I have a question. I am using PLECS 5.0.3 to analyze losses in a HERIC inverter that uses MOSFETs with diodes. To do this, I uploaded the thermal model for the IPW60R041P6 MOSFET and placed a Heat Sink block over the full-bridge switches, and another over the Bypass switches. Using a PROBE and a SCOPE, I analyzed the plot for conduction and switching losses across the switches.
The problem is that the average dissipated power shown by the HEAT SINK block is higher than the sum of the losses in the switches themselves. For example, the heat sink block over the full-bridge indicates a power dissipation of approximately 50 W. The conduction losses for a single switch are around 1 W, and its switching losses are around 8x10^-5 W. Since all four switches in the bridge behave similarly, the sum of the individual switch losses is much lower than the value reported by the heat sink.
What could be happening here? How does PLECS calculate these losses?
Hi @ErickGP
First of all, I would suggest having a look at the tutorial “PLECS: Thermal Simulation of a Buck-Converter”, which you can find on our website:
This tutorial introduces the fundamentals of how thermal simulations in PLECS should be performed. Another useful reference is the documentation:
As you correctly state, once steady state is reached, the heat flow out of the heat sinks should correspond to the total losses of the connected components, such as semiconductors and resistors. During the thermal transient, however, these values can differ because energy is still being stored in or released from the thermal network.
Normally, the system heats up from its initial temperature. In this case, the heat flow out of the heat sinks is initially lower than the losses generated by the semiconductors. Since your model shows the opposite behavior, it looks like the system is initially cooling down until it reaches thermal steady state.
Without having your model at hand, I cannot determine the exact reason. One possibility is that degrees Celsius and Kelvin have been mixed. PLECS can work with either unit because only temperature differences are relevant in the underlying equations. However, it is important that the thermal descriptions and all initial temperatures use the same convention.
For example, the thermal descriptions may use degrees Celsius, which is common for semiconductor manufacturer data, while the initial temperature of the heat sink may have been entered in Kelvin. This could explain why the system initially cools down.
Feel free to attach your thermal descriptions and PLECS model if you would like us to investigate this further.