Study And Design Of All Digital Feedforward Calibration Algorithm For TIADC

The development of electronic system equipment has facilitated the development of high-performance analog to digital converters.Time-Interleaved Analog-to-Digital Converter(TI-ADC)uses interleaved sampling techniques to achieve high sampling rates.However,factors such as process variation,device aging,and changes in the working environment will cause mismatch between TI-ADC sub-channels,including offset mismatch,gain mismatch,and sampling time mismatch.The dynamic performance of the inter-channel mismatched converter has a large impact,and as the process progresses,the mismatch between channels becomes more prominent,so the calibration techniques to solve the mismatch between channels have emerged.Firstly,the main three mismatch errors between channels are analyzed,investigated the development status of digital calibration technology;Then,an all-digital feedforward TI-ADC calibration algorithm is designed for the most difficult sampling time mismatch error.By constructing an error extraction function to estimate the sampling time phase difference between two adjacent sampling points of the TI-ADC,thereby forming a linear equation group to obtain the time mismatch,and based on the Taylor expansion principle for error compensation;Subsequently,a channel randomization technique of TI-ADC was designed to convert the remaining spurs after calibration into noise,SFDR performance can be improved without affecting the SNDR of the system.The calibration algorithm designed in this thesis has the characteristics of simple structure,wide applicable bandwidth,stable system,fast convergence,and has good scalability and can be used in any channel number TI-ADC system.The main problems and solutions encountered in the research include: For the problem that the calibration effect of the feedforward calibration algorithm is degraded when the frequency of the input signal is high,a method for constructing an inverse function for error adaptive correction is proposed,the effective calibration bandwidth can be increased from 0.19 to 0.39;In order to achieve channel randomization,the sampling clock of TI-ADC has different requirements than the order,a simple channel randomization control unit is designed in this thesis to allocate and control the sampling clock of the subchannel;For the problem of high hardware complexity of some arithmetic units in the calibration algorithm,the MMA filter is used to achieve the desired operation,and The LMS-based division and square operation unit is proposedand designed in this thesis.The effectiveness of the calibration algorithm was verified by MATLAB-based behavioral simulation,Verilog's circuit-level simulation,and FPGA-based hardware implementation.FPGA verification results show that,14-bit 4-channel TI-ADC system applies a sampling time mismatch error with an input signal normalized to a frequency of 0.35,the calibration algorithm and channel randomization can increase the SNDR of the system by 36.65 dB and the SFDR by 62.29 dB.