Wavelet-based differential nonlinearity testing of mixed signal system ADCs

Abstract

This paper discusses the work done in the testing and measurement of Differential Nonlinearity (DNL) of Analogto-Digital Converters (ADCs) by applying a novel testing technique of wavelet transform. The results of the new method is compared with conventional testing methods.The wavelet transform technique shows significant advantage over the conventional techniques.

The Application Wavelet Transform Algorithm in Testing ADC Effective Number of Bits

“…In real ADC's operation, quantization errors which cause nonlinearity effect on ADC's transfer function results into spectral frequencies other than the input frequency being tested. Quantization errors caused by IC internal or external noise [1,4] appear as random noise spread across the frequency spectrum of FFT, distort the conversion process, and result in harmonic distortions [10] and higher noise floor. The relation between nonlinearity, harmonics, and noise floor are used to address ADC's ENOB and SNR.…”

“…Wavelet transforms have been used in various fields of signal processing [10,12,17] due to the functionality of multi-resolutions that allow pinpointing of signal components. With special properties of dilation and translation [4,10,13], wavelets can create different scaled and shifted functions of signal transformation. In other words, unlike the Fourier transform, wavelet transform capability of dilation and translation allow the shift of a signal in the time domain (X axis), rescaling (to expand or compress a signal on Y axis) and produces flexible windows for analysis as shown in figure 1.…”

“…For proper system performance and digitizing accuracy, output digitized data need to be as close to the analog input signal. At most basic level, digitizer testing would seem simple matter; however, testing is extremely expensive, and time consuming as in [2][3][4][5] for both static and dynamic parameter characterizations.…”

“…For instance, some focused on improving classical methods such as Fourier transform and Sinusoidal Histogram in testing AC and DC performance [5][6][7][8][9]. While Fourier Transform is based on additive noise model and requires a large number of data samples [2], histogram testing has most samples occur near the ends of histogram (large number of samples must be collected to increase the height of bins around the center) [1,4].…”

“…In this paper, Wavelet based testing algorithms is proposed for evaluating an ADC's performance in term of instantaneous ENOB. With the focus on shortening testing process, reducing sample size, simplifying testing complexity and improving ENOB estimation for higher ADCs resolutions, Wavelet computation efficient was very suitable for this application in contrast to the mean value measured by the conventional methods [2,4,10].…”