《[ADC Characterization in Time Domain]》.pdf

1536 IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 55, NO. 5, OCTOBER 2006 [ADC Characterization in Time Domain] Frequency Estimation to Linearize Time-Domain Analysis of A/D Converters ˇ Dominique Dallet, Member, IEEE, David Slepicka, Yannick Berthoumieu, Djamel Haddadi, and Philippe Marchegay Abstract—Time-domain analysis is a powerful tool for testing of the ADC data record, a power estimation of each harmonic A/D converters (ADCs), particularly by incoherent sampling. This component as well as noise can be obtained. Then, it is possible sampling is encountered for instance in real application and sys- to compute all of the dynamic parameters as described in [1]. tems where the clock is generated by an on-board crystal oscillator. In the case of the sine-wave model for ADC output samples, The four-parameter sine-wave fitting is based on least fitting a sine-wave function to the data record is nonlinear with square methods requiring iterative algorithms (Newton, respect to the frequency. To overcome the convergence problem, Gauss–Newton, Levenberg–Marquard, ...) whose convergence the authors propose some alternative methods based on frequency depends on the precision of the initial estimate [3], [4], particu- estimation by means of fast Fourier transform and eigenanalysis of larly that of the initial frequency [5]–[7]. A globally convergent the autocorrelation matrix. This leads to a simple linear problem. The effectiveness of these methods is proven on both simulation method based on the behavior of the cost function taking into and experimental results.