Science ouverte et collaborative pour l'élaboration d'un banc automatisé de caractérisation de pertes en commutation par opposition
Nicolas Rouger, Luiz Villa, Matthieu Masson, Pauline Kergus, Joseph Kemdeg, Lorenzo Leijnen, Jean Alinei, Adrien Colomb, Ayoub Farah-Hassan, Arnauld Biganzoli
TL;DR
The paper tackles measuring switching losses via the opposition method, a precise but thermally sensitive approach. It presents an open-science workflow built around a two-day hackathon and ongoing open sessions to create an automated test bench for generating switching signals, configuring instruments, and managing hardware. Key contributions include an open-source Python-based supervisor, a JSON-configurable test-sequence framework, and a hardware workflow centered on Infineon evaluation boards and Si MOSFETs, with initial experimental validation. The work demonstrates the value of open, collaborative development for complex instrumentation and sets the stage for open data on Si, SiC, and GaN devices while addressing calibration and thermal effects in future work.
Abstract
The switching losses of power transistors are generally measured using the so-called double pulse method. Measuring the opposition of two switching cells is a complementary method that is more accurate but indirect. However, implementing this method can be more complex and requires calibration steps and comprehensive control, with the added issue of thermal management. In this context, we proposed to address this topic through open and collaborative science, first in the form of a two-day hackathon, followed by monthly open sessions. More than 20 participants contributed to the two-day hackathon, followed by monthly sessions for those wishing to continue working together. This enabled us to set up an automated bench, in open science, including the generation of switching commands, the configuration and control of measuring instruments, and the hardware part. Here we present and share our work and this open approach.