Résumé
Testing of Analog-to-Digital Converters is classically composed of two successive and independent phases: the histogram-based test technique evaluating static specifications and the spectral analysis technique evaluating the dynamic performances. Consequently, the fundamental objective here is to investigate the feasibility of an alternative test flow involving exclusively spectral analysis to replace these two time consuming and expensive phases. The viability of this solution depends on the ability of spectral analysis to detect static specifications. In this context, this paper presents a new methodology based on a statistical approach to quantitatively evaluate the efficiency of detecting static errors from dynamic parameter measurements. This methodology has been implemented in an in-house automatic tool allowing one to process any ADC specifications. It is then possible to choose a priori the best test flow for a given application considering the most adequate trade-off between test time and test efficiency.