I’m trying to determine the number of parallel FETs required for my application and had trouble making sense of the temperature and losses values I was getting.
I noticed an error with the PLECS models for the EPC2305 (probably present for other models): the number of parallel FETs num_par is not correctly implemented in the thermal description. I think it should be dividing the conduction losses and multiplying the switching losses.
I edited the thermal description to include num_par and it seems to be a perfect match with a manually paralleled connection (at least for conduction losses).
Hello,
thank you for bringing this to our attention. We are investigating this in more detail, and for the moment we have disabled the feature in the PLECS you can download from our website.
In the meantime, you should be able to parallel the devices in the PLECS schematic.
We will do our best to put the feature back in…
Hello David, this is Kris from Plexim. Indeed something looks strange here, but it’s not 100% obvious from the images along.
Let’s rule something out first. I am wondering if you are accounting for the paralleling in the RthBOT and RthTOP thermal impedances as well? If not then you are only dealing with the loss calculations when considering the paralleling, but not the thermal circuit. Also, for whatever path (top or bottom) you consider to be for the heatsink, are you including a case-sink aspect to this? There are really two impedances in series and Rth_ch should be paralleled, while Rth_ha should not. Anyway, if you are able to post your model file or email it to me (mylastname @ plexim.com) I’d be happy to further analyze and provide feedback.
