Adaptive Blind Correction of TIADC Mismatch Based on Cyclic Autocorrelation

Abstract

A blind and adaptive method is proposed for time-interleaved analog-to-digital converter (TIADC) online mismatch correction. Firstly, a TIADC model is derived by transforming mismatches to FIR filters with different amplitudes and lags. Subsequently, residual error measure of corrected output for wide sense stationary input is built following cyclic autocorrelation function. Mismatches estimator could be obtained by minimizing the measure, in consequence, gain and timing mismatches correction could be achieved accurately via adjust parameterized filter adaptively. It is experimentally verified that the proposed method can enhance spurious-free dynamic range of TIADC output by 27~32dB. The proposed method can be used for online correction of gain and timing mismatch of TIADC with zero-mean wide stationary input.