A Blind Timing Skews Estimation Algorithm of Time-interleaved Analog-to-digital Converters Based on Spectra Sparsity

The blind Timing Skews Estimation (TSE) issue for the Time-Interleaved Analog-to-Digital Converters (TIADC) is investigated in this paper. Accordingly to the TIADC configuration, the spectra sparsity of the analog input signal, and the non-overlapping frequency points, the spectra relations among the input signal of the TIADC, the output of the TIADC and the output of the sub-ADCs are explored based on the sampling theorem and the undersampling theory. A novel blind timing skews estimation algorithm is developed by using the phase information of the resultant sequence of the Inverse Discrete Fourier Transform (IDFT) of the relative output spectrum ratio at the non-overlapping frequency point. The simulation results show that the proposed TSE algorithm has the comparable parameter estimation performance of the sin-fit algorithm. Moreover, the proposed algorithm offers some desired properties, such as robustness to the additive noise, no limitation to the frequency of the input signal, no constrains on the channel number of the TIADC and no requirement on oversampling of the input signal. The experimental result using the captured data of a developed TIADC prototype validates further the high accuracy and effectiveness of the proposed TSE algorithm.