Using logical decision in DFT method selection

Describes how logical decision analysis can be used to find the most cost-effective design-for-testing (DFT) method among numerous choices. These days, due to the chip complexity following Moore's Law, testing is getting more difficult and more important. As a result, testing-related cost increases dramatically. To solve this problem, the concept of DFT was suggested, and there are several DFT methods. But, these DFT methods require some additional circuits and a little more complex register than an original register. This area overhead causes more production cost and fewer yields. Other problems occurred while implementing DFT, are performance degradation, EDA tool cost, increase of pin number, etc. Thus, it is very hard to decide which DFT method is most suitable. Optimization technique can be used to find the cost-effective test method if we have all cost models. But, defining cost models is very difficult. Some parameters of the cost models are not available yet because we need DFT selection sometimes before even a real design starts. Thus, we suggest using logical decision analysis to find cost-effective DFT methods. In this method, developing cost models is not needed.