Evaluation of economics of testing in a manufacturing environment

In the manufacture of complex electro-mechanical devices, testing operations are instituted at many points in the process. The primary purpose of these test operations is to ensure a minimum number of faulty products propagate to the next level of assembly and that the corresponding test yield at that level of assembly is sufficient. The recurring and non-recurring expense of each of these test operations can be high but often the real cost is not. well understood. Similarly, a reduction of yield at an assembly operation is qualitatively understood to be costly but a quantitative assessment of this cost is not commonly performed. This paper proposes a model to assist in performing the cost-benefit trades that must be made when developing or evaluating a manufacturing test flow. Using a ground assumption defining the conditions required to justify a test operation, a general mathematical model is developed. This model seeks to quantify the cost of the test operation, the incremental cost of yield changes and the relationship of the topology of the assembly network in order to predict whether the specific set of conditions warrant a test. The general model is then used to explore some realworld scenarios, resulting in some surprising conclusions. A model template is provided to guide others in evaluating test economics in their environments.