A Practical Method for Module-Level Capacitance Estimation in Automotive Power Module Using Datasheet-Based Modeling

This paper proposes a practical method for efficiently estimating the module-level capacitance of high-voltage power modules used in automotive applications. Traditionally, acquiring the complete C–V characteristics of a module required expensive high-voltage measurement equipment such as curve tracers. In contrast, this study presents a low-cost alternative that combines low-voltage measurement data with vendor datasheets to estimate high-voltage characteristics. The proposed approach begins by measuring the total capacitance of the module under low-voltage conditions (below several tens of volts) using instruments such as an LCR meter. The measured data is then compared with device-level datasheet values to extract the package-induced parasitic capacitance. This parasitic component, obtained under low-voltage conditions, is added to the high-voltage C–V data from the datasheet to estimate the total capacitance in the high-voltage region. Additionally, for the collector-emitter capacitance, which shows voltage dependence due to the parallel diodes, a nonlinear model based on PN junction theory is applied. Using the capacitances estimated by the proposed method, noise source modeling is performed and compared against models based on direct curve tracer measurements. Verification using a double pulse test shows that the proposed method achieves high accuracy in predicting high-frequency ringing and switching noise. This study demonstrates that accurate module-level capacitance estimation is possible even without high-voltage measurement equipment, relying solely on simple low-voltage measurements and datasheet information. The proposed method is applicable to various practical scenarios such as custom module design, EMI analysis, and noise source modeling.