Power MOSFET

What

What "What"?

It might help if you have a schematic and a little more info on what you're doing. ; )

Yes, 65,000 volts, 7 mA. Unfortunately, schematic is not available. So, we can only guess from the input/output spec.

From 110V AC (around 6A) to 170V DC (around 4A) to 65,000V DC. The G40 switching the primary of a transformer would need close to 4A. A G50 would be better with 6A. Perhaps they did not have the G50 when this equipment was built.

It's likely that more than a MOSFET failed.

Thanks. Hopefully someone with more experience can chime in, but I'm wondering is this a "mod" that's even possible? 170V to 65kV is a lot and I would guess there's more to worry about than just the ratings of the switching devices. Is the rest of the board even equipped to handle it? Transformer insulation breakdown voltages, creepage distances on the secondary side? I'm a bit out of my depth here but these are things I'd be thinking about...

Oh, I'm totally stupid, maybe? So this is a 2 stage thing. I thought you were trying to mod 110AC to 65kV from something originally designed for just AC-low voltage DC... ; )

Forgive me, no coffee this morning.

What else are more likely? All the passives (R & C) seems fine. Perhaps replacing all MOSFET with the 350W STW13NK100Z.

IRFPG40 1000V 4.3A 3.5rds 150W IRFPE50 800V 7.8A 1.2rds 190W IRFP450 500V 14A 0.4rds 190W

IRFPF50 900V 6.7A 1.6rds 190W IRFPG50 1000V 6.1A 5.0rds 190W STW13NK100Z 1000V 13A 0.7rds 350W SCT2450KEC 1200V 10A 0.6rds 85W STW9N150 1500V 8A 2.5rds 320W

As much of a core as it may be, you may save time and trouble in the end by reverse-engineering yourself a schematic. That'll give you some understanding of what you have and what should be there.

Well... Drivers Controllers Aux Supply Feedback etc.? What kind of design? Is it discrete, integrated or mixed? Is it self-oscillating, resonant, class D, quasi-resonant...?

Many stories can be told from a quick inspection of the PCB, even if it's not the complete story.

As for "350W" rating, what does that matter? Can the heatsink even maintain a temperature of 25C at 350W, to achieve the transistor's 350W dissipation rating? Is there a heatsink at all?

More importantly: should it /need/ a heatsink (or much of one)? If it's a switching supply, the efficiency should be high, and the answer should be no.

Using an overly large transistor almost always makes things worse, primarily because of excessive gate charge.

Tim

The logic and controller/drivers are external to this board. The logic boa rd has already been replaced; so we can rule it out for the moment. This p ower board is more difficult to find for replacement; so it is more likely to be the problem. The power board contains mostly discrete components: MO SFET, DIODE, RESISTOR, CAPACITOR and INDUCTOR/TRANSFORMER only. There is a

No heat sink. Temperature is probably not a problem, since it is pulsed fo r 2 seconds only. For a ratio of 600:1 of 7mA, the current is over 4.2A wi th 100% efficiency. It may or may not damage the MOSFET. But a 6A G50 wou ld be safer than the 4A G40.

'Seem fine' depends on how they might have failed. At 65,000V, some failures may not be detectable with common low voltage test gear. Insulation can be punched through or conductive arc paths created that won't show up until energized at the working voltage. And that might take out your brand new MOSFETs.

The MOSFETs are mostly switching the low voltage side. However, one of them is definitely stressing the current rating.

IRFPG40 1000V 4.3A 3.5rds 120nC 1600pF 150W (under current) IRFPE50 800V 7.8A 1.2rds 200nC 3100pF 190W IRFP450 500V 14A 0.4rds 150nC 2600pF 190W (under voltage)

I am tempted to replace them all with SiC MOSFET:

C2M0280120D 1200V 10A 0.4rds 20nC 259pF 62W