Research And Implementation Of Time Base Deviation Estimation And Correction In High Speed Multiplexing A/D Alternate Sampling Technique

Analog-to-digital converters(ADC)are widely used in modern electronic devices and calculator control systems.When an analog signal processor needs to interface with a digital signal processor,the latter is used to extract information from the signal or eliminate unwanted interference or distortion to adjust the signal.Conversion speed is still a key bottleneck for high-speed signal processing applications.Unfortunately,the design of high speed converters is limited by hardware.One way to achieve high sampling rates is to use time-interleaved ADC structures that use front-end parallel progressive sampling,back-end serial,and multiplexing techniques,ie,multiple A/D alternate sampling techniques.However,the accuracy of the data obtained by this technique is affected by the mismatches(bias mismatch,gain mismatch,time base deviation)between the sampling channels,where the time base deviation causes unwanted spectral components.This results non-uniform sampling.Therefore,how to provide more accurate sampling time base and reconstruct the sampled signal by using the estimated time base deviation.Issues to consider when using multi-channel A/D alternate sampling techniques.In this paper,combined with the practical application in the field of sampling,theoretical research and experimental analysis are carried out on the two aspects of time base deviation prediction algorithm and TIADC sampling system signal correction in high-speed multi-channel A/D alternate sampling technology:Firstly,an effective time-base deviation estimation algorithm,a fast iterative blind estimation algorithm,is proposed and analyzed to estimate the time base deviation in the time alternating ADC in the M channels.The algorithm uses sampling channel 0 as the reference channel.Based on any multi-channel system,it can be decomposed into a dualchannel time-base deviation prediction model for subsequent channels and reference channels.At the same time,in order to reduce the computational complexity in practical engineering,the iterative method is used,also,the accelerated convergence SteffensenNewton algorithm is used in the iterative process.The algorithm does not need to calculate the derivative value,and can reduce the number of iterations under the same precision requirement,thereby reducing the calculation amount.Secondly,the essence of the high-speed multi-channel A/D alternate sampling signal correction problem is to reconstruct the sampling signal by using non-uniform sampling data after the known time base deviation.In the theoretical analysis of the polynomial approaching reconstruction algorithm,it can be seen that the periodic non-uniform sampling signals have the same time interval every M sampling points,so the periodicity feature can be used to simplify the reconstruction algorithm.Theoretical analysis and experiments prove that this greatly reduces the complexity of the algorithm under the premise of ensuring the accuracy of the algorithm.Finally,based on the fast iterative blind estimation algorithm,the time-base deviation of the multi-channel A/D alternate sampling system is estimated,and the obtained sample is reconstructed using the improved polynomial approaching reconstruction algorithm.Experiments show that the improved polynomial approximation reconstruction algorithm based on fast iterative blind estimation algorithm can effectively reconstruct the sample sequence and improve the SNDR,ENOB and SFDR(SNDR can be improved by 13~18dB,ENOB can be improved by 1bit~2bit,SFDR can be increased by 5~14dB).