REDUCTION OF SYSTEMATIC ADC ERRORS BY OVERSAMPLING

摘要

Oversampling, successive noise shaping, and final low-pass filtering is a well-known approach to enhance dynamic range of analog-to-digital converters. However, the potential of this approach can be limited by systematic errors overriding the quantization error. The paper deals with the reduction of this limitation: an interpolating algorithm, based on a moving-average FIR filter with Bayesian coefficients is proposed, and its configuration inside some correction schemes are presented. Effectiveness, drawbacks, and the corresponding evaluation results of these schemes are discussed.