View M66332FP_217793.PDF datasheet online --- IC-ON-LINE

Datasheet File OCR Text:

MITSUBISHI DIGITAL ASSP MITSUBISHI DIGITAL ASSP
M66332FP M66332FP
FACSIMILE IMAGE DATA PROCESSOR FACSIMILE IMAGE DATA PROCESSOR
DESCRIPTION The M66332 is a facsimile image processing controller that converts analog signals that are photoelectrically converted by an image sensor into bi-level signals. It has image processing functions such as peak detection, uniformity correction, correction, MTF compensation, detector of background and object levels, dither control, separation of image data area, scale down, and area specification. This controller has a built-in 5-bit flash type A-D converter and interface circuits to image sensor, analog signal processing circuit, and CODEC (Coder & Decoder) to simplify control of the readout mechanism. FEATURES 0 High Speed Scan (MAX. 2 ms/line, TYP. 5 ms/line) 0 A3 (8 pixels/mm) Line Sensor Attachment 0 Image sensor (CCD,CIS) control signal generation CCD: SH, CK1, CK2, RS Contact sensor (CIS): SH, CK1 (or CK2)
0 Analog signal processing circuit control signal generation CLAMP, S/H, AGC, DSCH 0 Built-in 5-bit Flash Type A-D Converter 0 Bi-level data external input/output interface Serial output (M66330) 8-bit MPU bus output with external DMA control signal 0 Image data processing correction Uniformity correction (block correction in units of 8 pixels) MTF compensation (1 dimension) Detector of background and object level (programmable) Dithering control * Dither method (16 levels using 4 x 4 matrix) Separation of image data area (1 dimension) Scale down A3 B4, A3 A4, B4 A4 0 5V Single Power Supply APPLICATION Facsimiles
PIN CONFIGURATION (TOP VIEW)
D7 D6 D5 D4 MPU Interface D3 (DATA) D2 D1 D0 GND VCC A3 A2 MPU Interface A1 (ADDRESS) A0 DGND DVCC
44 43 42 41 40 39 38 37 36 35 34 33 32 31 30
GND VCC CS RD MPU Interface WR RESET GND GND DRQ DMA Interface DAK VCC SH Sensor Interface
29
45 46 47 48 49 50 51 52 53 54 55 56
M66332FP
28 27 26 25 24 23 22 21 20 19 18 17
TEST ASIG NC VWL VML3 VML2 GND VML1 VBL AGND AVCC DSCH
Test pin Analog Signal Processing Interface White Basic Supply Voltage Middle Basic Supply Voltage 3 Middle Basic Supply Voltage 2 Middle Basic Supply Voltage 1 Black Basic Supply Voltage
10
11
12
13
14
15
CK2 Sensor Interface CK1 RS Analog Signal CLAMP Prosessing S/H Interface GND Test pin TEST System clock SYSCK Test pin TEST VCC SVID CODEC SCLK Interface STIM SRDY Sensor Interface PTIM Analog Signal AGC Processing Interface
16
Analog Signal Processing Interface
1
2
3
4
5
6
7
8
9
NC: No Connection
Outline 56P6N-A
1
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
BLOCK DIAGRAM
System Clock SYSCK 8 PTIM SH Sensor CK1 Interface CK2 RS Analog Signal Interface CLAMP S/H AGC DSCH 15 56 2 1 3 4 5 16 17 ADC Analog Vcc AVCC 18 ADC Logic Vcc DVCC 29
VCC 10 35 46 55
Image Processing Sequence Control Signal
Sensor Control To each block
Detection of Image Data Area
Separation of Image Data Area
Convert to bi-level
Cur out/ Scale down
14 13 12 11
SRDY STIM CODEC SCLK Interface SVID
Analog Control Correction memory (304 words x 5bits)
MTF compensation
Simple Bi-level Conversion/Background and object Level Detection
DMA Control
54 DAK DMA 53 DRQ Interface RESET CS MPU Interface RD WR A0 MPU Interface 34 A3 (Address) 37 D0 MPU 44 D7 Interface (Data)
~ ~
Uniformity Correction
SRAM 16 words x 4bits
Collective Dithering (16 levels) MPU Bus Interface
50 47 48 49 31
ASIG 27
5bit A-D converter (flash type)
20 21 23 24 25 VBL VML2 VWL VML1 VML3 ADC Reference Voltage
19 AGND ADC Analog GND
30 6 22 36 45 51 52 DGND GND ADC Logic GND
Table 1 Image Processing Functions
Image Processing Function Read Width Resolution Read speed Uniformity Correction MTF Compensation Simple Bi-level Conversion Pseudo half-tone Separation of Image Data Area Scale down Correction Image Sensor Control Signal Analog Signal Processing Specifications * A4, B4, A3 * 8 pixels/mm (primary scanning direction) * 5ms/line Typ. 2ms/line maximum * White correction only * Block correction in units of 8 pixels * 50% Correction range * Laplacian filter circuit for 3 x 1 pixels in current line (1 dimension) * Floating threshold method using background and object level detection circuit * Dither method: 16 levels (4 x 4matrix) * Detection by brightness difference in 5 x 1 pixels area in current line * Selection method * Scale down: A3 B4 set to 13/15; B4 A4, 9/11; A3 A4, 12/17 * Logarithmic correction * Control signal generation for contact sensor (CIS) and scale down CCD * Generate control signals for external CLAMP circuit, sample/hold circuit, and AGC circuit Remarks
* Operated with system clock and PRE_DATA (registers 2, 3) * Built-in SRAM as correction memory (304 words x 5bits) (read/write allowed from MPU) * No need for compensation memory
* Built-in SRAM as dither memory (16 words x 4bits) (read/write allowed from MPU) * No need for processing memory
* Apply external voltage (resistor connection is also allowed) to A-D converter middle basic supply voltage pins.
* Built-in 5-bit flash A-D converter
2
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
PIN DESCRIPTIONS
Block Sensor Interface Pin Names SH CK1 I/O O O Description CCD: Shift pulse signal to transmit photo charges from the sensor to the transfer unit. CIS: Start signal for the sensor read circuit. CCD: Clock pulse signal for sequentially transmitting the transfer unit signal charge of the sensor. CIS: Clock pulse signal for the sensor read circuit shift register. Reverse of CK1. Pulse to reset the voltage of the CCD sensor floating capacitor to initial status. Read roller pulse motor control signal. CLAMP pulse to set the dark level of the sensor to reference voltage of the digital circuit. Sample-hold signal to smooth out sensor image signal waveform. Analog signals. External AGC circuit gain down signal. External AGC circuit gain up signal. Data transmission ready signal from CODEC. Data transmission bound signal for CODEC. Clock signal for transmitting image data to CODEC. Serial output of image data to CODEC. "H": Black; "L": White. DMA request signal to external DMA controller for parallel output of image data through MPU bus. DMA acknowledge signal from external DMA controller for the above DRQ signal. System clock input pin. System reset signal. Resets counter, register, F/F, and latch, sets internal memory in standby mode, and halts clock generation circuit. Chip select signal used by MPU to access M66332. Set to "H" in operating mode (AGC, UNIF, SCAN). Control signal used by MPU to read data from M66332. Control signal used by MPU to write data to M66332. Address signals used to access M66332 internal registers. 8-bit bidirectional buffer.
Analog Circuit Interface
CODEC Interface
DMA Interface
CK2 RS PTIM CLAMP S/H ASIG AGC DSCH SRDY STIM SCLK SVID DRQ DAK SYSCK RESET CS RD WR A0~A3 D0~D7
O O O O O I O O I O O O O I I I I I I I I/O
Clock MPU Interface
3
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
PIN DESCRIPTIONS (CONTINUED)
Block Others Pin Names VCC AVCC DVCC GND AGND DGND VWL VBL VML1 VML2 VML3 TEST(IN) TEST(OUT) I/O -- -- -- -- -- -- -- -- -- -- -- -- -- Description Plus supply voltage. Plus supply voltage for A-D converter analog units. Plus supply voltage for A-D converter logic units. GND pin. Ground for A-D converter analog units. Ground for A-D converter digital units. A-D converter white basic supply voltage pin. A-D converter black basic supply voltage pin. Middle basic supply voltage pin. VML1 =(VWL - VBL) /4 Middle basic supply voltage pin. VML2 =2 * (VWL - VBL) /4 Middle basic supply voltage pin. VML3 =3 * (VWL - VBL) /4 Test input pin. Fix to "L". Test output pin. Keep open.
FUNCTIONAL DESCRIPTION The following items which are necessary to use the image processing functions of the M66332 are described. (1) Operating mode (2) Line period and read sequence (3) Image processing function (4) Sensor unit/analog signal processing unit interface (5) CODEC interface (6) Read/write to dither memory and uniformity correction memory (7) Reset (8) Image quality control using registers
4
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
(1) Operating mode The M66332 performs three basic operations. * Peak value detection: The peak value of the analog signal output from the analog signal processing circuits is matched to the white reference voltage (VWL) of the M66332 internal A-D converter. (See also Figs. 19 to 22 in the M66333FP document.) * Uniformity correction data creation: White reference data is created for sensor unit uniformity correction and written to the correction memory (SRAM: 304 words x 5bits). * Read operation: A document is read and the image is processed to output bi-level data as serial or parallel output. These three basic operations are performed in the following sequence depending on whether the sensor is CCD or CIS. The sensor is selected with register 0 (SENS). When the sensor is CCD: UNIF mode Operation is started by setting the UNIF command in register 0 to "H". If the sensor is CCD, peak detection (16 line periods) and white uniformity correction data creation (8 line periods) are performed consecutively. To exit this operating mode, wait 30 line periods (at least 24 lines) from the start and set the UNIF command to "L". The read operation is started by setting the SCAN command in register 0 to "H". Set the SCAN command to "L" to exit this operation mode.
When the sensor is CIS: (Creation and transmission of uniformity correction data) UNIF mode This mode is started when the UNIF command in register 0 is set to "H". When the sensor is CIS, if white correction is started with the UNIF command, peak detection (16 line periods) and uniformity white correction data creation (8 line period) are started. To exit this operating mode, wait 30 line periods (at least 24 line periods) from the start and set the UNIF command to "L". The uniformity correction data pertaining to white correction created in UNIF mode are transferred to the backup memory.
v Data transfer
(Read operation) AGC mode Peak detection is performed for 16 line periods when the AGC command in register 0 is set to "H". To exit this operating mode, wait 20 line periods (at least 16 lines) from the start and set the AGC command to "L". The read operation is started by setting the SCAN command in register 0 to "H". Set the SCAN command to "L" to exit this operating mode.
v SCAN mode
v SCAN mode
The signal functions and data flow in each mode are shown on pages 4-123 and 4-124. Flowcharts are shown on pages 4-158 to 4-160.
5
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
Operation During Peak Detection
SYSCK Image sensor
Image Processing Sequence Control Signal PTIM SH CK1 CK2 RS Sensor Control To each block Detection of Image Data Area Separation of Image Data Area Convert to bi-level Cut out/ Scale down STIM SCLK SVID CODEC SRDY Analog Control Resolution compensation Simple Bi-level Conversion/Background and object level detection DAK DMA Control DRQ SRAM 16 words x 4bits Collective Dithering (16 levels) MPU Bus Interface RESET CS,RD WR A0~A3 MPU ASIG 5bit A-D converter (flash type) D0~D7 DMA
Analog signal processing circuit
Correction memory (304 words x 5bits) CLAMP S/H AGC DSCH
Uniformity Correction
VBL VML2 VWL VML1 VML3
Data Flow in Creation of Uniformity Correction Data
SYSCK Image sensor
Image Processing Sequence Control Signal PTIM SH CK1 CK2 RS Sensor Control To each block Detection of Image Data Area Separation of Image Data Area Convert to bi-level Cut out/ Scale down STIM SCLK SVID CODEC SRDY Analog Control Simple Bi-level Conversion/Background and object level detection DAK DMA Control DRQ SRAM 16 words x 4bits Collective Dithering (16 levels) MPU Bus Interface RESET CS,RD WR A0~A3 MPU ASIG 5bit A-D converter (flash type) D0~D7 DMA
Analog signal processing circuit
Correction memory (304 words x 5bits) CLAMP S/H AGC DSCH
Uniformity Correction
VBL VML2 VWL VML1 VML3
6
MTF Compensation
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
Date Flow During Read Operation (for serial output)
SYSCK
Image sensor
Image Processing Sequence Control Signal PTIM SH CK1 CK2 RS Sensor Control To each block Detection of Image Data Area Separation of Image Data Area Convert to bi-level Cut out/ Scale down STIM SCLK SVID CODEC SRDY Analog Control Simple Bi-level Conversion/ Background and object level detection DAK DMA Control DRQ RESET CS,RD WR A0~A3 MPU D0~D7 DMA
Analog signal processing circuit
Correction memory (304 words x 5bits) CLAMP S/H AGC DSCH
Uniformity Correction
MTF Compensation
SRAM 16 words x 4bits
Collective Dithering (16 levels)
MPU Bus Interface
ASIG
5bit A-D converter (flash type)
VBL VML2 VWL VML1 VML3
: Image date : Correction, Compensation date
Date Flow During Read Operation (for parallel output)
SYSCK
Image sensor
Image Processing Sequence Control Signal PTIM SH CK1 CK2 RS Sensor Control To each block Detection of Image Data Area Separation of Image Data Area Convert to bi-level Cut out/ Scale down STIM SCLK SVID CODEC SRDY Analog Control
Resolution compensation
Analog signal processing circuit
Correction memory (304 words x 5bits) CLAMP S/H AGC DSCH
Simple Bi-level Conversion/ Background and object level detection
DAK DMA Control DRQ RESET CS,RD WR A0~A3 MPU D0~D7 DMA
Uniformity Correction
SRAM 16 words x 4bits
Collective Dithering (16 levels)
MPU Bus Interface
ASIG
5bit A-D converter (flash type)
VBL VML2 VWL VML1 VML3
: Image date : Correction, compensation date
7
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
(2) Line period and read sequence Figure 1 shows the relationship between the M66332 line period and the read sequence. * 1 line period (1/ACCK): Defines the processing time per line for M66332. The line period is determined from the line period counter registers 2 and 3 (PRE_DATA) and pixel transmission clock (ADCK). ADCK is 1/ 16th of SYSCK. 1 line period (1/ACCK) [NS] = line period counter x pixel transmission clock period [NS] = (PRE_DATA + 1) x 1/ADCK [NS] = (PRE_DATA + 1) x 16/SYSCK [NS] The line period counter is counted down with the pixel transmission clock after loading the PRE_DATA value and generates the following addresses. * Sensor start pulse (SH): Image sensor start pulse. The position of the start pulse is determined by the value in register 4 (ST_PL) which is the offset from the uniformity correction range (UNIFG). Set ST_PL to the following values according to the type of image sensor. CCD: ST_PL = sensor dummy pixel + 2 CIS: ST_PL =2
* Uniformity correction range (UNIFG):
* AGC range (AGCG):
* Source document read width:
* Pulse motor control signal (PTIM):
Defines the uniformity correction range. This range corresponds to the sensor width (A3 to A4). Refer to Table 2 for the relationship between sensor width and uniformity correction range. Defines the peak detection range. This range corresponds to the sensor width (A3 to A4). Auto gain control is performed for the entire width (solid line) of the sensor in AGC mode and for the range inside the sensor width (dotted line) in SCAN mode. Refer to Table 2 for the relationship between sensor width and AGC range. Defines the source document read width. If the document width is less than the sensor width, the document should be centered on the sensor because the read range is set from the center of the sensor. Refer to Table 3 for the relationship between sensor width and source document read width. Generates the pulse motor control signals for the read roller.
PRE_DATA load Line period (ACCK) Sensor start pulse (SH) Uniformity correction range (UNIFG) AGC range (AGCG)
Down count
0
Relationship with registers Registers 2, 3 (PRE_DATA) Register 0 (SENS_W) Register 4 (ST_PL)
ST_PL
Register 0 (SENS_W)
Register 0 (SENS_W) Register 1 (SOURCE)
Source document read width Pulse motor control (PTIM)
Register 1 (SOURCE, DEST , REDU)
1 line period
Fig. 1 Line period and read sequence
8
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
1 line period (ACCK) Sensor start pulse (SH) SRDY (SSCAN) (BCSCAN) STIM SCLK SVID PTIM (SSCAN), (BCSCAN) : Internal signals : Output bound
Fig. 2 CODEC Interface and read sequence
Table 2 Sensor width and gate signal range
Sensor width Gate signal Uniformity correction range (UNIFG) AGC range (AGCG) AGC mode SCAN mode A3 2487/55 2487/55 2370/162 B4 2279/231 2279/231 2194/306 A4 2119/391 2119/391 1760/740
Table 3 Source document read width according to sensor width and source document size
Sensor width Source document size A3 B4 A4 A3 2487/55 2278/230 2118/390 B4 -- 2278/230 2118/390 A4 -- -- 2118/390
X/Y X : Left end address Y : Right end address
X
Y
9
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
(3) Image processing function The M66332 converts image signals from the image sensor to bi-level signals. Bi-level conversion can be either simple bilevel conversion or pseudo half-tone conversion which converts image shades into bi-levels. The signal output from the image sensor must be corrected and compensated to reduce distortion and degradation before it can be converted to bi-level signals. Furthermore, for reduction in transmission time, separation of image data area and optimum bi-level conversion must be performed. The functions necessary for image processing are described below. * Peak detection * Uniformity correction * MTF compensation * Background and object level detection (simple bi-level conversion) * Pseudo half-tone dither method * Separation of image data area * Image scale down/area specification
* Peak detection The A-D converter of the M66332 is used with its reference voltages (VWL, VBL) fixed. Normallly, VWL is set to VCC and VBL, is set to 0V to keep the dynamic range of the A-D converter wide. Peak detection must be performed for analog signals to match them with the full scale value of the A-D converter before they are input to the A-D converter. Peak detection is performed by reading white data in AGC mode, one of the three M66332 operating modes (AGC, UNIF, SCAN). In AGC mode, 8-line period worth of DSCH signal to raise gain--for gain control--and 16-line period worth of AGC signal to lower gain--for the overflowing of the A-D converter-- are generated after AGC command start (register 0: AGC) as shown in Fig. 3. This changes the gain as shown in Fig. 4.
Peak detection Line period DSCH (raise gain) AGC A-D converter overflow (lower gain) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Peak detection preprocessing
Peak gain control
: Generated in pixel units only when the A-D converter overflows.
Fig. 3 Peak detection
Peak detection (preprocessing) After completing peak detection preprocessing VWL White data Match the output level of the last pixel of a line to VWL VWL
Peak gain control After completing gain control for peak value Match the peak sensor output within a line toVWL
VBL
1 line
VBL
1 line
Fig. 4 Changes in gain during peak detection
10
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
* Uniformity correction Uniformity correction corrects the drop in lighting level at both ends of the light source, shading distortion due to drop in lighting level at the rim of the lens, and high frequency distortion caused by the scattering of pixel-unit image sensor characteristic (see Fig. 5). The M66332 creates uniformity correction data in UNIF mode, one of the three operating modes (AGC, UNIF, SCAN), handling 8 pixels as a unit as shown in Fig. 6. The created data is written to the internal correction memory (SRAM: 304 words x 5 bits). In SCAN mode, the correction data is read from the internal correction memory to successively correct the input image data in pixel units.
Black level High frequency distortion Shading distortion
White level
1 line
Fig. 5 Image sensor white data output waveform
Primary scanning direction 1 8 9 16 17 24
Pixels
Correction data 1
Correction data 2
Correction data 3
For pixels from 1 to 8, uniformity correction is carried out by correction data 1. For pixels from 9 to 16, uniformity correction is carried out by correction data 2. For pixels from 17 to 24, uniformity correction is carried out by correction data 3.
Fig. 6 Creation of uniformity correction data
11
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
* Correction The M66332 performs entire pixel correction for 50% correction range as shown in Fig. 7.
Correction is not possible if the white correction data exceeds the 50% correction range as shown in Fig. 7. Therefore, be sure to keep the input signal within the correction range.
White Correction Analog signal input White data VWL 25-1
50%
24-1
VBL 1 line
0
If correction range is exceeded (during white correction) Analog signal input White data VWL 25-1
50%
24-1 White data exceeding correction range
VBL 1 line
0
Portion exceeding correction range
Fig. 7 Uniformity correction
12
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
* MTF compensation As shown in Fig. 8, characters and photos that have been photoelectrically converted by the sensor unit are characterized by a drop in resolution. The MTF compensation per-
formed by the M66332 enhances the high frequency components with a Laplacian filter to maintain the resolution of the image data and creates a perception of increased dynamic range.
Photoelectric conversion
Photoelectric conversion
Source document (characters)
Image signal MTF compensation
Compensated data Photoelectric conversion
Image signal Source document (photo) MTF compensation
Compensated data MTF compensation X' = X + ((X - A) + (X - B)) = Primary scanning compensation coefficient, register 5 (MTF)
Secondary scanning direction Window 3 x 1 F
G Primary scanning direction F-G
A
X
B
Pixel
Fig. 8 MTF Compensation
13
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
* Background and object level detection The M66332 uses the floating threshold method rather than the fixed threshold method. This method successively generates a threshold for optimum simple bi-level conversion of the target pixel. Therefore, a threshold matching the picture data is generated without modifying the image data. This value is used as the threshold of the bi-level area when simple bi-level conversion or image separation is selected as bi-level conversion mode. : register 5 (MODE) Background level counter If an image data greater (brighter) than the current counter value is input, this counter is incremented to approach the image data. If an image data less (darker) than the current counter value is input, this counter is decremented to approach the image data.
The count up/down speed can be set with the following register. : register 9 (MAX_UP, MAX_DOWN) The lower limit of the background level can be set with the following register. : register B (LL_MAX) Object level counter If an image data greater (brighter) than the current counter value is input, this counter is incremented to approach the image data. It an image data less (darker) than the current counter value is input, the image data is set to this counter. The count down speed can be set with the following register. : register 9 (MIN_UP) The upper limit of the character level can be set with the following register. : register A (UL_MIN)
Image data
Background level counter Threshold generation Object level counter Compare Bi-level data
White level Background level
Threshold level Input data Object level Threshold level = (background level - object level) x K + object level K = bi-level threshold coefficient: register 9 (SLICE) Black level
Fig. 9 Background-object level
14
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
* Dither method The M66332 has a built-in 16 words x 4 bits SRAM which is used as a collective dithering memory. During initialization, threshold values are written in the dither memory, matching the desired dither pattern into 4 x 4 dither matrix. : register E (DITH_D)
Fig. 10 shows some examples of dither patterns. Refer to the section on dither memory and uniformity correction memory read/write for details on how to read/write the dither memory. This is used when dither method and image data area separation are selected for bi-level conversion mode during read. : register 5 (MODE)
4 x 4 matrix sequence
0 12 3 15 8 4 11 7 2 14 1 13 10 6 9 5 4 15 9 14 10 0 3 8 5 2 1 13 11 6 12 7 11 12 7 3 4 0 8 15 6 2 10 13 9 14 5 1
Table 4 Scanning line density and dither matrix size
Scanning Line Density Normal Fine Primary/Secondary Scanning Line (line/mm) 8 x 3.85 8 x 7.7 Level -- 16 Matrix Size -- 4x4
a)Diffused type
b) Concentrated type
c) Mesh type
Fig. 10 Collective dither pattern
15
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
* Separation of image data area In order to perform bi-level conversion appropriate for the image, a black and white image is separated into bi-level conversion area and gradation conversion area. Simple bi-level
conversion is applied to the bi-level conversion area and dither method is applied to the gradation area. : register 5 (MODE)
Area Black and white image Bi-level conversion area Gradation conversion area
Process Simple bi-level conversion Pseudo half-tone conversion
Example Character Photo
There is no significant change in illumination of the gradation conversion area (photo) when a black and white image is viewed through a 5 x 1 window. This characteristic is used to distinguish between gradation conversion area and bi-level conversion area.
Secondary scanning direction 5 x 1 window Lmax: maximum illumination in window Lmin : minimum illumination in window Current line Primary scanning direction X
Pixel
Identification equation 1 Lmax - Lmin > A (illumination difference in bi-level conversion area) : register 6 difference (SEPA_A) Identification equation 2 Lmin > B (area is entirely white) : register 7 minimum (SEPA_B) Identification equation 3 Lmax < C (area is entirely black) : register 8 maximum (SEPA_C) Simple bi-level conversion if identification equation 1, 2, or 3 is satisfied. Pseudo half-tone conversion when none of the identification equations is satisfied.
Difference White level = 15 Lmax
Minimum
Lmin
Maximum
A Input data Lmin Black level = 0 Lmax - Lmin Lmax B C
Fig. 11 Separation of image data area
16
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
* Image scale down/area specification Scale down function The image data input from the analog signal processing circuit can be scaled down (A3 B4, A3 A4, B4 A4) by leaving out pixels in the primary scanning direction for bi-level conversion. : register 1 (SOURCE, DEST, REDU) Scale down in secondary scanning direction can be performed in the same rate by MPU program. Table 5 Scaling rate
IN OUT B4 A4 A3 13/15 12/17 B4 1 9/11 A4 -- 1
Area specification function When area specification is selected, bi-level conversion is performed only in the specified area from the center of the source document as shown Fig. 12. : register 1 (SOURCE, DEST, REDU)
Source document width
Specified area
Fig. 12 Cut out function
17
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
(4) Sensor unit/analog signal processing unit interface CCD-bit clamp type
SH 1 2 RS When ST_PL is odd CK1= 1 CK2= 2 When ST_PL is even CK1= 2 CK2= 1
Sensor output (OS) S/H
CLAMP
OUT (ASIG) (A-D converter clock) Register 4: STPL = dummy pixel + 2 Dummy area Valid pixel area
Unit: 1/SYSCK SH 2 1 2 16 RS (OS) N 16 S/H 16 CLAMP 13 2 1 34 9 8 4 4 12 2 16
OUT (ASIG)
N
(A-D converter clock)
18
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
CCD-line clamp type
SH 1 2 RS When ST_PL is odd CK1= 1 CK2= 2 When ST_PL is even CK1= 2 CK2= 1
Sensor output (OS) S/H (Line clamp area) CLAMP
OUT (ASIG) (A-D converter clock)
8: fixed Register 4: STPL = dummy pixel + 2 Dummy area Valid pixel area
Note: Line clamp uses sensor output equivalent to (dummy area -8) pixels from the first pixel after SH.
Unit: 1/SYSCK SH 2 1 2 16 RS 8 4 4 12 2 16
(OS) S/H CLAMP
16 34 9
N
34
9
N OUT (ASIG)
(A-D converter clock)
19
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
CIS type
SH
CK1 CK2 Sensor output (SIG)
S/H CLAMP
OUT (ASIG) (A-D converter clock) Resister 4: STPL = 2 Unit: 1/SYSCK Select with register 5 SH_W 16 CK1 6 8 2
SH 16
CK2
6
8
2 16
(SIG) 16 4 S/H 2 16 16 CLAMP 2 1 1 1
N
Select with register 5 S/H_W
OUT (ASIG)
N
(A-D converter clock)
Note: CLAMP: In case of CIS, check with the sensor manufacturer for the use of CLAMP. SH and CK1, CK2: SH can be selected with register 5 and CK can be selected with CK1 and CK2 (2 choices each) to provide interface with various types of CIS.
20
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
(5) CODEC interface Serial output
SRDY
SH
A STIM 10 4 12
B
SCLK 2 10 2
SVID Unit: 1/SYSCK
Note: A is determined by register 4 (ST_PL), and B is determined by register 1 (SOURCE, DEST, REDU).
Parallel output
Pixel SCLK SVID DRQ DAK RD D0 D1 D2 D3 D4 D5 D6 D7 N-1 N-2 N-3 N-4 N-5 N-6 N-7 N-8 N-1 N N+1 N+2 N+3 N+4 N+5 N+6 N+7 N+8 1 2 3 4 5 6 7 8
Note: Handshaking of three lines SRDY, SH, and STIM, which are interface to the CODEC, is the same as serial output.
21
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR ing dither patterns in the 16 words x 4 bits collective dithering SRAM built in the M66332.
(6) Read/write to dither memory, uniformity correction memory The following figures show the sequence for writing and readDither memory write (MPU M66332)
Initial setting (1)
Initial setting (2)
Memory address (0)
Memory address (1)
CS
A3 ~ A0
1H
0H
EH
EH
~
WR
D7 ~ D0 (Input)
D4 = "0" 1
D7 = "1" 2
DATA (0) 3
DATA (1) 3
Dither memory read (M66332 MPU)
Initial setting (1) Initial setting (2) Memory address (0) Memory address (1)
CS
A3 ~ A0
1H
0H
EH
EH
~
WR D7 ~ D0 (Input)
D4 = "0"
D7 = "1"
RD D7 ~ D0 (Output) 1 2
DATA (0) 4
DATA (1) 4
Clear D4 (PO) in register 1 to "0" in order to set the MPU bus (D7 - D0) to dither matrix memory data output mode. Set D7 (RESET) in register 0 to "1" in order to reset the dither memory address counter. Select DITH_D with register E and write DATA (0) on the MPU bus (D5 - D0). Increment the address counter of the dither memory at the rising edge of WR. (during write) Select DITH_D with register E and read DATA (0) in dither memory to the MPU bus (D5 - D0). Increment the address counter of the dither memory at the rising edge of RD. (during read) Dither matrix address
A0 A4 A8 A1 A5 A2 A6 A3 A7
A9 A10 A11
A12 A13 A14 A15
4 x 4 matrix
22
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
The M66332 can read/write uniformity correction data in the external correction SRAM through the MPU bus. This enables the uniformity correction data to be temporarily saved Uniformity correction memory write (MPU M66332)
in backup memory during power off. The following figures show the uniformity correction data read/write sequence.
Initial setting CS A3~A0 WR D7~D0 (Input) D7 = "1" D4 = "0" 1 2 DATA 3 1H 2H 3H FH Repeats steps 1, 2, and 3 .
Closing setting
1H
D7 = "0"
Uniformity correction memory read (M66332 MPU)
Initial setting CS A3~A0 WR D7~D0 (Input) RD D7~D0 (Output) 1 2 DATA 4 D7 = "1" D4 = "0" 1H 2H 3H FH Repeats steps 1, 2, and 4 . 1H Closing setting
D7 = "0"
D7 D6 D5 D4 D3 D2 D1 D0 R2 A4 A3 A2 A1 A0 0 R3 0 0 0 0
0 A8 A7 A6 A5
D_LOAD : 0 for normal 1 for data load A8~A0 : UNIF memory address
The last 5 digits (A4 - A0) of an address in the UNIF memory are written in register 2. The initial 4 digits (A8 - A5) of the address in the UNIF memory and D_LOAD = "1" (D6) are written in register 3. Steps and identifies the address in the UNIF memory. The UNIF memory is selected with register F, and DATA on the MPU bus (D4 - D0) is written at the identified address. The UNIF memory is selected with register F, and DATA stored at the identified address is read to the MPU bus (D4 - D0). Initial setting: D7 (UM_R/W) and D4 (P0) of register 1 are set to "1" and "0", respectively, to select read/write mode of uniformity correction memory. Closing setting : D7 (UM_R/W) of register 1 is set to "0" while D4 (P0) is set to that taken in operation, to cancel read/write mode of uniformity correction memory.
Address Space
Sensor width A3 B4 A4 Left end address 310 284 264 Right end address 7 29 49
23
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
Dither memory write/read
SYSCK Image sensor
Image Processing Sequence Control Signal PTIM SH CK1 CK2 RS Sensor Control To each block Detection of Image Data Area Separation of Image Data Area Convert to bi-level Cut out/ Scale down STIM SCLK SVID CODEC SRDY Analog Control Simple Bi-level Conversion/ Background and object level detection DAK DMA Control DRQ RESET CS,RD WR A0~A3 MPU ASIG 5bit A-D converter (flash type) D0~D7 DMA
Analog signal processing circuit
Correction memory (304 words x 5bits) CLAMP S/H AGC DSCH
MTF Compensation
Uniformity Correction
SRAM 16 words x 4bits
Collective Dithering (16 levels)
MPU Bus Interface
VBL VML2 VWL VML1 VML3
: Dither date
Uniformity correction memory write/read
SYSCK Image sensor
Image Processing Sequence Control Signal PTIM SH CK1 CK2 RS Sensor Control To each block Detection of Image Data Area Separation of Image Data Area Convert to bi-level Cut out/ Scale down STIM SCLK SVID CODEC SRDY Analog Control Simple Bi-level Conversion/ Background and object level detection DAK DMA Control DRQ RESET CS,RD WR A0~A3 MPU ASIG 5bit A-D converter (flash type) D0~D7 DMA
Analog signal processing circuit
Correction memory (304 words x 5bits) CLAMP S/H AGC DSCH
MTF Compensation
Uniformity Correction
SRAM 16 words x 4bits
Collective Dithering (16 levels)
MPU Bus Interface
VBL VML2 VWL VML1 VML3
: Uniformity correction date
24
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
(7) Reset The M66332 has three types of reset. Each reset function is described below. Hard reset: Initializes the circuit. Hard reset also performs the following soft reset and standby reset. Soft reset: Used when cancelling a line read operation in the middle during SCAN mode. Read operation is resumed starting from the next line.
Standby :
Used as standby mode. The internal clock is stopped by stopping the clock generator which generates the internal clock from the system clock. Therefore, the internal circuit is stopped and power is saved. The period counter and register statuses are saved and the internal memory is placed in standby mode.
Table 6 Reset function
Function Reset Type Hard Reset RESET Soft Reset Register 0 (RESET) Standby Register1 (STNBY) Initialize Register Initialize Internal F/F Reset Period Counter Stop Clock Generator Operation Stop Line Read
(8) Image quality control using registers * MTF compensation If the sensor has high resolution, resolution compensation need not be performed for half-tone area. MTF compensation should be performed for bi-level area regardless of the sensor resolution in order to achieve good object reproduction. * Simple bi-level conversion, background and object level detection Set the background level detection and object level detection counters as follows in order to obtain clear output of objects that do not have completely white background and that are not entirely black. fast MAX_UP > MAX_DOWN > MIN_UP The output becomes darker as bi-level conversion threshold coefficient (SLICE) is increased. Select a large SLICE value for light source document.
* Pseudo half-tone conversion, dither method Select collective dithering (16 gradations using 4 x 4 dither matrix) for fine mode. Refer to the section on image processing function for details on providing dither pattern threshold. * Separation of image data area The optimum parameter is selected to perform the best bilevel conversion for each area: simple bi-level conversion for the object and pseudo half-tone conversion for half-tone.
25
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
Table 7 shows the recommended values for parameters related to picture quality.
Use these values as reference to determine the optimum parameter.
Table 7 Recommended parameter values
Image Uniformity Correction Resolution Compensation MTF Background and Object Level SLICE MAX UP 5/8 Normal MAX DOWN Normal MIN UP Normal UL MIN 04H LL MAX 0AH Correction Dither Pattern Separation of Image Data Area SEPA SEPA SEPA A B C -- -- --
Simple Bi-Level Conversion
Yes
1/2
No =0.9 VML1=1.1V VML2=2.2V VML3=3.5V =0.9 VML1=1.1V VML2=2.2V VML3=3.5V
-- 4x4 diffusion pattern, = 0.8 4x4 diffusion pattern, = 0.8
Dithering
Yes
MON
--
--
--
--
--
--
--
--
--
Separation of Image Data Area
Yes
MON
5/8
Normal
Normal
Normal
04H
0AH
06H
0DH
01H
0
8
2
10
1
12 4
6
14
2
6
8 4
9
A
3 11
3
1
C 2
13 5
2
15 7
9 5
7 3
D
B
Dither pattern ( = 0.8)
26
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR * correction correction is performed to simulate the sensitivity characteristics (exponential nature) of the human eye in order to make the image data more similar to natural image. = 0.45 is said to be the optimum correction when using a thermal head printer. The M66332, due to its capacity to handle 4-bit internal data, performs correction by means of both collective dithering and the middle reference voltage pins (VML1, VML2, and VML3) of the A-D converter. ( Correction by Collective Dithering ) correction is realized applying a characteristic to the threshold value to be written in the dither matrix as shown in Fig. 15. The example given in Fig. 15 is an approximation of characteristic, , to 0.8. ( Correction by the Middle Reference Voltage Pins of the AD converter) The example shown in Fig. 16 is an approximation of characteristic, , to 0.9 , which is carried out by applying VML1 = 1.1V, VML2 = 2.2V, and VML3 = 3.5V to the middle reference pins of the A-D converter. Fig. 23 in the M66332FP leaflet shows an example of circuits for applying voltages to middle reference voltage pins.
USAGE PRECAUTIONS * Peak detection in SCAN mode In SCAN mode, successive peak detection is performed for the image data being read as shown for the AGC range (dotted line) in Fig. 1. This enables better picture reproduction when picture data brighter than the white reference used during peak detection is input in SCAN mode. This is especially effective for sensor units such as CIS that do not have a built-in white reference. * Read operation with CIS sensor If the sensor is CIS, it is possible to select whether or not to use white correction in SCAN mode. Do not select white correction for the input of analog signals already processed by entire pixel correction. * Collective dithering Thresholds written in dither matrix should be between 1 and 15 excluding 0 as shown in Fig. 13. As the M66332 carries out block correction in 8-bit units for uniformity correction, a CIS sensor may generate background noises due to irregularity of pixels. It is possible to remove noises and gain a fine image quality by reducing the maximum threshold value as shown in Fig. 14.
15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Threshold
15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
Threshold sequence
Threshold sequence
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Threshold
Fig. 13 Thresholds for collective dithering : Example 1
15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 ;0.8
Fig. 14 Thresholds for collective dithering : Example 2
White 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 Black 0 Black = 0V
A-D converter output
Threshold sequence
=1 ;0.9
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Threshold
VML1 =1.1V
VML2 VML3 =2.2V =3.5V Analog input
White = 5V
Fig. 15 An example of correction by dither matrix
Fig. 16 An example of correction by middle reference pins
27
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
* TIME function When TIME = "1" is set in register 1, the processing time per line is doubled to 2 line periods, Data is read once every two line periods and processed.
When the read and write motors operate simultaneously during copy operation, this command can be used to reduce the processing speed to 1/2 in order to reduce the power load.
1 line period (ACCK) Sensor start pulse (SH) SRDY (SSCAN) (BCSCAN) STIM PTIM
N period
N + 1 period N + 2 period
Note: (SSCAN), (BCSCAN): internal signals Read is performed when (SSCAN) is "H". Data read is processed when (BCSCAN) is "H".
Fig. 17 When processing speed is 1/2
28
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
Register Structure
Address 0H R/W W D7 RESET D6 SENS D5 D4 Explanation D3 UMODE D2 AGC D1 UNIF D0 SCAN (Default is 00H)
SENS_W
D7 0 1
RESET System Reset Normal Mode Reset Mode
Reset while write pulse is "L" when D7 = "1"
D6 0 1
SENS Sensor Type CCD CIS
D5 0 0 1 1
D4 0 1 0 1
SENS_W Sensor Width A4 B4 A3 --
D3 0 1
UMODE CIS Uniformity Correction Mode With White Correction No White Correction
* For selecting with or without correction in SCANning (with CIS only)
D2 0 1
AGC AGC mode Stop Start
* Controls AGC mode start/stop.
D1 0 1
UNIF UNIF Mode Stop Start
* Controls UNIF mode start/stop.
D0 0 1
SCAN SCAN Mode Stop Start
* Controls SCAN mode start/stop.
29
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
Address 1H
R/W W D7 UM_R/W D6 STNBY D5 TIME D4 P_0
Explanation D3 D2 D1 DEST D0 REDU (Default is 00H)
SOURCE
D7 0 1
UM_R/W Uniformity Correction Memory Read/ Write Normal UNIF Memory Read/Write STNBY Standby Mode Normal Standby Mode * Standby mode stops the clock generation circuit. The period counter and register status are saved and the internal memory is placed in standby mode. * When read and write operations are performed together as in copy operation, the power load can be reduced by selecting 2 line period. The processing speed drops to 1/2 when 2 line period is selected. * D0 is output in LSB format and D7, in MSB format. When SCAN data is output in SCAN mode, D7 is output in LSB (left) format and D0, is output in MSB (right) format.
D6 0 1
D5 0 1
TIME Line Time 1 Line Period 2 Line Period
D4 0 1
P_O Parallel Output Without Parallel Output Parallel Output
D3 0 0 1 1
D2 0 1 0 1
SOURCE Source Width A4 B4 A3 --
D1 0 1
DEST Destination Width A4 B4
D0 0 1
REDU Scale down/Cut out Cut out Scale down
Scaling Rate A3 B4 13/15 B4 A4 9/11 A3 A4 12/17
* Refer to image scale down/ area specification for scale down/cut out.
30
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
Address 2H
R/W W D7 D6 D5 D4
Explanation D3 D2 D1 D0 (Default is 00H)
PRE_DATA <7:0>
D7~D0: PRE_DATA <7:0> Pre Data of Line period Counter (Lower part) D7~D4: If register 3 D_LOAD = "1" these bits will be the address denoted by lower 5 digits (A4~A0) used for read/write operations on the uniformity correction memory.
3H W D7 0 D6 D_LOAD D5 D4 D3 D2 D1 D0 (Default is 00H) PRE_DATA <12:8>
D6 0 1
Uniformity correction memory address setting mode. Normal Data load
* This bit is for address setting for the access form MPU to the uniformity correction memory. Set this bit to normal during access operation.
D4~D0: PRE_DATA <12:8> Pre Data of Line period Counter (Upper part) * 1 line period is determined from PRE_DATA and pixel transmission clock frequency (ADCK). ADCK is 1/16 of system clock. Refer to line period and read sequence section. D3~D0: If register 3 D_LOAD = "1" these bits will be the address denoted by upper 4 digits (A8~A5) used for read/write operations on the uniformity correction memory.
4H
W
D7
D6
D5
D4
D3
D2
D1
D0 (Default is 00H)
ST_PL <6:0>
D6~D0: ST_PL <6:0> Start Pulse of Line sensor ST_PL = (sensor dummy pixel + 2) Refer to line period and read sequence Section.
31
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
Address 5H
R/W W
D7 D6 S/H_W D5 SH_W D4 CLAMP
Explanation
D3 MODE D2 D1 MTF D0 (Default is 00H)
D6 0 1
S/H_W Pulse Width of S/H Normal (Sysclk period x 4) Normal x 0.5
D5 0 1
SH_W SH pulse width for CIS Normal Reverse of (normal x 2)
* In case of CCD, there is only one SH pulse width and this register is ignored.
D4 0 1
CLAMP Clamp Method of Analog Circuit Bit Clamp Line Clamp
* Line clamp does not apply to CIS.
D3 0 0 1 1
D2 0 1 0 1
MODE Bi-level Mode Simple Bi-level Dither Separation (Simple Bi-level + Dither) --
D1 0 0 1 1 6H W D7
D0 0 1 0 1
MTF Main Coefficient of MTF Compensation NON(0) A little less (1/4) Middle (1/2) A little over (1)
D6
D5
D4
D3
D2
D1
D0 (Default is 00H)
SEPA_A
D3~D0: SEPA_A Separation of Image Data Area (Difference)
7H
W D7 D6 D5 D4 D3 D2 D1 D0 (Default is 00H) SEPA_B
D3~D0: SEPA_B Separation of Image Data Area (MIN.)
32
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
Address 8H
R/W W D7 D6 D5 D4
Explanation D3 D2 D1 D0 (Default is 00H)
SEPA_C
D3~D0: SEPA_C Separation of Image Data Area (MAX.)
9H
W D7 SLICE D6 D5 D4 D3 D2 D1 D0 (Default is 00H)
MAX_UP
MAX_DOWN
MIN_UP
D7 0 0 1 1
D6 0 1 0 1
SLICE Detector of Background and Object levels (SLICE) Normal (4/8) Light (3/8) Dark (5/8) Darker (6/8)
D5 0 0 1 1
D4 0 1 0 1
MAX_UP Detector of Background level (Up Counter CLK) Normal (T = (1 pixel period) x 32) Slow (T = (1 pixel period) x 64) Fast (T = (1 pixel period) x 16) Faster (T = (1 pixel period) x 8)
D3 0 0 1 1
D2 0 1 0 1
MAX_DOWN Detector of Background level (Down Counter CLK) Normal (T = (1 pixel period) x 128) Slow (T = (1 pixel period) x 256) Fast (T = (1 pixel period) x 64) Faster (T = (1 pixel period) x 32)
D1 0 0 1 1
D0 0 1 0 1
MIN_UP Detector of Object level (Up Counter CLK) Normal (T = (1 pixel period) x 512) Slow (T = (1 pixel period) x 1024) Fast (T = (1 pixel period) x 256) Faster (T = (1 pixel period) x 128)
T: Counter clock period
33
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
Address AH
R/W W D7 D6 D5 D4
Explanation D3 D2 D1 D0 (Default is 06H)
UL_MIN
D3~D0: UL_MIN Detector of background and object levels (upper limit of object level)
BH W D7 D6 D5 D4 D3 D2 D1 D0 (Default is 07H) LL_MAX
D3~D0: LL_MAX Detector of background and object levels (lower limit of background level)
EH R/W
D7
D6
D5
D4
D3
D2 DITH_D
D1
D0
D3~D0: DITH_D Internal dither memory data * Refer to the section on dither memory and uniformity correction memory read/write for information concerning read/write method.
FH R/W D7 D6 D5 D4 D3 D2 UNIF_D D1 D0
D4~D0: UNIF_D Internal uniformity correction data * Refer to the section on dither memory and uniformity correction memory read/write for information concerning read/write method.
34
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
ABSOLUTE MAXIMUM RATING (Ta = -20 ~ 75C, unless otherwise noted)
Symbol VCC VI VO AVCC VWL VBL VML VAIN Tstg Parameter Supply voltage Input voltage Output voltage Analog supply voltage Reference voltage (White) Reference voltage (Black) Reference voltage (Middle) Analog Input voltage Storage temperature range Conditions Ratings -0.3 ~ +7.0 -0.3 ~ VCC + 0.3 0 ~ VCC VCC-0.3 ~ VCC+0.3 -0.3 ~ AVCC+0.3 -0.3 ~ AVCC+0.3 -0.3 ~ AVCC+0.3 -0.3 ~ AVCC+0.3 -65 ~ 150 Unit V V V V V V V V C
RECOMMENDED OPERATING CONDITIONS
Symbol VCC GND VI AVCC AGND VWL VBL VAIN Topr Parameter Supply voltage GND voltage Input voltage Analog supply voltage Analog GND voltage (Note) Reference voltage (White) Reference voltage (Black) Analog input voltage ASIG Operating temperature range Min. 4.5 0.0 4.5 3 0.0 VBL -20 Limits Typ. 5.0 0.0 5.0 0.0 0.0 Max. 5.5 VCC 5.5 AVCC 1.0 VWL 75 Unit V V V V V V V V C
Note: Connect AGND with GND externally.
35
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
ELECTRICAL CHARACTERISTICS (Ta = -20 ~ 75C, VCC = 5 V 10%, unless otherwise noted)
Symbol VIH VIL VT+ VT- VH VOH VOL VOH VOL IIH IIL IOZH IOZL IAIN RL SINL "H" Input voltage "L" Input voltage Parameter SYSCK, SRDY, DAK, CS, RD, WR, A0~A3, D0~D7 Test conditions Min. 2.0 Limits Typ. Max. Unit V 0.8 2.4 0.6 0.2 IOH=-12mA IOL=12mA IOH=-4mA IOL=4mA VCC=5.5V VI=5.5V VCC=5.5V VI=0V VCC=5.5V VI=5.5V VCC=5.5V VI=0V 1.0 VCC=5.0V VCC=5.5V VI=VCC, GND VCC=5.5V VI=VCC, GND VCC=5.5V VI=VCC,GND 0.5 1.0 VCC-0.8 0.55 VCC-0.8 0.55 1.0 -1.0 5.0 -5.0 10 V V V V V V V V A A A A A k LSB
Positive-going threshold voltage Negative-going threshold voltage RESET Hysteresis voltage "H" output voltage D0~D7 "L" output voltage "H" output voltage DRQ, SH, CK1, CK2, RS, PTIM, CLAMP, S/H, AGC, DSCH, "L" output voltage STIM, SCLK, SVID "H" input current "L" input current SYSCK, SRDY, DAK,RESET, CS, RD, WR, A0~A3
ICCS
ICCA ICC
Off-state "H" output current D0~D7 Off-state "L" output current ASIG (Standby) Analog input current Reference resistance A-D converter Non-linear error (Note 1) Quiescent supply current (Standby) (Note 2) Quiescent supply current (Active state) (Note 2) Dynamic supply current SYSCK=8MHz
10
20
mA
15 40
40
mA mA
Note 1: The A-D converter has a 5-bit resolution. 2: Current flowing in the reference resistor in the A-D converter is not included.
36
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
TIMING REQUIREMENTS (Ta = -20 ~ 75C, VCC = 5 V 10%, unless otherwise noted)
Symbol tc(SYS) tw+(SYS) tw-(SYS) tr(SYS) tf(SYS) tw(RD) tsu(CS-RD) tsu(A-RD) tsu(DAK-RD) th(RD-CS) th(RD-A) th(RD-DAK) tw(WR) tsu(CS-WR) tsu(A-WR) tsu(D-WR) th(WR-CS) th(WR-A) th(WR-D) th(STIM-SRDY) Parameter Period High-level pulse width Low-level pulse width Rise time Fall time Pulse width Setup time Setup time Setup time Hold time Hold time Hold time Pulse width Setup time Setup time Setup time Hold time Hold time Hold time Hold time Test conditions Min. Limits Typ. 125 62.5 62.5 Max. Unit ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns
System clock
20 20 CS A0~A3 DAK CS A0~A3 DAK CS A0~A3 D0~D7 CS A0~A3 D0~D7 SRDY 100 20 20 20 10 10 10 100 20 20 50 20 10 0 0
Read pulse
Write pulse
STIM
37
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
SWITCHING CHARACTERISTICS (Ta = -20 ~ 75C, VCC = 5 V 10%, unless otherwise noted)
Symbol tPZL(RD-D) tPZH(RD-D) tPLZ(RD-D) tPHZ(RD-D) tPHL(RD-DRQ) Parameter Output enable time to low-level and high-level (RD-D) CL=150pF Output disable time from low-level and high-level (RD-D) High-level to low-level output propagation time (RD-DRQ) CL=50pF 10 50 50 ns ns Test conditions Min. Limits Typ. Max. 75 Unit ns
Test Circuit
INPUT VCC OUTPUT VCC
RL=1k SW1 SW2 CL GND RL=1k
P.G
DUT
Parameter tPLH, tPHL tPLZ tPHZ tPZL tPZH
SW1 Open Closed Open Closed Open
SW2 Open Open Closed Open Closed
50
(1)The pulse generator (PG) has the following characteristics (10%~90%) : tr = 3 ns, tf = 3 ns (2)The capacitance CL = 150pF includes stray wiring capacitance and the probe input capacitance.
System Clock
tc(SYS) tf(SYS) tw+(SYS) tw-(SYS) 3V tr(SYS)
90% SYSCK 1.3V 1.3V 1.3V 10%
90%
10% 0V
38
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
MPU Interface (1) Read timing (M66332 MPU)
3V CS 1.3V tsu(CS-RD) 1.3V 0V th(RD-CS) 3V A0~A3 1.3V 1.3V 0V tsu(A-RD) tw(RD) th(RD-A) 3V RD 1.3V 1.3V 0V tPZL(RD-D) tPLZ(RD-D)
D0~D7
50% 10% tPZH(RD-D) 90% tPHZ(RD-D) VOH VOL
D0~D7
50%
(2) Write timing (MPU M66332)
3V CS 1.3V tsu(CS-WR) 1.3V 0V th(WR-CS) 3V A0~A3 1.3V tsu(A-WR) tw(WR) 1.3V 0V th(WR-A) 3V WR 1.3V tsu(D-WR) 1.3V 0V th(WR-D) 3V D0~D7 1.3V Valid data 1.3V 0V
39
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
DMA Timing Read timing (M66332 System bus)
VOH DRQ 50% 120/SYSCK 50% VOL tPHL(RD-DRQ) 3V DAK 1.3V tsu(DAK-RD) tw(RD) 1.3V 0V th(RD-DAK) 3V RD 1.3V tPZL(RD-D) 1.3V 0V tPLZ(RD-D)
D0~D7
50% 10% tPZH(RD-D) tPHZ(RD-D) 90% VOH VOL
D0~D7
50%
CODEC Interface
th(STIM-SRDY) 3V SRDY 1.3V 0V
VOH STIM 50% VOL
VOH SCLK VOL VOH SVID VOL
40
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
FLOWCHART Read Operation (Sensor: CCD)
Start Power ON Soft reset S/H, SH pulse width CLAMP Output format Period counter Start pulse Register 0 Register 5 Wait 1 line period x 2 Turn off light
A
Set standby mode Y
Register 1
Power OFF?
Initial settings
Register 1 Registers 2, 3 Register 4
N Y
Next document
Next document? N
Image data area separation parameter Registers 6, 7, 8 Background and object level parameter Registers A, B Dither pattern write N Next document Complete? Y Register 1 Register E Power OFF End
Reset standby mode Turn on light Stabilize (white reference) Start uniformity correction Wait 1 line period x 30 (24 or more)
Register 0
AGC + white compensation AGC: 16 times white compensation : 8 times End uniformity correction Register 0
Document width, specification width Register 1 Bi-level mode Image processing function Register 5 Registers 5, 9
Start document read N Complete? Y End document read Y
Register 0
Read a page
Register 0
Transfer? N A
Setting for each White compensation document
41
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
Uniformity correction data creation, transmission (sensor: CIS)
Start Power ON Soft reset S/H, SH pulse width CLAMP Output format Period counter Start pulse Turn on light stabilize Start uniformity correction Wait 1 line period x 30 (24 or more) Register 0 Transmit uniformity Uniformity correction correction data /white Register 0 Register 5 Initial settings
Register 1 Registers 2, 3 Register 4
AGC + white compensation AGC: 16 times white compensation : 8 times
Turn off light
End uniformity correction Transmit correction data N Complete? Y End
Register 0 Registers 1, 2, 3, F
42
MITSUBISHI DIGITAL ASSP
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
Read operation (sensor: CIS)
Start A Power ON Soft reset S/H, SH pulse width CLAMP Output format Period counter Start pulse Register 0 Register 5 Power OFF? Register 1 Registers 2, 3 Register 4
Initial settings
Wait 1 line period x 2 Turn off light Set standby mode Y Register 1
N Y
Next document
Next document? N
Image data area separation parameter Registers 6, 7, 8 Background and object level parameter Registers A, B Dither pattern write N Complete? Y Write to uniformity correction memory Register F N Next document Complete? Y Reset standby mode Turn on light stabilize (white reference) Register 1 Register E
Power OFF End
Wait 1 line period x 20 (16 or more) End AGC
(AGC: 16 times)
Document width, specification width Register 1 Bi-level mode Image processing function Register 5 Registers 5, 9
N
Complete? Y End document read Y Transmit? N A Register 0
Read a page
Start document read
Register 0
Settings for each document
Register 0
AGC
Start AGC
Register 0
43

▲Up To Search▲

Price & Availability of M66332FP

	To Download M66332FP Datasheet File
If you can't view the Datasheet, Please click here to try to view without PDF Reader .