Time domain interleaving analog-digital converter sample time mismatch calibration method and system

Abstract

The present invention discloses a time domain interleaving analog-digital converter sample time mismatch calibration method and system. The system includes N standard channels and a reference channel, a sample time mismatch coefficient extracting circuit, and a sample time adjustment control circuit. Adjustment circuit control coefficients of a to-be-calibrated standard channel and a to-be-calibrated reference channel are extracted by the sample time mismatch coefficient extracting circuit, the adjustment circuit control coefficients are transmitted to the sample time adjustment control circuit, and the sample time adjustment control circuit controls a sample clock of the to-be-calibrated standard channel. Through adoption of the time domain interleaving analog-digital converter sample time mismatch calibration method and system, sample time calibration is not affected by amplitude errors, the circuit structure is simple to implement, a differentiator is not needed, and the method is rapid in convergence, has high calibration accuracy, and is not affected by signal amplitude and gain mismatch.

Description

technical field [0001] The invention relates to the field of background calibration, in particular to a calibration method and system for sampling time mismatch of a time-domain interleaving analog-to-digital converter. Background technique [0002] Employing a time-domain interleaved pattern is an easy way to significantly increase the effective sampling rate of an analog-to-digital converter (ADC), figure 1 shows a typical time-domain interleaved ADC consisting of N identical ADC channels, each clocked at f ck , but the sampling time is staggered by Tck / N, therefore, the overall sampling rate of the N-channel ADC is equivalent to a clock frequency of f ck *N single-channel ADC with a sampling time of 1 / (f ck *N). Obviously, the chip area and power consumption of this interleaved ADC are N times that of the single-channel ADC. However, when the single-channel ADC is difficult to achieve a high sampling rate (limited and manufacturing process), in These are acceptable wh...

AI Technical Summary

Problems solved by technology

[0003] Compared with the single-channel ADC, although the time-domain interleaving ADC has the advantage of speed, it still has some defects: the mismatch between different channels produces clutter signals on the output spectrum, and the channel mismatch mainly includes : There are three types of offset mismatch, gain mismatch, and sampling time mismatch. For an N-channel interleaved analog-to-digital converter, the offset mismatch is at a frequency of m*f ck Noise spectral components are produced at ; the gain mismatch is at frequency m*f ck + / -f in Noise spectral components are generated at ; the sampling time mismatch is also at frequency m*f ck + / -f in Noise spectrum components are generated at (m=1, 2...N, f in is the input signal frequency); in actual design, 8-bit to 10-bit analog-to-digital converters can achieve accurate matching between channels through careful circuit design and layout design, but in more than 10-bit analog-to-digital converters , when higher performance is required, it is necessary to design a specific calibration scheme for mismatch error calibration. In order to avoid the interruption of the operation of the analog-to-digital converter, it is very necessary and desirable to run the calibration in the background

[0009] In order to reduce the calibration range and make the design simple, in the actual design, we only need to obtain the relative sampling time mismatch between different channels rather than the absolute sampling time mismatch relative to the reference channel (the reference channel may vary according to the specific implementation has large timing skew), the method is not sensitive to gain mismatch mainly due to the fact that S n *S ref is the cross-correlation of the derivative of the input signal and the input signal, and its value is close to zero if the sampling time of the reference channel and the calibration channel are exactly the same (the input signal can be seen as the sum of the differential sine wave and the cross-correlation of the sine wave and its derivative to zero), compared to figure 2 method, which is more robust, however, measuring the derivative of the input signal shown is not a trivial task and requires an analog or digital differentiator

But this not only increases the complexity of the circuit, but its calibration accuracy is limited and the accuracy of the differentiator

Images