Maximum likelihood estimation for linearity testing of ADCs stimulated by known constant signals

Abstract

A maximum likelihood (ML) estimator is derived for the problem of measuring the code transition levels of an analog-to-digital converter (ADC). The proposed method is intended to test the linearity of the ADC in the static regime, using only constant test signals, except for a small amount of additive noise. The measurement data are employed in a nearly optimal manner, due to the statistical properties of the ML estimator, which are thoroughly examined. The reported analysis allows the design of the test under a given uncertainty constraint.