ABSTRACT
This chapter provides guidelines to choose and design analog-to-digital converters (ADCs) for image sensors. Results can be extrapolated to other kinds of radiation detectors on the focal plane. The main architectures usually employed to read out image sensor pixels’ outputs are studied and compared. Modern ADC topologies are analyzed. The specific ADC requirements for image sensor applications are also 48described. The performance of relevant, recently published ADCs for image sensors is benchmarked. Finally, a discussion of the advantages and disadvantages of the different ADC architectures based on the state of the art is included. A specific figure of merit (FoM) to compare different ADCs for image sensors is proposed. ADCs are widely used for multiple purposes that are related to signal acquisition and data processing. Frame-based image sensors have traditionally included one ADC to read out their outputs and digitize them for further processing, representation, or storage. The image capture is a process with several steps. Image sensors transduce photons into an analog voltage signal that is proportional to the light intensity. Then, this analog output is converted into a digital signal, usually stored on a memory, and sent out for further representation. The quality of the analog-to-digital conversion affects the quality of the final image or frame. Our eyes are quite sensitive to the fixed-pattern noise (FPN) that is introduced during the analog-to-digital conversion. (Humans can detect a 0.5% change in mean intensity [1].) The number of bits of the ADC should be high enough to represent the output image with a number of gray levels that are similar to or higher than the number that our eyes can detect. The speed of the analog-to-digital conversion can limit the frame rate of our sensor. The noise introduced by the ADC should be controlled. Finally, the circuitry in charge of processing the output data flow provided by the converters has to be fast enough to avoid losing information.
