This week I was in our Nottingham office to help deliver some internal training. One of the subjects of the training course was setting up transient analysis in FloTHERM. The example model we used was a detailed model of a TO220 on a PCB in free air. We set the heat source in the die to pulse on for 50ms, off for 50ms and repeat for 1s total. The idea was to show how a transient model can be set up, things to look out for, solver settings pertinent to transient analysis in free convection and how to post process those results afterward. With a mini competition for the best looking animation that the trainees could come up with. Note: Simple often works better, trying to be too clever and do too much normally means that your processor and/or graphics card just can’t keep up!
Anyway whilst showing the post processing we noticed a cool bit of physics going on. Now the TO220 model we had has 7 leads, but only 1 was directly connected to the die, the other 6 were connected to the encapsulant.
When the die first powered on, the heat traveled through the die to the die flag and out to the top signal layer of the board, it also traveled from the die to the centre lead.
Heat was able to pass along the signal layer and back up into the base of the lead for a short time
until the heat traveling down the lead finally caught up and overrode it.
Now in the centre lead this took a very short time, however in the outer leads, heat was still traveling in the opposite direction up the leads at the end of 1s.
So what does this teach us - mainly that conduction paths may not be as simple as they first appear.
Just to finish off, here’s an animation of the whole analysis. See how the die cools down in the “off” cycle while the heat still spread throughout the rest of the model.