Difference between revisions of "Mitsubishi M64282FP"

From GbdevWiki
Jump to: navigation, search
Line 178: Line 178:
  
 
=== X<sub>ck</sub>, SIN timing ===
 
=== X<sub>ck</sub>, SIN timing ===
[[File:M64282FP-xck-sin-timing.png]]
+
[[Image:M64282FP-xck-sin-timing.png|none|frame|X<sub>ck</sub>, SIN timing]]
  
 
=== x<sub>ck</sub> LOAD timing ===
 
=== x<sub>ck</sub> LOAD timing ===
[[File:M64282FP-xck-load-timing.png]]
+
[[Image:M64282FP-xck-load-timing.png|none|frame|x<sub>ck</sub> LOAD timing]]
  
 
=== X<sub>ck</sub>, X<sub>rst</sub>, RESET timing ===
 
=== X<sub>ck</sub>, X<sub>rst</sub>, RESET timing ===
[[File:M64282FP-xck-xrst-reset-timing.png]]
+
[[Image:M64282FP-xck-xrst-reset-timing.png|none|frame|X<sub>ck</sub>, X<sub>rst</sub>, RESET timing]]
  
 
=== X<sub>ck</sub>, START timing ===
 
=== X<sub>ck</sub>, START timing ===
[[File:M64282FP-xck-START-timing.png]]
+
[[Image:M64282FP-xck-START-timing.png|none|frame|X<sub>ck</sub>, START timing]]
 +
 
 +
== Operation ==
 +
[[Image:M64282FP-operation-flowchart.png|none|frame|Fig. 11-1 Operation Flow Chart]]
 +
 
 +
 
 +
Figure 11-1 shows the image sensing sequence. First of all, all the registers must be reset and must be initialized to the appropriate values. The reset sequence completes when both X<sub>rst</sub> and RESET signals are set low. There are 8 sets of registers, each of which consists of 8 bits of data. Each input data consists of 11 bits; of these 11 bits, the first 3 bits are the address and the last 8 bits are the data. The input data is latched at the rising edge of X<sub>ck</sub> when LOAD signal is high, and the data of a register become valid at the falling edge of X<sub>ck</sub>.
 +
 
 +
After all register are set, START signal must be asserted. Then, image sensing sequence starts at the rising edge of X<sub>ck</sub>. Image sensing sequence consists of two different processes: the exposing process to adjust the light intensity and the image read process to put out the image data after converting optical signal into electrical signal. After the exposure time defined by the registers 2 and 3 has passed, analog image data (total 16k pixels) is read out. To read image signal, READ signal must be asserted. At this moment, it becomes possible to change the registers, because the registers are irrelevant to the image read sequence.
 +
 
 +
Once image sensing sequence starts, the chip will continue to put out image data until it is reset.
 +
 
 +
=== Parameter Register Assignments ===
 +
{|border="1"
 +
! Symbol ||Bit Assignment ||Operation
 +
|-
 +
|N
 +
|1 bit
 +
|Exclusively set edge enhancement mode
 +
|-
 +
|VH
 +
|2 bits
 +
|Select vertical - horizontal edge operation mode
 +
|-
 +
|E
 +
|4 bits
 +
|Edge enhancement ratio
 +
|-
 +
|Z
 +
|2 bits
 +
|Zero point calibration ( Set the dark level output signal to Vref )
 +
|-
 +
|I
 +
|1 bit
 +
|Select inverted/non-inverted output
 +
|-
 +
|C0, C1
 +
|8 bits x 2
 +
|Exposure time
 +
|-
 +
|O
 +
|6 bits
 +
|Output reference voltage ( In both plus and minus direction )
 +
|-
 +
|V
 +
|3 bits
 +
|Output node bias voltage (Vref)
 +
|-
 +
|G
 +
|5 bits
 +
|Analog output gain
 +
|-
 +
|P, M, X
 +
|8 bits x 3
 +
|1-D filtering kernel.
 +
|}
 +
 
 +
=== Image Acquisition Modes ===
 +
 
 +
{|border="1"
 +
|-
 +
|(a)
 +
|Positive Image
 +
|Set “P” register
 +
|-
 +
|(b)
 +
|Negative Image 1
 +
|Set “I” register
 +
|-
 +
|(c)
 +
|Negative Image 2
 +
|Set “M” register (optional)
 +
|-
 +
|(d)
 +
|Edge Extraction (V, H, 2-D)
 +
|Set “N” and “VH” register
 +
|-
 +
|(e)
 +
|Edge Extraction (1-D)
 +
|Set “P” and “M” register (optional)
 +
|-
 +
|(f)
 +
|Edge Enhancement
 +
|Set “N”, “VH” and “E” register
 +
|-
 +
|(g)
 +
|Offset Level Output
 +
|Set “0” to both “C0” and “C1”
 +
|}
 +
 
 +
=== Register Assignment ===
 +
 
 +
{|border="1"
 +
! Reg.No. ||Address ||7 ||6 ||5 ||4 ||3 ||2 ||1 ||0
 +
 
 +
|-
 +
!|1 ||001
 +
||N ||VH1 ||VH0 ||G4 ||G3 ||G2 ||G1|| G0
 +
 
 +
|-
 +
!|2 ||010
 +
||C17 ||C16 ||C15 ||C14 ||C13 ||C12 ||C11 ||C10
 +
 
 +
|-
 +
!3 ||011
 +
||C07 ||C06 ||C05 ||C04 ||C03 ||C02 ||C01 ||C00
 +
 
 +
|-
 +
!4 ||100
 +
||P7 ||P6 ||P5 ||P4 ||P3 ||P2 ||P1 ||P0
 +
 
 +
|-
 +
!5 ||101
 +
||M7 ||M6 ||M5 ||M4 ||M3 ||M2 ||M1 ||M0
 +
 
 +
|-
 +
!6 ||110
 +
||X7||X6||X5||X4||X3||X2||X1||X0
 +
 
 +
|-
 +
!7 ||111
 +
||E3||E2||E1||E0||I||V2||V1||V0
 +
 
 +
|-
 +
!0 ||000
 +
||Z1||Z0||O5||O4||O3||O2||O1||O0
 +
|}
 +
 
 +
=== Data LOAD sequence ===
 +
 
 +
[[Image:M64282FP-data-load-sequence.png|none|frame|Data LOAD sequence]]
 +
 
 +
=== About the image processing functions ===
 +
 
 +
[[File:M64282FP-neighboring-pixels.png|frame|Four connected neighboring pixels]]
 +
 
 +
Artificial retina chip can put out positive, negative, edge extracted, and edge enhanced
 +
image in accordance with the parameter register settings.
 +
 
 +
On-chip image processing is done with the 3 x 3 neighboring pixels. This chip executes
 +
subtraction between the central pixel P and the four neighboring pixels M<sup>N</sup>, M<sup>S</sup>, M<sup>W</sup>, and M<sup>E</sup>(see the right figure) to realize edge extraction as shown below: vertical edge (V edge), horizontal edge (H edge), and 2 dimensional edge (2-D edge). Moreover, this chip can program the weight value of the central pixel P and the other four pixels to produce edge enhanced images.
 +
 
 +
[[Image:M64282FP-edgemodes.png]]
 +
 
 +
{|border="1"
 +
!Edge Modes
 +
!Output Signal
 +
!Effective Pixels
 +
|-
 +
!V-Edge Extraction
 +
|{2P - (M<sup>N</sup>+M<sup>S</sup>)} x α P-M<sup>S</sup>
 +
|128(H) x 121(V) 128(H) x 123(V)
 +
|-
 +
!H-Edge Extraction
 +
|{2P - (M<sup>W</sup>+M<sup>E</sup>)} x α
 +
|128(H) x 123(V)
 +
|-
 +
!2D-Edge Extraction
 +
|{4P - (M<sup>N</sup>+M<sup>S</sup>+M<sup>E</sup>+M<sup>W</sup>)} x α
 +
|128(H) x 121(V)
 +
|-
 +
!V-Edge Enhancement
 +
|P + {2P - (M<sup>N</sup>+M<sup>S</sup>)} x α
 +
|128(H) x 121(V)
 +
|-
 +
!H-Edge Enhancement
 +
|P + {2P - (M<sup>W</sup>+M<sup>E</sup>)} x α
 +
|128(H) x 123(V)
 +
|-
 +
!2D-Edge Enhancement
 +
|P+{4P - (M<sup>N</sup>+M<sup>S</sup>+ M<sup>E</sup>+ M<sup>W</sup>)} x α
 +
|128(H) x 121(V)
 +
|}
 +
 
 +
α is the edge enhancement ratio set by “E” register. P and M indicate the signal value from each pixel.

Revision as of 06:02, 10 October 2010

PIN CONFIGURATION (TOP VIEW)

M64282FP-pinconfig.png

Outline : 16C 9-B

Mitsubishi M64282FP, a.k.a the "Artificial Retina" is the CMOS image sensor used in the Gameboy Camera. Below follows the data sheet in wiki formatting; the data sheet is also available as a PDF in /docs/misc/ of the file hub.

Name: Mitsubishi Integrated Circuit M64282FP Image Sensor (Artificial Retina LSI) Version: ver. 1.1E Date: 5/21/98

Description

M64282FP is a 128 x 128 pixel CMOS image sensor with built-in image processing and analog image output tuning functions. This device can detect an image and process the image simultaneously as human retinas can. M64282FP can achieve smaller system size, lower power consumption, and more intelligent image processing functions.

Features

  • Single 5.0V supply
  • Low power dissipation ( Typ. 15 mW )
  • Positive and negative image output
  • Edge enhancement / extraction
  • Output level & gain tuning

Application

Image input device, Gaming, Human interface for PC, etc

Block Diagram

M64282FP-blockdiagram.png

Pin Configuration

M64282FP-pinconfig2.png

PinNo. Symbol Function Description
1 START Start Input Image sensing start.

Pulled down internally by 10k ohm.

2 NC1 Non Connect
3 SIN Data Input Parameter input. Pulled down internally by 10k ohm.
4 DVDD Digital Power Supply Power for logic circuits. Must be connected to 5.0 V digital supply.
5 DGND Digital GND Ground for logic parts.
6 LOAD Data Set Input Parameter set enable. Pulled down internally by 10k ohm.
7 Xrst System Reset System reset terminal. Pulled up internally by 10k ohm. Low active.
8 Xck System Clock Input Clock input for MUX. Pulled down internally by 10k ohm.
9 RESET Memory Reset Input Parameter register reset. Pulled up internally by 10k ohm. Low active
10 READ Read Image Read image signal.
11 TSW Reserved NOTE: Don’t connect to this pin.
12 AGND1 Analog GND Ground for analog circuits.
13 AVDD1 Analog Power Supply Power for analog circuits. Must be connected to 5.0 V analog supply.
14 Vout Signal Output Analog image signal output in voltage.
15 AVDD2 Analog Power Supply Power for analog circuits. Must be connected to 5.0 V analog supply.
16 AGND2 Analog GND Ground for analog circuits.

Image Sensing Specifications 1

Item Specification
1 Resolution 128 x 123
2 Optical System 1/4 inch

Image Sensing Specifications 2

Item Specification
1 Detectable Illumination Range (Faceplate) 1 lx ~ 10000 lx [1]
2 Optical System 1/4 inch
3 System Clock (Xck) 500 KHz
4 Frame Rate 10 fps ~ 30 fps
5 Output Voltage Range (Vout) 2.0Vp-p
Note Note: illum
Under Halogen Light Valve Illumination

Electrical Specifications - Absolute Maximum Ratings

Symbol Parameter Limits Unit
Min. Typ. Max.
DVDD Digital Power Supply Voltage 4.5 5.0 5.5 V
AVDD Analog Power Supply Voltage 4.5 5.0 5.5 V

Electrical specifications - DC Specifications

Symbol Parameter Limits Unit
Min. Typ. Max.
VOH "H" Output Voltage (READ) 4.5 DVDD V
VOL "L" Output Voltage (READ) 0.0 0.5 V
VIH "H" Input Voltage 2.2 DVDD V
VOH "L" Input Voltage 0.0 0.8 V

AC Timing Requirements

See the waveforms.

Symbol Parameter Limits Unit
Min. Typ. Max.
tcr Xck cycle time 2 µs
tWHX Xck high pulse width 0.8 µs
tWLX Xck low pulse width 0.8 µs
tr Xck rise time 0.2 µs
tf Xck fall time 0.2 µs
tSS SIN setup time 0.4 µs
tHS SIN hold time 0.4 µs
tSL SIN setup time 0.4 µs
tHL SIN hold time 0.4 tWLX - 0.4 µs
tWHL LOAD high pulse width 0.8 µs
tSXR Xrst setup time 0.4 µs
tHXR Xrst hold time 0.4 µs
tSR RESET setup time 0.4 µs
tHR RESET hold time 0.4 µs

Xck, SIN timing

Xck, SIN timing

xck LOAD timing

xck LOAD timing

Xck, Xrst, RESET timing

Xck, Xrst, RESET timing

Xck, START timing

Xck, START timing

Operation

Fig. 11-1 Operation Flow Chart


Figure 11-1 shows the image sensing sequence. First of all, all the registers must be reset and must be initialized to the appropriate values. The reset sequence completes when both Xrst and RESET signals are set low. There are 8 sets of registers, each of which consists of 8 bits of data. Each input data consists of 11 bits; of these 11 bits, the first 3 bits are the address and the last 8 bits are the data. The input data is latched at the rising edge of Xck when LOAD signal is high, and the data of a register become valid at the falling edge of Xck.

After all register are set, START signal must be asserted. Then, image sensing sequence starts at the rising edge of Xck. Image sensing sequence consists of two different processes: the exposing process to adjust the light intensity and the image read process to put out the image data after converting optical signal into electrical signal. After the exposure time defined by the registers 2 and 3 has passed, analog image data (total 16k pixels) is read out. To read image signal, READ signal must be asserted. At this moment, it becomes possible to change the registers, because the registers are irrelevant to the image read sequence.

Once image sensing sequence starts, the chip will continue to put out image data until it is reset.

Parameter Register Assignments

Symbol Bit Assignment Operation
N 1 bit Exclusively set edge enhancement mode
VH 2 bits Select vertical - horizontal edge operation mode
E 4 bits Edge enhancement ratio
Z 2 bits Zero point calibration ( Set the dark level output signal to Vref )
I 1 bit Select inverted/non-inverted output
C0, C1 8 bits x 2 Exposure time
O 6 bits Output reference voltage ( In both plus and minus direction )
V 3 bits Output node bias voltage (Vref)
G 5 bits Analog output gain
P, M, X 8 bits x 3 1-D filtering kernel.

Image Acquisition Modes

(a) Positive Image Set “P” register
(b) Negative Image 1 Set “I” register
(c) Negative Image 2 Set “M” register (optional)
(d) Edge Extraction (V, H, 2-D) Set “N” and “VH” register
(e) Edge Extraction (1-D) Set “P” and “M” register (optional)
(f) Edge Enhancement Set “N”, “VH” and “E” register
(g) Offset Level Output Set “0” to both “C0” and “C1”

Register Assignment

Reg.No. Address 7 6 5 4 3 2 1 0
1 001 N VH1 VH0 G4 G3 G2 G1 G0
2 010 C17 C16 C15 C14 C13 C12 C11 C10
3 011 C07 C06 C05 C04 C03 C02 C01 C00
4 100 P7 P6 P5 P4 P3 P2 P1 P0
5 101 M7 M6 M5 M4 M3 M2 M1 M0
6 110 X7 X6 X5 X4 X3 X2 X1 X0
7 111 E3 E2 E1 E0 I V2 V1 V0
0 000 Z1 Z0 O5 O4 O3 O2 O1 O0

Data LOAD sequence

Data LOAD sequence

About the image processing functions

Four connected neighboring pixels

Artificial retina chip can put out positive, negative, edge extracted, and edge enhanced image in accordance with the parameter register settings.

On-chip image processing is done with the 3 x 3 neighboring pixels. This chip executes subtraction between the central pixel P and the four neighboring pixels MN, MS, MW, and ME(see the right figure) to realize edge extraction as shown below: vertical edge (V edge), horizontal edge (H edge), and 2 dimensional edge (2-D edge). Moreover, this chip can program the weight value of the central pixel P and the other four pixels to produce edge enhanced images.

M64282FP-edgemodes.png

Edge Modes Output Signal Effective Pixels
V-Edge Extraction {2P - (MN+MS)} x α P-MS 128(H) x 121(V) 128(H) x 123(V)
H-Edge Extraction {2P - (MW+ME)} x α 128(H) x 123(V)
2D-Edge Extraction {4P - (MN+MS+ME+MW)} x α 128(H) x 121(V)
V-Edge Enhancement P + {2P - (MN+MS)} x α 128(H) x 121(V)
H-Edge Enhancement P + {2P - (MW+ME)} x α 128(H) x 123(V)
2D-Edge Enhancement P+{4P - (MN+MS+ ME+ MW)} x α 128(H) x 121(V)

α is the edge enhancement ratio set by “E” register. P and M indicate the signal value from each pixel.