This method assesses packaging degradation in power semiconductors, suggesting a simpler monitoring approach.
Condition monitoring of power semiconductors is a useful tool for predicting their packaging health status under operation in power converters. To assess degradation in the packaging, it is imperative to measure or estimate the junction temperature of the dies. However, direct measurement of the junction temperature is rather impossible due to the highly integrated encapsulation structures. On the other hand, junction temperature estimation imposes the need for additional sensing circuits, which increase design and manufacturing complexity. To overcome these challenges, this paper proposes a low-invasive condition monitoring method for liquid-cooled power semiconductors, which does not require measurement or estimation of the junction temperature. The proposed method only requires thermal characterization of the liquid-cooled power semiconductor only once for the construction of the thermal network. This is done for the healthy stage of the power semiconductor. Packaging degradation is detected by measuring the actual increased power in the cooling liquid, with appropriate compensation for the share of heat dissipation to the ambient air and comparing these two values to the expected power dissipation at the given loading condition for the healthy stage using the thermal network. The proposed condition monitoring method has been experimentally validated in a down-scale laboratory setup, using both healthy and degraded Silicon Carbide metal oxide semiconductor field-effect transistors and under varying cooling and ambient conditions. • A low-invasive condition monitoring (CM) method for power semiconductors is proposed. • The need for junction temperature measurement or estimation is eliminated. • CM is performed by assessing the amount of heat dissipated to the cooling liquid. • Thermal characterization of the semiconductor is only done for the healthy stage. • This CM method detects solder layer, bond-wire and threshold voltage degradation.
No takes yet. Share an insight, caveat, or question.
Baumann et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: