Built-in self-diagnostic by space-time compression of test responses

Abstract

Presents two different methodologies for built-in self-diagnostic of boards and systems by space-time compression of test responses. The first method, soft decision, uses nonbinary multiple error-correcting codes to obtain space-time signatures. These obtained signatures and the corresponding precomputed references are compared and magnitudes of distortions in signatures are analyzed to identify faulty components. The second method, hard decision, makes use of the information indicating whether the corresponding signatures are distorted or not. Both approaches show considerable savings in hardware overheads when compared with a straightforward approach where a separate signature is required for every component. A transition from the soft to hard decision approach results in an increase in the number of signatures required for diagnostic but at the same time it results in a decrease in the complexity of a fault locating algorithm. Results pertaining to VLSI implementations are presented where the hardware overhead is estimated in terms of two-input equivalent gates.<>