Wavelet decomposition and reconstruction-based INL method for testing high-precision ADC

Abstract

In traditional analog-to-digital converter (ADC) linearity testing, the cost of testing escalates in tandem with the increase in ADC resolution. This research proposes a time-efficient method for testing the integral nonlinearity (INL). Drawing on the wavelet decomposition and reconstruction technique in conjunction with sine wave histogram, our approach facilitates swift and accurate testing of high-precision ADC linearity compared to conventional methods. Our approach acquiring the initial INL result by sampling the sinusoidal histogram using a modest number of samples. Subsequently, we employ wavelet decomposition to break down the initial result into high-frequency and low-frequency portions. Setting thresholds for these components and reconstruct them to derive an accurate INL result. Experimental findings demonstrate that our proposed method effectively enhances the accuracy of high-precision ADC linearity testing. Furthermore, it reduces the linearity testing time by over tenfold and diminishes estimation errors by more than 30%. Consequently, our method establishes a more dependable means of testing for ADC design and manufacturing.