Thermal Mapping of Power Modules Using Optical Fibers during AC Power Cycling Tests

Abstract

The power cycling withstand capability of power semiconductors is of great interest in determining the component's qualification and reliability performance. Compared to the traditional DC Power cycling test, the AC power cycling test can accelerate the component to fail under more realistic operating conditions. The wear process in the AC power cycling test depends on many parameters, among which the maximum junction temperature $T_{J}$, and temperature swing $\Delta T_{j}$ are crucial to be determined. The traditional way in DC power cycling test is to measure the $T_{j}$ indirectly using the temperature-sensitive electrical parameter (TSEP) method, which is not very applicable during the AC power cycling test as it will increase the circuit complicity and may interrupt the PWM operation. The purpose of this paper is to evaluate a direct junction temperature measurement in an IGBT power module through silicone gel using optical fibers, which enables a fast and accurate $T_{j}$ determination during AC power cycling. For this purpose, junction temperatures have been measured for both gel-filled modules and gel-removed modules under different experimental conditions. The experimental results presented concern about: the presence of silicone gel's impact on the $T_{J}$ measurement accuracy, the temperature difference inside the silicone gel while the optical fiber is being instrumented at different positions, and the spatial temperature distributions of the IGBT chip. Future work will also include the comparison of the $T_{J}$ measurements of a gel-filled power module between using the optical fibers and the well-established TSEP method.