A novel timing skew calibration method based on autocorrelation function for time-interleaved ADCs
Abstract
A novel timing skew calibration method using autocorrelation function is proposed, with adaptive calibration sequences introduced for four-channel and multi-channel TI-ADCs. For four-channel systems, the channel with the largest deviation of one iteration step is adjusted by the method in 99% of cases. For multi-channel systems, the four-channel method is cyclically applied for calibration, achieving residual timing skew within 3.3 iteration steps in 99% of cases. Superior SFDR and SNDR compared to existing methods are demonstrated in simulations on a 32-GS/s 8-bit TI-ADC behavioral model with random skews (-3ps to 3ps). Significant improvements in SFDR under near-Nyquist frequency inputs are achieved by the proposed method compared to previous methods, with enhancements of 6-10 dB in the four-channel case and 3-12 dB in the eight-channel case. Background calibration with limited additional hardware is enabled by the approach, overcoming bandwidth limitations and convergence issues in prior techniques.