Defect Level Estimation of Random and Pseudorandom Testing

In this work, sequential statistical analysis has been applied to determine the defect level of random and pseudorandom testing. Results derived using worst case analysis show that the defect level of pseudorandom testing is always no larger than the defect level of random testing. We also find that the defect level of random testing is a good approximation to that of pseudorandom testing, only if either the yield or circuit detectability is high.