pulse transformers?
 

As with many of you, I'm still scratching my head on things like high side drivers and associated expenses and complexities.

So the phrase "pulse transformers" keeps coming up, and I sort-of get it, and want to do some experiments on said same. Looking for some pointers.

I treated myself to one of those fancy shmancy DSO-2250 usb oscilloscopes so I can get a reasonably accurate picture and ensure the waveform edges are nice and vertical. And I figure I can do some experimenting with ferrites and wire and an igbt (starting with a 100 amp 6 pack) on the desk and see if the gate is charging/discharging quickly before hooking it up to power.

I should also be able to determine what the lowest frequency it can sustain 100% duty cycle is and plan on doing a refresh pulse at that interval (minus some safety factor).

I figure 12v input to the pulse transformer, what turns ratio for +16 and -8 gate voltages? 2:1?

We want the pulse transformer power switching right next to the gate I assume, minimal stray capacitance/inductance.

Has anyone been down this path that can offer some starting points? Or caveats? I sort of assume the idea is to make the transformer hold at least enough charge to saturate the gate plus whatever it needs to charge it quickly without "destroying" it.

Thanks in advance.

You won't hurt the gate if you give it the appropriate voltage, the transformer will set that according to the turns ratio. You just need to make sure your transformer can tolerate the Volts x Seconds product. I'm used to MOSFETs, but an IGBT gate should be similar, use a 1:1 transformer and drive it with 12V, the gate can tolerate more than 20V, probably 30V.

Multiply the 12V by your longest pulse width and ensure the transformer is designed to that much plus at least 20%-30% margin.

Ah, googling for pulse transformer and mosfet is much more productive than igbt.

hmm, I might be missing something obvious about duty cycle. I mean the gate simply can't hold its charge for long while hooked up to the secondary, without some extra circuitry.

I gotta rethink this :) Might actually be easier to measure opto delays and re-purpose some phone transformers and dial in the appropriate voltages.