Improving test time on a COTS based system

Abstract

The volume of unit under test (UUT) devices that can be tested on automatic test equipment (ATE) is ultimately determined by the test time for those devices. Conversely, the number of devices to be tested over a period of time establishes how much ATE is required. The amount of ATE required not only drives the acquisition costs but the costs associated with maintaining and supporting the ATE. As a result, no other topic seems to generate more attention in the ATE industry than test time. This focus has driven ATE manufacturers to produce faster test equipment and the test engineering community to identify efficient methods for detecting and isolating faults quickly. Today there is a large suite of specialized commercial-off-the-shelf (COTS) test equipment that is a result of the test time focus. Test system designers integrate a diverse set of specialized equipment to produce a sophisticated test system with the capability to test a large collection of dissimilar UUT devices. One of these test systems built with the freshest and fastest equipment may give one the impression that re-hosting a UUT from a legacy test system to the new test system should decrease the UUT test time from what it was before. The thought process would go as follows. "The new test system is twice as fast as the old. We should be able to test the UUT in half as much time as before. Therefore, the new requirement is old divided by two." This paper addresses this scenario where unrealistic test time requirements were imposed on several UUTs. The paper addresses specifically how the problem was approached, how the problem was resolved, and some of the specific techniques used to decrease test time on a COTS based test system.