A Signal-Independent Background Calibration Technique for Time-Interleaved ADCs

This paper proposes a signal-independent background calibration technique to compensate for offset, gain and timing mismatch in time-interleaved analog-to-digital converters. In this topology, a user-provided calibration signal is adopted and alternatively sampled by all the interleaved channels in the calibration mode, which can be turned on or off by a digital control signal. Based on the user-provided calibration signal, these mismatches can be extracted through various estimation algorithms, which is independent of the input signal. To avoid interrupting the normal signal conversion operation, an extra substitute sub-ADC channel is employed to cycle through and substitute for the interleaved channels to sample the analog input. Once the estimation process converges, the calibration mode can be turned off and the digital outputs of normal signal conversion are then corrected by utilizing the estimates of these mismatches. Simulation results show that the proposed calibration architecture combined with a correlation-based timing skew estimation method, successfully compensates for the offset, gain and timing mismatch, with various inputs of DC voltage, single tone near Nyquist and random signal, respectively. Compared with traditional reference-ADC-based calibration schemes, besides the signal-independent characteristic, the proposed scheme avoids the time-variant input impedance problem. Moreover, under the condition of homologous mismatch errors, the proposed calibration technique achieves identical mismatch estimation process regardless of the input.