Freewheeling diodes are burning (Now other problems)

[davekni]

Shorted-turn is another way to describe induction heating.  I hadn't heard of making a top-load that wasn't a shorted turn, but that should be possible by having an insulating connection between the start and end of the duct.  Do either that or add space between the top of the secondary and top load.

Have you double-checked conductivity just to rule out arcing?

[hip424]

Hi,
When I changed MOSFETs to IXFH20N50P3's two of them blew up at about 60VDC. I think it's because that I removed Zener diodes from gates. Now I will change MOSFETs to AOK60B65H2AL IGBTs and add 18V Zener to it. I want to know what you think about it and hear some other possible issues. I just want to finish this project (BTW it's for school teacher). I will leave it at Static medium BPS and run on 330VDC.

Best Regards
Daniel

P.S.
To avoid double posting, I'm editing my post. I've replaced my MOSFETs with IGBTs and I' curious if there is possibility to drive gates of MOSFETs directly from the driver (without GDT). I'v seen datasheet for UCC3732x and there is possibility to drive gates directly from the ICs. I know that the chip hasn't galvanic isolation between output and input but it has separate ground and I've ran ground traces correctly. I know that this chip has only one output so I can't drive fullbridge with it but maybe I will do the halfbridge. Please answer my question.

Answering to davekin question: conductivity is OK. If I will fix other problems then maybe I will take a look at this topload again. For now this is secondary case.

Best Regards
Daniel

[davekni]

Read the UCC3732x specification, especially the pin functions table.  That will answer your question.

[ritaismyconscience]

Don't skip the GDT.

You run into a lot of other problems (like voltage spikes and noise getting into your driver circuit and blowing things up since they aren't isolated anymore). Also, the upper IGBT might be at +330V and the lower one at 0V, and your UCC chip will blow up due to the voltage difference. There are specialized bootstrap driver ICs for this purpose but skipping the GDT is not worth it in the end.

I'm wondering if the MOSFETS survive because high avalanche energy, while IGBTs fail because they aren't very good at taking voltage spikes. If this is correct, adding a snubber cap or adding TVS diodes would fix this problem (try getting the 440V diodes, those should work)

[hip424]

Hi,
Thanks for answers. My Tesla Coil runs just fine with AOK60B65H2AL IGBTs at static medium BPS. Ritaismyconscience do you meant replacing freewheeling diodes by bipolar TVS? Maybe I know the reason why my MOSFETs were blowing at higher BPS: I current limited my power supply to about 450mA (driver board) and when testing it at low bus voltage I've noticed that power supply switches between CC/CV modes so it needed more current in pulse. But anyway I'm still afraid to run it at higher BPS so I will probably leave it just like it is. Also I have question: Can I safely run it on rectified 240VAC with softstart firstly turning on driver board? I mean I ran it with variac but I don't know if I stick it to mains the voltage rise would be too fast. So please help me with that.

Best Regards
Daniel


Update:
When i ran my coil at medium BPS and 330VDC the arcs got very small and I found the issue. I shielded feedback wire an connected shielding to mains earth. When I disconnected ground from shielding it was running OK but when I ran it one more time the arcs were as small as before and I found that the bottom-left IGBT got shorted but not completly (3.3ohms between Collector-Emitter) and I don't know what is causing this. In addition I have wire-wound gate resistors so maybe I should replace them with metalized ones.

Best Regards
Daniel