Optimization And Implementation Of ADC Dynamic Parameter Estimation Algorithm

The ADC traditional test is that static and dynamic parameters are obtained from the histogram and the dynamic test respectively.To abtain the whole parameters, the cost is very high. So, the relationship between static and dynamic parameters is studied. By using the static test to estimate the dynamic parameters, namely dynamic parameter estimation algorithm, all parameters can be abtained through one process. Compared with the tradirional test, the test cost can be saved in the premise of ensuring the accuracy.Dynamic parameters estimation method based on INL is found that very precise INL is needed which requires large samples. The test time increases greatly. Moreover, the noises only come form the quantization noise, the complexity of the actual noise source of the ADC circuit is not taken into account, which makes the dynamic parameter eatimation unaccurate. Based on the above defects, the original eatimation algorithm is optimized in the thesis. By using the static code to get dynamic parameters by spectrum analysis. The histogram test and estimation dynamic parameters test are independent processes, the precise INL is not in need which can save large test time. The whole parameters and noises can be obtained from the spectrum which result in more precise accuracy. The detail process is that extracting the static codes by the same interval to get the smaller codes for spectrum analysis and then the least squares fitting is used to abtain the codes with no truncation to eliminate the influence of the gain error. Finally, computing dynamic papameters by the spectrum analysis. In addition, the thesis uses the moving average filter for the the histogram test to optimize the test time. The testing platform is built with AD9258 chip which is fabricated by ADI corporation to verify the algorithm. The experiments indicate that the estimate error of maximum INL, SINAD, THD and SFDR, ENOB is 0.186LSB,0.211dB, 0.159dB and 0.119dB,0.035. Compared with the original algorithm, the estimation accuracy of SINAD and ENOB increases 0.343dB and 0.057dB. Compared with the traditonal method, the test time is reduced by 50.769% on the promise of the test precision.In the thesis, the optimized algorithm is widely applied in the mixed signal test field. All ADC parameters are obtained by one process and achieve the purpose of saving the test cost.