`
`Boyd Fowler
`
`Abbas El Gamal
`
`David X. D. Yang
`
`Information Systems Laboratory, Electrical Engineering Department, Stanford University, Stanford, CA 94305-4055
`
`email: fowler@isl.stanford.edu abbas@isl.stanford.edu dyang@isl.stanford.edu
`
`Phone: 1-415-723-3473 FAX: 1-415-723-8473
`
`Session Number: TP 13.5
`
`June 21, 2001
`
`Abstract
`
`A CMOS 64 × 64 pixel area image sensor chip using Sigma-Delta modulation at each pixel for A/D
`
`conversion is described. The image data output is digital. The chip was fabricated using a 1.2µm two
`
`layer metal single layer poly n-well CMOS process. Each pixel block consists of a phototransistor and 22
`
`MOS transistors. Test results demonstrate a dynamic range potentially greater than 93dB, a signal to
`
`noise ratio (SNR) of up to 61dB, and dissipation of less than 1mW with a 5V power supply.
`
`1
`
`Magna 2029
`TRW v. Magna
`IPR2015-00436
`
`
`
`Boyd Fowler, Abbas El Gamal, and David X. D. Yang
`
`2
`
`Charge-coupled devices (CCD) are at present the most widely used technology for implementing area
`
`image sensors. CCD image sensors have their shortcomings, however. They suffer from low yields, they
`
`consume too much power [3], and they are plagued with SNR limitations due to the shifting and detection
`
`of analog charge packets, and the fact that data is communicated off chip in analog form.
`
`Several alternatives to CCD area image sensors that use standard CMOS technology have been developed.
`
`Self scanned photodiode arrays have been used to produce both binary and grayscale image sensors [2, 3].
`
`Bipolar junction phototransistor arrays [6], and charge injection arrays [5] have also been used. However,
`
`these alternatives can suffer from low resolution due to limited pixel observation time, limited SNR due to
`
`analog sensing, and as with CCDs, data is communicated off chip in analog form.
`
`In this paper we describe an area image sensor that can potentially circumvent the limitations of CCDs
`
`and their alternatives. The proposed image sensor uses a standard CMOS process and can therefore be
`
`manufactured with high yield. Digital circuitry for control and signal processing can be integrated with
`
`the sensor. Moreover, CMOS technology advances such as scaling and extra layers of metal can be used to
`
`improve pixel density and sensor performance.
`
`The analog image data is immediately converted to digital at each pixel using a one-bit Sigma-Delta
`
`modulator [1]. The use of Sigma-Delta modulation also allows the data conversion circuitry to be simple
`
`and insensitive to process variations [1]. A global “shutter” provides variable light input attenuation to
`
`achieve wide dynamic range [2]. Data is communicated off chip in a digital form, thus eliminating the SNR
`
`degradation of analog data communication.
`
`To demonstrate the viability of our approach, an area image sensor chip has been designed and fabricated
`
`in a 1.2 µm 1 CMOS technology. A functional block diagram of the chip is given in Figure 2. It consists of
`an array of 64 × 64 pixel blocks, a clock driver, a 6:64 row address decoder, 64 latched sense amplifiers, and
`
`16 4:1 column multiplexers. The chip also contains data compression circuitry which will not be described
`
`in this paper. A die photograph is given in Figure 5 and a summary of the main characteristics of the chip
`
`are listed in Table 1.
`
`1µ is the symbol for 10
`
`−6.
`
`
`
`Boyd Fowler, Abbas El Gamal, and David X. D. Yang
`
`3
`
`A circuit schematic of the function implemented at each pixel is given in Figure 1. The phototransistor
`
`is a vertical bipolar PNP transistor; the emitter is formed using source-drain p+ diffusion, the base is the
`
`n-well surrounding the emitter and the collector is the p- substrate. The n-well is exposed to light, while
`
`the rest of the circuitry is covered with the second level of metal to reduce the chance of photon induced
`
`latch-up. The physical construction and operation of bipolar phototransistors are described in [7, 4]. Control
`
`of the input photocurrent is achieved by setting the duty cycle (the ratio between the on and off times) of
`
`the shutter input SHUTTER — the higher the duty cycle the larger the input photo current. Current from
`
`the phototransistor is integrated on C1 and quantized using a regenerative latch clocked via PHI2. The
`
`quantized value is converted into a current using a 1 bit D/A converter and fed back to the input capacitor
`
`C1. The duty cycle of PHI1 and the voltage VBIAS1 control the magnitude of the feedback signal Delta.
`
`PHI1 and PHI2 constitute a two phase nonoverlapping clock. At the completion of each two phase clock
`
`cycle a single bit is produced. The bit is read by enabling the word line WORD. If the bit is high the
`
`precharged bit line BIT is pulled down and sensed by a simple single-ended sense amplifier.
`
`The operation of the area image sensor chip is as follows: after an image is focused on the chip the Sigma-
`
`Delta modulators are reset via the global RESET signal. SHUTTER is then globally set to maximize
`
`image SNR without saturating the data conversion circuitry. Next the Sigma-Delta modulators are globally
`
`clocked at a rate Fs above the image frame rate 2Fd. This is necessary since a Sigma-Delta modulator
`
`reduces quantization error at the cost of extra data. At the end of each clock cycle the outputs of the
`Sigma-Delta modulators form a 64 × 64 array of bits referred to as a “bit plane.” Each bit plane is read out
`
`row by row. The image is fully captured using a number of bit planes determined by the target SNR. Using
`
`the theoretical analysis in [1] the number of bit planes L needed versus SNR is given by
`
`SNR = 9 log2 L − 5.2dB.
`
`The maximum achievable SNR has been measured at 61dB. SNR degradation due to charge injection of the
`
`digital circuitry in close proximity to the analog sensors is negligible since the frequency of operation is very
`
`low (1kHz) and the circuitry consumes less than 20nA per pixel.
`
`Figure 3 shows the output from a single pixel’s Sigma-Delta modulator.
`
`
`
`Boyd Fowler, Abbas El Gamal, and David X. D. Yang
`
`4
`
`The digitized pixel values are reconstructed using a decimation filter [1]. Depending on the type of
`
`application in which the sensor is used, this reconstruction may be implemented in software, using special
`
`purpose hardware external to the sensor, or integrated with the sensor.
`
`In a low resolution application
`
`where no local reconstruction is needed, e.g. video phone or surveillance camera, the sensor digital output
`
`is compressed and immediately transmitted. Reconstruction is done at the receiving end using general or
`
`special purpose hardware. If the image is to be displayed or processed locally one or more decimation filters
`
`are integrated with the sensor and an external RAM is used. The pixel values stored in the RAM are
`
`recursively updated by reading them into the sensor, updating their values using the decimation filters and
`
`the new bits from the corresponding sigma delta modulators, and storing the new values back into the RAM.
`
`This scheme appears feasible even for a sensor with as many as 1 million pixels operating at 30 frames per
`
`second at 8 bits per pixel resolution.
`
`The sensor can achieve a dynamic range2 potentially greater than 93dB. This is because the magnitude
`
`of the photocurrent can be varied by a factor of 1000, or 60dB, and the maximum measured SNR is approx-
`
`imately 33dB with the SHUTTER duty cycle set at 100%, the frame sampling rate set at 30Hz, and the
`
`oversampling ratio set at 64.
`
`Figure 4 shows a scan from a 35mm print, and the image obtained by the sensor when contact exposed
`
`to the 35mm negative.
`
`The sensor’s estimated total power of less than 1 mW is significantly lower than that of other types of
`
`image sensors.
`
`We would like to thank B. P. Wong and D. How for their contribution to the test bed, M. Godfrey, B.
`
`Wooley, and L. Hesselink for their support and encouragement, and MOSIS for fabrication.
`
`References
`
`[1] J. C. Candy. “A Use of Double Integration in Sigma Delta Modulation”. IEEE Trans. Comm., 33(3):249–
`
`258, March 1985.
`
`2The ratio of the maximum non-saturating photocurrent to the dark current.
`
`
`
`Boyd Fowler, Abbas El Gamal, and David X. D. Yang
`
`5
`
`[2] P.B. Denyer et al. “On-chip CMOS sensors for VLSI imaging systems”. In VLSI 91, Edinburgh, UK,
`
`August 1991.
`
`[3] EG&G Reticon, 345 Potrero Ave., Sunnyvale, CA 94086-4197 USA. Image Sensing Products 1992/1993,
`
`1992.
`
`[4] C. Mead. “A Sensitive Electronic Photoreceptor”. In 1985 Chapel Hill Conference on VLSI, Chapel Hill,
`
`NC, 1985.
`
`[5] G. J. Michon and H. K. Burke. “Charge Injection Imaging”. In ISSCC Digest of Technical Papers, pages
`
`138–139, February 1973.
`
`[6] N. Tanaka et al. “A 310k Pixel Bipolar Imager (BASIS)”. In ISSCC Digest of Technical Papers, San
`
`Fransisco, CA, February 1989.
`
`[7] N. Tanaka, T. Ohmi, and Y. Nakamura. “A Novel Bipolar Imaging Device with Self-Noise Reduction
`
`Capability”. IEEE Trans. Elec. Dev., 36(1):31–38, January 1989.
`
`
`
`Boyd Fowler, Abbas El Gamal, and David X. D. Yang
`
`6
`
`1 bit A/D
`
`2-D Mux
`
`VBIAS1
`
`PHI2
`
`SHUTTER
`
`Integrator
`
`Phototransistor
`
`VBIAS2
`
`C1
`
`RESET
`
`VBIAS1
`
`PHI1
`
`1 bit D/A
`
`WORD
`
`BIT
`
`Figure 1: Figure 1: Pixel Schematic. TP-13.5: ”A CMOS Area Image Sensor with Pixel-Level A/D Conver-
`
`sion”, Boyd Fowler, Abbas El Gamal, and David X. D. Yang
`
`
`
`Boyd Fowler, Abbas El Gamal, and David X. D. Yang
`
`7
`
`Two Phase Clock Inputs
`
`Clock Drivers
`
`64 x 64 Image Sensor Core
`
`6 bit Address
`
`6:64 Row Decoder
`
`4 Column Outputs
`
`4 Column Outputs
`
`Output
`Latch
`
`D
`Q
`
`D
`Q
`
`D
`Q
`
`D
`Q
`
`D
`Q
`
`D
`Q
`
`D
`Q
`
`D
`Q
`
`4:1 Multiplexer
`
`4:1 Multiplexer
`
`2 bit column
`decode
`
`16 bit Output Data
`
`Figure 2: Figure 2: Image Sensor Chip Functional Block Diagram. TP-13.5: ”A CMOS Area Image Sensor
`
`with Pixel-Level A/D Conversion”, Boyd Fowler, Abbas El Gamal, and David X. D. Yang
`
`
`
`Boyd Fowler, Abbas El Gamal, and David X. D. Yang
`
`8
`
`0.001
`
`0.002
`
`0.003
`
`0.004
`
`0.005
`Seconds
`
`0.006
`
`0.007
`
`0.008
`
`0.009
`
`0.01
`
`0.001
`
`0.002
`
`0.003
`
`0.004
`
`0.005
`Seconds
`
`0.006
`
`0.007
`
`0.008
`
`0.009
`
`0.01
`
`0246
`
`-2
`0
`
`Volts
`
`6
`
`4
`
`2
`
`0
`0
`
`Volts
`
`Figure 3: Figure 3: Single pixel Sigma-Delta modulator output from HP54601A. The top graph is PHI2
`
`versus time and the bottom graph is the pixel output waveform versus time. TP-13.5: ”A CMOS Area
`
`Image Sensor with Pixel-Level A/D Conversion”, Boyd Fowler, Abbas El Gamal, and David X. D. Yang
`
`
`
`Boyd Fowler, Abbas El Gamal, and David X. D. Yang
`
`9
`
`Main Characteristics of 64x64 Area Image Sensor
`
`Technology
`
`Die Area
`
`Pixel Area
`
`1.2µm, 2-layer metal, 1-layer poly, nwell CMOS
`6500µm × 5000µm
`60µm × 60µm
`
`Number of transistors per pixel
`
`22
`
`Phototransistor Area
`
`105µm2
`
`Package
`
`Supply Voltage
`
`Maximum SNR
`
`Dynamic Range
`
`Power Dissipation w/o Pads
`
`Measurement Temperature
`
`84pin PGA
`
`5v
`
`61dB
`
`93dB
`
`< 1 mW
`23◦C
`
`Table 1: Table 1: Area Image Sensor Characteristics. TP-13.5: ”A CMOS Area Image Sensor with Pixel-
`
`Level A/D Conversion”, Boyd Fowler, Abbas El Gamal, and David X. D. Yang
`
`
`
`Boyd Fowler, Abbas El Gamal, and David X. D. Yang
`
`10
`
`Figure 4: Figure 4: 300dpi scan of print from negative (left). 64 × 64 image produced by sensor using
`
`35mm negative contact exposure (right). TP-13.5: ”A CMOS Area Image Sensor with Pixel-Level A/D
`
`Conversion”, Boyd Fowler, Abbas El Gamal, and David X. D. Yang
`
`
`
`Boyd Fowler, Abbas El Gamal, and David X. D. Yang
`
`11
`
`Figure 5: Figure 5: Die Photograph of 64 × 64 Image Sensor Chip. TP-13.5: ”A CMOS Area Image Sensor
`
`with Pixel-Level A/D Conversion”, Boyd Fowler, Abbas El Gamal, and David X. D. Yang