LSB-first SAR ADC with bit-repeating for reduced energy consumption

Abstract

Least-significant bit first quantization (LSBFQ) is presented as an energy-efficient method for analog-to-digital conversion (ADC) when input signal activity is low. LSBFQ has been shown to conserve switching energy and comparator bitcycles, but certain code transitions will force unnecessarily large steps in the DAC output code and degrade the performance. A novel bit-repeating LSBFQ algorithm is proposed which prevents these undesirable code steps. It is shown that the proposed algorithm saves switching energy over previously published LSBFQ in all cases. For low signal activity (the intended target application for LSBFQ) the proposed LSBFQ method even outperforms merged capacitor switching (MCS), which is the most energy-efficient of the traditional MSB-first SAR methods.