Analyzing repair decisions in the site imbalance problem of semiconductor test machines

Abstract

Test machines can test multiple IC devices simultaneously. When testing the same group of devices, unusual deviations in yield rates of specific sites from the other sites (i.e., site imbalance) imply a fault in the corresponding sites and the machine. This study develops a decision analysis framework for maximizing profit and customer satisfaction under uncertain conditions. The proposed framework can provide the on-site operators specific decision rules to help decide whether they should continue the test, close specific sites, or shut the machine down to repair it. A numerical example is used for illustration.