A Sampling Time Mismatch Correction Method Based on Extremum Function

Abstract

Aiming at the problem of sampling time mismatch between channels in a time-interleaved analog-to-digital converter, the invention provides a sampling time mismatch correction method based on an extremum function. First, a reference channel is introduced, and the sampling clock of the reference channel is sequentially aligned with the time-interleaved sub-channels; then, the reference channel and the sub-channels are normally quantized, and the output at the time when the sampling clock of the reference channel is aligned with the sub-channel clock is selected as the calibration result. Then, subtract the selected data correspondingly, calculate the absolute value, and accumulate and sum to obtain an extreme value function about the sampling time mismatch, and the minimum value point of the function is no sampling time mismatch Then, according to the change of the function value of the extreme value function, adjust the clock delay of the corresponding sub-channel, so that the minimum value function converges to the direction of the minimum value point; finally, after several iterations, the extreme value function converges At the minimum point, the sample time mismatch is corrected.

Description

technical field [0001] The invention relates to a correction method for sampling time mismatch of a time-interleaved analog-to-digital converter (ADC) in microelectronic technology. Background technique [0002] With the improvement of ADC performance requirements, it becomes more and more important to solve the trade-off between speed and power consumption, and the way of time interleaving is an effective solution to this problem. The basic structure of an N-channel time-interleaved ADC, such as figure 1 shown, by setting N operating frequencies as f s / N sub-analog-to-digital converters are connected in parallel, and the operating frequency of the entire analog-to-digital conversion system is increased to f s . The time interleaving method can maximize the performance of a single channel, while improving the overall conversion rate, and can take power consumption into consideration, but its performance is also affected by the mismatch between channels, such as sampling ...

AI Technical Summary

Problems solved by technology

However, complex operations will introduce a multiplier, which has a large hardware overhead, and the corrected frequency range is also limited.

Images