SSTC FETs dying for no reason (was: MOSFET failure modes) / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

SSTC FETs dying for no reason (was: MOSFET failure modes)

previous 1 2 3 next

Move Thread

LAN_403

Dr. Dark Current

Sun Apr 11 2010, 12:26PM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

So... I've put the oscillator into a grounded metallic box (note that it's a fixed oscillator, just with a dynamic detuning circuit which modulates the frequency a little based on the mains waveform), removed the fast gate turn-off diodes, replaced the FETs with 20amp 600volt ones and put a 500watt lightbulb in series.
I turn it on, find the resonance and tune a bit around it. After a few seconds, the FETs are DEAD AGAIN

WTF??!?!! I'm completely lost now... Probably gonna blow some IGBTs too.

ScotchTapeLord

Sun Apr 11 2010, 12:51PM

Registered Member #1875 Joined: Sun Dec 21 2008, 06:36PM
Location:
Posts: 635

You may want to switch to a half-bridge to cut your losses *and to see how putting capacitors across the bridge changes things*...
Perhaps you should post a schematic? The DC block capacitors I always hear about and used in my SSTC are in series with the primary, whereas snubbers go across the entire legs of the bridge.

Your designs sound almost exactly like Richie's, but without the 470nF capacitors...

It sounded like your FETs were dying from overvoltage, and if it ate up an IGBT, then it is almost definitely overvoltage.

Dr. Dark Current

Sun Apr 11 2010, 01:29PM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

Okay, I'll try to give as much info as I can.
This is the schematic I'm using right now: (not shown in the schematic the 500W series lightbulb)

Picture of the bridge: (you can also see the CT with ultrafast rectifiers)

Gate driver:

The other transformer is for the dynamic detuning circuit, just a simple VCO modulation. The driver circuit is grounded to mains Earth, secondary is grounded to central heating radiator.

The GDT is 5:5:5 turns wound on an EI ferrite core, with stereo screened audio cable. The screenings are in parallel as the primary and inner conductors as the secondaries.

Gate waveform at 320khz, only gate driver running: (5V/div, 500ns/div)

The very last idea I have is that the capacitance of the GDT might be causing some problems, so I'll try to make a new one with the classic trifilar PVC insulated wires, even though it wont have such good parameters.

Thanks for any help...

ScotchTapeLord

Sun Apr 11 2010, 02:28PM

Registered Member #1875 Joined: Sun Dec 21 2008, 06:36PM
Location:
Posts: 635

Weird... If you have IGBTs to spare, you might want to try them, so as to remove some clutter and possible sources of error. Those are a lot of diodes to check after every chorus-line in the Silicon Searing Symphony. I see you have paralleled your mains line diodes. I have never had a successful attempt at paralleling diodes directly like that. Any one of the diodes in your power circuit going bad could lead to a chain reaction of semiconductor destruction. It's a cruel cycle.
Some swear against using non-toroidal gdts, but you your waveforms are better than any I've ever scoped. It's probably worth replacing, just to see.
One time I noticed that my IRF740s got much hotter when I was using 10 ohm resistors on the gate than when I used none, but the datasheet for them seems to suggest 9.1 ohms on the gate, so that's probably not it.
\
The pseudoswitch capacitors should definitely be sufficient for snubbing with the given headroom.

I've never used PLL so I can't really comment on it, but if you think that the driver could be the problem, then it can either be some fault (possibly within the breadboard?) or interference. You can try running just a primary without a secondary to see if it's interference, which I doubt it would be, but we're running out of ideas here... lastly, you could try to go with a simpler design, maybe just plain old clamped antenna feedback or CT on the secondary going right to the gate drivers.

At this point I feel like I'm shooting in the dark. I'm surprised it died with the lightbulb. I don't see a variac... Those are very helpful for this. I suggest finding a way to get a lower voltage, removing the secondary, and then scoping everything!

You're close to solving this and learning a valuable lesson about something, as well as enlightening the forum on what could possibly be a serious issue to consider with SSTCs, so keep it up!

brtaman

Sun Apr 11 2010, 02:31PM

Registered Member #2161 Joined: Fri Jun 05 2009, 03:36PM
Location:
Posts: 247

Based on your picture, your bridge has a LARGE amount of inductance. Place the mosfets MUCH closer together and take care of all those leads. That is what is causing your deaths.

All I ever used on my sstc bridges were the MUR1560 diodes and I never had a problem. However, I took great care in minimizing inductance.

Basically place the mosfets closer together and make all leads as short as possible any you should be good. As your bridge stands now, I am not too surprised that you are getting mosfet deaths.

Adding snubbers and protective diodes can only do so much, a good bridge design is the way you want to go. Right now all those diodes and snubbers are like a band-aid on a 50 cal gunshot wound.

Dr. Dark Current

Sun Apr 11 2010, 06:49PM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

Wound new GDT, so far no popFETcorn !!! (knock on wood)

Arcstarter

Sun Apr 11 2010, 07:29PM

Registered Member #1225 Joined: Sat Jan 12 2008, 01:24AM
Location: Beaumont, Texas, USA
Posts: 2253

Dr. Kilovolt wrote ...

Wound new GDT, so far no popFETcorn !!! (knock on wood)

That's good, but it might still be a very good idea to move some things closer together. Boy, don't you hate inductance when you don't need it?

ScotchTapeLord

Sun Apr 11 2010, 07:40PM

Registered Member #1875 Joined: Sun Dec 21 2008, 06:36PM
Location:
Posts: 635

Great! What do your new gate waveforms look like?

Dr. Dark Current

Sun Apr 11 2010, 08:00PM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

ScotchTapeLord: They are almost the same, but I know the leakage inductance is slightly worse because when I add the fast turn-off diodes, I get more overshoot than previously. But I'll use just the resistors without the diodes.

Luca

Sun Apr 11 2010, 10:00PM

Registered Member #2481 Joined: Mon Nov 23 2009, 03:07PM
Location: ITALY
Posts: 134

Dr. Kilovolt wrote ...

This is the schematic I'm using right now: (not shown in the schematic the 500W series lightbulb)

Hi Dr. kilovolt

In my opinion your mosfet failure is due to voltage spikes between S-D. Maybe adding 10uF between the DC bus (in parallel to the DC blocking caps, very close to the mosfets) will help to reduce such spikes... I would add TVS between S-D of each mos as well.

Regards,

Luca

previous 1 2 3 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob