Multi-chip parallel high-voltage high-power modules are widely used in rail transport

which requires a high long-term reliability of power modules. The power cycle test (PCT) is the most important test to assess the power modules’ reliability

in which the junction temperature is the most important factor affecting the module lifetime. VCE(T) method is commonly used for junction temperature measurement in PCT. As for the power module with multi-chip paralleled

whether the VCE(T) method can still used is an issue

and the physical significance of the junction temperature measured by VCE(T) method of a power module has not been studied. Therefore

the 6 500 V/750 A power module used in high-speed trains is selected for this paper. By calibrating the junction temperature of 16 parallel chips in the module and testing the temperature distribution

it is concluded that the junction temperature measured by the VCE(T) method is equal to the average junction temperature of multiple chips and the junction temperature calculated by the “1/3 method”. The accuracy of the experimental conclusions are furthermore verified from the conclusion that the average junction temperature of multiple chips has the same expression with junction temperature calculated by the “1/3 method”

which is obtained by the analysis from the finite element theory. Therefore

the VCE(T) method is still suitable for the junction temperature measurements in multi-chip parallel power modules

where the measured junction temperature can be physically characterised as the average temperature of the surface of multiple chips.