ON THE QUALITY OF ACCUMULATOR-BASED COMPACTION OF TEST RESPONSES

The accumulator-based compaction (APC) technique uses an accumulator t o generate a composite fault signature for a circuit under test. The e rror coverage for this method has been previously analyzed using Marko v chains. We describe an alternative technique for calculating the err or coverage of ABC using the asymmetric error model. This technique re lies on the central limit theorem of statistics and can he applied to other count-based compaction schemes. Our analysis shows that ABC prov ides very high coverage or asymmetric errors. Experiments on the actua l fault coverage for the ISCAS 85 benchmark circuits show that extreme ly high postcompaction fault coverage (close to 100%) is obtained with ABC. They also indicate that the use of a rotate-carry adder does not always improve the fault coverage;in some cases, the fault coverage i s actually reduced.