一种基于最小量化误差流水线ADC校准算法

A Pipeline ADC Calibration Algorithm Based on Minimum Quantization Error

现代高速高精度流水线ADC往往需要校准算法辅助工作。高速高精度流水线ADC中运算放大器无法同时满足信号建立速度和建立精度的要求,往往会牺牲信号建立精度以保证建立速度,从而使得流水线ADC关键模块乘法数模转换器(MDAC)产生增益误差。基于流水线ADC最小量化误差原理,通过逐级搜索流水线ADC每级增益使得流水线ADC总量化误差达到最小,从而使流水线ADC有效位数(ENOB)最大,完成校准。基于流水线ADC最小量化误差校准算法应用于一款14 bit 250 Msps的流水线ADC,经校准后流水线ADC ENOB可达13.20 bit,信噪失真比(SNDR)可达81.23 dB,无杂散动态范围(SFDR)可达87.99 dB。

Modern high speed and high resolution pipelined ADC often needs calibration algorithm towork effectively. The operational amplifier in high speed high resolution pipeline ADC cannot meet both the speed and establish precision requirement at the same time. So we often sacrifice signal establish precision to guarantee speed requirement, thus the gain error of MDAC which is the pipeline ADC key module is produced.Based on the principle of minimum quantization error, through searching for the gain of each stage in the pipeline ADC step by step, when the total quantization error of the pipeline ADC is minimized then the ENOB reaches the maximum, and so the calibration is completed. The calibration algorithm based on the principle of minimum quantization error is applied to a pipeline ADC with 14 bit resolution 250 Msps sampling rate. After calibration, the ADC ENOB can reach 13.20 bit, the SNDR can reach 81.23 dB, and the SFDR can reach 87.99 dB.