Efficient Determination of Fault Criticality for Manufacturing Test Set Optimization

Defective part levels of zero are almost impossible to achieve in an era of complex defects, process variations, and limited testing resources. It is important to ensure that any defects missed during test will impact the end user as little as possible. However, optimizing test sets for superior detection of critical defects requires an understanding of relative fault criticality, and this determination may be very expensive. In this paper, we will present a method which efficiently estimates this criticality under realistic usage conditions using a translation between combinational and sequential fault analysis. We will show that the resulting criticalities are sufficiently accurate to select an optimized test set which is as effective as one created with perfect criticality information.