How do you calculate the current in pulsed or square wave DC?

I will be specific with this question - it concerns the following schematic. I want to know, at least approximately, if the peak current flowing through C6 and the UCC373 chips. Because, for convenience reasons, I am using a single UCC27425 that has both non-inverting and inverting on the once chip. However these have a maximum current rating of 4.5A rather than 9A.

It has been rather surprising to me how bloody hard I have found it to find any SPECIFICS on this sort of thing, while searching the net.

One detailed worked example with this specific schematic, with some explanations about each step in the calculation, would be GREATLY appreciated and will set me up for all similar circuits in the future. Assuming that the supply voltage is 12V (as per the spec) and the duty cycle of U5 is 90% and the primary coil of the GDT or gate drive transformer has an inductance of 0.12 mH.

My problem is that I am not formally trained in electronics and there are significant gaps in my knowledge. So just a general description of how to do the calculation is not enough.

I am going to take a wild guess, here. You need to specify the frequency you are driving the IC' with because you will have both capacitive reactance with C6 and inductive reactance with the coil. .

Paul

Paul_KD7HB:
I am going to take a wild guess, here. You need to specify the frequency you are driving the IC' with because you will have both capacitive reactance with C6 and inductive reactance with the coil. .

Paul

Around about 432kHz

I'd add TVS's to protect gate-source of each MOSFET, which need to be heatsinked as the circuit has large
switching losses at the frequency. Don't skimp on the drivers, use one UCC27425 for each MOSFET and
it will need very good heatsinking and cooling (The DGN package is the only one that can handle the power, note)

If your current is rectangular - or can be assumed to be rectangular, you can simply measure the average current and divide by the duty cycle. If you have overshoot or other anomalies, the only way is to get a bunch of samples across the pulse.