Research On Continuous-wave Time-of-flight (TOF)3D CMOS Image Sensors

Continuous-wave time-of-flight 3D CMOS image sensors differ from the conventional CMOS image sensors due to their property of detecting the distance information of objects.3D CMOS image sensors have been widely used in consumer electronics,medical imaging,industrial defection and etc.As for a 3D CMOS image sensor,the performance of its pixel directly influences the measurement accuracy.The readout circuits of it convert the signals of pixel from analog domain into digital domain,which decides the final measurement accuracy of a 3D CMOS image sensor.Therefore,the research on the pixel and the low-noise readout circuits of a continuous-wave timeof-flight 3D CMOS image sensor is presented in this thesis.In the first part of this thesis,how the light source and shot noise of pixel effecting measurement accuracy are presented.After that,the analysis on the pixels of single FD or a double FDs used in a TOF 3D CMOS image sensor is presented.A capacitor-flipflop-integration pixel suitable for standard CMOS process is designed.The readout circuits are also designed.In order to lower the readout noise without lowering the speed,a fast correlated multiple sampling method based on a 12-bit SAR ADC is proposed,which is from the conventional correlated multiple sampling.The SAR ADC is designed with digital calibration and one-bit redundancy in order to enhance the tolerance of capacitor mismatch in SAR ADC.Finally,a switching scheme of DAC is proposed to lower the dynamic power consumption of the DAC in the SAR ADC.Besides,chopping,a high-speed comparator and digital calibration are used to lower the power of in the comparator in the SAR ADC.A TOF 3D CMOS image sensor has been designed and fabricated in a SMIC 0.18-?m CMOS process.The transistor-level simulation results show that the measurement error caused by the pixel is 6%.The ENOB of the SAR ADC with mismatch is improved from 10.06 bits to 11.76 bits with the help of calibration.A 16-times fast sampling and quantization based on this SAR ADC lowers the readout noise to one fourth and consumes 5.45?s,which is reduced by 31.9% compared with that based on traditional method,8?s.Finally,the power consumption of the comparators in SAR ADC is optimized and reduced from 1.023 mW to 0.3768 mW without decreasing the ENOB.