Options
2025
Journal Article
Title
Higher Order Thermal Impedance Extraction of GaN Power HEMTs by I – V Measurements
Abstract
This work presents a method for extracting higher order thermal models of GaN power high-electron mobility transistors (HEMTs) from I–V measurements using a typical commercial power analyzer. The approach involves deriving and fitting an electrothermal model func tion to measure data using a nonlinear least
quares solver, yielding the thermal parameters for a higher order thermal model. Measurements are conducted using a parameter analyzer and a controlled thermal chuck, with a transient drain current response signal. The method is employed to derive the thermal impedance parameters of a 5th-order thermal Foster model for a GaN power transistor. The Foster model parameters are presented in both time and frequency domains and are subsequently transformed into
Cauer model parameters. The results demonstrate strong agreement with data obtained from an on-chip temperature sensor, confirming the method’s validity. This new extrac tion method can be executed using standard laboratory equipment typically available for the electrical characteri zation of GaN power transistors.
quares solver, yielding the thermal parameters for a higher order thermal model. Measurements are conducted using a parameter analyzer and a controlled thermal chuck, with a transient drain current response signal. The method is employed to derive the thermal impedance parameters of a 5th-order thermal Foster model for a GaN power transistor. The Foster model parameters are presented in both time and frequency domains and are subsequently transformed into
Cauer model parameters. The results demonstrate strong agreement with data obtained from an on-chip temperature sensor, confirming the method’s validity. This new extrac tion method can be executed using standard laboratory equipment typically available for the electrical characteri zation of GaN power transistors.
Author(s)
Open Access
Rights
CC BY 4.0: Creative Commons Attribution
Language
English