Synchronization and resource locking in simultaneous UUT testing

Abstract

The testing of multiple units under test (UUT) is becoming more common as the need to increase production becomes a fact of doing business. Automated test equipment (ATE) needs to simultaneously handle several UUTs concurrently with limited test resources. The software needs to handle the synchronization of testing and locking of single resource items in the ATE to meet this goal. It will also need to ensure no communications collisions occur and to ensure resources are being used in an orderly and logical fashion. There are several synchronization and locking issues that need to be addressed when dealing with parallel testing. Some of these are: (1) Who initializes and shuts down the instrument and when? (2) How long should an instrument be locked for single use? (3) How smart (and complex) should a driver be for parallel use of an instrument? (4) How can threads be used to improve test performance? (5) How does reuse affect the rest of the ATE hardware and software, (i.e. noise between UUTs and communications collisions)? Some current test sequence execution software, National Instruments (NI) TestStand for example, offers some synchronization and locking features, but should they be used? Or should a more precise method be used. Issues such as efficiency and complexity need to be addressed. Some balance is needed when determining the methodology to be used in the in the parallel tests cases. This paper discusses these issues in more depth in the context of missile testing, the use of a common test set and how the issues are being resolved on this project.