Help debugging a PWM control

[awakephd]

I was starting to think that I need to go to a higher-voltage MOSFET. I've been looking on Digikey particularly at the R6530 family of MOSFETs from Rohm - 650V, 30A. There are small variations that I don't quite know how to choose between. For example, the R6530KNZ4C13 is described as a "high speed switch" version, while the R6530ENZ4C13 is described as a "low noise power" version. The specs seem nearly identical to me, with slight variations in total gate charge, rise or fall time, etc. I don't know how to judge whether or which would be better for my needs.

I've also wondered whether an IGBT would have any advantages for this purpose. My *extremely limited* understanding, which may well be incorrect, is that "in general" MOSFETs can handle higher switching speeds, while IGBTs can handle higher currents - but of course, when one starts looking at the hundreds or thousands of different options, there are plenty of what seem to be high-speed IGBTs and high-current MOSFETs. One thing that is distinctly different, from what I can tell, is the voltage drop across the IGBT C-E - a couple of volts - vs across the MOSFET D-S - only millivolts. Not sure if that makes any practical difference in my application.

I'm learning so much ... and yet the more I learn, the more I recognize that there is so much more to learn!

[moffy]

I found a nice article comparing MOSFETs to IGBTs: https://weishielectronics.com/igbt-vs-mosfet/
It seems fairly balanced, the IGBTs are meant for high power applications, voltage/current, which are appropriate for their slower switching compared to MOSFETs. In your application MOSFETs are probably the best choice as they will have lower conduction losses than an IGBT.
Regarding the two versions of the MOSFET, the R6530KNZ4C13, with the lower gate charge, will give faster switching, but also excite higher voltage spikes. The R6530ENZ4C13 with the higher gate charge will give slower edges and excite less aggressive voltage spikes but will also dissipate a little more in switching losses. It's all a compromise and just depends where your priorities lie for this design. 

P.S. You can slow down the switching of a MOSFET by reducing the gate drive current, and the simplest way to do that is to increase the value of the series gate drive resistor.

[moffy]

Just as an illustration of how the switching speed of a MOSFET can effect the spikes, I ran two simple LTSpice sims which are identical except the gate resistor in the 'fast' one is 5R and it is 33R in the 'slow' one. The ringing peak in the 'fast' one reaches close to 550V while the 'slow' peak is just 330V.

[awakephd]

Moffy, very helpful indeed. Many thanks! I have continued to work on this off and on, in the bits and moments of time available. I am close to finalizing a circuit board design which, hopefully, will bring together all of the things I have learned thus far. Chances are good that it will be a "first draft" needing further revision, but I am excited to be getting this close!