busy tone 05 770+1336 dtmf digit 5 15 400 busy, conference exterior, call waiting 06 770+1477 dtmf digit 6 16 400+450 audible ring tone 07 770+1633 dtmf digit b 17 425 dial tone 08 852+1209 dtmf digit 7 18 1400 intrusion tone 09 852+1336 dtmf digit 8 19 l1+h1 programmable 1 1 0a 852+1477 dtmf digit 9 1a l2+h2 programmable 2 1 0b 852+1633 dtmf digit c 1b l3+h3 programmable 3 0c 941+1209 dtmf digit * 1c l4+h4 programmable 4 0d 941+1336 dtmf digit 0 1d l5+h5 programmable 5 0e 941+1477 dtmf digit # 1e l6+h6 programmable 6 0f 941+1633 dtmf digit d 1f l7+h7 programmable 7 2
advance information MT90812 45 table 16 - tone ringer programmable frequencies 17.0 frequency shift keying (fsk) transmitter the fsk transmitter is a phase coherent fsk modulator that generates two output frequencies, representing the marks and spaces, of the digital data loaded into the fsk transmit memory. the mark and space frequencies are selectable to bell 202 or ccitt v.23 standards at 1200 baud (refer to table 17). the fsk transmitter output is a pcm coded signal that can be directed to any outgoing local tdm channel. table 17 - fsk signalling frequencies as with all outgoing channels, a gain from +3 to -27 db in steps of 1db, as well as - db, may be applied to the output of the fsk transmitter. refer to section 7.0 for further description of gain control. the signal output level is -6.12dbm0 ( m -law) and -12.12 dbm0 (a-law), when the programmable gains are set at zero db. the fsk transmit memory is a 20-byte fifo that is accessed at the address 07 h . refer to control register addresses indicated in table 20 on page 53. the fsk output is memory mapped to address 5f h as shown in table 3 on page 14. fig. 26 illustrates the data ?ow for the fsk transmitter in the MT90812. coef freq coef freq coef freq coef freq coef freq coef freq coef freq coef freq 0 na 32 242.42 64 123.08 96 82.47 128 62.02 160 49.69 192 41.45 224 35.56 1 4000.00 33 235.29 65 121.21 97 81.63 129 61.54 161 49.38 193 41.24 225 35.40 2 2666.67 34 228.57 66 119.40 98 80.81 130 61.07 162 49.08 194 41.03 226 35.24 3 2000.00 35 222.22 67 117.65 99 80.00 131 60.61 163 48.78 195 40.82 227 35.09 4 1600.00 36 216.22 68 115.94 100 79.21 132 60.15 164 48.48 196 40.61 228 34.93 5 1333.33 37 210.53 69 114.29 101 78.43 133 59.70 165 48.19 197 40.40 229 34.78 6 1142.86 38 205.13 70 112.68 102 77.67 134 59.26 166 47.90 198 40.20 230 34.63 7 1000.00 39 200.00 71 111.11 103 76.92 135 58.82 167 47.62 199 40.00 231 34.48 8 888.89 40 195.12 72 109.59 104 76.19 136 58.39 168 47.34 200 39.80 232 34.33 9 800.00 41 190.48 73 108.11 105 75.47 137 57.97 169 47.06 201 39.60 233 34.19 10 727.27 42 186.05 74 106.67 106 74.77 138 57.55 170 46.78 202 39.41 234 34.04 11 666.67 43 181.82 75 105.26 107 74.07 139 57.14 171 46.51 203 39.22 235 33.90 12 615.38 44 177.78 76 103.90 108 73.39 140 56.74 172 46.24 204 39.02 236 33.76 13 571.43 45 173.91 77 102.56 109 72.73 141 56.34 173 45.98 205 38.83 237 33.61 14 533.33 46 170.21 78 101.27 110 72.07 142 55.94 174 45.71 206 38.65 238 33.47 15 500.00 47 166.67 79 100.00 111 71.43 143 55.56 175 45.45 207 38.46 239 33.33 16 470.59 48 163.27 80 98.77 112 70.80 144 55.17 176 45.20 208 38.28 240 33.20 17 444.44 49 160.00 81 97.56 113 70.18 145 54.79 177 44.94 209 38.10 241 33.06 18 421.05 50 156.86 82 96.39 114 69.57 146 54.42 178 44.69 210 37.91 242 32.92 19 400.00 51 153.85 83 95.24 115 68.97 147 54.05 179 44.44 211 37.74 243 32.79 20 380.95 52 150.94 84 94.12 116 68.38 148 53.69 180 44.20 212 37.56 244 32.65 21 363.64 53 148.15 85 93.02 117 67.80 149 53.33 181 43.96 213 37.38 245 32.52 22 347.83 54 145.45 86 91.95 118 67.23 150 52.98 182 43.72 214 37.21 246 32.39 23 333.33 55 142.86 87 90.91 119 66.67 151 52.63 183 43.48 215 37.04 247 32.26 24 320.00 56 140.35 88 89.89 120 66.12 152 52.29 184 43.24 216 36.87 248 32.13 25 307.69 57 137.93 89 88.89 121 65.57 153 51.95 185 43.01 217 36.70 249 32.00 26 296.30 58 135.59 90 87.91 122 65.04 154 51.61 186 42.78 218 36.53 250 31.87 27 285.71 59 133.33 91 86.96 123 64.52 155 51.28 187 42.55 219 36.36 251 31.75 28 275.86 60 131.15 92 86.02 124 64.00 156 50.96 188 42.33 220 36.20 252 31.62 29 266.67 61 129.03 93 85.11 125 63.49 157 50.63 189 42.11 221 36.04 253 31.50 30 258.06 62 126.98 94 84.21 126 62.99 158 50.31 190 41.88 222 35.87 254 31.37 31 250.00 63 125.00 95 83.33 127 62.50 159 50.00 191 41.67 223 35.71 255 31.25 mark space bell 202 1200 2200 ccitt v.23 1300 2100
MT90812 advance information 46 figure 26 - data flow for fsk transmitter data is written to the fsk fifo. start and stop bits are added to each byte. the fsk modulator generates two output frequencies, representing the marks and spaces. at 1200 baud, one bit is transmitted every 6 2 / 3 frames. the resulting pcm encoded signal is written to dm where it can be switched to any outgoing channel. when an interrupt occurs the fts bit is set in the interrupt status register (ints). the interrupt may be enabled by setting the fte bit in interrupt enable register (inte). an fsk interrupt may occur for two reasons. when the fis bit, in the ringer and fsk control register (rfc), is low, an interrupt is generated when the tx fifo is empty, when the fis bit is high, an interrupt is generated at the end of transmission. the fifo empty interrupt will occur when the last byte has been read and before the byte has been transmitted. fsk transmitter will ?nish transmitting this last byte 11 bits (73 1/3 frames) later and start the idle state if the fifo has not been re?lled. the 11 bits include the start bit, 8 bits from the fifo and 2 idle bits. end of transmission must be con?rmed before disabling the fsk transmitter or switching to another output channel. the end of transmission can be selected by setting the fis bit. end of transmission interrupt will occur after 2 idle state bits have been transmitted. subsequently end of transmission interrupts occur once every 8 idle state bits. this interrupt can also be used by the system to count the number of bits sent in the channel seizure and mark preamble before the fsk data. when the fsk transmitter is enabled, by setting the fen bit in the fsk control register, and the fifo is not written to, the ?rst end of transmission interrupt is after 2 idle state bits (13 1/3 frames). subsequent interrupts are once every 8 bits (53 1/3 frames). in addition to selecting the type of interrupt the ringer and fsk control register (rfc) is used to select the idle state of continuous mark or continuous space, the mark and space frequency standard, and to turn on or off the transmitter. also, a channel seizure signal of alternating ones and zeros can be enabled. refer to the description of the ringer and fsk control register (rfc) on page 58. programming sequence for on-hook data transmission (channel seizure + mark + data packet): ? transmit channel seizure 1. fen initially 0 so that fsk transmit is disabled. select end of transmission interrupt and enable. fsk interrupt. select channel seizure as idle state. 2. enable transmission via fen=1. use end of transmission interrupt to count how many bits have been transmitted. the ?rst interrupt occurs after two idle state bits (13 1/3 frames) have been sent. fskm data memory fsk transmitter uport write connect memory control registers parallel to serial output stream channel m fsk fifo 0100 1001 0 pcm byte 1 2 1 - fsk modulator that generates two output frequencies, representing the marks and spaces. start and stop bits are added to each byte. 2 - fsk transmitter outputs pcm coded signal. 1 5f h 07 h 5f h
advance information MT90812 47 subsequently the interrupt occur once every 8 bits (53 1/3 frames). wait until the desired number of channel seizure bits have been sent. ? transmit mark 3. switch to mark as the idle state. again use end of transmission interrupt to count the number of mark bits. when switching the idle state the interrupt bit count is not affected, i.e. the bit count is not restarted but will interrupt 8 bit times after the last interrupt. wait until the desired number of mark bits have been sent. ? transmit data packet 4. select fifo empty interrupt. end of transmission interrupt is disabled by the selection. even though the fifo is empty there will be no interrupt because the empty interrupt occurs only when the fifo is read to empty, not when it is emptied via fifo clear. write to fifo for up to 20 bytes. the ?rst byte will be sent on the next bit boundary or the one after depending on when the byte is written. 5. when fifo empty interrupts, reload fifo. repeat as necessary until the entire message has been sent. when the fifo empty interrupts, there are 11 bit times (73 1/3 frames) before idle state transmission will commence. ? end transmission 6. after the fifo has been loaded for the last time, switch to end of transmission interrupt. select mark as the idle state. 7. wait for end of transmission interrupt which occurs after 2 idle state bits have been sent after the stop bit of the last fifo byte. if the system does not disable fsk via fen=0, end of transmission will interrupt again once every 8 bit times. 18.0 ringing generator a ringing generator is provided on pins r+ and r-. the output are two square waves 180 degrees out of phase. the frequency is selected with bits f1,f0 in ringer and fsk control register (rfc) and can be 16, 20, 25 or 50 hz. setting the bit re to 0 in the same register tri-states the drivers for both pins r+ and r-. 19.0 supervisory signal detection and cadence measurement two energy detect blocks, a and b, are provided for monitoring supervisory signalling during trunk calls. each of energy detect blocks can be assigned to an incoming channel, by programming one of the two connect memory low locations 70 h and 71 h . a low and high threshold level is programmed in the energy detect low and high threshold registers (edlta/b or edhta/b). the energy detect blocks are enabled by setting ena or enb bits in the tone generation and energy detect control register (tedc) described on page 61. the energy detect is implemented using a peak detector with an exponential attack and decay time constant of 2 msec and a 25 msec leaky hold time to bridge between the envelope peaks. the peak detector decays exponentially following the hold time limit. supervisory signalling cadence measurement is illustrated in fig. 27. a counter is used to time the cadence of the signal. when the signal envelope crosses the energy detect high threshold at point a, the counter value,t0, is transferred to the sscr register and the counter is reset and starts counting the next interval. the position of the signal envelope, now above the high threshold, is indicated with bit 7 of sscr (also labelled the p bit) set to 1. an interrupt is generated and the energy detect bit in the interrupt status register (ints) is set. at point b the low threshold limit is crossed, the sscr register is updated with the new count, t1, and an interrupt is generated. the position of the signal envelope, now below the low threshold, is indicated with the p bit of sscr set to 0.
MT90812 advance information 48 for a continuous signal, such as dial tone, where there is no off time, an interrupt occurs when the counter reaches a maximum count of 508 msec. when a maximum count of 508 msec is reached the p bit can be used to determine if the interrupt was a result of a transition or a counter over?ow. if the p bit remains unchanged from its previous value then the interrupt is a result of a counter over?ow. interrupts are generated with a minimum of 4 msec (32 frames) between consecutive irqs. if a threshold is crossed within 4 msec of the last interrupt, the interrupt will be held off and the sscr will be updated after the 4 msec time has transpired. refer to fig. 27. for example, if the signal envelope crosses the threshold levels at point b and point c, then there will be an interrupt generated at point d 32 frames from point a. the sscr is updated at the delayed irq point d with the last position and count, p=1 and t2, the time between points b to c. figure 27 - time between supervisory signal detection interrupts 3 4 ms table 18 lists the registers which are used for the energy detect blocks. table 18 - supervisory signal detection registers refer to msan-178 note for implementing an algorithm for interpreting the measured cadence of a call progress signal by the MT90812. register description page eda/b energy detect page 61 edalt/edaht energy detect a low and high threshold page 62 edblt/edbht energy detect b low and high threshold page 63 sscr1/2 supervisory signal cadence registers page 62 and page 64 t1 t3 a, p=1, t0 e, p=0, t3 high threshold level signal low threshold level b 4ms c t2 d, p=1, t2 f, p=1, t4 t4 4ms interrupt delayed interrupt interrupt 4ms interrupt envelope
advance information MT90812 49 20.0 microprocessor port the MT90812 provides a parallel microprocessor interface for non-multiplex or multiplexed bus structures. this interface is compatible with motorola non-multiplexed/multiplexed and intel/national multiplexed buses. if the im input pin is low or not connected, the device assumes its default mode (motorola non-multiplexed bus). in non-multiplexed mode, the microprocessor port consists of an 8-bit parallel data bus (ad0-ad7), 10-bit address lines (a0-a9) and four control lines ( cs, ds, r/ w and dt a). the parallel microprocessor port provides the access to the control registers and the connection and data memories of the MT90812. data memory is read only. the control register at location 61 h ( 3e1 h in motorola non-muxed, 061 h in in multiplexed mode) must be initialized to 080 h . if the im pin is high, the microprocessor port provides compatibility to motel interface. in motel interface, motorola, national, and intel multiplexed bus cpu can be connected to the device. in this mode, the interface pins are: ad<7:0> (data and address), as/ale (address latch enable/ address strobe), ds/ rd (data strobe/ read), r/ w \ wr (read/write\write), cs (chip select) and dt a (data acknowledgment). the motel circuit automatically identi?es the type of cpu bus connected to the MT90812. this circuit uses the level of the ds/ rd input pin at the rising edge of the as/ale to identify the appropriate bus timing connected to the MT90812. if ds/ rd is low at the rising edge of as/ale then motorola bus timing is selected. if ds/ rd is high at the rising edge of as/ale, then intel bus timing is selected. see figures 48 to 50 for each cpu interface timing. a MT90812 memory address, in multiplexed microport mode, consists of two portions. the higher order bits(3) originate from the control register. the lower order bits(8) originate from the address lines directly. the address lines a6-a0, on the control interface, give access to the control registers directly if a7 is zero, or depending on the contents of control register, to the high or low sections of the connection memory, or to the data memory. refer to address memory map on page 12. interrupts can occur from d-channel basic receive transmit (dbrt), fsk, energy detect, conference, or timing blocks. two registers are provided to help the microprocessor deal with interrupts. the interrupt enable register (inte), allows interrupts from each source to be enabled or disabled. the interrupt status register (ints), indicates which interrupt source has generated an interrupt. for further description on the ints and inte registers, refer to interrupt status register (ints) on page 56 and interrupt enable register (inte) on page 57, respectively.
MT90812 advance information 50 21.0 connection memory bits locations in the connection memory are associated with the local tdm output streams and the expansion bus streams. it also determines whether individual output channels are in message mode, allows individual output channel to go into a high-impedance state and speci?es the gain control for the outgoing channels. refer to section 5.2, data memory and connect memory for further description. 21.1 connection memory high bit name description 7-3 channel attenuation g4,g3, g2, g1, g0 de?nes gain from +3 to -27 db in steps of 1db, as well as - db for the outgoing channel 2 message channel/ conference inversion bit when 1, the contents of the corresponding location in connection memory low are output on the locations channel. when 0, the contents of the corresponding location in connection memory low act as an address for the data memory and so determine the source of the connection to the locations channel and stream. conference inversion bit: (for conference address locations 60-6e) the inversion bit allows for every other channel in a conference to be inverted. this reduces noise due to re?ections and line impedance mismatch. 1 cst (address locations other than 60-6e) this bit is used to select minimum(low) or constant(high) delay, on a per channel basis. 0 output enable (address locations other than 60-6e) if the ode pin is high and bit 5 of the control register is 0, then this bit enables the output driver for the locations channel and stream. this allows individual channels on individual streams to be made high-impedance, allowing switching matrices to be constructed. a 1 enables the driver and a 0 disables it. 1-0 ns1-ns0 (for conference address locations 60-6e) channel noise suppression 00 = no noise suppression 01 = 9/4096 (a-law), 9/8159 (u-law) 10 = 16/4096 (a-law), 16/8159 (u-law) 11 = 32/4096 (a-law), 32/8159 (u-law) 76543210 msg/ g0 g1 g4 g2 g3 inv cst/ ns1 oe/ ns0 0000 = +3 db 0001 = +2 db 0010 = +1 db 0011 = 0 db 0100 = -1 db 0101 = -2 db 0110 = -3 db 0111 = -4 db 1000 = -5 db 1001 = -6 db 1010 = -7 db 1011 = -8 db 1100 = -9 db 1101 = -10 db 1110 = -11 db 1111 = -12 db 10000 = -13db 10001 = -14db 10010 = -15db 10011 = -16db 10100 = -17db 10101 = -18 db 10110 = -19 db 10111 = -20 db 11000 = -21 db 11001 = -22 db 11010 = -23 db 11011 = -24 db 11100 = -25 db 11101 = -26 db 11110 = -27 db 11111 = - db
advance information MT90812 51 at locations 60 to 6e, connect memory high is used to specify the conference incoming channel attenuation and noise suppression. for these locations bits 7-3 are used as incoming channel attenuation, bits 1-0 are used for noise suppression bits and bit 2 is not used. 21.2 connection memory low the cml is de?ned as follows: connection memory low is used to specify the source for the outgoing channels and connect the conference, energy detect and dbrt blocks to incoming channels. table 19 lists the data memory block and channel address bits used in data memory addresses. at locations 60 to 6e, connect memory low is used to identify channels in a conference. at location 6f, connect memory low is used to identify which channel is to be routed to the d-channel rx fifo. locations 70 and 71h are used to identify the channel routed to the energy detect a and b blocks, respectively. table 19 - data memory addressing bit name description 7-5 data memory block address bits the number expressed in binary notation on these 3 bits is the number of the data memory block for the source of the connection. bit 7 is the most signi?cant bit. 4-0 channel address bits the number expressed in binary notation on these 5 bits is the number of the channel which is the source of the connection. bit 4 is the most signi?cant bit. e.g., if bit 4 is 1, bit 3 is 0, bit 2 is 0, bit 1 is 1 and bit 0 is 1, then the source of the connection is channel 19. hex address a7-a5 hex address a4-a0 local/expansion data memory description 000 00-1f sti0 32 channels sti0 32 channels 001 00-1f sti1 32 channels sti1 32 channels 010 00-1f tones(32) tone generator output 011 00-0e confout(15), conference output 011 0f unused unused(1) 011 10 dchout(1) output from the d-channel tx fifo buffer. allows d-channel tx buffer to be directed to any outgoing channel. 011 11-1f unused(14) unused(14) 100 00-1f ei1 32 channels expansion bus block 1 101 00-1f ei2 32 channels expansion bus block 2 110 00-1f ei3 32 channels expansion bus block 3 111 00-1f ei4 32 channels expansion bus block 4 76543210 a5 a2 a1 a0 a4 a6 a3 a7
MT90812 advance information 52 22.0 detailed register descriptions the ?rst page of 128 locations of memory contains the control registers. the control registers are accessed independent of the setting of the memory select bits when in multiplexed mode by setting external address bit a7 to 0. in non-multiplexed mode the control registers are accessed with a7,a8 and a9 set to 1 (as described in section 5.1). the control registers and their addresses are listed in table 20. hex address a6-a0 name description page 00 ams address memory select page 53 01 ctl control page 54 02 tc timing control page 55 03 occ output clocking control page 56 04 ints interrupt status page 56 05 inte interrupt enable page 57 06 rfc ringer and fsk control page 58 07 fskm fsk transmit memory page 59 08 confo conference over?ow status page 59 09 cc conference control page 60 0a-0f - unused (6) 10 tedc tone generation and energy detect control page 61 11 edalt energy detect a - low threshold page 62 12 edaht energy detect a - high threshold page 62 13 ssca supervisory signal cadence a page 63 14 edblt energy detect b - low threshold page 63 15 edbht energy detect b - high threshold page 63 16 sscb supervisory signal cadence b page 64 17-1f - unused (9) 20-26 ltc1-7 low tone coef?cient 1-7 page 64 27 unused 28-2e htc1-7 high tone coef?cient 1-7 page 64 2f unused 30-3e cpc1-15 conference party control 1 - 15 page 65 3f unused unused 40 drxit d-channel receive interrupt threshold page 66 41 drxc d-channel rx control page 66 42 drxs d-channel br status page 68 43 drxout d-channel rx fifo output page 68 44 dtxin d-channel tx fifo input page 68 45 dtxc d-channel tx control page 69 46-4f unused unused(10)
advance information MT90812 53 table 20 - control registers test register 1 and 2 are at locations 60 h and 61 h respectively. location 61 h must be initialized to 80 h . 22.1 address memory select register (ams) the address memory select register (ams) selects data/connect memory for read/write operations in microport multiplexed mode. the ams register also allows the microport read of speci?c data memory locations from their lower/upper bytes. 50 hc1 hra control 1 page 70 51 hc2 hra control 2 page 71 52 hc3 hra control 3 page 72 53 hlo1 hra lock out 1 page 72 54 hlo2 hra lock out 2 page 73 55 hs1 hra status 1 page 73 56 hs2 hra status 2 page 74 57 hs3 hra status 3 page 74 58 hs4 hra status 4 page 75 59-5f unused unused(7) 60 reserved reserved 61 reserved must be iniitialized to 80 h 62-7f unused unused(30) read/write address is: 000 h reset value is: 00 h bit name description 7-4 (unused) 3 ub when 1, the next reads of data memory locations associated with tones and conference output are from the upper bytes of these locations. if 0 the reads are from the lower bytes. 2-0 memory select when 000, local data memory is selected for read operation. when 001, expansion data memory is selected for read operation. when 010, local connection memory low is selected for read or write operations. when 011, expansion connection memory low is selected for read of write operations. when 100, local connection memory high is selected for read or write operations. when 101, expansion connection memory high is selected for read or write operations. note: setting of the memory select bits is required only when operating the microprocessor port in multiplexed mode. hex address a6-a0 name description page 76543210 ms0 ms1 ms2 -- - - - ub
MT90812 advance information 54 22.2 control register (ctl) the control register (ctl) selects data/connection memory and de?nes expansion bus position. read/write address is: 001 h reset value is: 00 h bit name description 7- unused . 6 serial stream output enable output enable for the serial outputs . if this input is low, sto0, sto1, est0, est1 are high impedance. if this input is high, each channel may still be put into high impedance state by using per channel control bit in the connection memory. 5 message mode when 1, the contents of the connection memory low are output on the serial output stream except when the oe bit in cmh for the corresponding channels is low or the ode pin is low. when 0, the connection memory bits for each channel determine what is output. 4 ebm ebus mode select . selects idx link mode if low or tdm link mode if high. 3 fmat pcm format select . selects ccitt pcm coding if high, or sign magnitude pcm if low. 2 alaw/ m la w companding law selection . a-law is selected when high. m -law is selected when low. 1 - 0 ebus position / ebus data rate when ebus is in idx link mode, ep0 and ep1 de?ne the sequence of channels on the expansion bus with respect to the frame pulse. (the four positions are shown as a to d in figure 7 on page 9) 00 selects the output to be the ?rst of every four channels. 01 selects the output to be the second of every four channels. 10 selects the output to be the third of every four channels. 11 selects the output to be the fourth of every four channels. when ebus is in tdm link mode, ep0 and ep1 de?ne the data rate of the expansion bus. 00 = 2.048 mb/s. 01 = 4.096 mb/s. 10 = 8.192 mb/s. the clock rate is determined by the clock mode selected by bits cr1-0 in the timing control register (tc). refer to description in timing and clock control starting on page 24. 76543210 ep0 ep1 a/u ebm msg stoe - fmat
advance information MT90812 55 22.3 timing control register (tc) the timing control register is con?gured as follows: read/write address is: 002 h reset value is: 20 h bit name description 7 wde watchdog enable . when 0, disables the clock watchdog circuit. when 1, the clock watchdog circuit will generate an interrupt in the event of the loss of the clock at the c8 input. see watchdog timer on page 28. 6 fpo fpo. selects st-bus or gci frame alignment and polarity for outgoing frame pulse generation. when 0, st-bus frame pulses are generated on f8o and f4o. when 1, gci frame pulses are generated and the polarities of c4o and c8 are inverted. the outgoing frame pulse mode is independent of the incoming frame pulse mode. 5-4 cr1-0 input clock reference. 00 c4 01 c8 10 c8p (default) 11 c16 selects one of four possible clock references, c4, c8, c8p, or c16. c4 is not valid when the pll is not enabled. the MT90812 requires at least an 8m clock internally. when the c4 input clock is selected the 8.192 mhz clock is derived from the pll. when c8p is selected as the input clock reference no frame pulse is used and the MT90812 generates f4o and f8o when they are enabled. refer to table 9, clock modes, on page 25. 3 hmvip hmvip select. with c16 as input clock reference, when hmvip=1, enables the hmvip frame alignment interface. otherwise, the device operates in st-bus/gci mode. 2pe pll enable . when 0, disables the pll. when 1, enables the pll. with the pll off, c10 is disabled and c4 as an input clock reference is not valid. 1 pms pll mode select . with pe=1, when pms = 1 the pll operates in master mode. when pms = 0, the pll operates in slave mode. default slave mode. 0 pcs pll clock select . with pe=1, when pcs = 1 selects clocks generated from the pll for use in sti/o0, sti/o1 and incoming est0/1 tdm streams, c2o, f4o and c4o. otherwise the clocks are derived directly from the input clock reference. with c4 as the input clock reference, est0/1, 4 and 8 mb/s timing, is generated from the pll independent of pcs. refer to table 9, clock modes, on page 25 and section 9.2.5.
MT90812 advance information 56 22.4 output clocking control register (occ) the register is con?gured as follows: 22.5 interrupt status register (ints) the ints register is con?gured as follows: read/write address is: 003 h reset value is: 00 h bit name description 7 pcos pll clock output select . with pe=1, when pcs = 1 selects clocks generated from the pll for use in outgoing est1/0 tdm streams, f8o and c8o. otherwise the clocks are derived directly from the input clock reference. with c4 as the input clock reference, est0/1, 4 and 8 mb/s timing, as well as f8o and c8o, are generated from the pll independent of pcos. refer to table 9, clock modes, on page 25 and section 9.2.5 6- unused . 5 c10e c10 output enable . when 0, c10o is high impedance. when 1 and the pll is enabled, c10o is enabled. 4 c8e c8 output enable . when 0, c8 is high impedance. when 1 and the input clock reference is not c8, then c8 is enabled. 3 f8e f8 output enable . when 0, f8 is high impedance. when 1 and the input clock reference is not c8, then f8 is enabled. 2 c4e c4 output enable . when 0, c4 is high impedance. when 1, then c4 is enabled. 1 f4e f4 output enable . when 0, f4 is high impedance. when 1, then f4o is enabled. 0 c2e c2 output enable . when 0, c2 is high impedance. when 1, then c2 is enabled. read address is: 004 h reset value is: 00 h bit name description 7 dre d-channel receive fifo error . status indicated in d-channel br status (drxs) on page 68. 6 drx d-channel receiver attained the message length in mli mode or the rx fifo interrupt trigger level (number of words) in fli mode. 5 dtx d-channel transmit fifo empty or 3/4 empty or transmission complete. 4 fts memory empty status for the fsk transmit memory or end of transmission. 3 edbs energy detect block b interrupt . 2edas energy detect block a interrupt . 76543210 c2e f4e c4e c8e c10e - pcos f8e 76543210 edas cfs c8f edbs fts dtx drx drxe
advance information MT90812 57 this register is a read only register, which would generally be read by the external up on a regular basis. the contents of this register indicates if an interrupt has occurred. this register is reset when read. 22.6 interrupt enable register (inte) the inte register is con?gured as follows: this register enables/disables the interrupts as speci?ed in the interrupt status register (ints). setting high the appropriate bits in this register enables the associated interrupt source. if a bit is set low in this register the bits in the interrupt status register (ints) are still valid but they do not cause the irq output to go low. 1 cfs over?ow status of the conference accumulators. 0 c8f c8 input not present . read/write address is: 005 h reset value is: 00 h bit name description 7 dree interrupt enable for d-channel receive fifo error . status indicated in d- channel br status (drxs) on page 68 6 drxe interrupt enable for d-channel receive message length or fifo level . 5 dtxe interrupt enable for d-channel transmit. fifo empty or 3/4 empty and transmission complete. 4 fte enable fsk transmit memory empty interrupt or end of transmission. 3 edbe energy detect block b interrupt enable. 2edae energy detect block a interrupt enable. 1 cfe interrupt enable for conference accumulator over?ow. 0 c8fe enable monitoring for presence of c8. read address is: 004 h reset value is: 00 h bit name description 76543210 edas cfs c8f edbs fts dtx drx drxe 76543210 edae cfe c8fe edbe fte dtxe drxe dree
MT90812 advance information 58 22.7 ringer and fsk control register (rfc) the ringer and fsk control register (rfc) controls the ringing source and fsk transmitter. the ringer and fsk control register is con?gured as follows: the ringer and fsk control register (rfc) is used to select the type of interrupt, the idle state, the mark and space frequency standard, and to turn on or off the transmitter. the idle state of continuous mark or continuous space or a channel seizure signal of alternating ones and zeros can be selected. changing s1, s0 bits will change the idle state on the next bit boundary. at 1200 baud, one bit corresponds to 6 2 / 3 frames. read/write address is: 006 h reset value is: 00 h bit name description 7re ringer enable bit . 6-5 f1,f0 selects the frequency of the square wave output of the ringing generator . the ringer outputs are the r+ and r- pins. 00 = 16 hz square wave 01 = 20 hz square wave 10 = 25 hz square wave 11 = 50 hz square wave 4 fis fsk interrupt select . 0 = interrupt generated when the tx fifo is empty. 1= interrupt generated at the end of transmission. 2-3 s1, s0 00 = idle state is continuous space (logic zero). 01 = idle state is continuous mark (logic one). 10 = turn on channel seizure signal. 11 = unused. 1 fen fen=0 disables the fsk transmitter and clears the fsk fifo . writing to fsk fifo while fen=0 will have no effect. when 1 the fsk transmitter is enabled, and the fsk transmit memory can be written to. when the memory is empty, the idle state set by bits s1-0, is output until the transmit memory is reloaded or the transmitter is turned off. 0 fms 0 = mark and space frequencies of 1200 hz and 2200 hz corresponding to bell 202 standard. 1 = mark and space frequencies of 1300 hz and 2100 hz corresponding to ccitt v.23 standard. 76543210 fms fen s0 fis f0 s1 f1 re
advance information MT90812 59 22.8 fsk transmit memory (fskm) the register is con?gured as follows: a system write to the tx fifo buffer is performed by addressing location 07 h of the control register page. up to 20-bytes can be written to the fifo. the length of the message is determined by the number of bytes written to the fifo. the fsk output is memory mapped to address 5f h as shown in table 3 on page 14 refer to frequency shift keying (fsk) transmitter on page 45. 22.9 conference over?ow status register (confo) the register is con?gured as follows: this register is a read only register. when a conference over?ow occurs the conference over?ow bit in the interrupt register will be set and the conference id will be stored in this register. the conference id will be that of the conference which had an accumulator over?ow. the conference id corresponds to the conference id numbers programmed in the conference party control registers. refer to section 22.20. the conference id in this register will not be updated again until it is reset. the register is reset following a read of the register or resetting the conference block or MT90812 device. following an irq being asserted by the conference circuit, the int would generally be read by the external up. reading the interrupt status register (ints) will clear the conference over?ow bit. further conference over?ows do not trigger an interrupt until the conference over?ow bit is cleared. the conference over?ow interrupt is maskable using the interrupt enable register (inte). the conference interrupt mask does not disable updates of the confo register. write address is: 007 h reset value is: 00 h bit name description 7-0 d7-d0 tx fifo buffer . d7 is the msb. read address is: 008 h reset value is: 00 h bit name description 7-3 unused unused . 2-0 c2-c0 conference id number . valid conference id numbers are: 001 = conference 1 010 = conference 2 011 = conference 3 100 = conference 4 101 = conference 5 76543210 d2 d1 d0 d4 d3 d7 d6 d5 76543210 c2 c1 c0 -- ---
MT90812 advance information 60 22.10 conference control register (cc) the conference control register is used in conjunction with the conference party control registers for setting up conferences. the cc register is con?gured as follows: the conference circuit can be reset with cfen bit =0. with the conference circuit reset there is no output generated for all the 15 conference output locations in data memory. read/write address is: 009 h reset value is: 00 h bit name description 7-4 unused 3-1 td2-0 tone duration . speci?es a tone duration from 0.125 seconds to 1.0 seconds in steps of 0.125 seconds. with the addition of a party to a conference, by programming one of the conference party control registers with the appropriate conference id, the insertion tone will be added to the conference if the it bit is set. the tone duration is set as follows: 000 = 0.125s 001 = 0.250s 010 = 0.375s 011 = 0.500s 100 = 0.625s 101 = 0.750s 110 = 0.875s 111 = 1.000s 0 cfen 0 = reset for conference block . 1= enables the conference circuit . 76543210 cfen td0 - - td2 td1 - -
advance information MT90812 61 22.11 tone generation and energy detect control register (tedc) the tone generation and energy detect register controls the two tone ringers and both a and b energy detect modules. the tedc register is con?gured as follows: of the 7 programmable tone generators the ?rst two can be used to provide dual frequency squarewave ringing signal. tre1/wr1 and tre2/wr2 are used to control tone ringer generator 1 and 2, respectively. refer to tone ringer description on page 44. each energy detect module monitors the signal for the selected incoming channel. locations 71 h and 72 h of cm are used to specify the incoming channel for eda and edb, respectively. an interrupt is generated and the energy detect ?ag is set when the signal exceeds the threshold set in the energy detect a/b low/high threshold register. a reset is performed by resetting the edena/b bit. during a reset the peak detector value and energy detect ?ag are set to zero. the tone generation circuit can be enabled with tge bit set to 1. with the circuit reset there is no output generated for all the 32 tones including the tone ringers and fsk transmitter output. read/write address is: 010 h reset value is: 00 h bit name description 7 unused 6 tge 0 = reset for tone generation circuit . 1 = enables tone generation . tone generation must be enabled for fsk transmission and tone ringer operation. 5 tre2 tone ringer enable bit . for programmable tone generator 2. 1 = the tone ringer generator is enabled using the coef?cients for programmable tone generator 2 as well as the wr2 control bit. 0 = the tone ringer generator 2 is disabled. with tone ringer 2 disabled the 2nd programmable coef?cients are used to generate dtmf signalling instead of the squarewave ringing tone. 4 wr2 warble rate bit . for programmable tone generator 2. 1 = the tone ringer circuit will toggle between the two programmed frequencies at a 5 hz rate. 0 = the tone ringer warble rate is 10 hz. 3 tre1 tone ringer enable bit . for programmable tone generator 1. 2 wr1 warble rate bit . for programmable tone generator 1. 1 edenb 1 = enable for energy detect b circuit . 0 edena 1 = enable for energy detect a circuit . 76543210 era erb tre2 wr2 tre1 wr1 tge
MT90812 advance information 62 22.12 energy detect a - low threshold (edalt) the edalt register is con?gured as follows: et0-et3 encode the pcm step number while et4-et6 encode the pcm chord number. 22.13 energy detect a - high threshold (edaht) the edaht register is con?gured as follows: et0-et3 encode the pcm step number while et4-et6 encode the pcm chord number. 22.14 supervisory signal cadence register a (ssca) read/write address is: 011 h reset value is: 00 h bit name description 7 unused 6-0 low threshold energy detect low threshold . pcm sign-magnitude format with no sign bit. bit 6 = msb. read/write address is: 012 h reset value is: 00 h bit name description 7 unused 6-0 high threshold energy detect high threshold . pcm sign-magnitude format with no sign bit. bit 6 = msb. read address is: 013 h reset value is: 00 h bit name description 7 p position with respect to high and low thresholds. 1 = above high threshold. 0 = below low threshold. 0-6 t6-t0 cadence of the supervisory signal ranging from 0 to 508 msec in units of 4 msec. 76543210 et5 et2 et1 et0 et4 et6 et3 - 76543210 et5 et2 et1 et0 et4 et6 et3 - 76543210 t5 t2 t1 t0 t4 t6 t3 p
advance information MT90812 63 the ssca register is used to store the cadence information for energy detect block a. refer to supervisory signal detection and cadence measurement on page 47. the counter is used to time the cadence of the signal. when the signal envelope crosses the energy detect threshold the counter starts. when the other threshold limit is crossed the p bit is updated, the counter value is transferred to bits t6-t0, in this ssc register and an interrupt is generated. the p bit speci?es the current envelope position with respect to the high and low thresholds. when p is set this indicates the signal envelope has crossed above the high threshold otherwise it had crossed below the low threshold. the count is stored in 4 msec intervals and ranges from 0 to 508 msec. for a continuous signal, such as dial tone, where there is no off time, an interrupt is generated when the counter reaches a maximum count of 508 seconds or 7f h . the continuous signal is indicated by the p bit maintaining the same value. 22.15 energy detect b - low threshold register (edblt) the edblt register is con?gured as follows: et0-et3 encode the pcm step number while et4-et6 encode the pcm chord number. 22.16 energy detect b - high threshold register (edbht) the edbht register is con?gured as follows: et0-et3 encode the pcm step number while et4-et6 encode the pcm chord number. read/write address is: 014 h reset value is: 00 h bit name description 7 unused reserved . 6-0 low threshold energy detect low threshold . pcm sign-magnitude format with no sign bit. bit 6 = msb. read/write address is: 015 h reset value is: 00 h bit name description 7 unused reserved . 6-0 high threshold energy detect high threshold . pcm sign-magnitude format with no sign bit. bit 6 = msb. 76543210 et5 et2 et1 et0 et4 et6 et3 - 76543210 et5 et2 et1 et0 et4 et6 et3 -
MT90812 advance information 64 22.17 supervisory signal cadence register b (sscb) the sscb register is used to store the cadence information for energy detect block b. refer to the description of the energy detect b - low threshold register (edblt) on page 63. 22.18 low tone coef?cient registers 1-7 (ltc1-7) the seven ltc registers are con?gured as follows: 22.19 high tone coef?cient registers 1-7 (htc1-7) the seven htc registers are con?gured as follows: each of the seven programmable tone generators has a low and high tone coef?cient register used to program the two tone frequencies. frequencies for the programmable tone generators can be speci?ed according to one of three formulas, depending on what range the coef?cient value is in, as shown in table 14. read address is: 016 h reset value is: 00 h bit name description 7p position with respect to high and low thresholds . 1 = above high threshold. 0 = below low threshold. 0-6 t6-t0 cadence of the supervisory signal ranging from 0 to 508 msec in units of 4 msec. read/write addresses are: 20-26 h reset value is: 00 h bit name description 0-7 l0-l7 low tone coef?cient . read/write addresses are: 28-2e h reset value is: 00 h bit name description 0-7 h0-h7 high tone coef?cient . 76543210 t5 t2 t1 t0 t4 t6 t3 p 76543210 l0 l1 l2 l3 l4 l5 l6 l7 76543210 h0 h1 h2 h3 h4 h5 h6 h7
advance information MT90812 65 the coef?cient is an integer value from 0 to 255. if a single tone is desired then one of the registers is programmed to zero. a tone output is disabled if both low and high coef?cient registers are programmed to zero. of the 7 programmable tone generators the ?rst two can also be used to provide dual frequency squarewave ringing signals. tone ringer enable bits, tre1 and tre2, in the tedc register, enable the generation of the squarewave ringing signals by the two tone ringer circuits. the tone generator and tone ringer functions are described in section 16.0 22.20 conference party control register (cpc1-15) the conference party control register contains the conference id, start bit, insertion tone enable and outgoing channel attenuation. the register is con?gured as follows: the conference block provides conference call capability in the MT90812 and supports a total of 15 parties maximum, distributed over up to 5 conferences, i.e. 1x15 parties, 3x5 parties, 5x3 parties etc. each of the 15 parties are associated to a conference by programming the corresponding conference id number. as a party is added to a conference, if the insertion tone bit (it) is set, all channels connected in a conference will have the tone added to the conference output. the dm address of the desired tone must be programmed at location 6f h of cml, (16f h non-mux mode, ef h for mux-mode addressing). the pcm data from the speci?ed data memory location will be added to the conference output for a speci?ed tone duration. read/write address is: 30-3eh reset value is: 00 h bit name description 7-5 c2-c0 conference id number 000 = null conference 001 = conference 1 010 = conference 2 101 = conference 3 100 = conference 4 101 = conference 5 111 = clear all conferences 4 unused 3st start conference setting . the st bit for the ?rst party programmed for a conference will remove any other parties that may have been previously programmed for that conference. this is implemented at the time the conference party control register is updated and the state of the st bit is not actually stored in the register. 2iti nsertion tone enable . 1-0 gc1-gc0 outgoing gain control 00 = 0 db 01 = -3 db 10 = -6 db 11 = -9 db 76543210 it* gc1 gc0 - st* c2 c1 c0
MT90812 advance information 66 the it bit will be set for the duration of the tone added to the conference. reading any of the cpc registers in a particular conference, will show the it bit set for that time. channel attenuation is provided on incoming and outgoing channels in a conference. the incoming channel attenuation is set in cmh for the speci?c conference party. the outgoing gain is applied to the output of the conference block before it is written to the data memory output conference location. refer to conferencing on page 18. 22.21 d-channel receive interrupt threshold (drxit) the register is con?gured as follows: the d-channel receiver interrupt threshold is speci?ed in two ways depending on whether message length interrupt mode or fifo level interrupt mode is used. 22.21.1 message length interrupt mode an interrupt is generated when a message of n bits is received, where n is the message length in bits speci?ed by d7-d0. d7 is the msb. for example, 00 h corresponds to a bit length of 1 and ff h corresponds to a bit length of 256. the message length count is the number of bits between the start bit and the parity bit or stop bit. 22.21.2 fifo level interrupt mode an interrupt is generated when there are n bytes in the fifo, where n is speci?ed by d5-d0. d5 is the msb. for example, 01 h corresponds to an interrupt level of 1 byte in the fifo and 20 h corresponds to an interrupt level of 32 bytes in the fifo. read/write address is: 40 h reset value is: 00 h bit name description 7-0 d7-d0 message length interrupt mode . d-channel receive message length 1-256 bits. fifo level interrupt mode: d5-d0 used to specify fifo level of 1-32. 76 543 21 0 d2 d1 d0 d4 d3 d7 d6 d5
advance information MT90812 67 22.22 d-channel rx control (drxc) the register is con?gured as follows: the enable bit for the rx fifo is used to disable all the dbr circuitry including the transfer of data from dm to the rx. if there is a read of the rx fifo while it is disabled then the data is unde?ned. in mli mode, when the dbr is enabled the receiver identi?es the ?rst low bit as the start bit and then collects bits as speci?ed by the data rate, 1,2 or 8 bits per frame. in fli mode when the er bit is set the receiver transfers either 1,2 or 8 bits to the rx from data memory. if s=1 then the start and stop bits are expected on a per byte basis. if s=0 reception starts immediately after er is set. read/write address is: 41h reset value is: 00 h bit name description 7 txbo transmitter bit order . when 0 the ?rst bit transmitted on the tdm channel is the lsb read on the microport data bus (d0). when 1 the rx-fifo will maintain the same bit ordering as an access to dm. that is ?rst bit transmitted in the tdm channel is the msb read on the d7 of the microport data bus. 6 rxbo receiver bit order . when 0 the ?rst bit received on the tdm channel is the lsb read on the microport data bus (d0) when 1 the rx-fifo will maintain the same bit ordering as an access to dm. that is ?rst bit received on the tdm channel is the msb read on the d7 of the microport data bus. 5-4 w1-w0 data rate 00 = 1 bit per frame 01 = 2 bits per frame 10= 8 bits per frame 3m 1 = message length interrupt mode , i.e. message oriented, or message oriented with parity. 0 = fifo level interrupt mode , i.e. unframed, byte oriented, or byte oriented with parity. 2er enable receiver . if 0, clears rx fifo and resets counter. when er=0, data is unde?ned when a rx fifo read is performed. er= 1, the receiver is enabled. 1pe parity enable . in mlim, if 0 the receiver does not expect the parity bit. if 1 the receiver expects the parity bit following the message bits and before the stop bit. in flim, when se=1 and pe = 1 the receiver expects a parity bit per message byte. otherwise no parity bit is expected. if parity is enabled then an even number of logic 1s is expected in the data words and parity bit. 0se start and stop bit enable . in fli mode, if se=0 then no start and stop bits are expected. if se=1 then start and stop bits are expected in each message byte. in mli mode start and stop bits are always expected. 76543210 mer se w1 w0 rxbo pe txbo
MT90812 advance information 68 22.23 d-channel br status (drxs) the register is con?gured as follows: 22.24 d-channel rx fifo output (drxout) the register is con?gured as follows: the receiver transfers incoming data for a speci?ed channel to the rx fifo. the d-channel rx channel is identi?ed in address 70 hex of connect memory low (refer to section 5.3). the data in this channel is transferred to the rx fifo buffer. a read of control register d-channel rx fifo output (drxout) at 43 hex accesses the next data byte in the rx fifo buffer. 22.25 d-channel tx fifo input (dtxin) the register is con?gured as follows: read address is: 42 h reset value is: 00 h bit name description 7-3 - unused . 2oe rx overrun error . gets set when writing to a full fifo. 1pe rx parity error . 0se rx stop bit error . read address is: 43 h bit name description 7-0 d7-d0 next data byte in the rx fifo buffer . d7 is the msb. write address is: 44 h bit name description 7-0 d7-d0 tx fifo buffer . d7 is the msb. 76543210 ov pe se -- --- 76543210 d2 d1 d0 d4 d3 d7 d6 d5 76543210 d2 d1 d0 d4 d3 d7 d6 d5
advance information MT90812 69 a system write to the tx fifo buffer is performed by addressing location 44 hex of the control register page. up to 32 bytes can be written to the fifo. the length of the message is determined by the number of bytes written to the fifo. if n-bytes are written to this location and the st bit in the d-channel receive interrupt threshold (drxit)is set, the message transmitted will be n-bytes long. refer to transmitter operation on page 33. 22.26 d-channel tx control (dtxc) the register is con?gured as follows: the transmitter bit order (txbo) bit resides in the drxc register described in section 22.21. read/write address is: 45 h reset value is: 00 h bit name description 7is tx fifo interrupt select . when dtxe=1 and is=0 then when the tx fifo is empty or 3/4 empty (as selected by il bit) an interrupt is generated.when dtxe=1 and is=1 then an interrupt is generated when the transmission has ended. 6ili nterrupt level . 0 = interrupt generated when the tx fifo is empty. 1= interrupt generated when the tx fifo is 3/4 empty, when there are 8 bytes remaining. 5-4 w1-w0 tx data rate 00 = 1 bit per frame 01 = 2 bits per frame 10= 8 bits per frame 3m 1 = message length interrupt mode , i.e. message oriented, or message oriented with parity. 0 = fifo level interrupt mode , i.e. unframed, byte oriented, or byte oriented with parity. 2pe parity enable . if 0 disable parity bit. if pe =1 then enable parity bit. in fli mode parity can be enabled only if the start and stop bits are used, i.e. s=1. 1se start and stop bit enable . in fli mode, if 0 transmit message without including start, parity and stop bits. if 1 transmit message with start and stop bits; also transmit parity according to pe bit. in mli mode start and stop bits are always used. 0st start transmitter . st=1 starts transmission of the message following a write to the tx fifo. st=0 clears the fifo and resets its pointers. 76543210 mpe st w1 w0 is il se
MT90812 advance information 70 22.27 hra ctrl register 1(hc1) the register is con?gured as follows: read/write address is: 50 h reset value is: 01 h bit name description 7 sflag software controlled flag detect (sflag) . sflag is used to allow the system to inject a ?ag (indicating that a peripheral request-to-send has been detected). this bit can be written asynchronously, has exactly the same effect as the normal ?ag generated by the auto-hunt circuitry, and is cleared automatically by the rx circuitry when it has been used. on reset, this bit is cleared. 6 prxsel present receive channel select (prxsel) . normally low. this bit is set high if the receive channel is to be selected via drxi; unlike ddrx, however, cts is not disabled and rxchnl and rxact are not held high. instead, the receiver can be exercised by manipulating sreop and sflag in a manner similar to normal multiplexed operation. primarily intended as a test mode. on reset, this bit is cleared. 5-4 - unused . 3en enable hra . if 0, disables hra block. if 1 the hra block is enabled. 2 cd 0 = d before c-channel. the d-channel is transferred before the c-channel following the frame pulse. the d-channels are located in the ?rst 16 of the 32 channels of sti/ o1 stream. 1 = c before d-channel. the c-channel is transferred before the d-channel following the frame pulse. the d-channels are located in channels 16-31 of sti/o1 stream. 1-0 brsel1-0 baud rate select . the baud rate select ?eld controls the number of bit times actually used for d-channel activity within a selected transmit or receive channel time. it also controls the rate at which the auto-hunt circuitry monitors peripherals for request-to-sends. during inactive bit times within an active channel, idles are generated and output. one active bit gives a baud rate of 8k, while eight active bits gives a baud rate of 64k. 00 8k 01 16k 10 64k 11 unused. on reset, 16k baud rate will be selected. 76543210 prxsel cd brsel1 brsel0 sflag - - en
advance information MT90812 71 22.28 hra ctrl register 2 (hc2) the register is con?gured as follows: read/write address is: 51 h reset value is: 00 h bit name description 7 sreop software controlled rx end of packet (sreop). the system can at any time inject a receive end-of-packet by setting this bit. when rx circuitry uses this bit, sreop will be automatically cleared. sreop has exactly the same effect and internal timing as reop. on reset, this bit is cleared. 6 rects re-initiate clear to send (rects) . by setting this bit, the system can initiate another go- ahead ?ag to the peripheral currently enabled for transmission to the system. this is used when the cts, which is automatically generated upon receive channel selection, is not detected at the peripheral. rects is used with the same internal timing as the normal source of cts initiation. when rects has been used by the rx circuitry it is automatically cleared. a rects request will be accepted only if the receiver is active; the request will be ignored, and the rects bit cleared, if the receiver is inactive. on reset, this bit is cleared. 5 steop software controlled tx end of packet (steop) . the system can at any time inject a transmit end-of-packet by setting this bit. when the tx circuitry uses this bit, steop will be automatically cleared. steop has exactly the same effect and internal timing as teop. on reset, this bit is cleared. 4 ddrx dedicated rx control . this bit when high enables dedicated reception from the channel selected by the drxi bits. at the same time, the generation of cts is disabled, and ?ags generated by the auto-hunt circuitry are ignored. rxchnl and rxact are held high in dedicated receive mode. if this bit is low, normal rx operation is enabled. on reset, this bit is cleared. 3-0 drx4-1 dedicated rx channel number . if the ddrx bit or the prxsel bit is high, then these bits control the selection of the current receive channel, overriding the output of the auto-hunt circuitry. the rx channel number selects one of the ?rst 16 timeslots of sti1, with hc1 register bit cd=0, or the last 16 timeslots with bit cd=1. drx4 is the msb and drx1 is the lsb. on reset, these bits are cleared. 76543210 rects steop ddrx drx4 drx3 drx2 drx1 sreop
MT90812 advance information 72 22.29 hra ctrl register 3 (hc3) the register is con?gured as follows: 22.30 hra lock out register 1 (hlo1) the register is con?gured as follows: 22.31 hra lock out register 2 (hlo2) the register is con?gured as follows: read/write address is: 52 h reset value is: 00 h bit name description 7-4 unused unused . 3-0 ntx4-1 next tx channel number . these bits represent the channel number for the next packet to be transmitted and selects one of the ?rst 16 timeslots of sto1, with hc1 register bit cd=0, or the last 16 timeslots with bit cd=1. ntx4 is the msb and ntx1 is the lsb. on reset, this register is cleared. a write to this register causes the status bit txchnl, readable from the hra status 2 register, to be reset to low. read/write address is: 53 h reset value is: ff h bit name description 7-0 loc7-0 setting any bit in this register disables reception from the corresponding channel number. loc0-7 correspond to channels 0 to 7 on stream sti1, with hc1 register bit cd=0, or channels 16 to 23 with bit cd=1. on reset, this register is set, so that all channels are disabled. example: loc2 is channel 2 on sti1 read/write address is: 54 h reset value is: ff h bit name description 7-0 loc15-8 setting any bit in this register disables reception from the corresponding channel number. loc8-15 correspond to channels 8 to 15 on stream sti1, with hc1 register bit cd=0, or channels 24 to 31 with bit cd=1. on reset, this register is set, so that all channels are disabled. 76543210 - - ntx4 ntx3 ntx2 ntx1 - - 76543210 loc6 loc5 loc4 loc3 loc2 loc1 loc0 loc7 7 654321 0 loc14 loc13 loc12 loc11 loc10 loc9 loc8 loc15
advance information MT90812 73 22.32 hra status 1 (hs1) the register is con?gured as follows: read address is: 55 h reset value is: 00 h bit name description 7 rxchnl receive channel latched (rxchnl) . when this bit is high it indicates that a new receive channel number has been latched internally, and is as shown in the currently read byte. a system read of hra status register 3 will automatically clear this status bit. in dedicated receive mode, this bit will be held high. on reset, this bit will be low. 6 rxact receiver active (rxact) . when high, this status bit indicates that the receiver is currently active. on reset, this bit will be low, and rxcen will be disabled. in dedicated receive mode, this bit will be held high. 5 ctsact clear-to-send active (ctsact) . if high, this status bit indicates that the receiver is currently transmitting a clear-to-send go-ahead pattern to the peripheral on the transmit channel denoted by prxi. in dedicated receive mode, this bit will be held low, disabling cts generation and transmission. on reset, this bit will be low. 4 flag flag detect (flag) . when high, flag indicates that the auto-hunt circuitry has detected a request-to-send from a peripheral, but has not yet acted upon it. flag will also go high if the system injects an sflag via the micro interface. as soon as the rx circuitry selects the corresponding peripheral for system receive, this status bit is cleared. on reset, this bit will be low. this bit is primarily used for test purposes. 3-0 prx4-1 present receive channel (prx) . the current receive channel number is contained in these ?ve bits. reading the channel number via hra status register 3 clears the status bit rxchnl. prx4-1 represent channels 0-15 on sti1 stream, with hc1 register bit cd=0, or channels 16 to 31 with bit cd=1. on reset, channel 0 will be selected, but will be inactive. 76543210 rxact rxchnl prx4 flag ctsact prx3 prx2 prx1
MT90812 advance information 74 22.33 hra status 2 (hs2) the register is con?gured as follows: 22.34 hra status 3 (hs3) the register is con?gured as follows: reading the channel number via this register clears the status bit rxchnl. read address is: 56 h reset value is: 80 h bit name description 7 txchnl tx channel number latched (txchnl) . the next tx channel number in write register 3 has been latched by the tx control circuitry and may be rewritten when this bit is high. the bit is automatically reset when a write to register 3 occurs. on reset, this bit will be high, although the transmitter will be inactive. 6 txact transmitter active (txact) . when high, this status bit indicates that the transmitter is currently active. on reset, this bit will be low, and txcen will be disabled. 5 ctsact clear-to-send active (ctsact) . see description in hs1 register. 4 flag flag detect (flag) . see description in hs1 register. 3-0 prx4-1 present receive channel (prx) . see description in hs1 register. read address is: 57 h reset value is: 80 h bit name description 7 txchnl tx channel number latched (txchnl) . 6 rxchnl receive channel latched (rxchnl) . see description in hs1 register. 5 unused unused . 4 flag flag detect (flag) . see description in hs1 register. 3-0 prx4-1 present receive channel (prx) . see description in hs1 register. 76543210 txact ctsact txchnl flag prx4 prx3 prx2 prx1 76543210 txchnl flag prx4 prx3 prx2 prx1 rxchnl
advance information MT90812 75 22.35 hra status 4 (hs4) the register is con?gured as follows: read address is: 58 h reset value is: 00 h bit name description 7-4 - unused . 3-0 ptx4-1 present transmit channel (ptx) . the current transmit channel number is contained in these four bits. reading the channel number via this register does not clear the status bit txchnl.ptx4-1 represent channels 0-15 on sto1 stream, with hc1 register bit cd=0, or channels 16 to 31 with bit cd=1. on reset, channel 0 will be selected, but will be inactive. 76543210 -- ptx4 ptx3 ptx2 ptx1 - -
MT90812 advance information 76 23.0 applications 23.1 local tdm channel assignment fig. 28 shows the channel assignment for the local tdm streams used to support the stations, trunks and analog ports for a typical con?guration. there are 8 stations supported by mt9171/72 dnic transceivers, 8 analog ports supported by 8 codecs and 4 trunks supported by another 4 codecs. in this example, the b-channel information is on the sti/o0 streams. the c- and d-channel information is placed on the sti/o1 streams (dual port mode). figure 28 - MT90812 local tdm channel assignment for typical con?guration the local sti/o0 and sti/o1 channel assignment is summarized in table 21. table 21 - tdm channel assignment 23.2 dnic channel assignment the MT90812 supports direct connection to the dnics placed in digital network (dn), dual port mode. the dnic transfers four channels of information via the dv and cd ports. they are b1, b2, c and d-channels. the b1 and b2 channels each have a bandwidth of 64 kbit/s and are used for carrying pcm encoded voice or data. the c-channel, having a bandwidth of 64kbit/s, provides a means for the system microprocessor to control the dnic and for the dnic to pass status information back to the system microprocessor. the d-channel can be transmitted and received on the line with either 8, 16, or up to 64 kbit/s bandwidth. to support this, the MT90812 provides buffering of 1, 2 or 8 bits per frame to support d-channel end to end signalling or low speed data transfer. channel # of channels device sti/o0 sti/o1 0-3 4 trunk codec 1-4 b1 c 4-11 8 dnic 1-8 b1 d 12-15 4 codec 1-4 b1 c 16-19 4 codec 5-8 b1 c 20-27 8 dnic 1-8 b2 c 28-31 4 unused. - - i+16 i+32 frame n i f0i sti/o1 sti/o0 channel 0-3 b1 trunk 1-4 4-7 b1 dnic 1-4 8-11 b1 dnic 5-8 12-15 codec 1-4 16-19 codec 5-8 20-23 b2 dnic 5-8 24-27 b2 dnic 5-8 28-31 unused c trunk 1-4 d dnic 1-4 d dnic 5-8 c codec 1-4 codec 5-8 c dnic 1-4 c dnic 5-8 unused
advance information MT90812 77 in dn, dual port mode, the dnic receives a d-channel on cdsti while transmitting a d-channel on cdsto. fifteen channel times later (halfway through the frame) a c-channel is received on cdsti while a c-channel is transmitted on cdsto. the timeslot assignment used by the dnic is shown in fig. 29. figure 29 - dnic time slot assignment for sti/o0 and sti/o1 tdm streams in figure 28, on page 70, timeslots 4-11 and 20-27 are allocated for the 8 dnics. timeslots 4-11 contain the b1 channels on sti/o0 and d-channels on sti/o1. timeslots 20-27 contain the b2 channels on sti/o0 and c- channels on sti/o1. f0i sti/o1 dc sti/o0 b1 b2 channel 1 i+16 i+32 frame n frame n+1 f0o sti/o1 dc sti/o0 b3 b4 channel i+1 i+17 i+32
MT90812 advance information 78 24.0 ac/dc electrical characteristics * exceeding these values may cause permanent damage. functional operation under these conditions is not implied. ? characteristics are for clocked operation over the ranges of recommended operating temperature and supply voltage. ? typical ?gures are at 25?c and are for design aid only: not guaranteed and not subject to production testing. absolute maximum ratings* - voltages are with respect to ground (v ss ) unless otherwise stated. parameter symbol min max units 1v dd - v ss v dd - 0.3 7.0 v 2 voltage on any pin i/o (other than supply pins) v i/o v ss - 0.3 v dd + 0.3 v 3 current at any pin other than supply pins i pin 40 ma 4 package power dissipation p d 2w 5 storage temperature t s - 65 + 150 o c recommended operating conditions - voltages are with respect to ground (v ss ) unless otherwise stated. parameter symbol min max units 1 operating temperature t op - 40 + 85 o c 2 positive supply v dd 4.5 5.5 v 3 input voltage v i 0v dd dc electrical characteristics ? - voltages are with respect to ground (v ss ) unless otherwise stated. characteristics sym min typ ? max unit test conditions 1 supply current i dd 914ma pll off, outputs unloaded i dd 10 15 ma pll on, outputs unloaded 2 input high voltage v ih 2.0 v dd v 3 input low voltage v il 0 0.8 v 4 c 8 p input high voltage v ih 3.6 v dd v 5 input low voltage v il 0 0.8 v 6 leakage current i lk 10 ua 0 v v dd 7 pin capacitance c p 10 pf 8 output voltage high (digital outputs) v oh 2.4 v i oh = 12ma 9 output voltage low (digital outputs) v ol 0.4 v i ol = 10ma 10 output high sourcing current (digital outputs) i oh 12 ma v oh = 2.4v 11 output low sinking current (digital outputs) i ol 10 ma v ol = 0.4v 12 r e s e t positive threshold voltage hysteresis negative threshold voltage v+ v h v- 3.7 1.0 1.3 v v v
advance information MT90812 79 ? typical ?gures are at 25?c and are for design aid only: not guaranteed and not subject to production testing. ac electrical characteristics - input frame pulse parameters ? typical ?gures are at 25?c and are for design aid only: not guaranteed and not subject to production testing. ? see "notes" following ac electrical characteristics tables. ac electrical characteristics - timing parameter measurement voltages levels characteristics symbol ttl pin cmos pin units ttl reference level v tt 1.5 - v cmos reference level v ct - 0.5*v dd v input high level v h 2.4 0.9*v dd v input low level v l 0.4 0.1*v dd v rise/fall high meas. pt. v hm 2.0 0.7*v dd v rise/fall low meas. pt. v hl 0.8 0.3*v dd v ac electrical characteristics - input clock parameters characteristics sym min typ ? max units test conditions c4i clock period t c4ip 220 244 ns c4i clock width (high) t c4ih 110 122 140 ns c4i clock width (low) t c4il 110 122 140 ns c4i clock rise/fall time t c4ir/f 10 ns c8 clock period t c8p 110 122 ns c8 clock width (high) t c8h 50 61.0 70 ns c8 clock width (low) t c8l 50 61.0 70 ns c8 clock rise/fall time t c8ir/f 10 ns c16 clock period t c16ip 50 61.0 70 ns c16 clock width (high) t c16ih 25 30.5 35 ns c16 clock width (low) t c16il 25 30.5 35 ns c16 clock rise/fall time t c16ir/f 10 ns characteristics sym min typ ? max units test conditions ? f4i setup time (st-bus, gci) t f4is 10 150 ns 17 f4i hold time (st-bus, gci) t f4ih 20 150 ns 17 f4i width (st-bus, gci) t f4iw 195 244 295 ns f8 setup time (st-bus, gci) t f8s 10 ns 18 f8 hold time (st-bus, gci) t f8h 10 ns 18 f8 width (st-bus, gci) 8.192 mhz 16.384 mhz t f8w 50 50 122 61 145 70 ns
MT90812 advance information 80 ? see "notes" following ac electrical characteristics tables. figure 30 - frequency locking for the c4o and c10o clock ? typical ?gures are at 25?c and are for design aid only: not guaranteed and not subject to production testing. ? typical ?gures are at 25?c and are for design aid only: not guaranteed and not subject to production testing. ac electrical characteristics - 8.192 mhz master clock input characteristics sym min max units conditions/notes ? tolerance -32 +32 ppm duty cycle 40 60 % ac electrical characteristics - output clock parameters characteristics sym min typ ? max units test conditions c10o clock frequency f c10 10.24 mhz c10o output clock period t c10p 97.65 ns c10o clock duty cycle dc c10 40 50 60 % refer to table on page 95 c10o clock jitter (w.r.t. c4o) j c10 -15 +15 ns ac electrical characteristics - output frame pulse parameters characteristics sym min typ ? max units test conditions f4o output delay (high to low) from c4o t f4ol 015ns f4o output delay (low to high) from c4o t f4oh 015ns f4o width t f4ow 244 ns f8 output delay (high to low) from c8 output clock t f8ol 015ns f8 output delay (low to high) from c8 output clock t f8oh 015ns f8 width t f8ow 122 ns c4o c10o 2.0v 0.8v 3.0v 2.0v j c10o f * * the relative phase between these two clocks is not critical and may vary from 0 to tc4op for mt9171/72/73 dnic devices.
advance information MT90812 81 figure 31 - c4/f4 input clock reference - st-bus notes: 1. cmos output 2. ttl input figure 32 - c4/f4 input clock reference - gci notes: 1. cmos output 2. ttl input f4i 2 v h v l v tt c4i 2 v h v l v tt f8 1 v ct c8 1 v ct f4o 1 v ct c4o 1 v ct c2o 1 v ct t f4iw t f4is t f4ih t c8od t c4od t c2od t f8ol t f8oh t f4ol t f4oh f4i 2 v h v l v tt c4i 2 v h v l v tt f8 1 v ct c8 1 v ct f4o 1 v ct c4o 1 v ct c2o 1 v ct t f4iw t f4is t f4ih t c8od t c4od t c2od t f8ol t f8oh t f4ol t f4oh
MT90812 advance information 82 figure 33 - c8/f8 input clock reference - st-bus notes: 1. cmos output 2. ttl input figure 34 - c8/f8 input clock reference - gci notes: 1. cmos output 2. ttl input f8i 2 v h v l v tt c8 2 v h v l v tt f4o 1 v ct c4o 1 v ct c2o 1 v ct t f4ol t f8w t f4oh t f8s t f8h t c4od t c2od f8 2 v h v l v tt c8 2 v h v l v tt f4o 1 v ct c4o 1 v ct c2o 1 v ct t f4ol t f8w tf4oh t f8s t f8h tc4od tc2od
advance information MT90812 83 figure 35 - c8p input clock reference - st-bus (timing control, fpo bit =0) notes: 1. cmos output 2. ttl input figure 36 - c8p input clock reference - gci (timing control, fpo bit =1) notes: 1. cmos output 2. ttl input c8p 2 v h v l v tt f8 1 v ct c8 1 v ct f4o 1 v ct c4o 1 v ct c2o 1 v ct t f8oh t f4ol t f4oh t c8od t c4od t c2od t f8ol c8p 2 v h v l v tt f8 1 v ct c8 1 v ct f4o 1 v ct c4o 1 v ct c2o 1 v ct t f8oh t f4o l t f4oh t c8od t c4od t c2od t f8ol
MT90812 advance information 84 figure 37 - c16/f8 input clock reference - st-bus notes: 1. cmos output 2. ttl input figure 38 - c16/f8 input clock reference - gci notes: 1. cmos output 2. ttl input fp 2 v h v l v tt c16 2 v h v l v tt f8 1 v ct c8 1 v ct f4o 1 v ct c4o 1 v ct c2o 1 v ct t f8w t f8s t f8h t c8od t c4od t c2od t f8ol t f8oh t f4o l t f4oh fp 2 v h v l v tt c16 2 v h v l v tt f8 1 v ct c8 1 v ct f4o 1 v ct c4o 1 v ct c2o 1 v ct t f8w t f8s t f8h t c8od t c4od t c2od t f8o l t f8oh t f4ol t f4oh
advance information MT90812 85 figure 39 - c16 hmvip input clock reference notes: 1. cmos output 2. ttl input 24.1 timing references for tdm streams the expansion bus operates in two modes; idx link and tdm link modes (refer to section 3.0). in tdm link mode there are three data rates supported, 2.048, 4.096, and 8.192 mb/s. the idx link mode supports 8.192 mb/s data rate. depending on the clock control mode selected the serial stream ac parameters are speci?ed with respect to a given clock reference. table 22 lists the timing references used for the tdm streams given the clock control mode, pcs and pcos settings. the timing diagrams to be used for the given tdm stream and timing mode are listed for both st-bus and gci modes. the clock reference is labelled, clk, in each diagram and is either c4i, c4o, c8 (as an input or output), or c16, where c16 refers to 16.384 mhz signal applied to c8p_c16 pin. the frame pulse reference is labelled fp in each diagram and is either f4i, f4o, f8(input) or f8(output). tdm and data rate clock control mode pcs pc0s clk fp clk rate/ data rate timing diagram local 2 mb/s all 0 0 c4i f4i 2 fig. 40 (st-bus) & fig. 41 (gci) local 2 mb/s all 1 1 c4o f4o 2 fig. 40 (st-bus) & fig. 41 (gci) ebus 2 mb/s (tdm link) all 0 0 c4i f4i 2 fig. 40 (st-bus) & fig. 41 (gci) ebus 2 mb/s (tdm link) all 1 1 c4o f4o 2 fig. 40 (st-bus) & fig. 41 (gci) ebus 4mb/s (tdm link) c4/f4 x x c8 (output) f8 (output) 2 fig. 40 (st-bus) & fig. 41 (gci) f4i2 vh vl v tt c4i 2 vh vl v tt c16 2 vh vl v tt f8 1 v ct c8 1 v ct f4o 1 v ct c4o 1 v ct c2o 1 v ct t f4iw t f4is t f4ih t c8od t c4od t c2od t f8ol t f8oh t f4ol t f4oh t diff
MT90812 advance information 86 table 22 - timing references for tdm streams * clock rate/data rate=1, however the incoming data is clocked with the pll generated clock at 3 quarters into the bit cell. figure 40 - st-bus timing for local tdm bus at 2.048 mb/s and expansion bus interface at 2.048 mb/s, 4.096 mb/s and 8.192 mb/s (8.192 mb/s with 16.384 mhz clock) notes: 1. cmos output 2. ttl input c8/f8, c8p 0 0 c8 (input) f8 (input) 2 fig. 40 (st-bus) & fig. 41 (gci) c16 0 0 c16 (input) f8 (input) 4 fig. 40 (st-bus) & fig. 41 (gci) clk shown as c8 c8/f8, c8p, c16 11 c8 (output) f8 (output) 2 fig. 40 (st-bus) & fig. 41 (gci) ebus 8mb/s c4/f4 x x c8 (output) f8 (output) 1* fig. 42 (st-bus) & fig. 43 (gci) c8/f8, c8p 0 0 c8 (input) f8 (input) 1 fig. 42 (st-bus) & fig. 43 (gci) c16/f8 0 0 c16 (input) f8 (input) 2 fig. 40 (st-bus) & fig. 41 (gci) c16/hmvip 0 0 c16 (input) f4 (input) 2 fig. 44 (hmvip) c8/f8, c8p, c16 11 c8 (output) f8 (output) 1* fig. 40 (st-bus) & fig. 41 (gci) tdm and data rate clock control mode pcs pc0s clk fp clk rate/ data rate timing diagram fp 2 v h v l v tt clk 2 v h v l v tt sto/est0 1 v ct sti/est1 2 v h v l v tt t clkp t clkh t fpw t clkl t fps t fph t stod t stis t stih * fp = f4o for local or ebus tdm at 2mb/s = f8 for ebus at 4 and 8 mb/s **clk = c4o, local or ebus tdm at 2mb/s = c8/16 for ebus at 4 and 8 mb/s refer to timing references for tdm streams and table 22 on page 85.
advance information MT90812 87 figure 41 - - gci timing for local tdm bus at 2.048 mb/s and expansion bus interface at 2.048 mb/s, 4.096 mb/s and 8.192 mb/s (8.192 mb/s with 16.384 mhz clock) notes: 1. cmos output 2. ttl input figure 42 - st-bus timing for expansion bus interface at 8.192 mb/s with 8.192 mhz clock notes: 1. cmos output 2. ttl input figure 43 - - gci timing for expansion bus interface at 8.192 mb/s (8.192 mhz clock) notes: 1. cmos output 2. ttl input fp 2 v h v l v tt clk 2 v h v l v tt sto/est0 1 v ct sti/est1 2 v h v l v tt t clkl t fpw t clkh t fps t clkp t fph t stod t stis t stih * fp = f4o for local or ebus tdm at 2mb/s = f8 for ebus at 4 and 8 mb/s **clk = c4o, local or ebus tdm at 2mb/s = c8/16 for ebus at 4 and 8 mb/s refer to timing references for tdm streams and table 22 on page 85. fp 2 v h v l v tt clk 2 v h v l v tt sto/est0 1 v ct sti/est1 2 v h v l v tt t clkp t clkh t fpw t clkl t fps t fph t stod t stis t stih * fp = f4o for local or ebus tdm at 2mb/s = f8 for ebus at 4 and 8 mb/s **clk = c4o, local or ebus tdm at 2mb/s = c8/16 for ebus at 4 and 8 mb/s refer to timing references for tdm streams and table 22 on page 85. f8 2 v h v l v tt c8 2 v h v l v tt sto/est0 1 v ct sti/est1 2 v h v l v tt t clkl t fpw t clkh t fps t clkp t fph t stod t stis t stih * f8, c8 are input signals for sync or slave to c8 modes. output signals for free run or sync or slave to c4 modes. refer to timing references for tdm streams and table 22 on page 85.
MT90812 advance information 88 figure 44 - - hmvip bus timing for serial interface (8.192mb/s) notes: 1. cmos output 2. ttl input ? typical ?gures are at 25?c and are for design aid only: not guaranteed and not subject to production testing. figure 45 - output driver enable (ode) figure 46 - serial output and external control ac electrical characteristics - hmvip bus timing characteristics sym min typ ? max units test conditions f4i setup time t fps 10 ns f4i hold time t fph 20 ns f4i width t fpw 244 ns delay between rising edge of c4i and rising edge c8/16. t diff -3 3 ns fp 2 v h v l v tt c4i 2 v h v l v tt c16 2 v h v l v tt est0 1 v ct est1 2 v h v l v tt t fpw t fps t fph t stod t stis t stih t c4ip t c4ih t c4il t diff t c16p t c16h t c16l (gci mode)) t dz clk sto t zd sto hiz hiz valid data valid data clk (st-bus mode) or (hmvip mode) est0/1 est0/1 sto ode t oed hiz hiz t oed valid data
advance information MT90812 89 24.2 ac parameters referenced to incoming clock signals ac parameters are either measured with respect to incoming or outgoing clock signals. a list of the clock control modes for the local or expansion bus streams where the ac parameters are measured with respect to incoming clock signals are listed in table 23. the parametric values are listed in the following tables. table 23 - timing referenced to input clock signals for tdm streams tdm and data rate clock control mode pcs pc0s clk fp clk rate / data rate timing diagram local 2 mb/s all 0 0 c4i f4i 2 fig. 40 (st-bus) & fig. 41 (gci) ebus 2 mb/s (tdm link) all 0 0 c4i f4i 2 ebus 4mb/s (tdm link) c8/f8, c8p 0 0 c8 (input) f8 (input) 2 fig. 40 (st-bus) & fig. 41 (gci) c16 0 0 c16 (input) f8 (input) 4 fig. 40 (st-bus) & fig. 41 (gci) clk shown as c8 ebus 8mb/s c8/f8, c8p 0 0 c8 (input) f8 (input) 1 fig. 42 (st-bus) & fig. 43 (gci) c16/f8 0 0 c16 (input) f8 (input) 2 fig. 40 (st-bus) & fig. 41 (gci) c16/hmvip 0 0 c16 (input) f4 (input) 2 fig. 44 (hmvip)
MT90812 advance information 90 ? see "notes" following ac electrical characteristics tables. ? see "notes" following ac electrical characteristics tables. ac electrical characteristics - output delay parameters referenced to input clock signals characteristics sym min max units test conditions ? output driver enable delay (2.048, 4.096, 8.192 mb/s) t oed 55 ns 3 sto delay, active to active, high- z to active, active to high-z 2.048 mb/s 4.096 mb/s 8.192 mb/s t stod , t zd, t dz 65 55 55 ns ns ns non-pll clock 3, 9, 10. sto delay, active to active, high-z to active, active to high-z 2.048 mb/s - sto0,sto1 2.048 mb/s - est0 4.096 mb/s 8.192 mb/s t stod , t zd, t dz 45 45 45 45 ns ns ns ns pll output clock 3, 11, 14, 15 3, 11, 14, 16 3, 11, 14, 16 3, 11, 14, 16 clock output delays c2o c4ob c8o t c2od t c4od t c8od 50 50 40 ns ns ns non-pll clock. 3, 9,10 clock output delays c2o c4ob c8o c10o t c2od t c4od t c8od t c10od 35 35 30 35 ns ns ns ns pll output clock 3, 11, 14, 15 3, 11, 14, 15 3, 11, 14, 16 3, 11, 14 ac electrical characteristics - input set and hold parameters referenced to input clock signals characteristics sym min max units test conditions ? sti set-up time before clk rising (st-bus mode) before clk falling (gci mode) 2.048 mb/s - sto0,sto1,est1 4.096 mb/s 8.192 mb/s t stis 5 5 5 ns ns ns non-pll clock 3, 9, 10 sti set-up time before clk rising (st-bus mode) before clk falling (gci mode) 2.048 mb/s - sto0,sto1,est1 4.096 mb/s 8.192 mb/s t stis 10 10 5 ns ns ns pll clock 3, 11, 14, 15 1, 3, 11, 14, 16 2, 3, 11, 14, 16 sti hold time before clk rising (st-bus mode) before clk falling (gci mode) 2.048 mb/s - sto0,sto1,est1 4.096 mb/s 8.192 mb/s t stih 20 20 20 ns ns ns non-pll clock. 3, 9, 10 pll clock. 3, 11, 14, 15
advance information MT90812 91 24.3 ac parameters referenced to outgoing c4 or c8 clock signals ac parameters are either measured with respect to incoming or outgoing clock signals. a list of the clock control modes for the local or expansion bus streams where the ac parameters are measured with respect to the outgoing clock signals are listed in table 24. the parametric values are listed in the following tables. table 24 - timing referenced to output clock signals for tdm streams * clock rate/data rate=1, however the incoming data is clocked with the pll generated clock at 3 quarters into the bit cell. ? see "notes" following ac electrical characteristics tables. ? see "notes" following ac electrical characteristics tables. tdm and data rate clock control mode pcs pc0s clk fp clk rate/ data rate timing diagram local 2 mb/s all 1 1 c4o f4o 2 fig. 40 (st-bus) & fig. 41 (gci) ebus 2 mb/s all 1 1 c4o f4o 2 ebus 4mb/s c4/f4 x x c8 (output) f8 2 c8/f8, c8p, c16 1 1 c8 (output) f8 2 ebus 8mb/s c4/f4 x x c8 (output) f8 1* fig. 42 (st-bus) & fig. 43 (gci) c8/f8, c8p, c16 1 1 c8 (output) f8 1* fig. 40 (st-bus) & fig. 41 (gci) ac electrical characteristics - output delay parameters referenced to output c4/c8 signals characteristics sym min max units test conditions ? output driver enable delay (2.048, 4.096, 8.192 mb/s) t oed 55 ns 3 sto delay from active to high-z, active to high-z 2.048 mb/s - sto0,sto1, est0 4.096 mb/s 8.192 mb/s t dz, t zd 35 40 40 ns ns ns 3, 11, 14, 15, 19 3, 11, 14, 16, 20 3, 11, 14, 16, 20 sto delay (high and low) from clk falling (st-bus mode) from clk rising (gci mode) 2.048 mb/s - sto0,sto1,est0 4.096 mb/s 8.192 mb/s t stod 35 40 40 ns ns ns 3, 11, 14, 15, 19 3, 11, 14, 16, 20 3, 11, 14, 16, 20 ac electrical characteristics - input set and hold parameters referenced to output c4/ c8 signals characteristics sym min max units test conditions ? sti set-up time before clk rising (st-bus mode) before clk falling (gci mode) 2.048 mb/s - sto0,sto1,est1 4.096 mb/s 8.192 mb/s t stis 20 15 10 ns ns ns 3, 11, 14, 15, 19 3, 11, 14, 15, 20 3, 11, 14, 15, 20 sti hold time before clk rising (st-bus mode) before clk falling (gci mode) 2.048 mb/s - sto0,sto1,est1 4.096 mb/s 8.192 mb/s t stih 30 25 25 ns ns ns 3, 11, 14, 15, 19 3, 11, 14, 15, 20 3, 11, 14, 15, 20
MT90812 advance information 92 24.4 hra timing figure 47 - the hdlc controller related signals notes: 1. cmos output 2. ttl input 3. refer to sto1 output delay parameters. ac electrical characteristics - hra timing ? see "notes" following ac electrical characteristics tables. characteristics sym min max units test conditions ? teop setup time t teops 0 450 ns 9 teop hold time t teoph 20 450 ns 9 txcen propagation delay t txcenp 70 ns 3, 9 reop setup time t reops 0 450 ns 9 reop hold time t reoph 20 450 ns 9 rxcen propagation delay t rxcenp 70 ns 3, 9 dper to sto1 propagation delay t sto1p 65 ns 3 f82 v h v l v tt c82 v h v l v tt f4o1 v ct c4o1 v ct c2o1 v ct teop2 v h v l v tt txcen1 v ct reop2 v h v l v tt rxcen1 v ct dper2 v h v l v tt sto11,3 v ct t fpis t fpih t teops t teoph t reops t reoph t txcenp t rxcenp t sto1p t txcenp t sto1p t rxcenp t c4od t c2od
advance information MT90812 93 ? typical ?gures are at 25?c and are for design aid only: not guaranteed and not subject to production testing. ? see "notes" following ac electrical characteristics tables. ? typical ?gures are at 25?c and are for design aid only: not guaranteed and not subject to production testing. ? see "notes" following ac electrical characteristics tables. ac electrical characteristics - pll typical intrinsic jitter ? characteristics master slave conditions/notes ? uipp uipp intrinsic jitter at c2o (2.048 mhz) 0.015 0.008 12,14-16,21,25,27-32 intrinsic jitter at c4o (4.096 mhz) 0.029 0016 12,14-16,22,25,27-32 intrinsic jitter at c8 (8.192 mhz) 0.057 0.033 12,14-16,23,25,27-32 intrinsic jitter at c10o (10.24 mhz) 0.072 0.041 12,14-16,24,25,27-32 ac electrical characteristics - pll typical input to output jitter transfer for master mode ? characteristics output jitter uipp conditions/notes ? input jitter frequency khz jitter at output for input jitter frequency at 0.10uipp 0.100 0.187 12,14-16,22,25,27-32,33 4 0.187 5 0.191 6 0.194 8 0.200 10 0.207 15 0.228 20 0.247 25 0.261 28 0.257 30 0.241 34 0.224 36 0.210 40 0.180 45 0.165 50 0.150 60 0.130 70 0.119 80 0.109 90 0.105 100 0.101 200 0.088
MT90812 advance information 94 ? typical ?gures are at 25?c and are for design aid only: not guaranteed and not subject to production testing. ? see "notes" following ac electrical characteristics tables. ac electrical characteristics - pll typical input to output jitter transfer for slave mode ? characteristics output jitter uipp conditions/notes ? input jitter frequency khz jitter at output for input jitter frequency at 0.09uipp 0.001 0.077 12,14-16,22,25,27-32,34 0.002 0.125 0.003 0.147 0.005 0.150 0.010 0.156 30 0.158 40 0.164 60 0.172 80 0.193 100 0.213 120 0.240 140 0.290 160 0.326 170 0.315 180 0.285 190 0.250 200 0.221 220 0.178 240 0.150 260 0.133 300 0.112 350 0.100
advance information MT90812 95 ? typical ?gures are at 25?c and are for design aid only: not guaranteed and not subject to production testing. ? see "notes" following ac electrical characteristics tables. ac electrical characteristics - pll typical input jitter tolerance for c10o 40/60 duty cycle ? characteristics input jitter uipp conditions/notes ? input jitter frequency khz master slave input jitter tolerance for c10o 40/60 duty cycle 1 20.0 20.0 5,12,14-16,22,25,27-32,34 3 8.0 8.0 5 5.0 3.0 10 1.0 0.9 20 1.0 0.5 40 0.7 0.4 60 0.4 0.5 80 0.2 0.3 100 0.1 0.2 120 0.07 0.2 140 0.07 0.2 160 0.04 0.2 180 0.04 0.2
MT90812 advance information 96 ? see "notes" following ac electrical characteristics tables. figure 48 - intel/national multiplexed bus timing ac electrical characteristics - intel/national - hpc multiplexed bus mode characteristics sym min max unit test conditions ? 1 ale pulse width t alw 20 ns 2 address setup from ale falling t ads 5ns 3 address hold from ale falling t adh 10 ns 4 rd active after ale falling t alrd 10 ns 5 data setup from dt a low on read t ddr 5ns5 6 cs hold after rd/ wr t csrw 0ns 7 cs setup from rd t csr 0ns 8 data hold after rd t dhr 5 30 ns 3, 4 9 wr delay after ale falling t alwr 10 ns 10 cs setup from wr t csw 0ns 11 valid data delay on write t swd 0ns 12 data hold after write t dhw 10 ns 11 acknowledgment delay: register data memory or connect memory data memory a6=1 dbr, dbt and fsk fifo read dbr, dbt and fsk fifo write t akd 95 135 200 + 1/2 c8p 2xc8p + 95 1.5xc8p + 85 ns ns ns ns ns 5 5 5,7 5 5 14 acknowledgment hold time t akh 15 ns 3, 4 ale ad0-7 cs rd wr dta address data t alw t ads t adh t alrd t csr t csw t alwr t swd t akd t ddr t akh t csrw t dhw t dhr 2.0v 0.8v 2.0v 0.8v 2.0v 0.8v 2.0v 0.8v 2.0v 0.8v 2.0v 0.8v
advance information MT90812 97 ? see "notes" following ac electrical characteristics tables. figure 49 - motorola multiplexed bus timing ac electrical characteristics - motorola multiplexed bus mode characteristics sym min max units test conditions ? 1 as pulse width t asw 20 ns 2 address setup from as falling t ads 5ns 3 address hold from as falling t adh 10 ns 4 data setup from dt a low on read t ddr 0ns5 5 cs hold after ds falling t csh 0ns 6 cs setup from ds rising t css 0ns 7 data hold after write t dhw 10 ns 8 valid data delay on write t swd 0ns 9r/ w setup from ds rising t rws 0ns 10 r/ w hold after ds falling t rwh 0ns 11 data hold after read t dhr 5 30 ns 3, 4 12 ds delay after as falling t dsh 10 ns 13 acknowledgment delay: register data memory or connect memory data memory a6=1 dbr, dbt and fsk fifo read dbr, dbt and fsk fifo write t akd 95 135 200 + 1/2 c8p 2xc8p + 95 1.5xc8p + 85 ns ns ns ns ns 5 5 5,7 5 5 14 acknowledgment hold time t akh 15 ns 3, 4 ds/ rd r/ w as/ale ad0-ad7 ad0-ad7 dta t rws t akd 2.0v 0.8v 2.0v 0.8v data 2.0v 0.8v t asw t dsh t rwh 2.0v 0.8v address 2.0v 0.8v 2.0v 0.8v data address 2.0v 0.8v t ads t adh t swd t dhw write read 2.0v 0.8v t css t dhr t csh cs 2.0v 0.8v t ddr t akh
MT90812 advance information 98 note: 1. ds pin is used as ds in motorola non-multiplexed mode. ? see "notes" following ac electrical characteristics tables. figure 50 - motorola non-multiplexed bus timing. note: 1. ds pin is used as ds in motorola non-multiplexed mode. ac electrical characteristics - motorola non-multiplexed bus mode characteristics sym min max unit test conditions ? 1 cs setup from ds 1 falling t css 0ns 2r/ w setup from ds 1 falling t rws 0 3 address setup from ds 1 falling t ads 5ns 4 address hold after ds 1 falling t adh 10 ns 5 cs hold after ds 1 rising t csh 0 6r/ w hold after ds 1 rising t rwh 0 7 data setup from dt a low on read t ddr 5ns5 8 data hold on read t dhr 5 30 ns 3, 4 9 valid data delay on write t swd 0ns 10 data hold on write t dhw 10 ns 11 acknowledgment delay: register data memory or connect memory data memory a6=1 dbr, dbt and fsk fifo read dbr, dbt and fsk fifo write t akd 95 135 200 + 1/2 c8p 2xc8p + 95 1.5xc8p + 85 ns ns ns ns ns 5 5 5,7 5 5 12 acknowledgment hold time t akh 15 ns 3, 4 ds 1 r/ w a0-a12 d0-d7 dta 2.0v 0.8v t ads t adh t css t csh cs t swd t rws 2.0v 0.8v t rwh 2.0v 0.8v 2.0v 0.8v 2.0v 0.8v read d0-d7 2.0v 0.8v t dhr write 2.0v 0.8v t akd t ddr t akh t dhw
advance information MT90812 99 notes: voltages are with respect to ground (vss) unless otherwise stated. supply voltage and operating temperature are as per recommended operating conditions. timing parameters are as per ac electrical characteristics - timing parameter measurement voltage levels 1: measured with respect to 3/4 bittime, which for 4.096 mb/s is 244ns x 3/4 = 183ns from beginning of bit cell 2: measured with respect to 3/4 bittime, which for 8.192 mb/s is 122ns x 3/4 = 91.5ns from beginning of bit cell 3: cl= 150pf, rl=1 k load. 4: high impedance is measured by pulling to the appropriate rail with rl with timing corrected to cancel time to discharge cl. 5: cl= 150pf capacitive load. 6: ds pin is used as ds in motorola non-multiplexed mode. 7: acknowledgment may be held off for read of dm page a6=1. 8: pll disabled. 9: timing control register bit pcs = 0. 10: output clock control register bit pcos = 0. 11: jitter on reference input is less than 1ns pp. 12: applied jitter is sinusoidal. 13: applied jitter is a squarewave of 8khz of 32ns pp. 14: pll enabled. 15: timing control register bit pcs = 1. 16: output clock control register bit pcos = 1. 17: measured w.r.t. to c4i. 18: measured w.r.t. to c8 input clock. 19: measured w.r.t. to c4ob. 20: measured w.r.t. to c8 output clock. 21: 1 uipp = 488ns for 2.048mhz signals. 22: 1 uipp = 244ns for 4.096mhz signals. 23: 1 uipp = 122ns for 8.192mhz signals. 24: 1 uipp = 97.65ns for 10.24mhz signals. 25: no ?lter 26: input clock reference selected is not c4f4. 27: c4f4 input clock reference selected. 28: c8p input clock reference selected. 29: c8f8 input clock reference selected. 30: c16f8 input clock reference selected. 31: c16 input clock reference and hmvip selected. 32: jitter on reference input is 0.008 uipp. 1ui=244ns. 33: pll is in master mode. 34: pll is in slave mode. w
package outlines plastic j-lead chip carrier - p-suf?x f d 1 d h e 1 i a 1 a g d 2 e e 2 dim 20-pin 28-pin 44-pin 68-pin 84-pin min max min max min max min max min max a 0.165 (4.20) 0.180 (4.57) 0.165 (4.20) 0.180 (4.57) 0.165 (4.20) 0.180 (4.57) 0.165 (4.20) 0.200 (5.08) 0.165 (4.20) 0.200 (5.08) a 1 0.090 (2.29) 0.120 (3.04) 0.090 (2.29) 0.120 (3.04) 0.090 (2.29) 0.120 (3.04) 0.090 (2.29) 0.130 (3.30) 0.090 (2.29) 0.130 (3.30) d/e 0.385 (9.78) 0.395 (10.03) 0.485 (12.32) 0.495 (12.57) 0.685 (17.40) 0.695 (17.65) 0.985 (25.02) 0.995 (25.27) 1.185 (30.10) 1.195 (30.35) d 1 /e 1 0.350 (8.890) 0.356 (9.042) 0.450 (11.430) 0.456 (11.582) 0.650 (16.510) 0.656 (16.662) 0.950 (24.130) 0.958 (24.333) 1.150 (29.210) 1.158 (29.413) d 2 /e 2 0.290 (7.37) 0.330 (8.38) 0.390 (9.91) 0.430 (10.92) 0.590 (14.99) 0.630 (16.00) 0.890 (22.61) 0.930 (23.62) 1.090 (27.69) 1.130 (28.70) e 0 0.004 0 0.004 0 0.004 0 0.004 0 0.004 f 0.026 (0.661) 0.032 (0.812) 0.026 (0.661) 0.032 (0.812) 0.026 (0.661) 0.032 (0.812) 0.026 (0.661) 0.032 (0.812) 0.026 (0.661) 0.032 (0.812) g 0.013 (0.331) 0.021 (0.533) 0.013 (0.331) 0.021 (0.533) 0.013 (0.331) 0.021 (0.533) 0.013 (0.331) 0.021 (0.533) 0.013 (0.331) 0.021 (0.533) h 0.050 bsc (1.27 bsc) 0.050 bsc (1.27 bsc) 0.050 bsc (1.27 bsc) 0.050 bsc (1.27 bsc) 0.050 bsc (1.27 bsc) i 0.020 (0.51) 0.020 (0.51) 0.020 (0.51) 0.020 (0.51) 0.020 (0.51) notes: 1) not to scale 2) dimensions in inches 3) (dimensions in millimeters) 4) for d & e add for allowable mold protrusion 0.010" e: (lead coplanarity) general-10
m mitel (design) and st-bus are registered trademarks of mitel corporation mitel semiconductor is an iso 9001 registered company copyright 1999 mitel corporation all rights reserved printed in canada technical documen t a tion - n o t for resale world headquarters - canada tel: +1 (613) 592 2122 fax: +1 (613) 592 6909 north america asia/paci?c europe, middle east, tel: +1 (770) 486 0194 tel: +65 333 6193 and africa (emea) fax: +1 (770) 631 8213 fax: +65 333 6192 tel: +44 (0) 1793 518528 fax: +44 (0) 1793 518581 http://www.mitelsemi.com information relating to products and services furnished herein by mitel corporation or its subsidiaries (collectively mitel) is believed to be reliable. however, mitel assumes no liability for errors that may appear in this publication, or for liability otherwise arising from the application or use of any such information, product or service or for any infringement of patents or other intellectual property rights owned by third parties which may result from such application or use. neither the supply of such information or purchase of product or service conveys any license, either express or implied, under patents or other intellectual property rights owned by mitel or licensed from third parties by mitel, whatsoever. purchasers of products are also hereby noti?ed that the use of product in certain ways or in combination with mitel, or non-mitel furnished goods or services may infringe patents or other intellectual property rights owned by mitel. this publication is issued to provide information only and (unless agreed by mitel in writing) may not be used, applied or reproduced for any purpose nor form part of any order or contract nor to be regarded as a representation relating to the products or services concerned. the products, their speci?cations, services and other information appearing in this publication are subject to change by mitel without notice. no warranty or guarantee express or implied is made regarding the capability, performance or suitability of any product or service. information concerning possible methods of use is provided as a guide only and does not constitute any guarantee that such methods of use will be satisfactory in a speci?c piece of equipment. it is the users responsibility to fully determine the performance and suitability of any equipment using such information and to ensure that any publication or data used is up to date and has not been superseded. manufacturing does not necessarily include testing of all functions or parameters. these products are not suitable for use in any medical products whose failure to perform may result in signi?cant injury or death to the user. all products and materials are sold and services provided subject to mitels conditions of sale which are available on request.

▲Up To Search▲

Price & Availability of MT90812

	To Download MT90812 Datasheet File
If you can't view the Datasheet, Please click here to try to view without PDF Reader .