Hello, I am a new user of plecs, and I have questions about probe’s calculation of switch loss.
What is the principle of plecs probe module to calculate the switching loss of the device? According to the official website of plecs, I have learned that the switch loss is determined according to the voltage, current and junction temperature before and after the switch by looking up the table, etc. May I ask which specific moment the voltage and current here refer to (specifically before or after the switch action)?
Thank you very much for your answer, but I have some questions about the simulation results. The picture shows the voltage, current and switching loss of SiC MOSFET measured by probe. Why is the waveform of switching loss triangular? I read on the website that it was Dirac Pulse.
Hello, actually the switching losses are Dirac pulses, but it depends on how you display the waveform type in the Scope. In the screenshot below, I have the same switching loss signal from your model going into all three channels. As you can see in the Data table at the bottom, you can click on the green signal symbols to the right of the signal names and choose options for Auto, Continuous, Discrete, or Impulse. So, in the first plot I set the signal type to Impulse and you see what you expect for the switching losses, but with the second one, the Continuous type, you see the waveforms are connecting the pulse peak from/to zero in the previous/next time step, and so on.
Ok, thank you for your answer, I understand how to set the pulse waveform, but how to determine the duration of the continuous and discrete options? I have marked the time in the picture below.
Hi Emma, to be clear, only the top plot (impulse signal type) is correct and relevant for the switching losses. But as for your question, the duration/width of the other two waveforms is simply based on the step sizes that are taken by the continuous/variable-step solver at any given time. PLECS is not actually integrating these waveforms though- the scope simply “connects the dots” artificially with those signal type options.
The Dirac impulse frequency is driven by the switching frequency.
If you have a constant switching frequency, one would typically set T_average = N/fs where is an integer greater than or equal to 1. If you want the instantaneous losses N=1 and if you desire the average losses over some interval then set N according to the desired interval.
If you have a variable switching frequency then typically N must be set so the averaging is over a longer period such that it is at least greater than the longest switching period, if not higher.
