Nonuniform mismatches compensation algorithm for time-interleaved sampling system using neural networks

Time-interleaved sampling system increase the overall sampling rate by combining multiple slow ADCs. However, its performance suffers from several mismatches. This paper introduces a BPNN-based compensation method to deal with the automatic compensation of timing skew, gain and offset mismatches simultaneously, and track the time-varying errors self-adaptively. Simulation results show that the calibration technique can greatly attenuate the spurs and the SFDR can be significantly improved by 27-58 dB, and it demonstrates the efficiency of proposed method.