Arranging a Pool of Functional Test Sequences for Variable In-Field Test Periods

High workloads applied to a system cause chips to be more susceptible to aging effects that may eventually result in hardware defects. The detection of the defects requires tests for delay faults to be applied in-field. Both scan-based tests and functional test sequences are important to apply. In-field test periods vary in length. Therefore, test sets of both types should be arranged such that every test period would be utilized for targeting the most likely to occur faults. This is preferred over the alternative where each test period is used for achieving the highest possible fault coverage since the highest possible fault coverage may be achieved without detecting the most likely to occur faults. This article considers the problem of arranging a pool of functional test sequences to match different in-field test periods when the goal is to ensure that the most likely to occur faults are detected in every test period. The procedure described in this article produces a series of solutions with subsets of increasing lengths of the pool (subpools) to detect subsets of transition faults of increasing sizes. The increase in the length of every subpool in the series is minimum or close-to-minimum relative to the length of the previous subpool. The procedure is implemented in an academic simulation environment and applied to benchmark circuits to demonstrate its effectiveness.