Automating testability analysis of analog circuits and systems

Increasing avionic design complexities and diagnostic requirements have focused attention on the need for performing a testability analysis early in the design cycle. The author describes the general principles underlying the testability analysis of analog circuits and systems performed by the automated testability expert system tool (AutoTEST) which is an artificial intelligence (AI) based CAE tool. The goal is to perform a testability analysis early in the design cycle to ensure conformance with diagnostic requirements. To meet this goal, AutoTEST was designed to perform an analysis with varying levels of component and circuit detail. AutoTEST supports both a top-down and a bottom-up approach. The analog and system level portions of AutoTEST provide test and design engineers with both quantitative measures and qualitative descriptions of the testability of a given design. Quantitative measures are testability figures of merit such as fraction of faults detected, fault isolation resolution, and average ambiguity group size. Qualitative descriptions of testability problems include identifying feedback loops and unique circuit configurations which are difficult to test. The qualitative analysis consists of a set of design rules that are applied to the circuit. At the completion of an AutoTEST session, selected analysis results are written to a set of report files.<>