ADC Parameter Test Based on Improved Maximum Likelihood Estimation

Wanyu Yan,Houjun Wang,Yixi Liao
DOI: https://doi.org/10.1109/icemi52946.2021.9679511
2021-10-29
Abstract:With the continuous improvement of ADC resolution, the implementation of existing ADC Standard Test methods becomes more and more difficult. The traditional test method based on sinusoidal fitting has high test accuracy, but the test time is long, which limits its application in practical test. The sine wave fitting algorithm based on maximum likelihood estimation (ML) can better take into account the test accuracy and test efficiency, and has a good application prospect. This paper mainly studies the optimization of ADC test algorithm based on maximum likelihood estimation. Firstly, the ADC conversion level is obtained by the parameter spectrum estimation method. When the conversion level is used as a known parameter, the accuracy and speed of the algorithm are guaranteed. The initial frequency is estimated by the frequency estimation method, and the signal parameters and noise standard deviation are estimated by the three parameter sinusoidal fitting method. After improvement, only one data acquisition is needed to test the dynamic and static parameters of high-speed and high-precision analog-to-digital converter, such as INL, SINAD, ENOB, etc. Through simulation and test, the test time is shortened on the premise of ensuring the test accuracy, and the test efficiency of high-speed and high-precision ADC is improved. It has great practical value.
What problem does this paper attempt to address?