An Adaptive and Universal Timing Mismatch Estimation Method for TIADCs
Lin Sun, Zhenwei Zhang, Lili Lang, Tong Kang, Wei Xiong, Yu Liu, Wei Zhong, Yemin Dong
Abstract
This brief proposes a fully digital background estimation technique for timing mismatch in time-interleaved analog-to-digital converters (TIADCs). The improved estimation technique is a feedback method based on statistics and Taylor series expansion. Besides, an exponential moving average (EMA) with adaptive coefficient is introduced to prevent performance degradation near specific frequencies, which improves the robustness of calibration. This estimation method can maintain high performance of accuracy and convergence speed across nearly the entire input frequency band. Furthermore, it is not restricted to the number of channels, amplitude, and frequency of the input signal. The effectiveness of this technique is demonstrated through OFF-chip calibration of a commercial 12-bit four-channel 3 GS/s TIADC using a 1.01 GHz input signal. Results indicate that the proposed method converges after approximately 3000 samples, reducing spurious signals stemming from timing mismatch by 20.6, 19.3, and 18.5 dB, respectively.