On the comparison of ΔIDDQ and IDDQ testing

The purpose of this paper is to compare differential Iddq to Iddq itself; in their capabilities of detecting active current defects/faults, as bridging faults. To perform this comparison, a mathematical framework is developed, based on assumptions validated with data from the Sematech experiment S121. The comparison results reveal that the probability of false test decision can be reduced by orders of magnitude by using differential Iddq instead of Iddq. Based on the same framework, we also show that a wise vector selection may contribute to decrease by orders of magnitude the probability of a false test decision. Finally, we investigate the applicability of differential Iddq with deep-submicron technologies, by estimating the number of (pairs of) vectors required to achieve a given test qualify level. The proposed framework and the results help to understand the tendencies and to identify the requirements in order to meet the challenges of deep-submicron current testing