IBIS4-1300
1.3 MPxl Rolling Shutter CMOS Image
Sensor
Features
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Overview
The IBIS4-1300 is a digital CMOS active pixel image sensor
with SXGA format.
Due to a patented pixel configuration a 60% fill factor and 50%
quantum efficiency are obtained. This is combined with an
on-chip double sampling technique to cancel fixed pattern
noise.
Cypress Semiconductor Corporation
Document Number: 38-05707 Rev. *B
•
SXGA resolution: 1280 x 1024 pixels
High sensitivity 20 µV/eHigh fill factor 60%
Quantum efficiency > 50% between 500 and 700 nm.
20 noise electrons = 50 noise photons
Dynamic range: 69 dB (2750:1) in single slope operation
Extended dynamic range mode (80…100 dB) in double
slope integration
On-chip 10 bit, 10 mega Samples/s ADC
Programmable gain & offset output amplifier
4:1 sub sampling viewfinder mode (320x256 pixels)
Electronic shutter
7 x 7 µm2 pixels
Low fixed pattern noise (1% Vsat p/p)
Low dark current: 344 pA/cm2
(1055 electrons/s, 1 minute auto saturation)
RGB or monochrome
Digital (ADC) gamma correction
Part Number
Part Number
Package Glass lid
RGB/B&W
CYII4SC1300AA-QSC
LCC
S8612
RGB Bayer
pattern
CYII4SM1300AA-QDC
LCC
D263
B&W
198 Champion Court
•
San Jose, CA 95134-1709
•
408-943-2600
Revised January 2, 2007
[+] Feedback
IBIS4-1300
TABLE OF CONTENTS
Overview ............................................................................................................................. 1
Features .............................................................................................................................. 1
Part Number ........................................................................................................................ 1
Architecture of the image sensor.......................................................................................... 4
Image sensor core - focal plane array ...................................................................................... 5
Output amplifier ........................................................................................................................ 11
Analog to digital converter ........................................................................................................ 17
ADC timing ............................................................................................................................... 18
Operation of the image sensor ............................................................................................ 20
Set configuration & pulse timing ............................................................................................... 20
Illumination control ................................................................................................................... 26
"Double slope" or "High-dynamic range" mode ........................................................................ 27
Electrical parameters ............................................................................................................... 28
Pin configuration .................................................................................................................. 29
Pin list ....................................................................................................................................... 29
Bonding pad geometry for the IBIS4-1300 ............................................................................... 31
Package .............................................................................................................................. 33
Cover glass .............................................................................................................................. 33
Ordering Information ........................................................................................................... 35
FAQ ..................................................................................................................................... 35
Temperature dependence of dark signal ................................................................................. 35
Useful range of "double slope" ................................................................................................. 35
Skipping rows or columns ........................................................................................................ 36
Disclaimer ............................................................................................................................ 36
Document History Page ...................................................................................................... 37
LIST OF FIGURES
Block diagram ................................................................................................................................... 4
Architecture of the image sensor core .............................................................................................. 4
Pixel selection - principle .................................................................................................................. 5
Spectral response * fill factor of the IBIS4-1300 pixels ..................................................................... 7
Near infrared spectral response ....................................................................................................... 8
IBIS4-1300 response curve - two pixels - lowest gain setting (0000) ............................................... 9
Output amplifier architecture ............................................................................................................. 12
Offset adjustment: fast offset adjustment mode ............................................................................... 13
Slow offset adjustment mode ............................................................................................................ 13
Output amplifier DC gain .................................................................................................................. 15
Output amplifier bandwidth for different gain settings ....................................................................... 16
Typical transfer characteristic of the output amplifier (no clipping, Voffset = 2 V, input signal during offset
adjustment is 1.2 V) ........................................................................................................................... 16
Suggested circuit for high and low references of DAC ..................................................................... 17
Document Number: 38-05707 Rev. *B
Page 2 of 37
[+] Feedback
IBIS4-1300
ADC timing ........................................................................................................................................ 18
Linear and non-linear ADC conversion characteristic ....................................................................... 19
Typical operation mode (readout of a frame) .................................................................................... 20
Timing of Y shift registers (for row selection) .................................................................................... 22
End-of-scan pulse ............................................................................................................................. 22
Timing constraints for row readout initialization (blanking time) ....................................................... 23
Pulse on 'CALIB_F'& 'UNITYGAIN' to be given once per frame, or on CALIB_S once per line ....... 24
Timing of X shift register and pixels read-out ................................................................................... 25
Pixel timing ....................................................................................................................................... 25
Pulse sequence used in IBIS4 breadboard v. jan 2000 .................................................................... 26
Schematic representation of the curtain type electronic shutter ....................................................... 26
response curve of the pixels in dual slope integration ...................................................................... 27
Linear long exposure time ................................................................................................................ 27
Linear short exposure time ............................................................................................................... 28
Double slope integration ................................................................................................................... 28
Pin layout and package, top view ..................................................................................................... 33
Transmission characteristics of the BG39 glass used as NIR cut-off filter for the FUGA1000 color image
sensors .............................................................................................................................................. 34
Transmission characteristics of the D263 glass used as protective cover for the IBIS4-1300 monochrome
image sensors .................................................................................................................................... 34
LIST OF TABLES
Optical & electrical characteristics .................................................................................................... 5
Calculation of sensitivity in [V/lx.s]..................................................................................................... 8
Pins of the image sensor core .......................................................................................................... 10
Summary of output amplifier specifications ...................................................................................... 11
Pins involved in output amplifier circuitry .......................................................................................... 14
DC gain of output amplifier for different gain settings ....................................................................... 15
ADC specifications ............................................................................................................................ 17
Pins of the ADC ................................................................................................................................ 18
Co-ordinate of the row or column selected by the Y/X shift registers after a # clock periods in viewfinder
mode and full image mode ................................................................................................................. 21
Timing constraints on row initialization pulses sequence ................................................................. 23
FillFactory and Cypress part numbers .............................................................................................. 35
Document Number: 38-05707 Rev. *B
Page 3 of 37
[+] Feedback
IBIS4-1300
Architecture of the image sensor
Block diagram
1280 x 1024 pixel array
10 bit ADC
output
amplifier
The IBIS4-1300 is an SXGA CMOS image sensor. The chip is
composed of 3 modules: an image sensor core, a
programmable gain output amplifier, and an on-chip 10 bit
ADC.
1286 x 1030 pixels
Y reset shift register
Pixel array
Reset row pointer
Readout row pointer
Y readout shift register
Figure 1. Architecture of the image sensor core
Clock_YR
Clock_YL
Column amplifiers
Sync_YL
Sync_YR
Document Number: 38-05707 Rev. *B
Sync_X
Clock_X
X shift register
Page 4 of 37
[+] Feedback
IBIS4-1300
Image sensor core - focal plane array
Figure 1. shows the architecture of the image sensor core.
The core of the sensor is the pixel array with 1280 x 1024
(SXGA) active pixels.
The name 'active pixels' refers to the amplifying element in
each pixel. This type of pixels offer a high light sensitivity
combined with low temporal noise. The actual array size is
1286 x 1030 including the 6 dummy pixels in X and Y. Although
the dummy pixels fall outside the SXGA format, their
information can be used e.g. for color filter array interpolation.
Y readout shift register
Figure 2. Pixel selection - principle
X shift register
Next to the pixel array there are two Y shift registers, and one
X shift register with the column amplifiers. The shift registers
act as pointers to a certain row or column. The Y readout shift
register accesses the row (line) of pixels that is currently
readout. The X shift register selects a particular pixel of this
row. The second Y shift register is used to point at the row of
pixels that is reset. The delay between both Y row pointers
determines the integration time -thus realizing the electronic
shutter.
A clock and a synchronization pulse control the shift registers.
On every clock pulse, the pointer shifts one row/column
further. A sync pulse is used to reset and initialize the shift
registers to their first position.
The smart column amplifiers compensate the offset variations
between individual pixels. To do so, they need a specific pulse
pattern on specific control signals before the start of the row
readout.
Table 1. summarizes the optical and electrical characteristics
of the image sensor. Some specifications are influenced by the
output amplifier gain setting (e.g. temporal noise, conversion
factor,...). Therefore, all specifications are referred to an output
amplifier gain equal to 1.
Table 1. optical & electrical characteristics
Pixel characteristics
Pixel structure
3-transistor active pixel
Photodiode
High fill factor photodiode
Pixel size
7 x 7 µm2
Resolution
1286 x 1030 pixels
SXGA plus 6 dummy rows & columns
Pixel rate with on-chip ADC
Nominal 10 MHz (note 1) (note 2)
Frame rate with on-chip ADC
About 7 full frames/s at nominal speed
Document Number: 38-05707 Rev. *B
Page 5 of 37
[+] Feedback
IBIS4-1300
Table 1. optical & electrical characteristics
Pixel characteristics
Frame rate with analog output
Up to 23 full frames per second (see table1.1)
Table 1.1: In this table you find achievable values using the analog output
X pixelsY pixels
#
#
1286 1030
1286 1030
1286 1030
1286 1030
1286
512
X Freq
Hz
1,00E+07
2,00E+07
3,00E+07
3,75E+07
3,75E+07
X Clock X Blanking
sec
sec
1,00E-07 6,25E-06
5,00E-08 6,25E-06
3,33E-08 6,25E-06
2,67E-08 6,25E-06
2,67E-08 6,25E-06
line time
sec
0,000134850
0,000070550
0,000049117
0,000040543
0,000040543
frame time frame rate
sec
per sec
0,138895500
7,20
0,072666500
13,76
0,050590167
19,77
0,041759633
23,95
0,020758187
48,17
pixel rate pixel rate freq
sec
Hz
1,049E-07
9536522
5,486E-08
18228207
3,819E-08
26182559
3,153E-08
31719148
3,153E-08
31719148
Note 1. The pixel rate can be boosted to 37.5 MHz. This requires a few measures.
• increase the analog bandwidth by halving the resistor on pin Nbias_oamp
• increase the ADC speed by the resistors related to the ADC speed (nbiasana1, nbiasana2, pbiasencload)
• experimentally fine tune the relative occurrence of the ADC clock relative to the X-pixel clock.
Note 2. The pure digital scan speed in X and Y direction is roughly 50 MHz. This is the maximum speed for skipping rows and
columns.
Light sensitivity & detection
Spectral sensitivity range
400 - 1000 nm
Spectral response * fill factor
0.165 A/W @ 700 nm
Quantum efficiency * fill factor
> 30% between 500 & 700 nm
Fill factor
60%
Charge-to-voltage conversion gain
20 µV/e-
Output signal amplitude
1.2 V
Full well charge [electrons]
IBIS4-1300: about 90000 saturation, 50000 linear range
Noise equivalent flux at focal plane (700 nm)
1.1e-4 lx*s (at focal plane)
6.3 e-7 s.W/m2
Sensitivity
7 V/lx.s
1260 V.m2/W.s
MTF @ Nyquist frequency
0.4-0.5 @ 450 nm
0.25-0.35 @ 650 nm
Optical cross talk
10% to 1st neighbor
2% to 2nd neighbor
Image quality
Temporal noise (dark, short integration time)
20 noise electrons = 50 peak noise photons (*) 400 µV RMS
Dynamic range
(analog output, before ADC conversion)
2750:1
69 dB
Dark current
344 pA/cm2 @ 21ºC
19 mV/s
1055 electrons/s
Dark current non-uniformity
Typically 15% RMS of dark current level.
Fixed pattern noise
(dark, short integration time)
9.6 mV peak-to-peak
1-2 mV RMS
Photo-response non-uniformity (PRNU)
10% peak-to-peak @ ½ of saturation signal
Document Number: 38-05707 Rev. *B
Page 6 of 37
[+] Feedback
IBIS4-1300
Table 1. optical & electrical characteristics
Pixel characteristics
Yield criteria
No missing columns nor rows
Less than 100 missing pixels, clusters=<4 pixels
Anti-blooming
Overexposure suppression > 105
Smear
Absent
(*) peak noise photons are defined as (noise electrons) / (FF*peak QE)
Features & general specifications
Electronic shutter
Rolling curtain type
Increment = line time = 135 us
Viewfinder mode
4 x sub-sampling (320 x 256 pixels)
Digital output
10 bit
Color filter array
Primary colors (Red, Green, Blue)
RGB diagonal stripe pattern or Bayer pattern
Die size
10.30 x 9.30 mm2
Package
84 pins LCC chip carrier
0.460 inch cavity
Supply voltage
5 V stabilized (e.g. from a 7805 regulator)
Power supply feed trough (dVout/dVdd)
< 0.3 for low-frequencies (< 1 MHz)
< 0.05 for high frequencies (> 1 MHz)
Power dissipation (continuous operation, 10 MHz, ADC
outputs loaded)
Min. 50 mA, Typ. 70 mA, Max. 90 mA
Light sensitivity
Figure 3. Spectral response * fill factor of the IBIS4-1300 pixels
0.2
0.18
Response [A/W]
0.16
30%
40%
0.14
20%
0.12
0.1
0.08
10%
0.06
0.04
0.02
0
400
500
600
700
800
900
1000
Wavelength [nm]
Document Number: 38-05707 Rev. *B
Page 7 of 37
[+] Feedback
IBIS4-1300
Figure 3. shows the spectral response characteristic. The
curve is measured directly on the pixels. It includes effects of
non-sensitive areas in the pixel, e.g. interconnection lines. The
sensor is light sensitive between 400 and 1000 nm. The peak
QE * FF is more than 30% between 500 and 700 nm. In view
of a fill factor of 60%, the QE is thus larger than 50% between
500 and 700 nm.
Figure 4. Near infrared spectral response
1.00E+00
800
850
900
950
Wavelength
1000 [nm]
1050
1100
Spectral Response (* FF) [A/W]
50% QE
1.00E-01
10% QE
Pixel array
plain diode
1.00E-02
1% QE
IBIS4 Near-IR spectral response
1.00E-03
calculation of sensitivity in [V/lx.s]
Pixel area A
49 E-12 m2
Fill factor FF
60%
Spectral response SR
0.22 A/W (average)
FF*SR
0.13 A/W (average over wavelength)
Pixel capacitance Ceff
5E-15 F
Sensitivity = FF*SR*Ceff/A
1.27E+3 [V.W/s.m2]
Conversion to lux: 1W/m2 =
About 180 lux, visible light only
About 70 lux, including Near Infrared
Sensitivity in lux units:
7.08 [V/lx.s] visible light only
18 [V/lx.s] if near IR included.
Document Number: 38-05707 Rev. *B
Page 8 of 37
[+] Feedback
IBIS4-1300
Color sensitivity
2.00E+05
1.80E+05
1.60E+05
1.40E+05
1.20E+05
1.00E+05
Ibis4b
8.00E+04
B&W curve: glass window
RGB curves: BG39 IR cut off
6.00E+04
4.00E+04
2.00E+04
Wavelength [nm]
0.00E+00
400
500
600
700
800
900
1000
Charge conversion - Conversion of electrons in an output signal
Figure 5. IBIS4-1300 response curve - two pixels - lowest gain setting (0000)
Output signal [V]
1,2
1
0,8
0,6
0,4
0,2
# electrons
0
0
Document Number: 38-05707 Rev. *B
20000
40000
60000
80000
100000
Page 9 of 37
[+] Feedback
IBIS4-1300
Figure 5. shows the pixel response curve in linear response
mode. This curve is the relation between the electrons
detected in the pixel and the output signal. This curve was
measured with light of 600 nm, with an integration time of
138.75 ms (10 MHz pixel rate), at minimal gain setting 0000.
The resulting voltage/electron curve is independent of these
parameters. The conversion gain is 18 uV/electron for this gain
setting.
Note that the upper part of the curve (near saturation) is
actually a logarithmic response, similar to the FUGA1000
sensor. The level of saturation can be adjusted by the voltage
on GND-AB. However note also that this logarithmic part of the
response is not FPN corrected by the on-chip offset correction
circuitry.
The signal swing (and thus the dynamic range) is extended by
increasing the VDD_RESET (pins 59/79) to 5.5 V. This is
mode of operation is not further documented.
Table 2. shows the pins of the IC that are related to the image
sensor core, describing their functionality.
Table 2. Pins of the image sensor core
Digital controls
SYNC_YR\
5
Reset right Y shift register (low active, 0 = sync)
CLK_YR
6
Clock right Y shift register (shifts on falling edge)
EOS_YR\
7
(output) low 1st CLK_YR pulse after last row (low active)
SYNC_X\
28
Reset X shift register (low active, 0 = sync)
CLK_X
29
Clock X shift register (shifts on falling edge)
EOS_X\
8
(output) Low 1st CLK_X pulse after last active column (low active)
SYNC_YL\
36
Reset left Y shift register (low active, 0 = sync)
CLK_YL
37
Clock left Y shift register (shifts on falling edge)
EOS_YL\
38
Low 1st CLK_YL pulse after last row
SHY
30
Parallel Y track & hold (1 = hold, 0 = track) apply pulse pattern - see
sensor timing diagram
SIN
35
Column amplifier calibration pulse
1 = calibrate - see sensor timing diagram
SELECT
40
Selects row indicated by left/right shift register high active (1= select
row)
Apply 5 V DC for normal operation
RESET
41
Resets row indicated by left/right shift register high active (1 = reset)
Apply pulse pattern - see timing diagram
L/R\
80
Use left or right register for SELECT and RESET
1 = left / 0 = right - see sensor timing
SUBSMPL
84
Activate viewfinder mode (1:4 sub sampling = 320 x 256 pixels) high
active, 1 = sub sampling
DCCON
31
Control voltage for the DCREF voltage generation Connect to ground
by default
DCREF
32
Reference voltage (output), to be decoupled to GND
Should be about 1.2V, can be adjusted by DCCON
NBIASARRAY
1
1 MegaOhm to VDD and decouple to ground by 100 nF capacitor
PBIAS2
2
1 MegaOhm to ground and decouple to VDD by 100 nF capacitor
PBIAS
3
1 MegaOhm to ground and decouple to VDD by 100 nF capacitor
XMUX_NBIAS
4
100K to VDD and decouple to ground by 100 nF capacitor
Reference voltages
Document Number: 38-05707 Rev. *B
Page 10 of 37
[+] Feedback
IBIS4-1300
Table 2. Pins of the image sensor core
Digital controls
GND_AB
54
Anti-blooming drain control voltage
• Default: connect to ground. The anti blooming is operational but
not maximal.
• Apply about 1 V DC for improved anti-blooming
VDD_RESETL
59
Power supply for left reset line drivers apply 5 V DC (default) or about
4…4.5 V for dual slope mode
VDD_RESETR
79
Power supply for right (default) reset line drivers 5 V DC
VDD_ARRAY
55
Power supply for the pixel array 5 V DC
VDD
11
34
53
77
Power supply of image sensor core & output amplifier 5 V DC
GND
10
33
52
78
Ground of image sensor core & output amplifier
Power & ground
Output amplifier
The offset setting is independent of the gain setting.
The output amplifier stage is user-programmable for gain and
offset level. Gain and offset are controlled by 4-bit wide words.
Gain settings are on an exponential scale. Offset is controlled
by a 4-bit wide DAC, which selects the offset voltage between
2 reference voltages (Vhigh_dac & Vlow_dac) on a linear
scale.
The gain setting is independent of amplifier bandwidth.
The amplifier is designed to match the specifications like the
output of the imager array. This signal has a data rate of 10
MHz and is located between 1.2 and 2.4 V. Table 3.
summarizes the specifications of the amplifier.
Table 3. Summary of output amplifier specifications
Min.
Typ
Max
Gain
1.2 (gain setting 0)
2.7 (setting 4)
16 (setting 15)
Output signal range
1V
Bandwidth
(40 pF load)
12 MHz
(gain setting 15)
22 MHz
(gain setting 0...8)
33 MHz
(gain setting 0)
Output slew rate
(40 pF load)
40 V/ µs
50 V/µs
80 V/µs
The range of the output stage input is between 1 and 4 V. A
lowest gain the sensor outputs a signal in between 1.2 and 2.2
V, which fits into the input range of the amplifier. The range of
the output signal is between 1 and 4.5 V, dependent on the
gain and offset settings of the amplifier. This range should fit
Document Number: 38-05707 Rev. *B
4.5 V
to the input range of the ADC, external or internal. The
on-chip ADC range is between 2 and 4 V. A minimal gain
setting of "3" seems necessary for the internal ADC, and the
offset voltage should be set to the low-reference voltage of the
ADC.
Page 11 of 37
[+] Feedback
IBIS4-1300
Figure 6. output amplifier architecture
pixel array
A
extin
Clip
+
1
sel_extin
gain [0..3]
unity gain
Calib_f
1.1.1.1.1
Calib_s
Vhigh_dac
offset [0..3]
D
A
C
Vlow_dac
Figure 6. shows the architecture of the output amplifier. First
of all, there is a multiplexer which selects either the imager
core signal or an external pin EXTIN as the input of the
amplifier. EXTIN can be used for evaluation, or to feed
alternative data to the output.
SEL_EXTIN controls this switch.
Then, the signal is fed to the first amplifier stage. This stage
has an adjustable gain, controlled by a 4-bit word
('gc_bit0...3').
Then, the upper level of the signal must be clipped in some
situations (clipping sometimes is necessary when the imager
signal is highly saturated, which affects the calibration level.
This is visible as black banding at the right side of bright
objects in the scene). In order to do this, a voltage should be
applied to the 'Clip' pin. The signal is clipped if it is higher than
Vclip - Vth,pmos, where Vth,pmos is the PMOS threshold
voltage and is typically -1 V. If clipping is not necessary, 5 V
should be applied to 'Clip'.
After this, the offset level is added. This offset level is set by a
DAC, controlled by a 4-bit word (DAC_bit0...3). The offset level
can be calibrated in two modes: fast offset adjustment or slow
offset adjustment. This is controlled by 'calib_s' and 'calib_f'.
The slow adjustment yields a somewhat cleaner image.
After this, the signal is buffered by a unity feedback amplifier
and it leaves the chip. This 2nd amplifier stage determines the
maximal readout speed, i.e. the bandwidth and the slew rate
Document Number: 38-05707 Rev. *B
of the output signal. The whole amplifier chain is designed for
a data rate of 10 Mpix/s (@ 40 pF). (It is up to the experimenter
to increase this speed by reducing the various setting
resistors)
Table 4. shows the IBIS4-1300 pins used by the output
amplifier with a short functional description. Power and ground
lines are shared between the output amplifier and the image
sensor.
Output amplifier offset level adjustment
The purpose of this adjustment is to bring the pixel voltage
range as good as possible within the ADC range. The offset
level of the output signal is controlled by a 4-bit resistive DAC.
This DAC selects the offset level on a linear scale between 2
reference voltages. These reference voltages are applied to
Vlow_dac and Vhigh_dac.
This offset level is adjusted during the calibration phase.
During this phase, the amplifier input should be constant and
refers to the 'zero' signal situation. The IBIS4-1300 outputs a
dark reference signal after a row has been read out
completely. This signal can be used as the 'zero signal'
reference. Alternatively one can apply an external reference
on pin EXTIN, which is applied to the output amplifier when
SEL_EXTIN is 1.
Offset adjustment can be done during row or frame blanking
time.
Page 12 of 37
[+] Feedback
IBIS4-1300
Figure 7. offset adjustment: fast offset adjustment mode
min. 500 ns
calib_f
stable input
dark
signal
unitygain
>0
> 100 ns
There are 2 modes of offset calibration for the output amplifier:
slow and fast adjustment. Figure 7. shows the timing and
signal waveforms for fast offset adjustment mode. Closing
both 'calib_f' and 'unitygain' operates it. After 'calib_f' is
opened again, the offset level is adjusted to the desired value
in a single cycle. The signal applied to the output amplifier
should be stable just before and during the adjustment phase.
The same is true for the DAC output.
The signal applied to the output amplifier can be either:
• The signal generated by the electrical dark reference in the
imager core itself, i.e. the pixels named "dark" inFigure 20.
• Apply the reference from outside on the pin EXTIN,
controlled by SEL_EXTIN.
If this fast offset adjustment is used, it should be done once
each frame, before the readout of the frame starts, e.g. during
the blanking time of the first line.
Figure 8. slow offset adjustment mode
min. 100 ns
calib_s
stable input
dark
signal
min 100 ns
Figure 8. shows the timing and signal waveforms for slow
offset adjustment mode. It is operated by pulsing 'calib_s'. The
amplifier input signal must be stable and refer to 'dark' signal
at the moment when calib_s goes low. The offset is slowly
Document Number: 38-05707 Rev. *B
adjusted with a time constant of about 100 of these pulses.
One pulse is then generated during each row blanking time.
The baseline is to use the fast calibration once per image. The
slow calibration is intended as alternative if, for very slow
readout, the offset drifts during the image.
Page 13 of 37
[+] Feedback
IBIS4-1300
Table 4. Pins involved in output amplifier circuitry
Name
No.
Function
Extin
12
External input of the output amplifier
Active if Sel_extin = 1
Output
13
Analog output signal
To be connected to the input of the ADC (in_adc, pin 73)
Sel_extin
9
1 = external input pin (extin) is applied at the input of the amplifier
0 = output amplifier is connected to the image sensor array
gc_bit0
17
LSB
gc_bit1
18
gc_bit2
19
Control bits for output amplifier gain setting
Gain adjustment between 1.2 (0000) & 16X (1111)
MSB
gc_bit3
20
unitygain
21
1 = output amplifier in unity feedback mode
0 = output amplifier gain controlled by gc_bit0...3
calib_s
16
Slow (or incremental) output offset level adjustment (calibration of output
amplifier). Offset adjustment converges after about 100 pulses on calib_s
Amplifier input should refer to a 'zero signal' at the moment of the 1->0
transition on calib_s
0 = connect to capacitor (of stage 2) and in- (of stage 1)
1 = connect to DAC output (of stage 2) and out (of stage1)
calib_f
22
Fast (=in 1 cycle) output offset level adjustment (calibration of output
amplifier)
Offset level is adjusted when both calib_f and unitygain are high
Amplifier input should refer to 'zero signal' when calib_f is high
1 = connect DAC output to offset of capacitor
0 = DAC output disconnected
dac_b0
26
LSB
dac_b1
25
dac_b2
24
Control bits for output offset level adjustment
Between Vlow_dac (0000) & Vhigh_dac (1111)
MSB
dac_b3
23
Analog signals
Digital Controls
Reference voltages
Vlow_dac
14
Vhigh_dac
15
Nbias_oamp
27
Output amplifier speed/power.
Connect with 100 K to VDD and decouple with 100 nF to GND. This setting
yields 10 MHz nominal pixel rate. Lowering the resistance does increasing
this rate.
Clip
83
Voltage that can be used to clip the output signal
Clips output if output signal > 'Vclip - Vth, PMOS' with Vth,PMOS=-1V
Default: 5 V (no clipping)
Document Number: 38-05707 Rev. *B
Low and high references for offset control DAC of the analog output.
The range of this resistive division DAC should be about 1V to 2.5V. If the
range is not OK, one will notice that it is not possible to adjust the output
voltage to the appropriate level of the ADC. As the internal division resistor is
about 1.3 Kohm, we suggest to tie Vlow_dac with 1K to GND and Vhigh_dac
with 2K7 to VDD.
Page 14 of 37
[+] Feedback
IBIS4-1300
Output amplifier gain control
Figure 9. output amplifier DC gain
20,00
18,00
y=1.074*2
16,00
0.246*x
DC gain (< 1MHz)
14,00
12,00
10,00
8,00
6,00
4,00
2,00
0,00
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
gain setting
Table 5. DC gain of output amplifier for different gain settings
gain setting
DC gain (<1MHz)
gain setting
DC gain (<1MHz)
0000
1.28
1000
5.33
0001
1.51
1001
6.37
0010
1.82
1010
7.41
0011
2.13
1011
8.91
0100
2.60
1100
10.70
0101
3.11
1101
12.65
0110
3.71
1110
15.01
0111
4.40
1111
17.53
The output amplifier gain is controlled by a 4-bit word. In
principle, the output amplifier can be configured in unity
feedback mode by a permanent high signal on UNITYGAIN,
but the purpose of this mode is purely diagnostic. The "normal"
gain settings vary on an exponential scale. Figure 9. and Table
5. report all gain settings.
Document Number: 38-05707 Rev. *B
In first approximation, the gain setting is independent of
bandwidth, as the amplifier is a 2-stage design. The first stage
sets the gain, and the second stage is a unity gain buffer, that
determines bandwidth and slew rate. There is however some
influence of gain setting on bandwidth. Figure 10. shows the
output amplifier bandwidth for all gain settings.
Page 15 of 37
[+] Feedback
IBIS4-1300
Figure 10. output amplifier bandwidth for different gain settings
35
Bandwidth [MHz]
30
25
20
15
10
5
0
Unity
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Gain setting
Figure 11. typical transfer characteristic of the output amplifier (no clipping, Voffset = 2 V, input signal during offset
adjustment is 1.2 V)
6
5
Output [V]
4
3
Unity
2
3
7
1
11
0
0
1
2
3
4
5
Input [V]
Document Number: 38-05707 Rev. *B
Page 16 of 37
[+] Feedback
IBIS4-1300
Figure 11. shows the output characteristic curve in a typical
case for the imager. The offset voltage is adjusted to 2 V, which
corresponds to the low-level voltage of the ADC. Clipping is
off, and the input signal is changed between 0 and 5 V. During
offset adjustment (when calib_s is switched from 1 -> 0 or
when calib_f is on), the input signal is at 1.2 V. This level corresponds to the imager dark reference output. The input signal
is transferred to the output by adding a 2V offset and multiplication with the appropriate gain. The input signal of dark pixels
(at 1.2 V) corresponds with 2 V at the output. Higher input
signals are amplified. The curves for 3 typical gain settings are
shown (unity gain, setting 3, 7 & 11).
Again, as can be seen on the above figure, the applied input
signal during the output amplifier calibration (by 'CALIB_S' or
'CALIB_F') is the reference level to which the signal is
amplified. During this calibration, a stable input is required.
Setting of the VLOW_DAC & VHIGH_DAC reference voltages
Figure 12. suggested circuit for high and low references of DAC
on-chip
resistor
1K3
VHIGH_DAC
About 2.3 V
VLOW_DAC
About 1 V
VLOW_DAC & VHIGH_DAC are the reference voltages for the
DAC. They represent the 0000 resp. 1111 code. The internal
series resistance is about 1.3 kOhms. They can be connected
as in Figure 12., and decoupled to ground.
Analog to digital converter
The IBIS4-1300 has a 10 bit flash analog digital converter
running nominally at 10 Msamples/s. The ADC is electrically
separated from the image sensor. The input of the ADC
("IN_ADC") should be tied externally to the OUTPUT of the
output amplifier.
Table 6. ADC specifications
Input range
2-4V
Quantization
10 Bits
Nominal data rate
10 Msamples/s (*)
DNL (linear conversion mode)
INL (linear conversion mode)
Input capacitance
< 20 pF
Power dissipation @ 10 MHz
107 mA
535 mW
Delay of digital circuitry (Td, 40 pF load)
< 50 ns after falling edge of clock
Input setup time (Ts) for a stable LSB
< 100 ns before falling edge of clock
Conversion law
Linear / Gamma-corrected
(*) Project partners have demonstrated 20 MHz data rate by
careful timing and by decreasing some or all of the resistors
Document Number: 38-05707 Rev. *B
on NBIAS* and PBIAS*.
Page 17 of 37
[+] Feedback
IBIS4-1300
ADC timing
The ADC converts on the falling edge of the CLK_ADC clock.
The input signal should be stable during a time Ts before the
falling clock edge. The digital output is available Td after the
falling clock edge (Figure 13., Ts = 100 ns, Td = 50 ns). These
values are the delays to obtain a stable LSB after a half-scale
swing of the input signal. For the MSB to become stable, Ts=20
ns is sufficient. For a full scale input swing (which normally
doesn't appear with image sensors), Ts is 140 ns for the LSB
and 20 ns for the MSB.
Figure 13. ADC timing
CLK_ADC
100 ns
IN_ADC
Ts
D0…D9
Td
TRI_ADC can be used to put the output bits in a tristate mode
(e.g. for bi-directional busses). If this is used, the output signal
becomes valid 50 ns after the falling edge on TRI_ADC.
When NONLINEAR is high, the ADC conversion is non-linear.
The contrast will be higher in dark image regions, and lower in
bright areas, similar to gamma correction.
BITINVERT can be used to invert the output word, if necessary
(one's complement).
Table 7. pins of the ADC
Name
No.
Description
Analog signals
IN_ADC
73
Input, connect to sensor's output (pin 13)
Input range is between 2 & 4 V (VLOW_ADC & VHIGH_ADC)
CLK_ADC
62
ADC Clock
ADC converts on falling edge
TRI_ADC
63
Tristate control of ADC digital outputs
1 = tristate; 0 = output
NONLINEAR
67
1 = non-linear analog-digital conversion
0 = linear analog-digital conversion
BITINVERT
39
1 = invert output bits
0 = no inversion of output bits
51…42
Output bits
D0 = LSB, D9 = MSB
Digital Controls
Digital output
DO… D9
Reference voltages
Document Number: 38-05707 Rev. *B
Page 18 of 37
[+] Feedback
IBIS4-1300
Table 7. pins of the ADC
Name
No.
Description
VLOW_ADC
71
VHIGH_ADC
61
Low reference and high reference voltages of ADC should be 2V to 4V.
The resistance between VLOW_ADC and VHIGH_ADC is about 1.5 K, thus this
range can be approximated by tying LOW with 2K to GND
And HIGH with 1K to VDD.
PBIASDIG1
64
Connect with 100K to GND and decouple to VDD
PBIASENCLOAD
65
Connect with 100K to GND and decouple to VDD
PBIASDIG2
66
Connect with 100K to GND and decouple to VDD
NBIASANA2
69
Connect with 100K to VDD and decouple to GND
NBIASANA
70
Connect with 100K to VDD and decouple to GND
These resistors determine the analog resp. digital speed /power of the ADC. Both
can be increased/decreased by lowering or increasing the resistance values.
Power & Ground
VDD_DIG
56, 76
Power supply of digital circuits of ADC, + 5 V
VDD_AN
58, 74
Power supply of analog circuits of ADC, + 5 V
GND_DIG
57, 75
Ground of digital ADC circuits
GND_AN
60, 72
Ground of analog ADC circuits
them. The appropriate 2 V and 4 V DC voltages can be
obtained as in Table 7. : pins of the ADC, and decoupled to
ground.
Control of the VLOW_ADC & VHIGH_ADC reference voltages
VLOW_ADC & VHIGH_ADC are the reference voltages for a
0 and 1023 code. A 2K-resistor ladder internally connects
Non-linear and linear conversion mode - "gamma" correction
Figure 14. linear and non-linear ADC conversion characteristic
1.400
linear
non-linear
1.200
Output
1.000
0.800
0.600
0.400
0.200
0.000
2.0
2.5
3.0
3.5
4.0
Input signal [V]
Figure 14. shows the ADC transfer characteristic. For this
measurement, the ADC input was connected to a 16-bit DAC.
The input voltage was a 100 KHz triangle waveform.
Document Number: 38-05707 Rev. *B
The non-linear ADC conversion is intended for
gamma-correction of the images. It increases contrast in dark
areas and reduces contrast in bright areas. The non-linear
Page 19 of 37
[+] Feedback
IBIS4-1300
curve is tolerant for external pixel offset error correction. This
means that pixel offset variations can be corrected by
changing the offset after the non-linear AD conversion. This is
so because the non-linear transfer function is
H(s) = 1-exp(-a*s)
by design, and neglecting the offset, the relation between the
non-linear output (y) and the linear output (x) is exactly:
Y = 1024 * (1 - exp(-x/713)) / (1 - exp(-1024/713))
This law yields an increased accuracy of about a factor 2 near
the zero end of the scale. It is thus possible to obtain an
effective 11 bit accuracy on a linear scale after post processing
by applying the reverse law to the non-linear output:
Z = -2 * 713 * ln(1 - y/(1024/(1-exp(-1024/713)))) = -1426 * ln(1-y/1343.5)
Then Z is an 11-bit linear output in the range 0...2047.
Operation of the image sensor
Set configuration & pulse timing
Figure 15. typical operation mode (readout of a frame)
Set configuration
-
output offset
amplifier gain
viewfinder on/off
determine integration
time of next frame
Set flag to restart YR shift
register with SYNC_YR
For Y = 0 till Ymax do
Y=
1026 integration
time ?
Yes
Set flag to restart YL shift
register with SYNC_YL
no
Execute row initialization
sequence
Start X shift register
SYNC_X
X
For X = 0 to Xmax do
Acquire pixel (X,Y)
Next Y
Document Number: 38-05707 Rev. *B
Page 20 of 37
[+] Feedback
IBIS4-1300
Figure 15. shows a typical operation mode of the image
sensor.
At the start of a new frame, the device may be (re-)configured.
If necessary, the output amplifier gain and offset are adjusted
or the device is put in viewfinder mode.
Then, the frame readout shift register is initiated by pulsing
"SYNC_YR". This pulse occurs once per frame, normally as a
part of the first row blanking sequence.
The readout of a row (line) starts with row blanking initialization
sequence. Here several pulses are applied for Y-direction
shift, the column amplifier S&H and nulling, and the start
(SYNC_X) of the X-direction shift register.
The frame reset shift register is started also once per frame by
"SYNC_YL", this pulse occurs once per frame, normally as a
part of the row blanking sequence of one particular row. The
time delay from the SYNC_YL to SYNC_YR is the integration
time. The integration is thus a multiple of the row readout time.
The reset shift register always leads the readout shift register.
Therefore, the integration time should be determined before
the start of the frame readout. The value that is fixed at that
moment will be the integration time of the NEXT frame. If the
value set for the integration time changes during frame
readout, the start pulse might be lost and the next frame might
be invalid. We will now discuss all steps in more detail.
Set configuration
Configuration of the image sensor implies control &
adjustment of the following points:
• output amplifier offset level, set by 'dac_bit[0...3]'
• output amplifier gain setting, set by 'gc_bit[0...3]'
• choose the integration time of the next frame
• set/clear viewfinder mode (pin 'subsampl')
• in case when the fast adjustment of the offset level is used,
plus 'calib_f' and 'unitygain' as described before in figures
Figure 7. and Figure 8.
Viewfinder mode vs. normal readout
Table 8. co-ordinate of the row or column selected by the Y/X shift registers after a # clock periods in viewfinder mode
and full image mode
Clock
Sync
1
2
3
4
5
6
Viewfinder mode
None
None
Row 1
Row 5
Row 9
Row 13
Row 17
Y reg.
Dark
Col. 1
Col. 5
Col. 9
Col. 13
X reg.
Row 1
Row 2
Row 3
Row 4
Row 5
Y reg.
Dark
Col. 1
Col. 2
Col. 3
Col. 4
X reg.
Full image mode
None
None
Clock
258
259
260
322
323
324
Viewfinder mode
Row 1025
Row 1029
EOS
Col. 1281
Col. 1285
EOS Dark
Clock
1030
1031
1032
1287
1288
1289
Full image mode
Row 1029
Row 1030
EOS
Col. 1285
Col. 1286
EOS Dark
Y shift register
In full image readout mode (pin 84, subsmpl = 0), the imager
is a 1280 x 1024 SXGA image sensor. There are 3 dummy
pixels read at all 4 borders of the image.
In viewfinder mode (subsmpl = 1), the imager acts as a 320 x
256 QVGA image sensor with one dummy pixel at the start of
a row/column.
Table 8. shows which column or row is selected after a
number of clock pulses.
Document Number: 38-05707 Rev. *B
X shift register
Start of the Y shift registers for row readout & row reset
The shift registers are put in their initial state by a
synchronization- or start pulse. (sync_x, sync_yr, sync_yl).
The synchronization signal is low-active and should only be
generated when the clock of the shift register is high. After the
synchronization pulse, two falling clock edges are needed to
skip dummy pixels/lines. On every falling clock edge, the shift
register selects a new row for readout or reset. Figure 16.
shows this timing.
Page 21 of 37
[+] Feedback
IBIS4-1300
Figure 16. Timing of Y shift registers (for row selection)
Dummy rows
Line
selected
n-
n
nil
0
1
2
0
1
3
4
5
CLCK_Y
SYNCY
Min 25 ns
Min 100 ns.
Figure 17. End-of-scan pulse
Clock
EOS
Col 1286
Row 1030
Col 321
Row 257
End-of-scan: EOS_YL, EOS_YR, EOS_X
All 3 shift registers are equipped with 'end-of-scan' pulses.
These pulses are low during the clock period after the last pixel
or row has been read out, also in viewfinder mode.
At the EOS_X pulse, the electrical dark reference level is put
on the readout bus. This voltage remains on the bus until the
SIN pulse goes high. During the row blanking time, this voltage
can be used for the offset adjustment of the output amplifier.
The SIN high forces the DCREF voltage on the output bus.
We advise not to use the EOS pulses as an input for the row
blanking time sequence generation, but to use simple
Document Number: 38-05707 Rev. *B
counters instead. If by some reasons the EOS signal is absent
or subject to glitches, the system would hang. EOS is intended
as diagnostic means.
Row initialization
During the row blanking time (which occurs at the beginning of
every row read), several tasks are executed: selection of a
new row, readout of this row by double sampling, reset of a
new row, and possibly (slow) offset adjustment of the output
amplifier. Therefore, a pulse patterns must be applied to
several signals during this time. There is some freedom to
make this pattern. The constraints are listed below:
Page 22 of 37
[+] Feedback
IBIS4-1300
Figure 18. Timing constraints for row readout initialization (blanking time)
SYNC_YL
SYNC_YR
Tc
CLK_YR
CLK YL
SHY
Ts
SIN
Tw
Tr
Th
Tr
RESET
To
To
To
L/R
Tn
Tm
SYNC_X
CLOCK_X
Table 9. Timing constraints on row initialization pulses sequence
Ta
Min 0
Delay between falling edge of CLK_Y* and SHY or SIN
Tc
Min 25 ns
CLK_YR & CLK_YL high time
Ts
Typ. 3 us
On-time of SIN (offset calibration pulse)
Delay between selection of new row and end of column amplifier calibration
Tw
Typ. 200 ns
Delay between end SIN and pixel reset
Tr
Typ. 1 us
On-time of reset pulse
Th
Typ 1 us
Th + Tr = Delay between pixel reset and column sample & hold
To
Typ 100 ns
Delay between SHY and L/R\
Overlap of L/R\ over 2nd reset pulse
Tm
Min 25 ns
On-time of one of the SYNC pulses. SYNC==low may only occur when the
associated CLOCK is high.
Tn
Min. 200 ns
Delay between SHY and start row readout
Figure 18. and Table 9. illustrate the timing constraints of the
row initialization/ blanking sequence.
Document Number: 38-05707 Rev. *B
• The EOS_X pulse flags the end of the scanning of previous
line, and should be considered as a diagnostic means only.
The blanking sequence could start earlier or later.
Page 23 of 37
[+] Feedback
IBIS4-1300
• The next row (=line) is selected after the falling edge of
CLK_YR and CLK_YL,
• The column amplifiers receive the signals on the pixels array
columns buses when SHY is low (transparent).
• The SIN pulse (high) forces the column amplifiers in an
“offset nulling" state.
• After 3 us, the column amplifiers have reached offset-free
equilibrium, and the SIN pulse is brought low again. The
pixel's signal level is thus stored in the column amplifier.
• After that the pixels in the selected row (line) are be reset
(first pulse on RESET).
• Consequently the reset level is frozen in the column
amplifiers when SHY goes high. Both signal level and reset
level have now been applied to the column amplifiers. The
sample hold (SHY) guarantees that this information will not
change anymore during readout of the line.
• Now, the row is ready for readout. A pulse on SYNC_X must
be given to start the row readout. SYNC_X initiates the
X-direction scanning register. The scanning itself is
controlled by CLOCK_X.
• During the beginning of the row readout, or possibly before,
the RESET pulse for the electronic shutter (ES) must be
given, if the ES is used. This is a pulse on RESET together
with a high level on L/R. If the ES is not used, L/R remains
low and the second RESET pulse is not generated.
During some or the entire row blanking times, the output
amplifier can be calibrated.
If the slow calibration method is used, pulse the 'CALIB_S' pin
once per line. The calibration happens on the rising edge of
the pulse.
If the fast calibration is used, the 'CALIB_F' should be pulsed
during the row blanking time of the first row only. This
calibration happens during the time that the pulse is high.
During this calibration, the input applied to the amplifier must
be the dark reference, which can either be the built-in electrical
dark reference, or an external dark reference on the pin
EXTIN.
Figure 19. Pulse on 'CALIB_F'& 'UNITYGAIN' to be given once per frame, or on CALIB_S once per line
Signal applied to
input of amplifier
"Dark"
reference
Calib_s
or
Calib_f
100ns
500ns (calib_f)
100ns (calib_s)
The X-direction shift register
The X shift register behaves like the Y shift registers.
The sequence if initiated by SYNC_X, which should occur
when CLOCK_X is high. As CLOCK_X is halted during the
Document Number: 38-05707 Rev. *B
blanking time, the SYNC_X pulse could occur anywhere, and
be taken equal to some other pulse (e.g. CLOCK_Y).
The first real (dummy) pixel is read out after the 3rd falling
edge on the clock. Dummy pixels are perfectly operational
pixels, but are added to shield the "real" pixels from the cross
talk of the periphery.
Page 24 of 37
[+] Feedback
IBIS4-1300
Figure 20. Timing of X shift register and pixels read-out
Dummy pixels
DCREF on bus
Output
n-1
n
nil
dark
0
1
0
2
3
1
4
Clock
Sync
Min 25 ns
Min 100 ns.
On-chip generated electrical dark references.
Pixel readout
The sensor outputs a electrical dark reference level after the
2nd falling edge on the clock (after sync).
The same continuous 10 MHz clock drives CLK_ADC and
CLK_X. On the falling edge of CLK_X, a new pixel is selected
and propagates to the output amplifier. At the same time, the
ADC input is frozen by the falling edge on CLK_ADC. The
digital output has a delay of one pixel compared to the analog
signal. The digital output becomes valid between 25 to 50 ns
after the falling edge on CLK_ADC.
At the end of the row readout, after EOS_X becomes low, the
sensor outputs the electrical dark reference voltage also, and
it remains present on the on the readout bus until the SIN goes
high.
Note that if the X-register is reset before the EOS is reached,
the dark reference is not put on the bus. Use the dark
reference of the beginning of the line instead.
Figure 21. pixel timing
Tp,adc
If the end of a row is reached, the sensor outputs an
end-of-scan (EOS) pulse during one pulse period. And the
electrical black reference level appears at the output for all
Document Number: 38-05707 Rev. *B
successive pulses. So, the same 10 MHz clock can drive
CLK_X and CLK_ADC.
Page 25 of 37
[+] Feedback
IBIS4-1300
Example: timing used on the IBIS4 breadboard
The next figure is the timing as used in the IBIS4 breadboard
version 12 jan 2000. In this baseline only CALIB_F is used
(pulsing once per frame). CALIB_S (pulse every line) is shown
as reference, but is actually not used in the baseline. The
UNITY_GAIN pulse is identical to CALIB_F.
Figure 22. pulse sequence used in IBIS4 breadboard v. jan 2000
1 usec
CLCK_Y
SIN
CALIB_S
Is calib_s is used, calib_f&unity are 0
Or CALIB_F&UNITY GAIN
Is calib_f&unity are used, calib_s is 0
RESET
SYNC_X
CLCK_X
SYNCY_L and SYNCY_R; once per frame per register (for electronic shutter L and R at different moments)
L/R
SHY
Illumination control
There are two means of controlling the illumination level
electrically. For high light levels, there is an electronic shutter.
For low light levels, the output signal can be amplified by
controlling the output amplifier gain. The offset level of the
signal can also be controlled digitally.
“Rolling curtain" electronic shutter
The electronic shutter can reduce the integration time (=
exposure time). This is achieved by an additional reset pulse
every frame. In this way, the integration time is reduced to a
fraction of the frame readout time.
There are two Y shift registers. One of them points at the row
that is currently being read out. The other shift register points
at the row that is currently being reset. Both pointers are
shifted by the same Y-clock and move over the focal plane.
The integration time is set by the delay between both pointers.
Reset
pointer
Integration
time
Figure 23. Schematic representation of the curtain type
electronic shutter
Readout
pointer
This is a so-called 'rolling curtain'-type shutter. It 'rolls' over the
focal plane.
The left and right shift registers can be used both for pointing
to the row that is readout or the row that is reset. The shift
register that is active for readout or reset is selected by the
signal on L/R. In the above timing diagrams, we use the R shift
register for readout, and the L shift register for electronic
shutter reset. We call them the readout shift register and reset
shift register.
The integration time is controlled by the delay between the
SYNCY_L and SYNCY_R pulse. The shorter this delay, the
shorter the integration time and the smaller the output signal
will be.
If the electronic shutter is not used, the L/R signal is not pulsed.
The integration time is then equal to the frame readout time.
For proper operation of the ES, the CLOCK_Y must come as
an uninterrupted pulse train. Also during the dead time
between frames the CLOCK_Y must be clocked. The reason
is that each line should see the same elapsed time between
the "ES-reset" and the reset of the line being read-out. If the
CLOCK_Y is halted, the lines between the two pointers will
have a longer effective integration time, and appear brighter.
Gain control
For low illumination levels, the electronic shutter is not used or set to its maximal value. Longer integration times can only
be obtained by decreasing the frame rate. As an alternative or
in complement, one can increase the output amplifier gain.
The gain is controlled by a 4-bit word. Gain values vary
between 1.2 and 16, and on an exponential scale, as the
F-stops of a lens.
Of course, increasing the signal amplitude by increasing the
gain, will also increase the noise level. The apparent increase
of sensitivity is at the cost of a lower dynamic range.
Document Number: 38-05707 Rev. *B
Page 26 of 37
[+] Feedback
IBIS4-1300
Offset level adjustment
"Double slope" or "High-dynamic range" mode
The offset level of the output signal is set by a 4-bit digital word.
The offset level voltage is selected between VLOW_DAC and
VHIGH_DAC on 16 taps.
IBIS4-1300 has a feature to increase the dynamic range. The
pixel response can be extended over a larger range of light
intensities by using a "dual slope integration" (patents
pending). This is obtained by the addition of charge packets
from a long and a short integration time in the pixel during the
same frame time.
1.8
1.6
1.4
1.2
Output signal [V]
Figure 24. response curve of the pixels in dual slope integration
1
0.8
0.6
Dual slope operation
0.4
Long integration time
Short integration time
0.2
Relative exposure (arbitrary scale)
0
0%
20%
40%
Fig.24 shows the response curve of a pixel in dual slope
integration mode. The curve also shows the response of the
same pixel in linear integration mode, with a long and short
integration time, at the same light levels.
60%
80%
100%
These
example
images
are
found
at
http://www.fillfactory.com/htm/technology/htm/dual-slope.htm
Figure 25. Linear long exposure time
Dual slope integration is obtained by
• Feeding a lower supply voltage to VDD_RESETL. E.g.
apply 4 to 4.5 Volts. The difference between this voltage
and VDD determines the range of the high sensitivity, thus
the output signal level at which the transition between high
and low sensitivity occurs.
• Put the amplifier gain to the lowest value where the analog
output swing covers the ADC's digital input swing.
Increasing the amplification too much will likely boost the
high sensitivity part over the whole ADC range.
• The electronic shutter determines the ratio of integration
times of the two slopes. The high sensitivity ramp
corresponds to "no electronic shutter", thus maximal
integration time. The low sensitivity ramp corresponds to
the electronics shutter value that would have been obtained
in normal operation.
Document Number: 38-05707 Rev. *B
Page 27 of 37
[+] Feedback
IBIS4-1300
Figure 26. Linear short exposure time
Electrical parameters
DC voltages
VDD and GND:
Nominal VDD-GND is 5V DC.
Overall current consumption for the different parts
• imager core + output amplifier analog
• imager core digital
• ADC analog
• ADC digital
Are quoted in the datasheets
The sensor works properly when using a 7805 type of
regulator.
Decoupling VDD to GND must happen close to the IC.
Other applied DC voltages
Should be clean as the VDD. Can be derived by resistive
division of VDD-GND, and decoupled to VDD or GND (as
indicated)
Figure 27. Double slope integration
External Resistors
Are used as current mirror settings. Should be decoupled to
the opposite rail voltage as the connection of the resistor (thus:
if the resistor is tied to VDD, the capacitor is tied to GND). In
practice the decoupling can be omitted for almost all signals to be experimented.
Input / output
Digital inputs:
Clean rail to rail CMOS levels. 10%-90% rise and fall times
between 10 ns and 40 ns
Digital outputs:
Deliver CMOS level, able to drive 40 pF capacitive loads
Analog output of imager core
Designed to drive a 40 pF capacitive load
Analog input of ADC
Is equivalent to a capacitive load of typ. 15 pF.
Document Number: 38-05707 Rev. *B
Page 28 of 37
[+] Feedback
IBIS4-1300
Pin configuration
I/O Symbols
Pin list
Signal type symbols
I
Input
O
Output
A
Analog
P
Power supply
D
Digital
G
Ground
W
Word bit
No.
Name
Type
I/O
Description
Signal
1
Nbiasarray
A
I
1MEG to VDD and decouple to GND
Pixel source follower bias current
2
pbias2
A
I
1MEG to GND and decouple to VDD
Column amp 1st source follower (after SHY)
bias current
3
Pbias
A
I
1MEG to GND and decouple to VDD
Column amp current source bias current
4
xmux_nbias
A
I
100K to VDD and decouple to GND
X-multiplexing bias current (/6)
5
Sync_yr\
D
I
low active (0=sync)
0 = reset right shift register
6
clk_yr
D
I
Shifts on falling edge
clock right shift register
7
Eos_yr\
D
O
Active low
low 1st clk_yr pulse after last row
8
Eos_x\
D
O
Active low
low 1st clk_x pulse after last active column
9
Selextin
D
I
input selector for output amplifier
1 = external input; [0] = imager core
10
Gnd
A
G
Analog GND
11
Vdd
A
P
Analog VDD
12
Extin
A
I
external input to output amplifier
13
Output
A
O
analog output of imager core
Connect to in_adc (p73)
14
Vlow_dac
A
I
low reference voltage offset DAC
+/- 1 V
15
Vhigh_dac
A
I
high reference voltage offset DAC
+/- 2.5 V
16
Calib_s
D
I
Slow dark offset level adjustment
0: connect to cap (st2) and in- (st1)
1: connect to rdac (st2) and output (st1)
17
gc_bit0
W
I
Lsb
gain control output amplifier
18
gc_bit1
W
I
19
gc_bit2
W
I
20
gc_bit3
W
I
Msb
21
Unitygain
D
I
sets output amplifier in unity gain
High active
22
Calib_f
D
I
fast dark offset level calibration
High active
23
Dac_b3
W
I
Msb
dac control for black offset level
24
Dac_b2
W
I
dac control for black offset level
25
Dac_b1
W
I
dac control for black offset level
26
Dac_b0
W
I
Lsb
dac control for black offset level
27
Nbias_oamp
A
I
100K to VDD and decouple to GND
output amplifier bias current
28
Sync_x\
D
I
low active (0=sync)
0 = reset X shift register
Document Number: 38-05707 Rev. *B
+ 5 V DC
Page 29 of 37
[+] Feedback
IBIS4-1300
No.
Name
Type
I/O
Description
Signal
29
clk_x
D
I
Shifts on falling edge
clock X shift register
30
shy
D
I
Column parallel track and hold
1 = hold; 0 = track
31
dccon
A
I
control voltage for DC reference
generation
Connect to GND (default)
32
dcref
A
O
reference voltage
Should be +/- 1.2 V, depends on dccon
33
gnd
A
G
34
vdd
A
P
35
sin
D
I
Column amplifier calibration signal
1 = calibrate, see timing diagram
36
sync_y\l
D
I
0 = start left shift register
low active (0=sync)
37
clk_yl
D
I
clock left shift register
Shifts on falling edge
38
eos_yl\
D
O
low 1st clk_yl pulse after last row
Active low
39
bitinvert
D
I
High active, 1 = invert bits
inverts ADC output bits
40
select
D
I
High active
selects row indicated by left/right shift
register
41
reset
D
I
High active
resets row indicated by left/right shift
register
42
d9
W
O
MSB
ADC output
43
d8
W
O
44
d7
W
O
45
d6
W
O
46
d5
W
O
47
d4
W
O
48
d3
W
O
49
d2
W
O
50
d1
W
O
51
d0
W
O
52
gnd
A
G
53
vdd
A
P
+ 5 V DC
54
gnd_ab
A
G
Anti-blooming drain voltage
GND or +1V for improved anti-blooming
55
vdd_array
A
P
+ 5 V DC
Pixel power supply
56
vdd_dig
D
P
+ 5 V DC
ADC digital power supply
57
gnd_dig
D
G
58
vdd_an
A
P
+ 5 V DC
ADC analog power supply
59
vdd_resetl
A
P
5 V DC default
(5.5 V for large output swing)
4…4.5 V for double slope mode
VDD for reset by left shift register
60
gnd_an
A
G
61
vhigh_adc
A
I
Document Number: 38-05707 Rev. *B
LSB
ADC ground of digital circuits
ADC ground of analog circuits
+ 4 V DC
High ADC reference voltage
Page 30 of 37
[+] Feedback
IBIS4-1300
No.
Name
Type
I/O
Description
Signal
62
clk_adc
D
I
ADC Clock
Converts on falling edge
63
tri_adc
D
I
ADC output tristate control
1=tristate; 0=output
64
pbiasdig1
A
I
100K to GND and decouple to VDD
current bias for comparator after encoder
65
pbiasencload
A
I
100K to GND and decouple to VDD
current bias for encoder
66
pbiasdig2
A
I
100K to GND and decouple to VDD
current bias for digital logic in columns
67
nonlinear
D
I
high active (1 = non-linear conversion) control for non-linear behavior of sensor
68
n.c.
69
nbiasana2
A
I
100K to VDD and decouple to GND
bias current 2nd comparator stage
70
nbiasana
A
I
100K to VDD and decouple to GND
bias current 1st comparator stage
71
vlow_adc
A
I
+ 2 V DC, +-2 K between P71 and P61 Low ADC reference voltage
72
gnd_an
A
G
73
in_adc
A
I
Converts between vlow and vhigh
(2-4V)
ADC input
74
vdd_an
A
P
+ 5 V DC
ADC analog power supply
75
gnd_dig
D
G
76
vdd_dig
D
P
+ 5 V DC
77
vdd
A
P
+ 5 V DC
78
gnd
A
G
79
vdd_resetr
A
P
5 V DC default
(5.5 for large signal swing)
Power supply for reset by right (readout)
shift register
80
L/R\
D
I
1=left; 0=right
Selects left or right shift register for 'select'
and 'reset'
81
Pixel diode
A
O
groups current of 24 x 18 pixels
Test structure for spectral response
measurement of pixels
82
Photodiode
A
O
168x126 um2 (eq. 24 x 18 pixels)
Test structure for spectral response
measurement of photodiode
83
clip
A
I
Clips if output > 'clip' - Vth (PMOS)
Clipping voltage for output amplifier
84
subsmpl
D
I
high active, 1 = subsampling
Selects viewfinder mode (1:4 = 320 x 256)
not connected
ADC ground of analog circuits
ADC ground of digital circuits
Bonding pad geometry for the IBIS4-1300
• The 84 pins are distributed evenly around the perimeter of
the Chip. At each edge there are 21 pins. Pin 1 is (in this
drawing) in the middle of the left edge.
• The opening in the bonding pads (the useful area for
bonding) is 200 x 150 um.
Document Number: 38-05707 Rev. *B
ADC digital power supply
• The centers of the bonding pads are at all four edges at 150
um distance from the nominal chip border.
• The scribe line (=the spacing between the nominal borders
of neighboring chips) is 250 um.
• The bonding pad pitch is 437 um in X-direction
• The bonding pad pitch in Y-direction is 393 um
Page 31 of 37
[+] Feedback
IBIS4-1300
• Relative position of pads in corners: see next figure
(measures in um).
635
627
150
393
9265
468
437
474
10158
Color filter geometry
B
Sensors with diagonal pattern have
• -pixel (1,1) is RED
• -first line sequence is BGRBGR
• -second line sequence is RBGRBG
sensors with Bayer pattern have
• -pixel (1,1) is GREEN
• -first line sequence is
.GRGRG
• -second line sequence is .BGBGB
Document Number: 38-05707 Rev. *B
Page 32 of 37
[+] Feedback
IBIS4-1300
Package
0.46" square cavity
Die thickness nominally 711 um +- 50um
84 pins ceramic LCC package (JLCC also available)
Clearance from top of die to bottom of glass lid: 400um
nominally
Standard 0.04 inch pitch outline
Figure 28. Pin layout and package, top view
74
54
75
53
1169 ȝm
ADC
0,0
Diagonal stripe pattern
84
Image sensor core
Pin1
568ȝm
588ȝm
1280,1024
11
885ȝm
12
33
32
Cover glass
Material: BG39
Size 18x18 mm for JLCC& LCC
This material acts a NIR cut-off filter. The transmission
characteristics are given in the figure below. The data used to
create the transmission curve of the BG39 material can be
obtained as an excel file upon simple request to
[email protected].
Color sensor
Refractive index: 1,55
Thickness: 0,75+-0.05 mm
Document Number: 38-05707 Rev. *B
Page 33 of 37
[+] Feedback
IBIS4-1300
Figure 29. Transmission characteristics of the BG39 glass used as NIR cut-off filter for the IBIS4-1300 color image sensors
Transmission
BG39 transmission characteristics
1
0,9
0,8
0,7
0,6
0,5
0,4
0,3
0,2
0,1
0
400
500
600
700
800
900
1000
Wavelength
Monochrome sensor
Material: D263
Refractive index: 1,52
The transmission characteristics are given in the figure below.
Thickness: 0,55+-0.05 mm
Figure 30. Transmission characteristics of the D263 glass used as protective cover for the IBIS4-1300 monochrome
image sensors
Transmission
D263 transmission characteristics
1
0,9
0,8
0,7
0,6
0,5
0,4
0,3
0,2
0,1
0
400
500
600
700
800
900
Wavelength
Document Number: 38-05707 Rev. *B
Page 34 of 37
[+] Feedback
IBIS4-1300
Ordering Information
Table 10. FillFactory and Cypress part numbers
FillFactory Part Number
Cypress Semiconductor Part Number
IBIS4-1300-C1-2
CYII4SD1300AA-QAC - Preliminary
IBIS4-1300-M-1
CYII4SM1300AA-HBC - Preliminary
IBIS4-1300-M-2
CYII4SM1300AA-QBC - Preliminary
FAQ
Temperature dependence of dark signal
IBIS 4C
1000
offset long tint
RMS long tint
% of saturation
nominal operation,
512 x 512 pixels corner,
160 ms tint
#pixels outside
6sigma
100
10
1
20
30
40
50
60
70
80
90
0,1
The above graph is measured on an IBIS4-1300 under
nominal operation, using breadboard. This particular sensor
has about 100 "bad pixels" at RT.
Useful range of "double slope"
Average offset (=dark signal) and RMS (=FPN of dark signal)
are measured versus temperature. Offset is referred to the
"short tint" offset at 20 C. Integration time was 160 ms (= "long
tint").
Assuming that the "regular" S/N is 2000:1, and that one can
put the knee point halfway the voltage range, the each
piecewise linear halve has 1000:1 S/N. If the ratio between
slopes is a, then the total dynamic range becomes
(1000+a*1000):1.
Y-axis is the output signal (100% = ADC range)
Which total dynamic range can reasonably be obtained with
the dual slope feature of the IBIS4-1300?
Example, is a=10, then the total dynamic range becomes
11000:1.
Document Number: 38-05707 Rev. *B
Page 35 of 37
[+] Feedback
IBIS4-1300
In practice, acceptable images are obtained with a up to 10.
Larger a's are useable, but near the knee, contrast artifacts
become annoying.
Skipping rows or columns
additional CLK_YR + CLR_YL, or CLK_X pulses. The
maximum clock frequency is not documented. But it is
probable that one can reach at least 10 MHz in Y and 40 MHz
in X.
Although these modes are not described in the datasheets, it
is possible to skip rows or columns by simply applying
Disclaimer
FillFactory image sensors are only warranted to meet the
specifications as described in the production data sheet.
Specifications are subject to change without notice.
Document Number: 38-05707 Rev. *B
Page 36 of 37
[+] Feedback
IBIS4-1300
Document History Page
Document Title: 1.3 MPxl Rolling Shutter CMOS Image Sensor
Document Number: 38-05707
REV.
ECN.
Issue Date
Orig. of
Change
Description of Change
**
310213
See ECN
SIL
*A
509557
See ECN
QGS
Converted to Frame file
*B
642577
See ECN
FPW
Ordering information update
Document Number: 38-05707 Rev. *B
Initial Cypress release
Page 37 of 37
[+] Feedback