Standard operating procedure for Highly Accelerated Life Testing (HALT) : design and standardization of fixture setup for circuit boards

Abstract

This thesis deals with the introduction of Highly Accelerated Life Testing (HALT) in a multinational corporation as part of a reliability improvement process for printed circuit boards (PCBs). More specifically, the thesis focuses on drafting a standard operating procedure (SOP) to standardize the fixture setup process so as to minimize variation in vibrational responses attributed to fixture setup. This study proposes a three phase approach to meet the objective. The first phase aims to minimize variation in the gauges and the vibration table. A gauge repeatability and reproducibility (gauge R&R) test is conducted on two triaxial accelerometers to determine the most suitable gauge for HALT. Next, a spatial uniformity test is conducted on the vibration table using design of experiments (DOE) method to determine the region with the most consistent vibration levels for testing. The second phase aims to reduce variation in fixture usage by introducing a generic fixture design base on fixture design principles. The last phase involves proposing a short-run statistical process control (SPC) method to detect setup variability. The results from the gauge R&R test show that mechanically mounted accelerometers have the least gauge variance. Using an SOP, switching from wax-mounting to threaded stud-mounting reduces gauge variance by approximately 90%. However, wax-mounted accelerometers are still relevant for measuring vibrational responses from PCBs due to its non-invasive properties. The drawback will be that the gauge becomes a key contributor to measurement variation. The spatial uniformity test reveals that the center of the vibration table has the least variation and is selected for the fixture setup. Further limitations on the HALT chamber are highlighted. Evaluation on the proposed generic fixture demonstrates that the fixture has met the key design criteria and manage to induce consistent failures to the PCBs. Limitations on the fixture design are also considered. The proposed SPC method is able to detect at least 88% of the assignable sources of variation in a simulation test. Finally, a custom SOP for the fixture setup is drafted for the company.