Efficient fault ordering for automatic test pattern generation for sequential circuits

Abstract

lhis paper analyzes fault dependency in sequential circuits to accelerate parallel automatic test pattern generation (ATPG). We present the new algorithms improved fault collapsing and fault m g i n g j b r an @cient f d t ordering to speedup ATPG. Experimental results obtained for sequential and fault parallel ATPG show the ejfciency of the proposed methods.

An effective fault ordering heuristic for SAT-based dynamic test compaction techniques

“…Signal probabilities or testability measures like SCOAP [25] are often used to judge the hardness to justify or propagate a certain fault [21]. However, these measurements are based on a static analysis of the structure and do not account for conflicts or reconvergences.…”

“…This paper proposes a new fault ordering technique based on the identification of hard-to-detect faults. Similar to the technique proposed in [21], hard-to-detect faults are ordered at the beginning of the fault list. In contrast, the hard-to-detect faults are not derived using static testability measurements, but extracted from a preprocessing phase.…”

“…Several techniques have been proposed to arrange faults in a way that the pattern count can be improved. The fault ordering technique proposed in [21] orders hard-to-detect faults at the beginning of the fault list. However, this technique uses static testability measurements in order to identify hard-to-detect faults.…”

The Accidental Detection Index as a Fault Ordering Heuristic for Full-Scan Circuits

“…Fault ordering was considered earlier in [8]- [10]. In [8] and [9], fault ordering is done to speed up fault simulation of synchronous sequential circuits by grouping together faults that cause similar events.…”

“…Here, we are not concerned with the events that a fault causes during fault simulation. In [10], fault ordering is done to support the generation of compact test sequences for synchronous sequential circuits by targeting hard-to-detect faults first. The accidental detection index defined here addresses more directly the issue of identifying faults whose tests will detect large numbers of other faults.…”