MITSUBISHI M66332FP

MITSUBISHI
MITSUBISHI
〈DIGITAL
〈DIGITAL
ASSP〉
ASSP〉
M66332FP
M66332FP
FACSIMILE
IMAGE
DATA
PROCESSOR
FACSIMILE
IMAGE
DATA
PROCESSOR
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
γ 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 × 4 matrix)
Separation of image data area (1 dimension)
Scale down A3 → B4, A3 → A4, B4 → A4
0 5V Single Power Supply
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, γ 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)
APPLICATION
Facsimiles
29
30
31
32
33
34
35
36
37
38
39
40
41
42
45
46
47
48
49
50
51
52
53
54
55
56
28
27
26
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
M66332FP
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
Analog Signal
Processing Interface
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
GND
VCC
CS
RD
MPU Interface
WR
RESET
GND
GND
DRQ
DMA Interface
DAK
VCC
SH
Sensor Interface
43
44
D7
D6
D5
D4 MPU
Interface
D3 (DATA)
D2
D1
D0
GND
VCC
A3
A2 MPU
Interface
A1 (ADDRESS)
A0
DGND
DVCC
PIN CONFIGURATION (TOP VIEW)
NC: No Connection
Outline 56P6N-A
1
MITSUBISHI 〈DIGITAL ASSP〉
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
BLOCK DIAGRAM
System Clock
ADC Analog
Vcc
AVCC
18
SYSCK
8
15
56
2
1
3
CLAMP
S/H
AGC
DSCH
4
5
16
17
Analog
Signal
Interface
VCC
10 35 46 55
Image Processing Sequence Control Signal
Sensor Control To each
block
Detection
of Image
Data Area
Analog Control
Correction memory
(304 words × 5bits)
Uniformity Correction
MTF compensation
PTIM
SH
Sensor
CK1
Interface
CK2
RS
ADC Logic
Vcc
DVCC
29
Convert
to
bi-level
Separation
of Image
Data Area
Simple Bi-level
Conversion/Background
and object Level Detection
SRAM 16
words
× 4bits
Cur out/
Scale
down
DMA Control
Collective
Dithering
(16 levels)
MPU Bus Interface
14
13
12
11
SRDY
STIM CODEC
SCLK Interface
SVID
54 DAK DMA
53 DRQ Interface
50
47
48
49
31
~
RESET
CS
MPU
Interface
RD
WR
A0 MPU
Interface
34 A3 (Address)
37 D0 MPU
44 D7 Interface
(Data)
5bit A-D converter
(flash type)
ASIG 27
~
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
Specifications
• A4, B4, A3
• 8 pixels/mm (primary scanning direction)
Read speed
• 5ms/line Typ. 2ms/line maximum
Uniformity Correction
MTF Compensation
Simple Bi-level Conversion
• White correction only
• Block correction in units of 8 pixels
• 50% Correction range
• Laplacian filter circuit for 3 × 1 pixels in current line
(1 dimension)
• Floating threshold method using background and
object level detection circuit
Pseudo half-tone
• Dither method: 16 levels (4 × 4matrix)
Separation of Image Data
Area
• Detection by brightness difference in 5 × 1 pixels
area in current line
• Selection method
• Scale down: A3 → B4 set
to 13/15; B4 → A4, 9/11; A3 → A4, 12/17
Scale down
γ Correction
Image Sensor Control Signal
Analog Signal Processing
2
• 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 ×
5bits) (read/write allowed from MPU)
• No need for compensation memory
• Built-in SRAM as dither memory (16 words × 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
MITSUBISHI 〈DIGITAL ASSP〉
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
PIN DESCRIPTIONS
Block
Sensor Interface
Pin Names
SH
I/O
O
CK1
O
DMA Interface
CK2
RS
PTIM
CLAMP
S/H
ASIG
AGC
DSCH
SRDY
STIM
SCLK
SVID
DRQ
O
O
O
O
O
I
O
O
I
O
O
O
O
Clock
MPU Interface
DAK
SYSCK
RESET
I
I
I
CS
I
Analog Circuit
Interface
CODEC Interface
RD
WR
A0~A3
D0~D7
I
I
I
I/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.
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 (V WL ) 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 × 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
v
SCAN 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
v
Data transfer
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.
(Read operation)
AGC mode
v
SCAN 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.
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
Image sensor
SYSCK
Image Processing Sequence Control Signal
PTIM
SH
CK1
CK2
RS
STIM
SCLK
SVID
Sensor Control
Detection
of Image
Data Area
To each
block
Separation
of Image
Data Area
Convert to
bi-level
Cut out/
Scale down
SRDY
Correction memory
(304 words × 5bits)
Uniformity Correction
Resolution compensation
Analog signal processing circuit
Analog Control
CLAMP
S/H
AGC
DSCH
CODEC
DAK
Simple Bi-level
Conversion/Background
and object level detection
DMA
DMA Control
DRQ
SRAM 16
words
× 4bits
Collective
Dithering
(16 levels)
MPU Bus Interface
RESET
CS,RD
WR
A0~A3
MPU
ASIG
5bit A-D converter
(flash type)
D0~D7
VBL
VML2
VWL
VML1
VML3
Data Flow in Creation of Uniformity Correction Data
Image sensor
SYSCK
Image Processing Sequence Control Signal
PTIM
SH
CK1
CK2
RS
STIM
SCLK
SVID
Sensor Control
Detection
of Image
Data Area
To each
block
Separation
of Image
Data Area
Convert to
bi-level
Cut out/
Scale down
SRDY
Correction memory
(304 words × 5bits)
Uniformity Correction
Simple Bi-level
Conversion/Background
and object level detection
DAK
DMA
DMA Control
DRQ
SRAM 16
words
× 4bits
Collective
Dithering
(16 levels)
MPU Bus Interface
RESET
CS,RD
WR
A0~A3
MPU
ASIG
5bit A-D converter
(flash type)
VBL
VML2
VWL
VML1
VML3
6
MTF Compensation
Analog signal processing circuit
Analog Control
CLAMP
S/H
AGC
DSCH
CODEC
D0~D7
MITSUBISHI 〈DIGITAL ASSP〉
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
Date Flow During Read Operation (for serial output)
Image sensor
SYSCK
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
Uniformity Correction
MTF Compensation
Analog signal processing circuit
Correction memory
(304 words × 5bits)
ASIG
CODEC
SRDY
Analog Control
CLAMP
S/H
AGC
DSCH
STIM
SCLK
SVID
DAK
Simple Bi-level
Conversion/
Background
and object
level detection
SRAM 16
words
× 4bits
DMA
DMA Control
DRQ
Collective
Dithering
(16 levels)
MPU Bus Interface
RESET
CS,RD
WR
A0~A3
MPU
5bit A-D converter
(flash type)
D0~D7
VBL
VML2
VWL
VML1
VML3
: Image date
: Correction, Compensation date
Date Flow During Read Operation (for parallel output)
Image sensor
SYSCK
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
Uniformity Correction
Resolution compensation
Analog signal processing circuit
Correction memory
(304 words × 5bits)
ASIG
CODEC
SRDY
Analog Control
CLAMP
S/H
AGC
DSCH
STIM
SCLK
SVID
DAK
Simple Bi-level
Conversion/
Background
and object
level detection
SRAM 16
words
× 4bits
DRQ
Collective
Dithering
(16 levels)
MPU Bus Interface
5bit A-D converter
(flash type)
VBL
VML2
VWL
VML1
VML3
DMA
DMA Control
RESET
CS,RD
WR
A0~A3
MPU
D0~D7
: 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 × pixel transmission clock period [NS]
= (PRE_DATA + 1) × 1/ADCK [NS]
= (PRE_DATA + 1) × 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
PRE_DATA load
• AGC range (AGCG):
• Source document read
width:
• Pulse motor control
signal (PTIM):
Down count
0
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.
Relationship with registers
Line period
(ACCK)
Registers 2, 3
(PRE_DATA)
Sensor start
pulse
(SH)
Register 0 (SENS_W)
Register 4 (ST_PL)
ST_PL
Uniformity correction
range
(UNIFG)
Register 0 (SENS_W)
AGC range
(AGCG)
Register 0 (SENS_W)
Register 1 (SOURCE)
Source document
read width
Register 1
(SOURCE, DEST , REDU)
Pulse motor
control
(PTIM)
Fig. 1 Line period and read sequence
8
• Uniformity correction
range (UNIFG):
1 line period
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 3 Source document read width according to
sensor width and source document size
Table 2 Sensor width and gate signal range
Sensor width
A3
B4
A4
Sensor width
Source document size
A3
B4
A4
2487/55
2279/231
2119/391
A3
2487/55
––
––
AGC mode
2487/55
2279/231
2119/391
B4
2278/230
2278/230
––
SCAN mode
2370/162
2194/306
1760/740
A4
2118/390
2118/390
2118/390
Gate signal
Uniformity correction
range (UNIFG)
AGC
range
(AGCG)
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
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Line period
DSCH
(raise gain)
AGC
A-D converter
overflow
(lower gain)
Peak detection
preprocessing
Peak gain control
: Generated in pixel units only when the A-D converter overflows.
Fig. 3 Peak detection
Peak detection (preprocessing)
Peak gain control
After completing peak
detection preprocessing
VWL
VWL
White data
VBL
Match the peak sensor
output within a line toVWL
Match the output
level of the last pixel
of a line to VWL
1 line
Fig. 4 Changes in gain during peak detection
10
After completing gain
control for peak value
VBL
1 line
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 × 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
Pixels
Primary scanning direction
1
8
9
Correction data 1
16 17
Correction data 2
24
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
0
1 line
If correction range is exceeded (during white correction)
Analog signal input
White data
VWL
25-1
50%
24-1
VBL
0
1 line
Fig. 7 Uniformity correction
12
White data exceeding
correction range
Portion exceeding correction range
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-
Photoelectric conversion
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
Image signal
Source document (characters)
MTF compensation
Compensated data
Photoelectric
conversion
Image signal
Source document (photo)
MTF compensation
Compensated data
MTF compensation
X’ = X + α ((X – A) + (X – B))
α = Primary scanning compensation coefficient, register 5 (MTF)
Secondary scanning direction
Window 3 × 1
F
G
A
X
B
Primary scanning
direction
F–G
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)
Background
level counter
Image data
Threshold
generation
Bi-level
data
Compare
Object level
counter
White level
Background level
Threshold
level
Input data
Object level
Threshold level = (background level - object level) × K + object level
K = bi-level threshold coefficient: register 9 (SLICE)
Fig. 9 Background-object level
14
Black level
MITSUBISHI 〈DIGITAL ASSP〉
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
• Dither method
The M66332 has a built-in 16 words × 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 × 4 dither
matrix.
: register E (DITH_D)
Table 4 Scanning line density and dither matrix size
4 × 4 matrix sequence
0
12
3
15
8
4
11
7
2
14
1
13
10
6
9
5
a)Diffused
type
4
15
9
14
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)
10
0
3
8
5
2
1
13
11
6
12
7
b) Concentrated
type
11
12
7
3
4
0
8
15
6
2
10
13
9
14
5
1
Scanning Line
Density
Normal
Fine
Primary/Secondary
Scanning Line
(line/mm)
8 × 3.85
8 × 7.7
Level
Matrix
Size
––
16
––
4×4
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
Black and white image
conversion is applied to the bi-level conversion area and
dither method is applied to the gradation area.
: register 5 (MODE)
Area
Process
Example
Bi-level
conversion area
Simple bi-level
conversion
Character
Gradation
conversion area
Pseudo half-tone
conversion
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 × 1 window.
This characteristic is used to distinguish between gradation conversion area and bi-level conversion area.
Secondary scanning direction
5 × 1 window
Lmax: maximum illumination in window
Lmin : minimum illumination in window
X
Current line
Primary scanning
direction
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
Minimum
Lmin
Maximum
Lmax
A
Input data
B
C
Lmin
Black level = 0
Lmax – Lmin
Fig. 11 Separation of image data area
16
Lmax
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
A3
B4
A4
B4
13/15
1
––
A4
12/17
9/11
1
OUT
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
When ST_PL
is odd
CK1=φ 1
CK2=φ 2
φ1
φ2
When ST_PL
is even
CK1=φ 2
CK2=φ 1
RS
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
12
SH
2
2
16
φ1
φ2
16
8
RS
4
4
(OS)
N
16
3 4
S/H
9
16
CLAMP
OUT (ASIG)
(A-D converter
clock)
18
13
2
1
N
MITSUBISHI 〈DIGITAL ASSP〉
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
CCD-line clamp type
SH
When ST_PL
is odd
CK1=φ 1
CK2=φ 2
φ1
φ2
When ST_PL
is even
CK1=φ 2
CK2=φ 1
RS
Sensor output
(OS)
S/H
(Line clamp
area)
CLAMP
OUT
(ASIG)
(A-D converter
clock)
8: fixed
Register 4: STPL = dummy pixel + 2
Valid pixel area
Dummy area
Note: Line clamp uses sensor output equivalent to
(dummy area –8) pixels from the first pixel after
SH.
Unit: 1/SYSCK
12
SH
2
2
16
φ1
φ2
16
8
RS
(OS)
S/H
CLAMP
4
4
N
16
3 4
9
3 4
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
SH
16
16
CK1
CK2
6
8
2
6
8
2
16
(SIG)
N
16
4
1
Select with
register 5
S/H_W
S/H
2
1
16
16
CLAMP
OUT (ASIG)
2
1
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
B
STIM
10 4
12
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
1
2
3
4
5
6
7
8
N+3
N+4
N+5
N+6
N+7
SCLK
SVID
N–1
N
N+1
N+2
N+8
DRQ
DAK
RD
D0
N–1
D1
N–2
D2
N–3
D3
N–4
D4
N–5
D5
N–6
D6
N–7
D7
N–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
(6) Read/write to dither memory, uniformity correction
memory
The following figures show the sequence for writing and readDither memory write (MPU → M66332)
ing dither patterns in the 16 words × 4 bits collective dithering
SRAM built in the M66332.
Initial setting (1)
Initial setting (2)
Memory address (0)
Memory address (1)
1H
0H
D4 = “0”
D7 = “1”
DATA (0)
DATA (1)
1
2
3
3
CS
A3 ~ A0
EH
EH
~
WR
D7 ~ D0
(Input)
Dither memory read (M66332 → MPU)
Initial setting (1)
Initial setting (2)
1H
0H
D4 = “0”
D7 = “1”
Memory address (0)
Memory address (1)
CS
A3 ~ A0
EH
EH
~
WR
D7 ~ D0
(Input)
RD
D7 ~ D0
(Output)
1
2
DATA (0)
DATA (1)
4
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
A1
A2
A3
A4
A5
A6
A7
A8
A9 A10 A11
A12 A13 A14 A15
4 × 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
in backup memory during power off. The following figures
show the uniformity correction data read/write sequence.
Uniformity correction memory write (MPU → M66332)
Initial setting
Closing setting
CS
A3~A0
1H
2H
3H
1
2
FH
WR
D7~D0
(Input)
D7 = “1”
Repeats
steps
1, 2,
and 3 .
DATA
D4 = “0”
1H
D7 = “0”
3
Uniformity correction memory read (M66332 → MPU)
Initial setting
Closing setting
CS
A3~A0
1H
2H
3H
FH
WR
D7~D0
(Input)
1H
Repeats
steps
1, 2,
and 4 .
D7 = “1”
D4 = “0”
D7 = “0”
RD
D7~D0
(Output)
DATA
1
2
4
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
Image sensor
SYSCK
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
Correction memory
(304 words × 5bits)
Uniformity Correction
MTF Compensation
Analog signal processing circuit
CODEC
SRDY
Analog Control
CLAMP
S/H
AGC
DSCH
STIM
SCLK
SVID
DAK
Simple Bi-level
Conversion/
Background
and object
level detection
SRAM 16
words
× 4bits
DMA
DMA Control
DRQ
Collective
Dithering
(16 levels)
MPU Bus Interface
RESET
CS,RD
WR
A0~A3
MPU
ASIG
5bit A-D converter
(flash type)
D0~D7
VBL
VML2
VWL
VML1
VML3
: Dither date
Uniformity correction memory write/read
Image sensor
SYSCK
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
Uniformity Correction
MTF Compensation
Analog signal processing circuit
Correction memory
(304 words × 5bits)
DAK
Simple Bi-level
Conversion/
Background
and object
level detection
SRAM 16
words
× 4bits
DMA
DMA Control
DRQ
Collective
Dithering
(16 levels)
MPU Bus Interface
RESET
CS,RD
WR
A0~A3
MPU
ASIG
5bit A-D converter
(flash type)
VBL
VML2
VWL
VML1
VML3
24
CODEC
SRDY
Analog Control
CLAMP
S/H
AGC
DSCH
STIM
SCLK
SVID
D0~D7
: Uniformity correction date
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
(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.
Reset Period
Counter
Stop Clock
Generator
Operation
Stop Line Read
• Pseudo half-tone conversion, dither method
Select collective dithering (16 gradations using 4 × 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
Uniformity
Correction
Resolution
Compensation MTF
Simple
Bi-Level
Conversion
Yes
1/2
Dithering
Yes
Separation
of Image
Data Area
Yes
Image
Background and Object Level
SLICE MAX
UP
MAX
DOWN
MIN
UP
UL
MIN
LL
MAX
5/8
Normal
Normal
Normal
04H
0AH
MON
––
––
––
––
––
––
MON
5/8
Normal
Normal
Normal
04H
0AH
0
8
1
12
2
6
4
A
3
14
3
11
2
15
C
9
4
9
2
13
5
8
6
1
7
D
10
2
7
5
B
3
Dither pattern (γ = 0.8)
26
γ
Correction
Dither
Pattern
No
––
γ=0.9
VML1=1.1V
VML2=2.2V
VML3=3.5V
γ=0.9
VML1=1.1V
VML2=2.2V
VML3=3.5V
4×4
diffusion
pattern,
γ = 0.8
4×4
diffusion
pattern,
γ = 0.8
Separation of Image
Data Area
SEPA SEPA SEPA
A
B
C
––
––
––
––
––
––
06H
0DH
01H
MITSUBISHI 〈DIGITAL ASSP〉
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
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
γ;0.8
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
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
Fig. 13 Thresholds for collective dithering : Example 1
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
• γ correction
γ 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.
γ = 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 γ correction by means of both collective dithering and the middle reference voltage pins (VML1, VML2,
and VML3) of the A-D converter.
(γ Correction by Collective Dithering )
γ correction is realized applying a γ 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 γ characteristic, γ, to 0.8.
(γ Correction by the Middle Reference Voltage Pins of the AD converter)
The example shown in Fig. 16 is an approximation of γ characteristic, γ, 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.
Threshold sequence
Threshold sequence
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.
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Threshold
Fig. 15 An example of γ correction by dither matrix
γ=1
γ;0.9
VML1
=1.1V
VML2
VML3
=2.2V
=3.5V
Analog input
White
= 5V
Fig. 16 An example of γ 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.
1 line period
(ACCK)
N period
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.
N + 1 period N + 2 period
Sensor
start
pulse (SH)
SRDY
(SSCAN)
(BCSCAN)
STIM
PTIM
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
R/W
0H
W
Explanation
D7
D6
RESET
SENS
D5
D4
SENS_W
D7
0
1
RESET System Reset
Normal Mode
Reset Mode
D6
0
1
SENS Sensor Type
CCD
CIS
D5
0
0
1
1
D4
0
1
0
1
D3
D2
D1
D0
UMODE
AGC
UNIF
SCAN
(Default is 00H)
Reset while write pulse is “L” when D7 = “1”
SENS_W Sensor Width
A4
B4
A3
––
D3
0
1
UMODE CIS Uniformity Correction Mode
With White Correction
No White Correction
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.
• For selecting with or without correction in SCANning
(with CIS only)
29
MITSUBISHI 〈DIGITAL ASSP〉
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
Address
R/W
1H
W
Explanation
D7
D6
D5
D4
UM_R/W
STNBY
TIME
P_0
D2
SOURCE
D1
D0
DEST
REDU
(Default is 00H)
D7
0
1
UM_R/W Uniformity Correction Memory Read/
Write
Normal
UNIF Memory Read/Write
D6
0
1
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.
D5
0
1
TIME Line Time
1 Line Period
2 Line Period
• 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.
D4
0
1
P_O Parallel Output
Without Parallel Output
Parallel Output
• 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.
D3
0
0
1
1
30
D3
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.
MITSUBISHI 〈DIGITAL ASSP〉
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
Address
R/W
2H
W
Explanation
D7
D6
D5
D4
D3
D2
D1
D0
PRE_DATA <7:0>
(Default is 00H)
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
D6
0
D_LOAD
D6
0
1
D5
D4
D3
D2
D1
D0
PRE_DATA <12:8>
Uniformity correction memory address setting mode.
Normal
Data load
(Default is 00H)
• 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
ST_PL <6:0>
D1
D0
(Default is 00H)
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
R/W
5H
W
6H
Explanation
D7
D6
D5
D4
S/H_W
SH_W
CLAMP
D3
D2
D1
MODE
D6
0
1
S/H_W Pulse Width of S/H
Normal (Sysclk period × 4)
Normal × 0.5
D5
0
1
SH_W SH pulse width for CIS
Normal
Reverse of (normal × 2)
D4
0
1
CLAMP Clamp Method of Analog Circuit
Bit Clamp
Line Clamp
D0
MTF
(Default is 00H)
• In case of CCD, there is only one SH pulse
width and this register is ignored.
• 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
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)
W
D7
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
SEPA_B
D3~D0: SEPA_B Separation of Image Data Area (MIN.)
32
D0
(Default is 00H)
MITSUBISHI 〈DIGITAL ASSP〉
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
Address
R/W
8H
W
Explanation
D7
D6
D5
D4
D3
D2
D1
D0
(Default is 00H)
SEPA_C
D3~D0: SEPA_C Separation of Image Data Area (MAX.)
9H
W
D7
D6
SLICE
D5
D4
MAX_UP
D3
D2
MAX_DOWN
D1
D0
MIN_UP
(Default is 00H)
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) × 32)
Slow (T = (1 pixel period) × 64)
Fast (T = (1 pixel period) × 16)
Faster (T = (1 pixel period) × 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) × 128)
Slow (T = (1 pixel period) × 256)
Fast (T = (1 pixel period) × 64)
Faster (T = (1 pixel period) × 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) × 512)
Slow (T = (1 pixel period) × 1024)
Fast (T = (1 pixel period) × 256)
Faster (T = (1 pixel period) × 128)
T: Counter clock period
33
MITSUBISHI 〈DIGITAL ASSP〉
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
Address
R/W
AH
W
Explanation
D7
D6
D5
D4
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
D1
D0
DITH_D
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
D1
D0
UNIF_D
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 ~ 75˚C, 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 ~ 75˚C, VCC = 5 V ± 10%, unless otherwise noted)
Symbol
Parameter
VIH
“H” Input voltage
VIL
“L” Input voltage
VT+
VT–
VH
VOH
VOL
VOH
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
VOL
IIH
“H” input current
IIL
“L” input current
IOZH
IOZL
IAIN
RL
SINL
ICCS
ICCA
ICC
SYSCK, SRDY, DAK,
CS, RD, WR,
A0~A3, D0~D7
SYSCK, SRDY,
DAK,RESET, CS,
RD, WR,
A0~A3
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
Test conditions
SYSCK=8MHz
Limits
Typ.
Max.
0.8
V
2.4
V
V
V
V
V
V
0.2
IOH=–12mA
IOL=12mA
IOH=–4mA
Unit
V
0.6
VCC–0.8
0.55
VCC–0.8
0.55
V
1.0
µA
–1.0
µA
5.0
µA
–5.0
µA
±10
µA
kΩ
±0.5
±1.0
LSB
VCC=5.5V
VI=VCC, GND
10
20
mA
VCC=5.5V
VI=VCC, GND
15
40
mA
VCC=5.5V
VI=VCC,GND
40
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
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
Min.
2.0
mA
MITSUBISHI 〈DIGITAL ASSP〉
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
TIMING REQUIREMENTS (Ta = –20 ~ 75˚C, 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)
Test conditions
Parameter
System clock
Read pulse
Write pulse
STIM
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
Min.
Limits
Typ.
125
62.5
62.5
Max.
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
Unit
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
37
MITSUBISHI 〈DIGITAL ASSP〉
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
SWITCHING CHARACTERISTICS (Ta = –20 ~ 75˚C, VCC = 5 V ± 10%, unless otherwise noted)
Parameter
Symbol
tPZL(RD-D)
tPZH(RD-D)
tPLZ(RD-D)
tPHZ(RD-D)
tPHL(RD-DRQ)
Test conditions
Min.
Limits
Typ.
Output enable time to low-level and high-level (RD-D)
Max.
Unit
75
ns
50
ns
50
ns
CL=150pF
Output disable time from low-level and high-level (RD-D)
High-level to low-level output propagation time (RD-DRQ)
10
CL=50pF
Test Circuit
INPUT
VCC
OUTPUT
VCC
Parameter
tPLH, tPHL
tPLZ
tPHZ
tPZL
tPZH
RL=1kΩ
P.G
SW1
DUT
SW2
CL
50Ω
SW1
Open
Closed
Open
Closed
Open
SW2
Open
Open
Closed
Open
Closed
(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.
RL=1kΩ
GND
System Clock
tc(SYS)
tf(SYS)
tw+(SYS)
tw-(SYS)
90%
SYSCK
1.3V
tr(SYS)
1.3V
90%
3V
1.3V
10%
10%
0V
38
MITSUBISHI 〈DIGITAL ASSP〉
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
MPU Interface
(1) Read timing (M66332 → MPU)
3V
CS
1.3V
1.3V
0V
tsu(CS-RD)
th(RD-CS)
3V
A0~A3
1.3V
1.3V
0V
tsu(A-RD)
tw(RD)
th(RD-A)
3V
1.3V
RD
1.3V
0V
tPZL(RD-D)
tPLZ(RD-D)
50%
D0~D7
10%
tPZH(RD-D)
VOL
tPHZ(RD-D)
VOH
90%
50%
D0~D7
(2) Write timing (MPU → M66332)
3V
CS
1.3V
1.3V
0V
tsu(CS-WR)
th(WR-CS)
3V
A0~A3
1.3V
1.3V
0V
tsu(A-WR)
tw(WR)
th(WR-A)
3V
WR
1.3V
1.3V
0V
tsu(D-WR)
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%
50%
VOL
120/SYSCK
tPHL(RD-DRQ)
3V
DAK
1.3V
1.3V
0V
tsu(DAK-RD)
tw(RD)
th(RD-DAK)
3V
1.3V
RD
1.3V
0V
tPZL(RD-D)
tPLZ(RD-D)
50%
D0~D7
10%
tPZH(RD-D)
VOL
tPHZ(RD-D)
90%
VOH
50%
D0~D7
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
A
Power ON
Soft reset
Wait 1 line
period × 2
Turn off light
Register 0
Register 1
Set standby mode
S/H, SH pulse width
CLAMP
Register 5
Output format
Register 1
Period counter
Registers 2, 3
Register 4
Initial settings
Start pulse
Power OFF?
Y
N
Next document
Y
Next document?
N
Image data area separation parameter Registers 6, 7, 8
Power OFF
Background and object level parameter Registers A, B
Dither pattern write
N
End
Register E
Complete?
Next document
Y
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
Start document read
N
Register 5
Registers 5, 9
Register 0
Complete?
Y
End document read
Y
Setting for each White compensation
document
Start uniformity correction
Wait 1 line
period × 30
(24 or more)
Register 1
Register 0
Read a page
Reset standby mode
Turn on light
Stabilize
(white reference)
Transfer?
N
A
41
MITSUBISHI 〈DIGITAL ASSP〉
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
Uniformity correction data creation, transmission (sensor: CIS)
Start
Soft reset
Register 0
S/H, SH pulse width
CLAMP
Register 5
Output format
Register 1
Period counter
Registers 2, 3
Start pulse
Start uniformity correction
Wait 1 line
period × 30
(24 or more)
Turn off light
Register 0
AGC + white compensation
AGC: 16 times
white compensation : 8 times
End uniformity correction
Register 0
Transmit correction data
Registers 1, 2, 3,
F
N
Complete?
Y
End
42
Register 4
stabilize
Transmit uniformity Uniformity correction
correction data
/white
Turn on light
Initial settings
Power ON
MITSUBISHI 〈DIGITAL ASSP〉
M66332FP
FACSIMILE IMAGE DATA PROCESSOR
Read operation (sensor: CIS)
Start
A
Power ON
Soft reset
Register 0
S/H, SH pulse width
CLAMP
Register 5
Wait 1 line
period × 2
Turn off light
Set standby mode
Power OFF?
Output format
Register 1
Period counter
Registers 2, 3
Background and object level parameter Registers A, B
Dither pattern write
N
Next document
Register 4
Image data area separation parameter Registers 6, 7, 8
Y
N
Y
Next document?
N
Initial settings
Start pulse
Register 1
Register E
Power OFF
End
Complete?
Y
Write to uniformity correction memory Register F
N
Complete?
Next document
Y
Reset standby mode
Register 1
End AGC
Register 0
Document width, specification width Register 1
Bi-level mode
Image processing function
Start document read
N
Register 5
Registers 5, 9
Register 0
Complete?
Y
End document read
Y
Settings for each
document
Register 0
(AGC: 16 times)
Read a page
Start AGC
Wait 1 line
period
x 20 (16 or more)
AGC
Turn on light
stabilize
(white reference)
Register 0
Transmit?
N
A
43