Author Options:

70-100V powered ZVS? Answered

So I was playing around with my ZVS driver and I turned up the variac to 40V, getting roughly 100kV and boom the flyback caught on fire and my ZVS driver fried. So I want to build a new one that could handle around 70-100V input. I know that for higher inputs some component values need to be changed but for voltages that high, I do not know exactly what would need to be changed. I was thinking of using two STY34NB50 MOTFET's, but wasn't sure if they were the right kind.


I get my zvs to drive straight from the wall voltage. I use FQA19N60 mosfets. you can see proof on https://www.youtube.com/watch?v=s_HP_is5m_E. has uplink to ltspice schematic there also

You might try an IGBT ZVS. Get some nice IGBTs (600V 48A and up) and replace the MOSFETs with those. Then take a 30-35V supply & connect + to the 470Ω resistors. Then take + of your main supply (70-100V in your case) and connect that to the inductor. Hook the - of both together & to ground. Tinker with that & please let me know how it goes.

Thanks to lokeycmos & juniortore1 for the idea.

Also, use 17V zeners.

I have two transformer that's 25 Volt each. If I'm going to combine both of those. like that (+ TR - +TR-) (TR=Transformer) Would diode recommended on each + for 470 Ohm and Inductor?


5 years ago

Zero Voltage Switching a flyback gives an idea of what your doing but
it is not possible to decide if a 500V un-heat-sunk 20A MOSFET will
work for you without a schematic of the whole circuit of your
Frankenstein energizer :-)
You could be using Plasmana's driver where the STY34NB50 will work.

You also have a EE's appreciation of power MOTs, I like that.

Here is a thought about voltage insulation.  Your system fire started when
the voltage exceeded the insulation break-down and created a
carbon track followed by lots of current doing the fire damage.

Your in the range of high voltage where there is no insulation even glass that can work for you,  as static charge will build up along any
insulator and travel towards the exposed edge with similar results.

Look at the air separation in Tesla primary coils and other mega KV
generators as the guide.. 
The poor flyback cannot handle such potentials without making its
own circuit crossovers wherever insulating material must fail..

What I'm saying is that you have reached the limit of your FB,
not the driver and must look into a different primary / secondary
insulating separation...  Good luck ;-/   Take lots of videos for us to see.


Never did I say that it was un-cooled, I did in fact have them on heat sinks and they worked quite well. The flyback only caught fire from arcing between uninsulated pins, I assume the transformer still works fine, I suppose that part was unnecessary to tell. I'm just looking for something that can handle more power than the standard ZVS driver.

Likewise never implied that your un-cool, I was simply describing
your chosen FETs which can actually handle over 34 amps but
probably could do 20 amps when hot as 125°

And my read on your question's main point, was how to increase voltage
output, possibly without the fire your avatar is clearly avoiding :)

You said "un-heat-sunk" I assumed you though I wasn't using one:P read that wrong. Perhaps just using better FET's and keeping everything else the same I could power it higher.

 If you are using Plasmana's final driver,  it should increase the frequency
because as Steve points out the core will saturate sooner..
( you don't have a good scope do you ? )
The Plasmana design FET  and your FET compare well
  • MOSFET       AMPS   VOLTS   ON_time    OFF_time
  • IRFP250         33 A    200 V      210 ns       220 ns
  • STY34NB50   34 A    500 V     109 ns        152 ns

If the FETs get hot then I would suspect the switching frequency
of your ZVS  has reached the upper limit  or serious RMS  ringing.. 
But the higher voltage output may be destroying the insulation.. 
A simple case of  WWFF     Which Will Fail First ? .

Amazing all the stuff that happens at a simple turning of your variac :-)
Take care not to over whelm your gate protection zener diodes.


I actually decided to buy some IXGR40N60CD1's (40N60) for their higher ratings and lower cost. They have an on_time of 25ns and off_ime of 100n. I do not have a scope so I have no way of measuing frequency and what not, perhaps I should look into buying. Any suggestions? (I'd prefer a built in function generator). Usually when building resonate electronics I just go by the schematic and hope it works rather cause I simply have no way of testing frequency.

Brave man you ;-)

I have scopes that use that little known CRT (Cathode Ray Tube )
and I can't afford a really fast nanosec scope so most time I use
a generic though I did buy a PC USB dual channel for under $200
with great features.

I would be content with a CRT scope. I work at a tv repair shop actually and we use CRT scopes for testing the frequency of transformer drivers in CRT TV's lol I like the Tektronix 475 Oscilloscope.

You're saturating the core. There's no need for these things to even run lukewarm.