SMPS failure, the Gates of the MOSFETs exploded??????

I've just build an SMPS using a TL494 chip using the power circuit from this circuit from uzzors2k, it's just a standard half bridge in my case using IRFP460's. The MOSFETs are driven through a small transformer that itself is driven with a H-bridge. 

Yesterday I was trying it's first mains power test slowly ramping up the voltage with a variac when at about 180VAC i could hear some fizzling from the circuit and at 200V the GATE's of the MOSFET's exploded violently, at close inspection there was no shorts or flashovers and the gates were the only thing that was damaged at all.

One thing i noticed later was the output waveform of the power transformer, unlike the signal waveform it didnt exhibit the same 96% duty cycle but a 100% with a lot of ringing on the rising peaks of the square wave, i know this can probably be fixed by reducing the duty cycle but i cant see how it could've blew up the gates.

I've included some pictures so you can see what i mean by explode, as you can see on the left Fet the gate pin has completely blown off and the copper has melted slightly and the right image the same happened but i cracked the Fet open with a hammer to see if there were any clues inside.

I really dont have a clue why this happened and would appreciate some help, thanks in advance.

Picture of SMPS failure, the Gates of the MOSFETs exploded??????
sort by: active | newest | oldest
The MadScientist (author) 2 years ago

Thanks guys, I put in those 470 ohm resistors and reduced the duty cycle a tich and got the circuit working just fine now powers the filament of one of my tubes nicely.


I'm not surprised it blew up.

First that is a full bridge rectifier not an H bridge, in fact an H bridge made with transistors or mosfets was a full bridge for 60 years until about 2004 when the term H bridge started to be used. (I did the research.)

The Gates need current limiting resistors, the Drain and the Source can take 17 and 20 amps but the gates can only take a few mA and the gates are rated for 20 volts gate to source, I would say they are getting a lot more than that from your schematic.

They are 250 watt mosfets at 20 amps there max voltage it 12 volts and at 500 volts the max current is 500 mA so at 200 volts the max current is 1.25 amps.

What current were you getting at 200 volts?

You can get the datasheet for the mosfets here:

You can bang several amps into the gate usually, provided Vgs isn't exceeded. Example:

Several devices there can push 4 AMPS into the gate. Designed to turn the gate on and off VERY fast, to minimise dissipation inswitching losses.

I'm more inclined to blame the maximum Vgs being exceeded, like you.

Nice low voltage 2 and 4 amp drivers shame they are 18 volts and his circuit is 200, they look like they would be good in a power inverter.

With a Max VGS = 20 volts and a RGS = 25 Ω I wouldn’t exceed 0.8 amps with those mosfets they might interpret the excess current as excess voltage.

Yes, you have to get creative to drive them at 200V, its not impossible with optoisolation, and a flexible attitude to "Ground".

I wonder if you put a voltage divider circuit for the driver supply if it might work.

Hi Josehf, thanks for your response.

Prior to running this test I ran a low power test at 30V on the power side of the circuit, I had this running perfectly for nearly an hour, so what I dont understand is how turning up the voltage on the power side resulted in the gates of all things blowing up.

I will put 470 ohm resistors on the gates and give it another test today, im not sure what the current was at 200V but im pretty sure it was within its limits if I do find that it's over powered then I have a couple FCA47N60F's or FDH44N50 I can use in their place.

I can see that at 30 volts it would only pull about 8.5 amps well within its tolerances.

I should have been more concise I meant the current on the gate.

470 ohm should be good, keep an eye on the gate to source voltage as you increase the voltage its max is 20 volts if you exceed that you will blow the gates again. Try to keep it between 10 to 15 volts.

Rebuild, limit the input volts, and measure Vgs

The MadScientist (author)  steveastrouk2 years ago

Hi, I thought the problem might've been caused by an over voltage on the gate but I just cant figure out where the voltage could've come from if this is the problem, the driver signal is at 16V which should be easily within limits.

Got a scope ? With systems like these its all too easy to get nasty spiky crap around that will blow your FETS. Try a zener clamp on the gates.

Are you sure the GDT is wired correctly?

The MadScientist (author)  Downunder35m2 years ago

Yep it's wound correctly 1:1:1 with a 16V input.