Chapter 13 - Conclusions

Abstract

An effective test strategy depends as much on declaring and understanding test goals as on implementing them. Every company, every department, every product-planning activity should include targets for testability, test coverage, and acceptable quality levels in the factory, as well as reliability, testability, and repairability after shipment to customers. Evaluating a strategy requires determining the effectiveness of each step or tactic. One common approach assesses a particular step by examining the board (or system) next-step yield. This method assumes that there are no new failures from product handling and transfer between test stations and, therefore, that escapes from the preceding test step are the only failures. If there is no subsequent step, customers perform this analysis. Another technique is fault injection. Primarily useful in high-throughput facilities, fault injection “seeds” the production process with boards containing known problems and then determines how many of those problems the test strategy can catch. For many operations, fault simulation represents the best means to determine test-strategy effectiveness. Its success depends on the efficiency and accuracy of both fault simulators and the design models on which they operate.