Redundant Double Conversion based Digital Background Calibration of SAR ADC with Convergence Acceleration and Assistance

Abstract

A redundant double conversion (RDC) based digital background technique for successive approximation analogue-to-digital converters (SAR ADCs) with convergence acceleration and assistance is presented. The convergence time of the RDC based calibration is reduced by monitoring the error between two raw conversion results of a sub-2 SAR ADC and increasing the step-size parameter of the least mean square (LMS) filter, if the detected error is less than a threshold. The convergence assistance is accomplished by injecting a random perturbation signal to a stationary input. Behavioral simulation shows that the proposed technique reduces convergence time of the RDC calibration by 30% while maintains a comparable resolution and power consumption. With the random perturbation signal generated by as-bit digital-to-analog converter (DAC) RDC calibration can be started and an ENOB about 10 is obtainable.