View M66335FP_2355969.PDF datasheet online --- IC-ON-LINE

Datasheet File OCR Text:

1 mitsubishi digital assp ? M66335FP facsimile image data processor mitsubishi digital assp ? M66335FP facsimile image data processor nc a0 a1 a2 a3 a4 cs rd wr reset dak drq int sysck gnd nc mpu interface dma interface system clock nc dgnd dvcc vbl adc black reference output vwl adc white reference output adin adc input avcc vri+ adc white reference input agnd vri- adc black reference input bcmo bcmv bcmi c2 c1 nc control signal for analog signal processing nc gnd v cc d7 d6 d5 d4 d3 d2 d1 d0 gnd v cc test6 test5 test4 test3 test2 test1 test0 testi testo v cc gnd mpu interface test pin nc nc v cc acck svid sclk stim srdy ptim gnd v cc rs ck1 ck2 sh gnd v cc av cc ain levaj agnd gcao nc nc M66335FP 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 40 39 38 37 36 35 34 33 32 31 30 29 28 27 26 25 63 62 61 60 59 58 57 56 55 54 53 52 51 50 49 48 47 46 45 44 43 42 41 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? y ? ? ? t ? ? ? ? ? t ? t ? t ? t ? ? ? ? ? ? ? t y t ? y ? t single-line cycle clock codec interface sensor interface sensor interface control signal for analog signal processing pin configuration (top view) outline 80p6nCa nc: no connection description the m66335 is a facsimile image processing controller to turn into binary signals analog signals which have been output through photo- electric conversion by the image sensor. the image processing functions includes peak value detection, uniformity correction, resolution change, mtf compensation, g cor- rection, detection of background/character levels, error diffusion, separation of image zones, and designation of regions. this controller contains not only the analog processing circuit, the a/ d converter of a 7-bit flash type and image processing memory, but also the image sensor and the interface circuit to the codec (coder and decoder). therefore, this lsi alone is capable of image pro- cessing. features ? high speed scan (max. 2 ms/line, typ. 5 ms/line) ? compatibility with up to the b4 (8 pixels/mm, 16 pixels/mm) image sensor ? generation of control signals for the image sensor (ccd, cis) for ccd: sh, ck1, ck2, rs for the contact sensor (cis): sh, ck1, ck2 ? built-in analog processing circuit (equivalent to the m64291) sample and hold circuit gain control circuit black level clamping circuit reference internal power supply for the a/d converter ? built-in a/d converter of a 7-bit flash type ? built-in image processing memories uniformity correction memory, line memory, error memory, g cor- rection memory ? external output interface for converted binary data serial output ( ? m66330) dma output ? external output interface for multivalued data dma transfer of data compensated for uniformity ? various image processing functions uniformity correction resolution change from 50% to 200% (by the 1% step) mtf compensation (2-dimensional processing, capable of correc- tion for each character/photo) g correction (capable of correction for each character/photo) detection of background/character levels change to pseudo-halftone ? error diffusion (64 tone steps through 6-bit processing) ? organized dither (64 tone steps through the 8 8 matrix) image zone separation (2-dimensional processing) ? 5v single power supply application facsimile, word processor and image scanner
mitsubishi digital assp ? M66335FP facsimile image data processor 2 block diagram sensor control analog control analog signal processing circuit 7-bit a/d converter uniformity correction resolution change g correc- tion table error diffusion detection of background/ character levels dither matrix image bus interface mpu bus interface dma control image processing sequence control signal mtf compensa- tion image zone separation correction data memory conversion table memory line memory organized dither error memory selection of conversion- to-binary processing ain sh vwl vbl vri+ vri- d0 ~ d7 resetb drq dakb srdyb acck sysck adin 9 12 13 14 15 ptimb rs ck2 ck1 19 20 26 27 28 levaj c1 c2 bcmi bcmv 29 35 22 30 gcao bcmo 73 72 71 54 61 66 70 a0 ~ a4 rdb wrb csb 74 75 76 8 5 6 7 stimb sclk svid 78 4 36 37 33 31 18 34 av cc 38 dv cc 3 11 v cc 17 42 52 62 21 32 agnd 39 dgnd 10 16 gnd 41 53 63 79 } } int 77 table 1 image processing functions specifications ? a4, b4 ? 8 pixels/mm, 16 pixels/mm (for the horizontal scanning direction) ? typ.: 5 ms/line; max.: 2 ms/line ? white correction, black correction ? correction range: 50% ? logarithmic correction ? laplacian filter circuit through 2-dimensional processing ? floating slice system through the detection circuit for background/character levels ? error diffusion: 6-bit processing (for 64 tone steps) ? organized dither: 8 8 matrix (for 64 tone steps) ? 2-dimensional processing through luminance difference ? range of the reduction rate: 50% to 100% (by the 1% step) ? range of the enlargement rate: 100% to 200% (by the 1% step) ? cis image sensor (clock duty: 75%) ? ccd image sensor ? the sample/hold circuit, gain control amplifier, black level clamping circuit, and 7-bit a/d con- verter are built-in. remarks ? controlled through the system clock. ? correction memory is built in. ? readable from/writable in mpu ? g correction memory is built in. ? capable of correction for each character/photo. ? correction memory is built in. ? capable of correction for each character/photo. ? error buffer memory is built in. ? 64w 6 bits dither memory is built in. ? capable of outputting the average line of a dropped line and the subsequent line instead of both lines ? capable of outputting the average line of a repeated line and the subsequent line instead of the repeated line image processing functions reading range resolution reading speed uniformity correction g correction mtf compensation simple conversion to binary pseudo-halftone image zone separation image reduction image enlargement image sensor control signal analog processing
3 mitsubishi digital assp ? M66335FP facsimile image data processor description of digital pin functions item sensor interface codec interface dma interface clock mpu interface others pin name sh ck1 ck2 rs ptim srdy stim sclk svid drq dak int sysck acck reset cs rd wr a0 ~ a4 d0 ~ d7 v cc gnd testi, 0~6 testo function outputs the shift pulse signal to transfer electric charges from the sensors photoconductor component to its transferring component for ccd and the start signal to start the sensor reading circuit for cis. outputs the clock pulse signal to sequentially transfer out signaling electric charges from the sensors transferring component for ccd and the clock pulse signal for the shift register of the sensor reading circuit for cis. reversed-phase pulses of ck1 outputs the reset pulse to return the voltage at the floating capacitor of the ccd sensor to the initial one. outputs the pulse motor control signal for the reading roller. transfer start ready signal for data from codec defines the data transfer section to codec. clock signal to transfer image data to codec outputs image data in serial to codec dma request signal to the external dma controller to output in parallel image data through the mpu bus dma acknowledge signal from the external dma controller in response to the above drq signal single-line termination interrupt system clock input pin single-line cycle clock input of the system reset. the cycle counter, register, f/f, and latch are reset. chip select signal for mpu to access the m66335 control signal for mpu to read data from the m66335 control signal for mpu to write data to the m66335 address signal to access various registers inside the m66335 8 bit two way buffer positive power supply pin gnd pin test input pin. hold this at l. test output pin. set this open. input/output output output output output output input output output output output input output input output input input input input input input/output C C input output
mitsubishi digital assp ? M66335FP facsimile image data processor 4 item power supply gnd sensor signal input part gain control circuit black level clamping circuit a/d converter pin name avcc dvcc agnd dgnd ain c1, c2 levaj gcao bcmi bcmv bcmo vri+ vriC adin vwl vbl input/output C C C C input input input output input input output input input input output output function analog power supply pin (rated supply voltage: 5v) digital power supply pin (rated supply voltage: 5v) analog ground pin digital ground pin pin to input analog signals output from ccd or cis (signals from ccd are input through capacity coupling and those from cis, with no clamping levels, are input directly.) pin to control the frequency characteristic of the gain control circuit pin to control the dc level of output signals of the gain control circuit. the output voltage, v gcao , is obtained by the following equation: v gcao = v levaj + g v v in , where, v levaj : voltage at levaj v in : input signal g v : gain of the gain control circuit v in is the signal element corresponding to the signal level clamped through the input clamping circuit for ccdcis3 input or to the gnd level for cis1cis2 input. signal output pin of the gain control circuit signal input pin to the black clamping circuit. use this with capacity coupling with the gcao pin. pin to set the black level clamping voltage. sets the black level of signals output from the bcmo pin for ccd signal processing. signal output pin of the black level clamping circuit output of the circuit to generate the a/d full-scale point reference voltage (3.8v). connected with vwl through the buffer inside the ic. to change the a/d reference voltage range, input a dc voltage from this pin. output of the circuit to generate the a/d zero point reference voltage (1.8v). connected with vbl through the buffer inside the ic. to change the a/d reference voltage range, input a dc voltage from this pin. signal input pin to the a/d converting circuit. use this by connecting with the bcmo pin for ccd or with the gcao pin for cis. input signals in the voltage range (1.8v to 3.8v) set through vwl and vbl. output of the circuit generating the a/d full-scale reference voltage (3.8v). connected inside the ic with the a/d converter. output of the circuit generating the a/d zero point reference voltage (1.8v). connected inside the ic with the a/d converter. description of analog pin functions (cont.)
5 mitsubishi digital assp ? M66335FP facsimile image data processor for the ccd sensor agc mode the peak value of the 16 line cycle is de- tected by setting the agc command in the register 00 at h. to escape this mode, set the agc com- mand at l after a 20 line cycle (or a cycle of 16 lines or more) passed since the start. this operation mode is started by setting the unif command in the register 00 at h after setting umode: h (white cor- rection) in the register 00 and unifm: l (only white correction) in the register 01. starting by the unif command also makes the system generate data for non- uniformity correction for white correction (for the 8 line cycle). to escape this mode, set the unif com- mand at l after a 10 line cycle (or a cycle of 8 lines or more) passed since the start. the read operation mode is started by setting the scan command in the regis- ter 00 at h. to escape this mode, set the scan command at l. unif mode (white) scan mode (1) operation mode the m66335 has three basic operations. ? peak value detection: adjusting the peak value of analog signals output from the analog circuit to the white reference voltage (vwl) of the a/d converter built in the m66335. ? generation of data for uniformity correction: generating data on a white reference original sheet for uniformity correction by the sensor unit and writing them to the memory for correction built in the m66335. ? read: reading original sheets, performing image processing of the read image data, and outputting in serial or parallel the indi- cated converted binary data. the m66335 is capable of performing the dma transfer of multivalued data (6-bit data=d7~d2, d1=d0=0) after correction about uniformities. these three basic operations are performed in the following mode sequences for the ccd sensor and cis sensor. the sensor is set through the register 00 (sens). for the cis sensor agc mode the peak value of the 16 line cycle is detected by setting the agc command in the register 00 at h. to escape this mode, set the agc com- mand at l after a 20 line cycle (or a cycle of 16 lines or more) passed since the start. when this operation mode is started by the unif command after setting umode: l (black correction) in the register 00 and unifm: h (black and white correc- tion) in the register 01, the system also generates black data for non-uniformity correction for black correction (for the 8 line cycle). to escape this mode, set the unif com- mand at l after a 10 line cycle (or a cycle of 8 lines or more) passed since the start. in the case of only white correction, the setting is not necessary. follow the instruction below. when this operation mode is started by the unif command in the register 00 af- ter setting umode: h (white correction) in the register 00 and unifm: l (only white correction) in the register 01, the system also generates white data for non- uniformity correction for white correction (for the 8 line cycle). to escape this mode, set the unif com- mand at l after a 10 line cycle (or a cycle of 8 lines or more) passed since the start. the reading operation is started by set- ting the scan command in the register 00 at h. to escape this mode, set the scan mode at l. the signal operations and data flow in each basic operation are shown in the page 4-217 and 4-218, and the flow- chart is in the page 4-260 and 4-261. unif mode (black) unif mode (white) scan mode
mitsubishi digital assp ? M66335FP facsimile image data processor 6 operations of signals in the peak value detection operation mpu dma codec acck sysck stimb srdyb drq int dakb rdb wrb csb resetb sclk svid a0~a4 d0~d7 bcao gcao adin vwl,vbl vri+,vri- levaj bcmi bcmv c2 c1 ptimb rs ck2 ck1 sh ain resolution change image processing sequence control signal analog control analog signal processing circuit correction data memory conversion table memory 7bit a /d converter detection of background/ character levels selection of processing for conversion to binary image bus interface mpu bus interface dither matrix organized dither g correction table sensor control image sensor uniformity correction mtf compensation line memory error memory image zone separation error diffusion dma control resolution change image processing sequence control signal analog control analog signal processing circuit correction data memory conversion table memory 7bit a /d converter detection of background/ character levels selection of processing for conversion to binary image bus interface mpu bus interface dither matrix organized dither g correction table sensor control image sensor uniformity correction mtf compensation line memory error memory image zone separation error diffusion dma control mpu dma codec acck sysck stimb srdyb drq int dakb rdb wrb csb resetb sclk svid a0~a4 d0~d7 bcao gcao adin vwl,vbl vri+,vri- levaj bcmi bcmv c2 c1 ptimb rs ck2 ck1 sh ain flow of data in the creation of data for uniformity correction
7 mitsubishi digital assp ? M66335FP facsimile image data processor mpu dma codec acck sysck stimb srdyb drq int dakb rdb wrb csb resetb sclk svid a0~a4 d0~d7 bcao gcao adin vwl,vbl vri+,vri- levaj bcmi bcmv c2 c1 ptimb rs ck2 ck1 sh ain resolution change image processing sequence control signal analog control analog signal processing circuit correction data memory conversion table memory 7bit a /d converter detection of background/ character levels selection of processing for conversion to binary image bus interface mpu bus interface dither matrix organized dither g correction table sensor control image sensor uniformity correction mtf compensation line memory error memory image zone separation error diffusion dma control flow of data in the reading operation (for output in serial: binary data) flow of signals in the reading operation (for multivated data) mpu dma codec acck sysck stimb srdyb drq int dakb rdb wrb csb resetb sclk svid a0~a4 d0~d7 bcao gcao adin vwl,vbl vri+,vri- levaj bcmi bcmv c2 c1 ptimb rs ck2 ck1 sh ain resolution change image processing sequence control signal analog control analog signal processing circuit correction data memory conversion table memory 7bit a /d converter detection of background/ character levels selection of processing for conversion to binary image bus interface mpu bus interface dither matrix organized dither g correction table sensor control image sensor uniformity correction mtf compensation line memory error memory image zone separation error diffusion dma control : image data : correction or compensation data mpu dma codec acck sysck stimb srdyb drq int dakb rdb wrb csb resetb sclk svid a0~a4 d0~d7 bcao gcao adin vwl,vbl vri+,vri- levaj bcmi bcmv c2 c1 ptimb rs ck2 ck1 sh ain : image data : correction or compensation data resolution change image processing sequence control signal analog control analog signal processing circuit correction data memory conversion table memory 7bit a /d converter detection of background/ character levels selection of processing for conversion to binary image bus interface mpu bus interface dither matrix organized dither g correction table sensor control image sensor uniformity correction mtf compensation line memory error memory image zone separation error diffusion dma control flow of data in the reading operation (for output in parallel: binary data)
mitsubishi digital assp ? M66335FP facsimile image data processor 8 (2) line cycle and reading sequence the relationship between the line cycle and the reading sequence of the m66335 is shown in fig. 1. the relationship between the codec interface operations and the reading sequence is shown in fig. 2 and that between the dma in- terface operations and the reading sequence is shown in fig. 3. ? single-line cycle (1/acck): defines the processing time per line of the m66335. the single-line cycle is decided by the line cycle counter value registers 03 and 04 (pre_data), and the pixel transfer clock. the pixel transfer clock is 1/16 of sysck. 1 line cycle (1/acck) [ns] = line cycle counter value pixel transfer clock cycle [ns] = (pre_data + 1) pixel transfer clock cycle [ns] = (pre_data + 1) 16/sysck [ns] after loading the pre_data value, the line cycle counter gen- erates the addresses of the following gate signals while count- ing down with the pixel transfer clock. ? sensor start pulse (sh): image sensor start pulse. the point of the start pulse is decided by the uniformity correction range (unifg) and the value of the reg- ister 05. [st_pl] the st_pl value must be set according to the following formu- las for each image sensor type. ccd: st_pl = dummy pixels of the sensor + 2 cis: st_pl = 2 fig. 1 line cycle and the reading sequence line cycle (acck) registers 03 and 04 (pre_data) sensor start pulse (sh) pulse motor control (ptim) register 00 (sens_w) register 11, 12 (offset) agc range (agcg) register 00 (sens_w) register 01 (source) register 00 (sens_w) register 05 (st_pl) original sheet reading range 1 line cycle register 00 (sens_w) uniformity correction range (unifg) pre_data loading st_pl countdown relationship with the registers 0 ? uniformity correction range (unifg): defines the range where uniformity correction is performed. this range corresponds to the width of the sensor (b4 to a4). for the relationship between the sensor width and the uniformity correction range, see table 2. ? agc range (agcg): defines the range where peak value detection is performed. this range corresponds to the sensor width (b4 to a4). auto gain control is performed for the whole width of the sensor (solid line) in the agc mode and for the narrower width (dashed line) than the sensor width in the scan mode. for the relationship between the sensor width and the agc range, see table 2. ? original sheet reading width: defines the reading width for original sheets. for original sheet widths narrower than the sensor width, the read- ing range (dashed line) is set, using the sensor center as the base center point. therefore, the points for the original sheet should be based on the sensor center. for the relationship between the sensor width and the original sheet reading width, see table 3. ? pulse motor control signal (ptim): generates control signals for the pulse motor for the reading roller.
9 mitsubishi digital assp ? M66335FP facsimile image data processor fig. 2 codec interface operations and the reading sequence (binary data output : sirial output) fig. 3 dma interface operations and the reading sequence (multivated data output) acck sh srdyb (sscan) int sclk svid ptimb stimb 1 srdyb : ??is taken in with a - flow of sh, when scanning is started and ptimb is output. (sscan : ?? 2 during the period that stimb is ?? converted binary data are output. 3 srdyb : ??is taken in with a - flow of acck, when the reading of one line ends. (sscan : ?? 4 int is asserted with a flow of sscan. (int : ?? 5 when cpu is ready for reading the next line, intclr is generated and int is negated, and then srdyb is set ?? (sscan) : internal signal 1 2 3 4 : output section : register setting 5 acck sh (sscan) int drq dakb rdb ptimb (dmafin) (counter reset) 1 srdyb : ??is taken in with a - flow of sh, when scanning is started and ptimb is output. (sscan : ?? 2 srdyb : ??is taken in with a - flow of acck, when the reading of one line ends. (sscan : ?? 3 the internal counter reset signal is generated with a flow of sscan, and drq is asserted with a - flow of sscan. 4 after the internal counter is reset, dma transfer is started. (the internal counter is counted up by one each time a pixel is transferred.) 5 when the value of the internal counter reaches the output pixel number, dmafin shifts to ?? and drq is negated with a - flow form dmafin and int is asserted with a flow of dmafin. 6 when cpu is ready for reading the next line, intclr is generated and int is negated, and then srdyb is set ?? (sscan), (dmafin), (counter reset) : internal signal , : register setting 1 2 3 4 5 6 : output section
mitsubishi digital assp ? M66335FP facsimile image data processor 10 1942/214 3886/430 200dpi 400dpi 200dpi 400dpi 200dpi 400dpi 2103/55 4207/111 2103/55 4207/111 2018/130 4037/261 1943/215 3887/431 1943/215 3887/431 1584/564 3169/1129 b4 table 2 gate signal ranges for the sensor widths sensor width a4 uniformity correction range (unifg) agc range (agcg) scan mode agc mode gate signal resolution table 3 original sheet reading widths according to the original sheet widths for the sensor widths 200dpi 400dpi 200dpi 400dpi 2102/54 4206/110 2102/54 4206/110 b4 sensor width a4 resolution b4 a4 when original sheets narrower than the sensor width, cut out the original sheet width with the registers 11 to 14. (offset, outlength): (region designation function) x y x/y x : left end address y : right end address original sheet width
11 mitsubishi digital assp ? M66335FP facsimile image data processor (3) image processing function the m66335 converts image signals input from the image sensor into binary data. this includes the simple conversion of characters and the change of images with various densities into pseudo-half- tone. before the conversion, distortions and characteristic degradations which signals from the image sensor almost always have must be corrected or compensated. image zone separation must also be performed to realize optimal conversion-to-binary of the image for the possible shortest trans- mission time. functions required for image processing are as follows. ? peak value detection ? uniformity correction ? resolution change (enlargement, reduction and averaging) ? mtf compensation ? g correction ? background/character level detection (simple conversion to binary) ? change to pseudo-halftone organized dither error diffusion ? image zone separation ? designation of regions peak value detection because the a/d converter of the m66335 uses the input dynamic range at 2 vp-p, the reference voltages (v wl , v bl ) corresponding to the peak value are fixed. the peak value of analog signals output from the analog processing circuit must be detected before those signals are input to the a/d converter in order to adjust the analog signal peak value to the full-scale value of the converter. the peak value detection is performed by reading white data from the sensor in the agc mode selected from its three modes (agc, unif and scan) of the m66335. as shown in fig. 4, preprocessing of peak value detection to increase the gain at the gain control is performed for a 8 line cycle and gain control processing to decrease the gain when the a/d converter over- flows is performed for another 8 line cycle after the start command (register 00: agc) in the agc mode. as a result, the gain changes as shown in fig. 5. preprocessing of peak value detection gain control on the peak value after the completion of preprocessing of peak value detection after the completion of gain control on the peak value one line one line white data v wl v bl v wl v bl the output level of the last pixed of the line is adjusted to v wl . the peak value of the sensor output in the line is adjusted to v wl . fig. 5 changes of the gain in peak value detection fig. 4 peak value detection 1 2 3 4 5 6 7 8 9 10111213141516 peak value detection preprocessing of peak value detection (increasing the gain) gain control on the peak value (decreasing the gain) line cycle
mitsubishi digital assp ? M66335FP facsimile image data processor 12 uniformity correction uniformity correction is to correct shading distortion due to less light at each end of the light source and faded light around the lens, or high frequency distortion due to characteristic variations pixel by pixel in the image sensor. as shown in fig. 7, the m66335 makes blocks each of two pixels, creates a set of uniformity correction data for each block, and write them to the built-in correction memory (sram: 1024 word 6 bits) in the unif mode selected from its three modes (agc, unif and scan). the correction data created each for two pixels are read from the built-in correction memory to correct the input image data consecu- tively in the scan mode. with the register 01 (unifs) set at 1, the uniformity is not implemented. with the register 02 (res) set at 1, uniformity correction is per- formed on a block for 4 pixels. for uniformity correction, white correction or the combination of black correction and white correction can be selected according to the types of image sensors as shown in table 4. this is set in the register 00 (sens, umode) and register 01 (unifm). to perform both black correction and white correction, the black cor- rection must be done first. the m66335 implements the correction in the correction range of 50% as shown in fig. 7. if a set of white correction data is beyond the correction range of 50%, the correction are not exactly performed as shown in fig. 7. therefore, ensure that input signals are within the range. fig. 6 waveform of white data output from the image sensor black level white level high frequency distortion shading distortion 1 line white correction + black correction analog signal input v wl 2 7 ? 2 6 ? v bl white data black data 1 line correction on over-range data (in white correction) analog signal input 2 7 ? 2 6 ? white data 1 line white correction analog signal input 2 7 ? 2 6 ? white data 1 line 50% 0 v wl v bl 50% 0 v wl v bl 50% 0 section over the correction range white data over the correction range fig. 7 uniformity correction selection of correction mode register 01 (unifm) table 4 uniformity correction due to the image sensor ccd cis correc- tion image sensor white correction white correction type of the sensor register 00 (sens) 0 1 creation of uniformity correction data register 00 (umode) 1 1 period of black correction : 0 period of white correction : 1 0 0 1 1 register black correction white correction
13 mitsubishi digital assp ? M66335FP facsimile image data processor l resolution change resolution change is controlled through h/w in the horizontal scan- ning direction and through s/w in the vertical scanning direction. the sequence for resolution change is shown in fig. 8. horizontal scanning direction the scaling factor is written from the register 15 (cnv_d) to the built-in resolution change memory (100w 1 bit) bit by bit by 100 operations. mssel of the register 6 must be set at 0 (which specifies the hori- zontal scanning direction) before the scaling factor is written in the memory. the procedure to specify cnv_d is as follows. u in the case of reduction data written in the resolution change memory have the following meaning. 0: 1 pixel is output. 1: no pixel is output. (example of reduction to 75%) 75 0s and 25 1s are written in the memory. the intervals of 1s should be as equal as possible to obtain the image with better qual- ity. u in the case of enlargement data written in the resolution change memory have the following meaning. 0: 1 pixel is output. 1: 2 pixels are output. (example of enlargement to 150%) 50 0s and 50 1s are written in the memory. the intervals of 1s should be as equal as possible to obtain the image with better quality as in the reduction. vertical scanning direction processing of lines to implement the scaling factor in the vertical scanning direction is decided for each line through the register. mssel of the register 6 must be set at 1 (which specifies the verti- cal scanning direction), and either 0 or 1 written in the register 15 (cnv_d) before the processing of each line. the timing for this setting is in the period between the first transition of the int signal (synchronized with that of acck) and that of the sh signal (the start of taking the srdy signal in). the procedure to specify cnv_d is as follows. u in the case of reduction cnv_d indicates the current line read. 0: 1 line of data are output. 1: no line of data are output. u in the case of enlargement cnv_d indicates the next line read. 0: 1 line of data are output with ptim generated (paper driven). 1: 1 line of data are output with ptim generated (paper not driven). (paper not driven: the same line is read again.) resolution change end yes yes no no enlargement/reduction is set in convx/convy. mssel is set at 0. data setting in cnv_d (100 bits in quantity) mssel is set at 1. specifying enlargement/reduction for horizontal/vertical scanning setting the scaling factor for resolution change in the horizontal scanning direction setting the scaling factor for resolution change in the vertical scanning direction specifying horizontal scanning start of reading a single line end of reading a single line specifying vertical scanning data setting in cnv_d (1 bit in quantity) setting of srdy int generated? page end? fig. 8 sequence of resolution change setting
mitsubishi digital assp ? M66335FP facsimile image data processor 14 use the ptimb signal as control signals for the pulse motor for the reading roller. the sequence for reduction is shown in fig. 9 and that for enlargement in fig. 10. fig. 9 reduction processing sequence fig. 10 enlargement processing sequence acck sh (start), (sscan) : internal signals : output section , , , : register setting 1 2 (start) int stimb sclk ptimb svid (sscan) 3 5 6 4 7 01 1 00 reduced line reduced line 1 at the initial setting, the enlargement/reduction setting (cnv_d) in horizontal scanning is implemented. then, after the system is switched into the setting mode for enlargement/reduction in vertical scanning, the first line is set. 2 with a flow of acck, the scan command is taken in, when the system comes into the standby mode for srdyb. (start: ?? 3 with a - flow of sh, srdyb: ??is taken in, when scanning starts and ptimb is output. (sscan: ?? 4 during the period that stimb is at ?? converted binary data are output while the data for reduced lines are not output because stimb for them are at ?? 5 with a - flow of acck, srdyb: ??is taken in, when the reading of the single line is completed. (sscan: ?? 6 with a flow of sscan, int is asserted. (int: ?? 7 with cpu ready for reading the next line, the enlargement/reduction setting (cnv_d) in vertical scanning is implemented; intclr is generated; int is negated; and then srdyb is set at ?? 1 2 3 5 6 4 7 01 1 00 acck sh (start) int stimb sclk ptimb svid (sscan) enlarged line enlarged line (start), (sscan) : internal signals : output section , , , : register setting 1 at the initial setting, the enlargement/reduction setting (cnv_d) in horizontal scanning is implemented. then, after the system is switched into the setting mode for enlargement/reduction in vertical scanning, the first line is set. 2 with a flow of acck, the scan command is taken in, when the system comes into the standby mode for srdyb. (start: ?? 3 with a flow of sh, srdyb: ??is taken in, when scanning starts and ptimb is output while it is not output for enlarged lines. (sscan: ?? 4 during the period that stimb is at ?? converted binary data are output. 5 with a - flow of acck, srdyb: ??is taken in, when the reading of the single line is completed. (sscan: ?? 6 with a flow of sscan, int is asserted. (int: ?? 7 with cpu ready for reading the next line, the enlargement/reduction setting (cnv_d) in vertical scanning is implemented; intclr is generated; int is negated; and then srdyb is set at ??
15 mitsubishi digital assp ? M66335FP facsimile image data processor mtf compensation as shown in fig. 11, image data of characters or pictures photoelec- trically converted by the sensor unit show degradation in resolution. mtf compensation function of the m66335 restores the resolution of x = x + a ((x C a) + (x C b) + (x C c) + (x C d)) where, a : mtf compensation coefficient in the register 08 (mtf_c, mtf_i) in the above equation, a is set according to the register 07 : mode (selection of conversion-into-binary mode) as follows: mode : 00 (simple binary) a = mtf_c mode : 01 (organized dither) a = mtf_i mode : 10 (image zone separation) separation (character) a = mtf_c for image zone separation (photo) a = mtf_i for image zone mode : 11 (error diffusion) a = mtf_i photoelectric conversion original (character) original (photo) photoelectric conversion photoelectric conversion image signal mtf compensation data after compensation image signal mtf compensation data after compensation fig. 11 mtf compensation those data and expands the apparent dynamic range by strengthen- ing the high-pass frequency constituent with the laplacian filter. resolution compensation
mitsubishi digital assp ? M66335FP facsimile image data processor 16 g correction g correction according to the sensitivity characteristics (logarithmic characteristics) of human eyes is implemented to approximate the image data to natural images. to do this, the m66335 writes the g correction table to the built-in sram and read the corrected values corresponding to read image data values from the sram. g = 0.45 is considered to be the optimal for g correction for thermal head printers. fig. 12 shows a characteristics example at g = 0.45. g correction processing is set through the register 06 : gamma as follows. gamma : 00 g =1 gamma : 01 g = conversion table value gamma : 10 g = 1 for image zone separation (character) g = conversion table value for image zone separation (photo) gamma : 11 g = conversion table value for image zone separation (character) g = 1 for image zone separation (photo) for the procedures of inputting/outputting of data, refer to the section on writing to/reading from the g correction memory. image data 6 data after g correction g correction memory 6 white 63 black 0 black : 0 white : 63 25 34 47 56 816 32 48 image data (address) image data after g correction (memory output) a <5,0> (address) do <5,0> (output) g = 1 g = 0.45 fig. 12 g correction by means of the conversion table if(d in < d low ) d out = 0 if(d low d in < d up ) d out ( ) if(d up d in ) d out 1.0 d out 0 d low d in d up 1.0 d in -d low d up -d low g = 1.0 =
17 mitsubishi digital assp ? M66335FP facsimile image data processor background/character level detection the m66335 uses not the fixed threshold system but the floating threshold system, where the optimal threshold for simple conversion- to-binary of objective pixels are continually generated by constantly detecting background/character levels. accordingly, the threshold value proper for image data is generated without processing the data. the threshold value is used for the areas to be converted to binary when simple conversion-to-binary or image zone separation is se- lected as the mode of conversion to binary in reading data. : register 07 (mode) background level counter when image data greater (lighter in light) than the current value are input, this counter counts up to approximate to the data. when image data smaller (darker in light) than the current value are input, this counter counts down to approximate to the data. image data background level detection counter character level detection counter generation of the threshold value comparison converted binary data this slope is decided through max_up. this slope is decided through min_up. this slope is decided through max_down. fixing of the background level white level background level lowest limit of the background level (ll_max) highest limit of the character level (ul_min) threshold level input data fixing of the character level character level black level threshold level = (background level point C character level) k + character level k = threshold factor for conversion to binary: register 07 (slice) lowest limit of the background level (ll_max) > highest limit of the character level (ul_min) fig. 13 background/character levels setting of the rate of count-up/count-down following data input : register 0c (max_up, max_down) setting of the lowest limit for background levels : register 0e (ll_max) character level counter when image data greater (lighter in light) than the current value are input, this counter counts up to approximate to the data. when image data smaller (darker in light) than the current value are input, this counter counts down to approximate to the data. setting of the rate of count-down following data input : register 0c (min_up) setting of the highest limit for character levels : register 0d (ul_min)
mitsubishi digital assp ? M66335FP facsimile image data processor 18 error diffusion the error diffusion, which is a conditional determination method, lo- cally diffuses density errors between the original image and the re- sult to obtain the best approximation. this generates images with good compatibility of gradation and resolution. this is operated by selecting the error diffusion in conversion-into- binary mode selection. : register 07 (mode) in error diffusion, dithers as well as density errors are added to image data. the dithers are data as commonly used for the dither matrix. : register 08 (error) g correction must be performed in the error diffusion. dither matrix m m n n 17 9 122025 10 11 8 15 19 12421 24 25 C32 k 2 k 1 fmn gmn fmn fmn > 32 ? gmn = 63 (white) fmn < 32 ? gmn = 0 (black) fmn = fmn + k 1 (1/ sa kl) sa klemCk, nC1 + k 2 (dither C32) integrated error sa klemCk, nC1 (note 2) error emn = fmn C gmn (note 1) weighting of the error filter a kl error buffer memory preceding line current line k,1 k,1 k 1 = register 08 (error) k 2 = register 08 (dither addition factor) note 1: characterized by using the difference from the corrected value fmn rather than that from the original pixel fmn. 2: errors before the point of remark are integrated. fig. 14 error diffusion method ? organized dither the m66335 has built-in sram with a configuration of 64 words 6 bits for organized dither memory. in the initial setting, write the threshold value proper for the preferred dither pattern to the dither memory after setting the dither matrix size. : register 07 (dith) : register 10 (dith_d) an example of dither patterns is shown in fig. xx. for the procedure of inputting/outputting data, refer to the section on writing to/reading from the dither memory. +C
19 mitsubishi digital assp ? M66335FP facsimile image data processor image zone separation to make data conversion fit for each image zone, a black and white image is separated into the zones to be converted to binary and the gradation zones. the binary zone is processed through simple con- version to binary and the gradation zone through the error diffusion. : register 08 to 0e in the black and white image, each window of the gradation zone (photo) does not have a large difference of luminance in it. with this characteristic of the gradation zone, it is distinguished from the conversion-into-binary zone through the following method. difference minimum maximum determining inequality 1: l max C l min >a (because the zone to be converted to binary has a large difference in luminance in it.): register 09 difference (sepa_a) determining inequality 2: l min >b (for the wholly white area): register 0a minimum (sepa_b) determining inequality 3: l max mitsubishi digital assp ? M66335FP facsimile image data processor 20 region designation function the sensor width is fixed for a4 and b4. the region designation function is to output only the data for a region defined and designated in terms of output data after resolution change (or after uniformity correction for multivalued data). registers 11 to 14 (offset, outlength) offset outlength designated region output width fig. 16 cut-out function
21 mitsubishi digital assp ? M66335FP facsimile image data processor pixel sclk svid drq dak rd d 0 d 1 d 2 d 3 d 4 d 5 d 6 d 7 12345678 n ?1 n n + 1 n + 2 n + 3 n + 4 n + 5 n + 6 n + 7 n + 8 n ?1 n ?2 n ?3 n ?4 n ?5 n ?6 n ?7 n ?8 (4) codec interface (binary data output) serial output srdyb sh stimb sclk svid 10 12 4 210 sclk svid 2 2 4 44 4 2 (equal scale, reduced scale) (enlarged scale) 2 unit : 1/sysck ab note: a is decided through the registers 05 (st_pl) 11 and 12 (offset), and b through the registers 13 and 14 (outlength). parallel output note: the 3-line handshake of srdy, sh and stim, which is the interface with codec, is the same as serial output.
mitsubishi digital assp ? M66335FP facsimile image data processor 22 (5) dma interface (multivalued output) the dma transfer of data after non-uniformity correction can be per- formed by setting p_o) of the register 01: at 1 (existence of dma sscan drq (dma counter reset signal) (dma counter signal) dakb rdb ( dmafin ) int d<7 : 2> 1 2 3 4 5 6 (xxxx): internal signal 1 on completion of reading one line, with a flow of sscan, the reset signal is entered in the dma counter. 2 with a - flow of the reset signal, drq shifts to h, when the dma transfer becomes ready. 3 with dakb at l and a flow of rdb, drq shifts to l, when multivalued data are output to d <7 : 2> during the period that rdb is at l. 4 with a - flow of dakb, the dma counter counts up and drq shifts to h, when the dma transfer becomes ready again. 5 the cycle of the above 3 and 4 is repeated until the dma counter counts up to reach the number of output pixels set in the registers 13 and 14 outlength subtracted by one. by that repetitive operation, dmafin shifts to h to terminate the dma transfer when it reaches the set number. 6 with a flow of dmafin, int shifts to h, when cpu has an interrupt. 7 reading is resumed from the next line by negating the int signal through the register 17 (intclr). output) and m_b of that register at 1 (multivalue). with this setting, neither enlargement, nor reduction, nor 400 dpi of resolution can be set.
23 mitsubishi digital assp ? M66335FP facsimile image data processor (6) writing to/reading from the dither memory, g correction memory, uniformity correction memory, and resolution change memory the sequences of writing a dither pattern to and reading it from sram with a configuration of 64 words 6 bits which is built in the m66335 for organized dither are shown below. initial setting (1) 07h d6,d5 1 initial setting (2) 01h d0=? 2 memory address (0) memory address (1) initial setting (1) initial setting (2) memory address (0) memory address (1) 10h data (0) data (1) 3 10h 3 124 4 csb a4 ~ a0 d7 ~ d0 (input) wrb ~ 07h d6,d5 01h d0=? 10h data (0) data (1) 10h csb a4 ~ a0 d7 ~ d0 (input) d7 ~ d0 (output) wrb rdb ~ writing to the dither memory (mpu ? m66335) reading from the dither memory (m66335 ? mpu) 1 d6 and d5 (dith) of the register 07 are set to define the dither matrix size. 2 d0 (cntrst) of the register 01 is set at 1 to reset the address counter of the dither memory. 3 dith_d is selected in the register 10, and data (0) of the mpu bus (d5 to d0) is written in the memory. the address counter of the dither memory is incremented at the edge of the first transition of wr. (for writing) 4 dith_d is selected in the register 10, and data (0) of the dither memory is read into the mpu bus (d5 to d0). the address counter of the dither memory is incremented at the edge of the first transition of rd. (for reading) dither matrix addresses a0 a4 a8 a12 a1 a5 a9 a13 a2 a6 a10 a14 a3 a7 a11 a15 4 4 matrix 4 8 matrix 8 8 matrix a0 a1 a2 a3 a4 a5 a8 a9 a6 a10 a11 a14 a15 a12 a13 a18 a19 a16 a17 a22 a23 a20 a21 a26 a27 a24 a25 a30 a31 a28 a29 a7 a0 a1 a9 a2 a3 a8 a18 a19 a20 a21 a26 a27 a24 a16 a25 a34 a35 a32 a33 a17 a42 a43 a40 a41 a50 a51 a48 a49 a58 a59 a56 a57 a4 a5 a6 a7 a22 a23 a10 a11 a12 a13 a14 a15 a33 a31 a28 a29 a38 a39 a36 a37 a46 a47 a44 a45 a54 a55 a52 a53 a62 a63 a60 a61
mitsubishi digital assp ? M66335FP facsimile image data processor 24 01h d0=1 1 0fh data (0) 2 0fh data (1) 2 133 csb a4 ~ a0 d7 ~ d0 (input) wrb ~ 01h 0fh 0fh data (1) data (0) ~ d0=1 writing to the g correction memory (mpu ? m66335) reading from the g correction memory (m66335 ? mpu) initial setting (2) memory address (1) memory address (0) initial setting (1) memory address (1) memory address (0) csb a4 ~ a0 d7 ~ d0 (input) d7 ~ d0 (output) wrb rdb the sequences of writing g correction table to and reading it from sram with a configuration of 64 words 6 bits which is built in the m66335 for g correction are shown below. 1 d0 (cntrst) of the register 01 is set at 1 to reset the address counter of the g correction memory. 2 gamma_d is selected in the register 0f, and data (0) of the mpu bus (d5 to d0) is written in the memory. the address counter of the g correction memory is incremented at the edge of the first transition of wrb. (for writing) 3 gamma_d is selected in the register 0f, and data (0) of the g correction memory is read into the mpu bus (d5 to d0). the address counter of the g correction memory is incremented at the edge of the first transition of rdb. (for reading)
25 mitsubishi digital assp ? M66335FP facsimile image data processor uniformity correction data can be written to and read from sram for uniformity correction built in the m66335 through the mpu bus. with this operation, the uniformity data can be temporarily saved in the backup memory when the power is off. the sequences of writing and reading uniformity correction data are shown below. 00h d1 d1 01h d0=? 19h data (0) data (1) 19h ~ 00h 01h d0=? 19h data (0) data (1) 19h ~ initial setting (1) 1 initial setting (2) 2 memory address (0) memory address (1) initial setting (1) initial setting (2) memory address (0) memory address (1) 3 3 124 4 csb a4 ~ a0 d7 ~ d0 (input) wrb csb a4 ~ a0 d7 ~ d0 (input) d7 ~ d0 (output) wrb rdb writing to the uniformity correction memory (mpu ? m66335) reading from the uniformity correction memory (m66335 ? mpu) 1 0 (black correction) or 1 (white correction) is set in d1 (umode) of the register 00. 2 d0 (cntrst) of the register 01 is set at 1 to reset the address counter of the uniformity correction memory. 3 unif_d is selected in the register 19, and data (0) of the mpu bus (d5 to d0) is written in the memory. the address counter of the uniformity correction memory is incremented at the edge of the first transition of wrb. (for writing) 4 unif_d is selected in the register 19, and data (0) of the uniformity correction memory is read into the mpu bus (d5 to d0). the address counter of the uniformity correction memory is incremented at the edge of the first transition of rdb. (for reading)
mitsubishi digital assp ? M66335FP facsimile image data processor 26 the sequences of writing a resolution change table to and reading it from sram with a configuration of 100 words 1 bit which is built in the m66335 for resolution change are shown below. 1 0 (horizontal scan) is set in d7 (mssel) of the register 06. 2 d0 (cntrst) of the register 01 is set at 1 to reset the address counter of the resolution change memory. 3 cnv_d is selected in the register 15, and data (0) of the mpu bus (d0) is written in the memory. the address counter of the resolution change memory is incremented at the edge of the first transition of wrb. (for writing) 4 cnv_d is selected in the register 15, and data (0) of the resolution change memory is read into the mpu bus (d0). the address counter of the resolution change memory is incremented at the edge of the first transition of rdb. (for reading) 06h 01h 15h data (0) data (1) 15h 06h 01h d0=1 15h data (0) data (1) 15h d0=1 d7=0 d7=0 initial setting (1) 1 initial setting (2) 2 memory address (0) memory address (1) initial setting (1) initial setting (2) memory address (0) memory address (1) 3 3 124 4 csb a4 ~ a0 d7 ~ d0 (input) wrb csb a4 ~ a0 d7 ~ d0 (input) d7 ~ d0 (output) wrb rdb writing to the resolution change memory (mpu ? m66335) reading from the resolution change memory (m66335 ? mpu) ~ ~ ~
27 mitsubishi digital assp ? M66335FP facsimile image data processor r/w r/w r/w w w w w w w w w w w w w w r/w r/w w w w w w r/w w r/w r/w a4 ~ a0 00h 01h 02h 03h 04h 05h 06h 07h 08h 09h 0ah 0bh 0ch 0dh 0eh 0fh 10h 11h 12h 13h 14h 15h 16h 17h 18h 19h default 00h 00h 00h 00h 00h 00h 00h 00h 00h 00h 00h 00h 00h 1fh 20h C C 00h 00h 00h 00h C 00h C 00h 00h d7 reset source res pre_data (7 : 0) st_pl (7 : 0) mssel pol d6 sens s/h_w lcmps d5 sens_w sh_w bls offset <7 : 0> outlength <7 : 0> gain <7 : 0> ave dith <1 : 0> pre_data (13 : 8) convx <1 : 0> error <1 : 0> sepa_a (5 : 0) sepa_b (5 : 0) sepa_c (5 : 0) max_up <1 : 0> ul_min <5 : 0> ll_max <5 : 0> gamma_d (5 : 0) dith_d (5 : 0) unif_d <5 : 0> intclr d4 agc unifs blcmps d3 unif p_o ccd d2 scan m_b cis3 d1 umode unifm cis2 d0 l cntrst cis1 mode <1 : 0> offset <12 : 8> outlength <12 : 8> convy <1 : 0> mtf_c <1 : 0> max_down <1 : 0> slice <2 : 0> gamma <1 : 0> mtf_i <1 : 0> min_up <1 : 0> agcstp srdys cnv_d srdyb list of the M66335FP registers
mitsubishi digital assp ? M66335FP facsimile image data processor 28 address r/w description 00 h r/w reset d7 d6 d5 d4 d3 d1 d0 sens sens_w agc unif scan umode l (default value: 00 h ) d7 0 1 reset : system reset normal mode reset mode l with d7 = 1, the system is reset during the period that the write pulse is l. d6 0 1 sens : sensor type ccd cis: (75% of clock duty) d5 0 1 sens_w : reading width of the sensor a4 b4 d4 0 1 agc : agc mode stop start d3 0 1 unif : unif mode stop start d2 0 1 scan : scan mode stop start d1 0 1 black correction + white correction black correction white correction only white correction C white correction umode : uniformity correction in the unif mode l controls start/stop of the agc mode. l controls start/stop of the unif mode. l controls start/stop of the scan mode. d2 source d7 d6 d5 d4 d3 d2 d1 d0 s/h_w sh_w unifs p_o m_b unifm cntrst (default value : 00 h ) d6 0 1 s/h_w : s/w pulse width normal (quadruple the system clock cycle) normal multiplied by 0.5 d7 0 1 source : reading width of the original a4 b4 01 h r/w register structure ( * ) write only
29 mitsubishi digital assp ? M66335FP facsimile image data processor address r/w description 01 h r/w 02 h w d5 0 1 sh_w : sh pulse width normal (16 times the system clock cycle) reverse of normal multiplied by 2 d4 0 1 unifs : uniformity correction valid invalid d3 0 1 p_o : dma output without dma output with dma output d2 0 1 m_b : processing mode binary multivalue d1 0 1 unif : uniformity correction in scan white correction black correction + white correction d0 0 1 cntrst : address counter reset normal mode reset mode l with d0 = 1, the counter is reset during the period that the write pulse is l. l all the built-in ram addresses are reset. l with the multivalue selected, data (6 bit) after non- uniformity correction can be output through the dma transfer. l d0 is output in the form of lsb and d7 in the form of msb. res d7 d6 d5 d4 d3 d2 d1 d0 lcmps bls blcmps ccd cis3 cis2 cis1 (default value : 00 h ) d7 0 1 res : resolution 200dpi 400dpi d5 0 1 bls : bit clamping invalid valid d4 0 1 blcmps : black level line clamping invalid valid d3 0 0 0 1 d2 0 0 1 0 d1 0 1 0 0 d0 1 0 0 0 sensors compatible with image sensor interfaces cis1 : sensors with the input level of 2v or higher cis2 : sensors with the input level of under 2v cis3 : sensors capable of line clamping ccd d6 0 1 lcmps : line clamping invalid valid ( * ) write only
mitsubishi digital assp ? M66335FP facsimile image data processor 30 address r/w description 03 h w 04 h w 05 h w 06 h w pre_data <7 : 0> d7 d6 d5 d4 d3 d2 d1 d0 (default value : 00 h ) d7 d6 d5 d4 d3 d2 d1 d0 (default value : 00 h ) d7 d6 d5 d4 d3 d2 d1 d0 (default value : 00 h ) mssel d7 d6 d5 d4 d3 d2 d1 d0 ave convx convy gamma (default value : 00 h ) d7 0 1 mssel : horizontal and vertical setting horizontal vertical d6 0 1 ave : averaging processing with averaging without averaging d4 0 1 0 1 convx : enlargement/reduction mode in the horizontal scanning direction original scale enlargement reduction d3 0 0 1 1 d2 0 1 0 1 convy : enlargement/reduction mode in the horizontal scanning direction original scale enlargement reduction d7 to d0 : pre_data <7 : 0> the lowest order 8 bits of the single-line cycle counter value d5 to d0 : pre_data <13 : 8> the highest order 6 bits of the single-line cycle counter value pre_data <13 : 8> l when with averaging selected : for enlargement : inserted lines are the average of the preceding one and the current one; and for reduction : the subsequent lines from removed lines are the average of the removed one and the current one. l res = 1 with the setting of 400dpi, enlargement cannot be set. d7 to d0 : st_pl <7 : 0> start pulse position to the sensor l set st_pl = (dummy pixels of the sensor + 2). st_pl <7 : 0> d5 0 0 1 1
31 mitsubishi digital assp ? M66335FP facsimile image data processor address r/w description d2 0 0 0 0 1 1 1 1 d1 0 0 1 1 0 0 1 1 d0 0 1 0 1 0 1 0 1 slice : threshold factor for conversion to binary 6/16 7/16 8/16 9/16 10/16 11/16 12/16 13/16 d4 0 0 1 1 d3 0 1 0 1 mode : selection of the conversion-to-binary mode simple binary organized dither image zone separation (simple binary + error diffusion) error diffusion d6 0 0 1 1 d5 0 1 0 1 dith : dither matrix size 4 4 4 8 8 8 C d7 0 1 pol : conversion-to-binary output mode white : 1; black : 0 white : 0; black : 1 pol d7 d6 d5 d4 d3 d2 d1 d0 dith mode slice (default value : 00 h ) d1 0 0 1 1 d0 0 1 0 1 gamma : g correction processing character, photo : g = 1 character, photo : g = download value character : g = 1; photo : g = download value character : g = download value ; photo : g = 1 06 h w 07 h w note: judgment between character and photo is based on the result of image zone separation.
mitsubishi digital assp ? M66335FP facsimile image data processor 32 address r/w description 08 h w d7 d6 d5 d4 d3 d2 d1 d0 error mtf_c mtf_i (default value : 00 h ) d5 0 0 1 1 d4 0 1 0 1 rate of dither addition to errors weak (1/8) strong (1/4) weak (1/8) strong (1/4) d7 d6 d5 d4 d3 d2 d1 d0 sepa_a (default value : 00 h ) d7 d6 d5 d4 d3 d2 d1 d0 sepa_b (default value : 00 h ) d7 d6 d5 d4 d3 d2 d1 d0 sepa_c (default value : 00 h ) 09 h w 0a h w 0b h w d3 0 0 1 1 d2 0 1 0 1 mtf_c : mtf compensation factor 1/4 1/2 1 0 error (base) strong (7/8) strong (7/8) weak (3/4) weak (3/4) error d1 0 0 1 1 d0 0 1 0 1 mtf_i : mtf compensation factor 1/4 1/2 1 0 note: this is valid when mode is simple binary or image zone separation (character). note: this is valid when mode is organized dither, error diffusion or image zone separation (photo). d5 to d0 : sepa_b image zone separation parameter (minimum) d5 to d0 : sepa_a image zone separation parameter (differential) d5 to d0 : sepa_c image zone separation parameter (maximum)
33 mitsubishi digital assp ? M66335FP facsimile image data processor address r/w description 0c h w 0e h w 0f h r/w 10 h r/w d7 d6 d5 d4 d3 d2 d1 d0 max_up max_down min_up (default value : 00 h ) d5 0 0 1 1 d4 0 1 0 1 max_up : background level detection clock for the up counter ordinary (t = (single pixel cycle) 32) slow (t = (single pixel cycle) 64) fast (t = (single pixel cycle) 16) fastest (t = (single pixel cycle) 8) d3 0 0 1 1 d2 0 1 0 1 max_down : background level detection clock for the down counter ordinary (t = (single pixel cycle) 128) slow (t = (single pixel cycle) 256) fast (t = (single pixel cycle) 64) fastest (t = (single pixel cycle) 32) d1 0 0 1 1 d0 0 1 0 1 min_up : character level detection clock for the up counter ordinary (t = (single pixel cycle) 128) slow (t = (single pixel cycle) 256) fast (t = (single pixel cycle) 64) fastest (t = (single pixel cycle) 32) d7 d6 d5 d4 d3 d2 d1 d0 ul_min (default value : 1f h ) d7 d6 d5 d4 d3 d2 d1 d0 ll_max (default value : 20 h ) d7 d6 d5 d4 d3 d2 d1 d0 gamma_d <5 : 0> d7 d6 d5 d4 d3 d2 d1 d0 dith_d <5 : 0> 0d h w d5 to d0 : dith_d built-in dither memory data d5 to d0 : ll_max detection of background/character levels lowest limit of background levels lowest limit of background levels (ll_max) > highest limit of character levels (ul_min) d5 to d0 : ul_min detection of background/character levels highest limit of character levels d5 to d0 : gamma_d built-in g memory data
mitsubishi digital assp ? M66335FP facsimile image data processor 34 address r/w description 11 h w 12 h w w 13 h 14 h w 15 h r/w offset <7 : 0> d7 d6 d5 d4 d3 d2 d1 d0 (default value : 00 h ) d7 d6 d5 d4 d3 d2 d1 d0 offset <12 : 8> (default value : 00 h ) outlength <7 : 0> d7 d6 d5 d4 d3 d2 d1 d0 (default value : 00 h ) d7 d6 d5 d4 d3 d2 d1 d0 outlength <12 : 8> (default value : 00 h ) d7 d6 d5 d4 d3 d2 d1 d0 cnv_d d7 to d0 : offset <7 : 0> offset for cut-out lowest order 8 bits d7 to d0 : outlength <7 : 0> no. of output pixels lowest order 8 bits d0 : cnv_d indication of enlargement/reduction d3 to d0 : outlength <12 : 8> no. of output pixels highest order 5 bits note: outlength <12 : 8> must be a multiple of 8. if a number of output pixels is not a multiple of 8, the remainder of the division must be omitted. d3 to d0 : offset <12 : 8> offset for cut-out highest order 5 bits
35 mitsubishi digital assp ? M66335FP facsimile image data processor address r/w description 16 h r/w 17 h w 18 h r/w 19 h r/w d7 d6 d5 d4 d3 d2 d1 d0 srdyb (default value : 00 h ) d2 0 1 agcstp : gain control counter gain control counter valid. gain fixed. d1 0 1 srdys : srdy control srdy control through the register srdy control through the external pin d0 0 1 srdyb : data transfer start ready. transfer allowed. transfer not allowed. intclr d7 d6 d5 d4 d3 d2 d1 d0 gain <7 : 0> d7 d6 d5 d4 d3 d2 d1 d0 d7 d6 d5 d4 d3 d2 d1 d0 unif <5 : 0> agcstp srdys l in the case of data control through the register, the sdryb input pin must be always set at h. for the control through the register, the srdy register must be controlled line by line. l with umode = 0, access to the uniformity correction memory for black correction is available. l with umode = 1, access to the uniformity correction memory for white correction is available. int signals are negated by accessing to this address. d5 to d0 : unif_d built- in uniformity correction memory data in reading : the current gain value of the gain control counter can be read. in writing : the gain value of the gain control counter can be set. however, this is valid only if agcstp = 1. ( * ) write only
mitsubishi digital assp ? M66335FP facsimile image data processor 36 sensor type ccd sensor cis sensor which outputs light voltages (white voltage) of 3.5v or lower cis sensor which outputs light voltages (white voltage) of 2v or lower cis sensor which output shielding pixels for each line description of the operations of the analog circuits the configuration of the analog processing circuits is shown in fig. 17. (1) sensor selection circuit the four types of sensors in the table can be connected to the circuit. register 02 h ccd cis1 cis2 cis3 max.500mvpp. signaling element blanking element white black shielding pixel part effective pixel part signaling element white max.3.5v black 200mv signaling element white max.2v black 200mv signaling element clamping level white black 2vpp. shielding pixel part effective pixel part the amplitudes of sensor signals are multiplied by C4 through the two operating amplifiers directly after the switch to select the ccd mode. (the waveforms of the signals are inverted at the same time.) as a result, the sensor signals input to the sample and hold circuit have a dark voltage of 2.2v. the amplitude of signals input from the sensor are halved. then, their reference potential is shifted up to 2.2v. as a result, the sensor signals input to the sample and hold circuit have a dark voltage of 2.2v. the reference potential of signals input from the sensor is shifted up to 2.2v. as a result, the sensor signals input to the sample and hold circuit have a dark voltage of 2.2v. sensor signals with a dark voltage of 2.2v clamped by line clamping input are directly input to the sample and hold circuit.
37 mitsubishi digital assp ? M66335FP facsimile image data processor (2) line clamping circuit this circuit is used for ccd (line clamping mode) and cis3. the reference voltage (dark voltage) output in the shielding pixel part of the sensor is sampled by lcmp (line clamping pulses) and shifted up to the internal reference voltage of 2.2v. this is not used for the cis1 or cis2 input sensor (set off constantly). : register 02 (lcmps) (3) sample and hold circuit and bit clamping circuit in the ccd mode, bit clamping, as well as line clamping, can be performed. the blanking elements of each pixel of sensor output is sampled by btcmp (bit clamping pulses). the differences of signals from the reference potential sampled by the bit clamping circuit are input to the gain control circuit of next step as signaling elements. to turn off bit clamping, set bls invalid, so that the reference potential will be fixed at the internal reference potential of 2.2v. : register 02 (bls) (4) gain control circuit the amplifying factor (gain) must be adjusted so that the amplitudes of sensor signals can come within the dynamic range of the a/d con- verter. the gain is set through the automatic gain control in the agc mode (register 00) or directly through the register 18 (gain <7 : 0>). the gain changes within the following ranges according to the sen- sor used. in the agc mode, the gain control counter is set at the greatest gain in the initial state and then counted down each time an overflow bit is output from the a/d converter. the count (gain) of the gain control counter is directly read/written through the register 18 (gain <7 : 0>). the counting operation of the counter can be controlled through the register 16 (agcstp). (5) internal reference voltage internal reference voltage source for the analog circuits: this generates the reference voltage (2.2v) for the line clamping circuit, the sample and hold circuit, and the bit clamping circuit. a/d converter reference voltage generation circuit: this generates vwl (white level reference voltage of 3.8v) and vbl (black level reference voltage of 1.8v) for the a/d converter. amplifying factor of signals (gain) 4 to 20 0.5 to 2.5 1 to 5 1 to 5 mode ccd cis1 cis2 cis3 (6) black level clamping circuit this circuit adjust the level of reference voltage to the a/d converter from analog circuits. the black clamping circuit is used in the ccd or cid3 mode. (see figs. 18, 19 and 22) the gcao pin and the bcmi pin are capacity- coupled. the output reference potential in the shielding pixel part of sensor signals are applied to the bcmv pin as the vbl (black level reference voltage of 1.8v) for the a/d converter. blcmp (black level clamping pulses) are generated concurrently with the shielding pixel part of each line. to turn off this circuit, set blcmps invalid and apply the black level reference voltage of the a/d con- verter to the bcmv pin. : register 02 (blcmps) in the cis1 or cis2 mode, the levaj pin is used. (see figs. 20 and 21) voltage is applied to the levaj pin so that the reference poten- tial of output at the gcao pin can be adjusted to the vbl (black level reference voltage of 1.8v) of the a/d converter. set voltage input to the levaj pin as follows. vlevaj = vvbl C a g v + 0.2 [v] vgcao = vlevaj + g v vin [v] where, a: the lowest limit of dark voltage of the sensor [v] g v : gain (multiplying factor) of the gain control circuit v in : signals input from the sensor [v]
mitsubishi digital assp ? M66335FP facsimile image data processor 38 + C C + C + + C av cc c1 c 2 gcao levaj bcmv bcmi bcmo v cc dv cc v cc avdd mcis<3:1>,mccd s/h btcmpbls c1 c2 gcao levaj bcmv blcmp bcmi bcmo ain lcmp agnd agnd dgnd gnd vri- vri+ vbl vwl adin agnd reset adck adin vrefl vrefh of dgnd ain mcis1 1.8v 8 3.8v 2.2v v vbl =1.8 [v] av cc dv cc vref- vri+ vri- vref+ b <7:1> of gain<7:0> agcsel <7:0> adck reset blcmp bls s/h btcmp mcis<3:1>,mccd lcmp din<6:0> gnd gca mcis2 mcis3 mccd level shift circuit(2.2v) bit clamping circuit sample and hold circuit black level clamping circuit gain control circuit a/d converter reference voltage generating circuit for the a/d converter digital circuit internal reference voltage source for the analog circuits input clamping circuit + C + C + C + C + C + C + C + C fig. 17 circuit configuration of the analog part of the M66335FP + C
39 mitsubishi digital assp ? M66335FP facsimile image data processor analog circuit timing chart (for ccd mode/bit clamping) address bit signal setting mode ccd (bit clamping) 00 h d 6 sens 1 02 h d 6 d 5 d 4 d 3 d 2 d 1 d 0 lcmps bls blcmps ccd cis 3 cis 2 cis 1 11110 00 register 2 2 12 16 16 16 39 1 2 13 9 4 8 3 n n n n 8 16 4 non-signaling part shielding pixel part effective pixel part sh ck 1 gcao signal output ccd signal output bcmo signal output lcmp s/h btcmp blcmp a/d clock a/d output non-signaling part shielding pixel part effective pixel part unit : 1/sysck gcao signal output bcmo signal output a/d clock a/d output sh ck1 ccd signal output lcmp s/h btcmp blcmp
mitsubishi digital assp ? M66335FP facsimile image data processor 40 analog circuit timing chart (for ccd mode/line clamping) 00 h d 6 sens 1 02 h d 6 d 5 d 4 d 3 d 2 d 1 d 0 lcmps bls blcmps ccd cis 3 cis 2 cis 1 10110 00 address bit signal setting mode ccd (line clamping) register 2 2 12 16 16 16 39 9 4 8 3 n n n n 8 4 lcmp btcmp=? sh ck 1 gcao signal output ccd signal output bcmo signal output s/h blcmp a/d clock a/d output non-signaling part shielding pixel part effective pixel part sh ck1 lcmp s/h btcmp=? blcmp non-signaling part shielding pixel part effective pixel part unit : 1/sysck gcao signal output bcmo signal output a/d clock a/d output ccd signal output
41 mitsubishi digital assp ? M66335FP facsimile image data processor analog circuit timing chart (for cis1 mode) cis 1 00 h d 6 sens 0 02 h d 6 d 5 d 4 d 3 d 2 d 1 d 0 lcmps bls blcmps ccd cis 3 cis 2 cis 1 00000 01 address bit signal setting mode register 2 4 10 16 16 16 7 n n n 41 sh ck1 cis signal output lcmp=? s/h btcmp=? gcao signal output a/d clock a/d output unit : 1/sysck sh ck1 cis signal output lcmp=? s/h btcmp=? gcao signal output a/d clock a/d output
mitsubishi digital assp ? M66335FP facsimile image data processor 42 analog circuit timing chart (for cis2 mode) cis 2 00 h d 6 sens 0 02 h d 6 d 5 d 4 d 3 d 2 d 1 d 0 lcmps bls blcmps ccd cis 3 cis 2 cis 1 00000 10 address bit signal setting mode register 2 4 10 16 16 16 7 n n n 41 unit : 1/sysck sh ck1 cis signal output lcmp=? s/h btcmp=? gcao signal output a/d clock a/d output sh ck1 cis signal output lcmp=? s/h btcmp=? gcao signal output a/d clock a/d output
43 mitsubishi digital assp ? M66335FP facsimile image data processor analog circuit timing chart (for cis3 mode) 2 4 10 16 16 16 7 88 n n n n 41 41 sh ck1 cis signal output lcmp s/h btcmp=? blcmp non-signaling part shielding pixel part effective pixel part gcao signal output bcmo signal output a/d clock a/d output non-signaling part shielding pixel part effective pixel part unit : 1/sysck cis signal output sh ck1 lcmp s/h btcmp=? blcmp gcao signal output bcmo signal output a/d clock a/d output cis 3 00 h d 6 sens 1 02 h d 6 d 5 d 4 d 3 d 2 d 1 d 0 lcmps bls blcmps ccd cis 3 cis 2 cis 1 10101 00 address bit signal setting mode register
mitsubishi digital assp ? M66335FP facsimile image data processor 44 fig. 18 external pin connections of the analog part (for the ccd mode/bit clamping) av cc c1 c 2 gcao levaj bcmv=vbl levaj=vbl bcmv bcmi bcmo v cc dv cc v cc h 0.1f avdd mcis <3:1> , 4 s/h btcmpbls c1 c2 gcao levaj bcmv blcmp bcmi bcmo 0.1f ain lcmp agnd agnd dgnd gnd vri- vri+ vbl vwl adin agnd reset adck adin vrefl vrefh of dgnd ain mcis1 1.8v 8 3.8v 2.2v av cc dv cc vref- vri+ vri- vref+ b <7:1> of gain <7:0> agcsel <7:0> adck reset blcmp bls s/h btcmp mcis <3:1>,mccd lcmp din <6:0> gnd gca mcis2 mcis3 mccd max. 500mvpp. mccd reference voltage generating circuit for the a/d converter gain control circuit sample and hold circuit (bold line) : signal line (dashed line) : clock line bit clamping circuit internal reference voltage source for the analog circuits a/d converter digital circuit sensor output blanking element signaling element black white level shift circuit (2.2v) input clamping circuit black level clamping circuit + C + C + C C + C + + C C + + C C + C + C + C + C +
45 mitsubishi digital assp ? M66335FP facsimile image data processor av cc c1 c 2 gcao levaj bcmv=vbl levaj=vbl bcmv bcmi bcmo v cc dv cc v cc l h 0.1f avdd mcis <3:1> ,mccd 4 s/h btcmpbls c1 c2 gcao levaj bcmv blcmp bcmi bcmo 0.1f ain lcmp agnd agnd dgnd gnd vri- vri+ vbl vwl adin agnd reset adck adin vrefl vrefh of dgnd ain mcis1 1.8v 8 3.8v 2.2v av cc dv cc vref- vri+ vri- vref+ b <7:1> of gain <7:0> agcsel <7:0> adck reset blcmp bls s/h btcmp mcis <3:1>,mccd lcmp din <6:0> gnd gca mcis2 mcis3 mccd max. 500mvpp. reference voltage generating circuit for the a/d converter gain control circuit input clamping circuit sample and hold circuit level shift circuit (2.2v) bit clamping circuit internal reference voltage source for the analog circuits black level clamping circuit a/d converter digital circuit blanking element (bold line) : signal line (dashed line) : clock line black white signaling element fig. 19 external pin connections of the analog part (for the ccd mode/line clamping) + C + C + C + C + C + C + C + C + C + C + C + C + C
mitsubishi digital assp ? M66335FP facsimile image data processor 46 fig. 20 external pin connections of the analog part (for the cis1 mode) + av cc c1 c 2 gcao levaj bcmv bcmi bcmo v cc dv cc v cc l l h h avdd mcis<3:1>,mccd 4 s/h btcmpbls c1 c2 gcao levaj bcmv blcmp bcmi bcmo ain lcmp agnd agnd dgnd gnd vri- vri+ vbl vwl adin agnd reset adck adin vrefl vrefh of dgnd mcis1 1.8v 8 3.8v 2.2v av cc dv cc vref- vri+ vri- vref+ b<7:1> of gain<7:0> agcsel<7:0> adck reset blcmp bls s/h btcmp mcis<3:1>,mccd lcmp din<6:0> gnd gca mcis2 mcis3 mccd max.3.5v ?00mv r1 r 2 v levaj max.5pf bcmv=vbl a/d converter white black set r1 and r2 so that the following equation will hold. v levaj = v vbl ?a 5 b + 0.2 [v] where, a : minimum limit for dark voltage of the sensor g r : gain of the gain control circuit (1.8v) reference voltage generating circuit for the a/d converter gain control circuit input clamping circuit sample and hold circuit level shift circuit (2.2v) bit clamping circuit internal reference voltage source for the analog circuits black level clamping circuit sensor output sensor output digital circuit (bold line) : signal line (dashed line) : clock line + + + + + + + + + + + +
47 mitsubishi digital assp ? M66335FP facsimile image data processor fig. 21 external pin connections of the analog part (for the cis2 mode) + av cc c1 c 2 gcao levaj bcmv bcmi bcmo v cc dv cc v cc l l h h avdd mcis<3:1>, 4 s/h btcmpbls c1 c2 gcao levaj bcmv blcmp bcmi bcmo ain lcmp agnd agnd dgnd gnd vri- vri+ vbl vwl adin agnd reset adck adin vrefl vrefh of dgnd ain mcis1 1.8v 8 3.8v 2.2v av cc dv cc vref- vri+ vri- vref+ b<7:1> of gain<7:0> agcsel<7:0> adck reset blcmp bls s/h btcmp mcis<3:1>,mccd lcmp din<6:0> gnd gca mcis2 mcis3 mccd max.2v ?00mv r1 r 2 v levaj max.5pf mccd bcmv=vbl (in the case of the pixel clock of 1 mhz) a/d converter white black reference voltage generating circuit for the a/d converter gain control circuit input clamping circuit sample and hold circuit level shift circuit (2.2v) bit clamping circuit internal reference voltage source for the analog circuits black level clamping circuit sensor output digital circuit set r1 and r2 so that the following equation will hold. v levaj = v vbl ?a 5 b + 0.2 [v] where, a : minimum limit for dark voltage of the sensor g r : gain of the gain control circuit (1.8v) (bold line) : signal line (dashed line) : clock line + + + + + + + + + + + +
mitsubishi digital assp ? M66335FP facsimile image data processor 48 fig. 22 external pin connections of the analog part (for the cis3 mode) av cc c1 c 2 gcao levaj bcmv bcmi bcmo v cc dv cc v cc l h 0.1f avdd mcis <3:1> ,mccd 4 s/h btcmpbls c1 c2 gcao levaj bcmv blcmp bcmi bcmo 0.1f ain lcmp agnd agnd dgnd gnd vri- vri+ vbl vwl adin agnd reset adck adin vrefl vrefh of dgnd ain mcis1 1.8v 8 3.8v 2.2v av cc dv cc vref- vri+ vri- vref+ b <7:1> of gain <7:0> agcsel <7:0> adck reset blcmp bls s/h btcmp mcis <3:1>,mccd lcmp din <6:0> gnd gca mcis2 mcis3 mccd max. 2vpp. bcmv=vbl levaj=vbl sensor output reference voltage generating circuit for the a/d converter gain control circuit input clamping circuit sample and hold circuit level shift circuit (2.2v) white black bit clamping circuit internal reference voltage source for the analog circuits black level clamping circuit a/d converter digital circuit shielding pixel part signalding part (bold line) : signal line (dashed line) : clock line + C + C + C + C + C + C + C + C + C + C + C + C + C
49 mitsubishi digital assp ? M66335FP facsimile image data processor flowchart: reading operations (for the ccd sensor) n y y n specifying the scaling factor for vertical scanning uniformity correction ends. n y peak value detection start power on software reset s/h: sh pulse width sensor control cycle counter start pulse image processing parameters writing the dither pattern writing g correction table register 00 register 01 register 02 register 03 and 04 register 08 to 0e register 00 register 06 registers 01 and 11 to 14 register 05 register 10 register 0f register 00 register 00 white correction : 8 times register 00 register 15 agc: 16 times initial setting white correction setting for the original sheet image processing function reading an original sheet starts. reading the original sheet ends. register 06 and 07 register 15 register 00 register 00 a transfer to be continued? register 16 power off end power off? next original sheet? 1 line cycle 2 wait the light source is turned off. next original sheet agc starts. agc ends. uniformity correction starts. specifying the horizontal scanning resolution 1 line cycle 20 (or 16 or more) wait 1 line cycle 10 (or 8 or more) wait reading a single page register 17 n y page end? y int generated? n n y completed? completed? n y completed? a original sheet width and output width y n next original sheet srdy setting the light source is turned on. (white reference) becomes stable. register 06 specifying the vertical scanning resolution specifying the vertical scanning resolution
mitsubishi digital assp ? M66335FP facsimile image data processor 50 start power on software reset s/h: sh pulse width sensor control cycle counter start pulse image processing parameters writing the dither pattern register 00 register 01 register 02 registers 03 and 04 registers 08 to 0e register 05 register 10 register 0f initial setting writing the resolution change table registers 01 and 11 to 14 register 06 register 06 register 15 a writing g correction table uniformity correction starts. uniformity correction ends. agc starts. agc ends. registers 00 and 01 registers 00 and 01 register 00 register 00 register 00 register 00 black correction: 8 times agc: 16 times 1 line cycle 10 (or 8 or more) wait uniformity correction starts. uniformity correction ends. register 00 register 00 white correction: 8 times 1 line cycle 10 (or 8 or more) wait 1 line cycle 20 (or 16 or more) wait black correction white correction setting for the original sheet peak value detection image processing function reading an original sheet starts. reading the original sheet ends. int generated? n y page end? register 06 and 07 register 15 register 00 register 00 a register 16 register 17 srdy setting power off end 1 line cycle 2 wait next original sheet reading a single page specifying the scaling factor for vertical scanning n y y n transfer to be continued? y n y n power off? next original sheet? n y completed? n y completed? n y completed? specifying the horizontal scanning resolution specifying the vertical scanning resolution original sheet width and output width uniformity correction mode (black) uniformity correction mode (white) next original sheet the light source is turned on. (white reference) becomes stable. the light source is turned on. (white reference) becomes stable. the light source is turned off. the light source is turned off. becomes stable. reading operations (for the cis sensor)
51 mitsubishi digital assp ? M66335FP facsimile image data processor symbol v cc v i v o av cc dv cc v wl v bl v ain t stg parameter supply voltage input voltage output voltage analog supply voltage supply voltage reference voltage (white) reference voltage (black) analog input voltage storage temperature conditions ratings C0.3 ~ +6.5 C0.3 ~ v cc +0.3 0 ~ v cc v cc C0.3 ~ v cc +0.3 v cc C0.3 ~ v cc +0.3 C0.3 ~ av cc +0.3 C0.3 ~ av cc +0.3 C0.3 ~ av cc +0.3 C55 ~ +150 unit v v v v v v v v c limits symbol v cc gnd v i av cc a gnd dv cc d gnd v ain t opr parameter supply voltage (for the digital system component) gnd voltage input voltage analog supply voltage analog gnd voltage supply voltage (for the digital system component) gnd voltage input range: v wl avcc; v bl 3 a gnd operating temperature max. 5.25 v cc 5.25 5.25 2.2 +75 min. 4.75 0 4.75 4.75 1.8 C20 unit v v v v v v v v pCp c typ. 5.0 0.0 5.0 0.0 5.0 0.0 2.0 recommended operational conditions note: connect the analog system component and the digital system component separately to power supply on the evaluation board for noise prevention. absolute maximum ratings (t a = C20 ~ 75c unless otherwise noted)
mitsubishi digital assp ? M66335FP facsimile image data processor 52 h input voltage l input voltage positive direction input threshold negative direction input threshold hysteresis value h output voltage l output voltage h output voltage l output voltage h input current l input current h input current in the off state l input current in the off state analog input current reference resistance differential non-linear error static current dissipation (during standby) symbol v ih v il vt+ vt C v h v oh v ol v oh v ol i ih i il i ozh i ozl i ain r l e d i ccs ratings unit v v v v v v v v v m a m a m a m a m a w lsb ma parameter test conditions i oh = C12ma i ol = 12ma i oh = C4ma i ol = 4ma v cc = 5.25v v i = 5.25v v cc = 5.25v v i = 0v v cc = 5.25v v o = 5.25v v cc = 5.25v v o = 0v v cc = 5.25v v i = v cc , gnd typ. 0.2 120 1.0 21 min. 2.0 0.6 v cc C0.8 v cc C0.8 max. 0.8 2.4 0.55 0.55 1.0 C1.0 5.0 C5.0 1.0 35 electrical characteristics (t a = C20 ~ 75c, v cc = 5v5% unless otherwise noted)
53 mitsubishi digital assp ? M66335FP facsimile image data processor symbol t c (sys) t w+ (sys) t wC (sys) t r (sys) t f (sys) t w (rd) t su (cs-rd) t su (a-rd) t su (dak-rd) t h (rd-cs) t h (rd-a) t h (rd-dak) t w (wr) t su (cs-wr) t su (a-wr) t su (d-wr) t h (wr-cs) t h (wr-a) t h (wr-d) t h (stim-srdy) system clock cycle system clock h pulse width system clock l pulse width system clock rise time system clock fall time read pulse width set-up time before read set-up time before read set-up time before read hold time after read hold time after read hold time after read write pulse width set-up time before write set-up time before write set-up time before write hold time after write hold time after write hold time after write hold time after stim ratings unit ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns parameter typ. min. 50 25 25 100 20 20 20 10 10 10 100 20 20 50 20 10 0 0 max. 20 20 cs a0 ~ a4 dak cs a0 ~ a4 dak cs a0 ~ a4 d0 ~ d7 cs a0 ~ a4 d0 ~ d7 srdy test conditions timing conditions (t a = C20 ~ 75c, v cc = 5v5% unless otherwise noted)
mitsubishi digital assp ? M66335FP facsimile image data processor 54 symbol __ t pzl (rd-d) __ t pzh (rd-d) __ t plz (rd-d) __ t phz (rd-d) __ t phl (rd-dro) enable time for data output after read disable time for data output after read propagation time of dro output after read ratings unit ns ns ns ns ns parameter test conditions c l = 150pf c l = 50pf typ. min. 10 max. 75 50 50 parameter t plh , t phl t plz t phz t pzl t pzh sw1 open closed closed closed open sw2 open open closed open closed test circuit system clock t w+(sys) sysck 1.3v 1.3v 1.3v 3v 0v 10% 10% 90% 90% t c(sys) t w-(sys) t f(sys) t r(sys) input p.g tested device gnd v cc r l=1k w c l 50 w r l=1k w sw1 sw2 v cc output (1) characteristics (10% to 90%) of the pulse generator (pg): t r = 3ns; t f = 3ns (2) capacitance cl (= 150pf) includes the stray capacitance of connections and in- put capacitance of the probe. switching characteristics (t a = C20 ~ 75c, v cc = 5v5% unless otherwise noted)
55 mitsubishi digital assp ? M66335FP facsimile image data processor 1.3v 1.3v 1.3v 1.3v 1.3v 1.3v 1.3v 1.3v t su(cs-wr) t su(d-wr) t su(wr-d) t h(wr-cs) t su(a-wr) t w(wr) t h(wr-a) effective data 3v 0v 0v 0v 0v 3v 3v 3v cs a0~a4 wr d0~d7 1.3v 1.3v 1.3v 1.3v 50% 10% 90% 50% 1.3v 1.3v t su(cs-rd) t h(rd-cs) t su(a-rd) t pzh(rd-d) t pzl(rd-d) t phz(rd-d) t plz(rd-d) t w(rd) t h(rd-a) 3v 0v 0v 0v v oh 3v 3v v ol cs a0~a4 rd d0 ~ d7 d0~d7 2) timing for write operation (mpu ? m66335) mpu interface 1) timing for read operation (m66335 ? mpu)
mitsubishi digital assp ? M66335FP facsimile image data processor 56 1.3v 50% t h(stim-srdy) 3v 0v v ol v oh v ol v oh v ol v oh srdy stim sclk svid 50% 50% 1.3v 1.3v 50% 10% 90% 50% 1.3v 1.3v 120/sysck t su(dak-rd) t phl(rd-drq) t pzh(rd-d) t pzl(rd-d) t phz(rd-d) t plz(rd-d) t w(rd) t h(rd-dak) v oh v ol 0v 0v v oh 3v 3v v ol drq dak rd d0~d7 d0~d7 dma timing timing for read operation (m66335 ? system bus) timing of codec
57 mitsubishi digital assp ? M66335FP facsimile image data processor cautions for use (1) access to address 00h to gain access to address 00h, the value of built-in gcc (gain control counter) may be set to ffh. this requires to read gain value at address 18h before access to address 00h and write the gain value at address 18h after the access (see flowchart a). start read gain value at address 18h access address 00h write gain value at address 18h end flowchart a. address 00h access flow (2) reset the M66335FP adopts the two types of reset. these reset func- tions are provided in table a. table a. reset functions register initialization reset type hardware reset (reset) software reset register 0 (reset) function internal f/f initialization gcc initialization

▲Up To Search▲

Price & Availability of M66335FP

	To Download M66335FP Datasheet File
If you can't view the Datasheet, Please click here to try to view without PDF Reader .