ZVS heating
..., Tue May 23 2006, 01:32AMI have been messing around with my HPMBRBDCPS and things seem to be going well, but for the fact that the primary is smoking
I am using the Mazzilli driver with a pair of hgtg40n60's running off <80v (above that the primary starts smoking instantly) feeding into a variety of primaries I have tried. The tank cap is ~1ufd, and I am running ~100khz.
The first primary I was using was a piece of 10awg copper wire, which got just barley smoking hot at 80v after a little while, but it was usable for short periods... That was a 4+4 turn primary. But eventually it melted through so I decided to make a new one.
So I took ~25 strands of 30awg wirewrap wire (Ag plated Cu, kynar insulated) and wrapped them together and then into a 4+4 turn primary and hooked it up. This started to smoke immediately with around 70v in, so I decided to try a 5+5 turn, but that was even worse
So how do I keep my primaries from melting? Everything else in the circuit is warm, and I really hate to have to run lower power because my primary is melting... My final goal is to get around 750-1000w out running directly off rectified mains (160v)
Re: ZVS heating
Desmogod, Tue May 23 2006, 01:52AM
Are you using a seperate power supply for your gates?
Maybe it's not turning the IGBT's on and off correctly???
Desmogod, Tue May 23 2006, 01:52AM
Are you using a seperate power supply for your gates?
Maybe it's not turning the IGBT's on and off correctly???
Re: ZVS heating
..., Tue May 23 2006, 03:18AM
Yes separate power supply for the gates.
The waveform does get a little screwy above ~40v, looking at the scope the fets synchronize so they are both turning on/off at the same time (the traces look exactly identical save for a a little ringing around 5v) but that does not change at the insane power levels... And as soon as you draw some power it goes back to normal...
I did have some success...
I tried replacing the litz primary with a single strand of 14awg solid copper in a 7+7 turn figuring that I wanted sparks... But it actually runs surprisingly cool. I was able to run at 110v in (about 800w me thinks) and draw sparks for a while (well until one of my 40n60s decided to completely short out) and the primary wasn't even all that warm
Which seems really confusing to me, I had probably 2x the metal with my litz primary, and much more of it usable to to the skin effect at 100khz... And at that point most of it should have been in the silver coating... Seems very odd to me. I didn't get a chance to scope the primary current in the litz primary since it was to short, so I can't say how much lower (if lower at all) is was.
[me looks longingly at the pile of wire and then at the spanish homework I have to do
]
..., Tue May 23 2006, 03:18AM
Yes separate power supply for the gates.
The waveform does get a little screwy above ~40v, looking at the scope the fets synchronize so they are both turning on/off at the same time (the traces look exactly identical save for a a little ringing around 5v) but that does not change at the insane power levels... And as soon as you draw some power it goes back to normal...
I did have some success...
I tried replacing the litz primary with a single strand of 14awg solid copper in a 7+7 turn figuring that I wanted sparks... But it actually runs surprisingly cool. I was able to run at 110v in (about 800w me thinks) and draw sparks for a while (well until one of my 40n60s decided to completely short out) and the primary wasn't even all that warm
Which seems really confusing to me, I had probably 2x the metal with my litz primary, and much more of it usable to to the skin effect at 100khz... And at that point most of it should have been in the silver coating... Seems very odd to me. I didn't get a chance to scope the primary current in the litz primary since it was to short, so I can't say how much lower (if lower at all) is was.
[me looks longingly at the pile of wire and then at the spanish homework I have to do
]Re: ZVS heating
Sulaiman, Tue May 23 2006, 06:23AM
At 75 Vdc input the primary current will be about 100 A rms
with 1 uF at 100 kHz,
about 16.6 kVAR in the primary.
For 1 kW output that's a loaded "Q" of over 16
which is unneccessarily high.
If the primary capacitor is changed to say 250 nF
then the "Q" would be about 4 which is enough
and the primary current would be a much cooler 25 A rms.
Sulaiman, Tue May 23 2006, 06:23AM
At 75 Vdc input the primary current will be about 100 A rms
with 1 uF at 100 kHz,
about 16.6 kVAR in the primary.
For 1 kW output that's a loaded "Q" of over 16
which is unneccessarily high.
If the primary capacitor is changed to say 250 nF
then the "Q" would be about 4 which is enough
and the primary current would be a much cooler 25 A rms.
Re: ZVS heating
..., Tue May 23 2006, 06:49AM
I worry about a smaller capacitor since I am already running at 100khz, and those flyback cores don't like much more than that...
What would I have to do to 1/4 the tank cap while keeping the Fres arround where I started? Doubling the primary turns had almost no effect on its Fres, nor did changing the secondary turns...
..., Tue May 23 2006, 06:49AM
I worry about a smaller capacitor since I am already running at 100khz, and those flyback cores don't like much more than that...
What would I have to do to 1/4 the tank cap while keeping the Fres arround where I started? Doubling the primary turns had almost no effect on its Fres, nor did changing the secondary turns...
Re: ZVS heating
Dr. Shark, Tue May 23 2006, 07:33AM
With regard to the frequency, I suppose it is the leakage inductance that resonates with the capacitor, so you should change that. You could try a slightly looser coupling, for example with an air gap in the core, or you could add some extra leakage inductance by putting a small inductor in series with the primary.
Dr. Shark, Tue May 23 2006, 07:33AM
With regard to the frequency, I suppose it is the leakage inductance that resonates with the capacitor, so you should change that. You could try a slightly looser coupling, for example with an air gap in the core, or you could add some extra leakage inductance by putting a small inductor in series with the primary.
Print this page