摘要

In this paper, we present an adaptive calibration method for correcting channel mismatches in time-interleaved analog-to-digital converters (TIADCs). An equivalent signal recombination structure is proposed to eliminate aliasing components when the input signal bandwidth is greater than the Nyquist bandwidth of the sub-ADCs in a TIADC. A band-limited pseudorandom noise sequence is employed as the desired output of the TIADC and simultaneously is also converted into an analog signal, which is injected into the TIADC as the training signal during the calibration process. Adaptive calibration filters with parallel structures are used to optimize the TIADC output to the desired output. The main advantage of the proposed method is that it avoids a complicated error estimation or measurement and largely reduces the computational complexity. A four-channel 12-bit 400-MHz TIADC and its calibration algorithm are implemented by hardware, and the measured spurious-free dynamic range is greater than 76 dB up 90% of the entire Nyquist band. The hardware implementation cost can be dramatically reduced, especially in instrumentation or measurement equipment applications, where special calibration phases and system stoppages are common.