The defect level is a basic measurement corresponding to the proportion of faulty circuits in the circuits which have passed the test. It decreases as the yield increases. It decreases as the performance of the test sequence increases.
Now, how can we obtain a good measurement of the performance of the test sequence?
The usual measurement is the fault coverage of single stuck-at faults (introduced in Section 3.1.4). Some relations between this measurement and the defect level have been proposed. For random testing, the fault coverage can also be used (expected fault coverage). On the other hand, other measurements can be used in this context: the probability of detecting the worst case fault and the probability of having a 100% fault coverage.
The defect level and related entities are presented in Section 4.1, and the concepts of fault model and fault coverage are given in Section 4.2. The relations proposed in the literature, between fault coverage and defect level, are presented in Section 4.3. In Section 4.4, the measurements of the confidence level for random testing are defined are compared.