ZVS driver failure

[rodrigopires]

Hi, i made a zvs driver for my tesla coil project, using the schematic on the attachments. I used two 20N60S5, 600V 20A MOSFET's, and two STTA1206DI, 600V 12A, 0.28ns ultra fast diodes. In the quest of greater power, i conected it to 200V and it worked for a few second until my 10A breaker fired. After this, only one channel works. With 12V input i got about 11.5V on the output in both drains, but when i connect a load (coil) it triggers the short circuit protection on my PS. This only happends on one side, one of the mosfet drain. Is the mosfet burned? Any ideas? Here is a video of it working on 12V and outputting 7KV arcs!
Thanks!

[Simon]

Often mosfets failed closed. The diagram does say at least 4 times the supply voltage

[rodrigopires]

So i need new mosfets? Or i only need to replace the damaged one?

[Simon]

I'd change both. Are you planning to run it again at 200V ? you need to check the other parts, there are 12V zeners to protect the mosfet gates but the 470 ohm resistors probably need increasing if your running at 200V, they may have also blown and that will have blown the mosfets from the gates.

[rodrigopires]

no, i was just testing it to see what it could do! The maximum voltage i used, before burning, was 60V and it made huge arcs, way to much power for a tesla coil, i think running it on about 20V is enough. i might by some new mosfets, IRFP250 or IRFP460. With the current setup, 12V zener and the 470 ohms resistors how high can i go safely, 100V?

[T3sl4co1l]

Gate supply needs to be from a separate input.  If gate voltage falls, both transistors turn off, and one of them carries the supply inductor's current in avalanche.  Poof.

This circuit is 180 degrees completely flipped from most SMPS.  Where usually you want both transistors off during switching, here they must both be on.  As well, the idle "off" state has them both on (so the output voltage difference is zero: just as the output current flow is zero, in a half-bridge inverter with both devices off).  Thus, the default gate voltage must be high.

(By the way, the supply inductor is emulating a current source.  When you short out a current source, its power output is zero.  An inductor is not a CCS at DC, so this obviously would be a bit of a problem.  With a constant voltage supply, you must reduce that to zero instead.  While keeping the gates high.  Then once the energy decays, the gates can be turned off.)

Tim

[bktemp]

The peak drain voltage in this circuit is supply voltage * pi, in your case 200V * 3.1416 = 628V. To add a bit of safety margin, the schematic says 4 times.

[rodrigopires]

since fixing my old zvs is boring, i will build a new one! What is best, IGBTs or IRFP's?

[rickselectricalprojects]

since fixing my old zvs is boring, i will build a new one! What is best, IGBTs or IRFP's?

IGBTs dont work very well for zvs drivers unless you have monstrous input power.
i get 8-10cm arcs out of my zvs and i only run mine of 30v (although the voltage sags to about 27.8v when it draws a peak of 12a ) and i have 5+5 turns on my flyback, i use IRFP250N mosfets, uf4007 diodes, and the tank capacitance is 0.66uf and i forgot which zener diodes i am using.
if you want to run a zvs of 200v then maybe you should build a half bridge or a full bridge flyback driver instead otherwise you will most likely fry your flyback transformer.
also please, please, please use a screw driver to draw arcs instead of picking up the wire. a zvs driver can kill you. just get a screw driver and tape it to a bit of PVC pipe to insulate yourself from the high voltage.

[rickselectricalprojects]

i also forgot to add that i am using a 150uh inductor (i think its 150uh, i built my driver ages ago) but if you want a higher voltage then you use a higher inductance and if you want higher current then use a lower value inductor.

[rodrigopires]

Thanks, i go with the IRFP250 and go from there!