help!! snubber - dc dc converter with a push pull setup

[dynamag]

Hi
I am trying to design a dc dc converter with a push pull setup from 12vdc to 340vdc of 1000watts. The frequency of operation is 50khz. For trial and test purposes, I have hand wound - quite tightly - a push pull transformer on an ETD44 bobbin, with a foil of 25mm x 0.05mm x 4 turns with center tap as the primary, and 19SWG(about 18AWG) x 136 turns with center tap as secondary. With a tiny load on the secondary, I get a severe resonating signal - and spike - on the drain when the MOSFETs are switched OFF. I have included the waveforms and circuit for better understanding of my issue. You can also see the two waveforms are not the same, so that is another issue I need to solve
The probes are X10 probes, so the values shown on the scope display are correct.
What should I do so I get similar, clean square waveforms, with no resonance or spikes for current loads ranging from 1 to 80 amps, for both sides, with no differences.

 

[mtwieg]

Snubbers aren't going to help you much, you have other problems. First I see overlap between the ON time of the FETs, so you're getting cross conduction. As for the lower frequency ringing, that might be from insufficient bypassing of the DC rail. If you probe your DC supply I bet you'll see the same ringing show up. Snubbers are only useful for the very high frequency ringing occurring immediately after switching. But try fixing the other issues first.

 

[dynamag]

Thank you mtwieg
My supply rail was ringing, though with a lower amplitude. I am changing the present setup to reduce all possible causes of ringing and ripple, including length of wires and bypass caps.

The switching overlap on the drains seems to happen only on one side of the push pull setup.
I think I need more advice. I will upload my circuit and the waveforms from controller, gate drive, and drain, after some more investigation on my own.

 

[schmitt trigger]

In addition to mtwieg's very good points:

Measure both the primary and leakage inductance. The ratio has to be as high as possible. Construct not one, but several transformer samples, each with different winding techniques and probably core sizes.

Resonant behavior is normal at minimal load. This occurs because the load is way too low to maintain continuous current and the magnetizing inductance runs "dry". Use at least 10% of your maximum load for testing...after you've measured your leakage. Otherwise you could blow your Fets.

 

[dynamag]

Thank you schmitt trigger
I am getting these values for 2t+2t primary winding with foil, foil on the inside, secondary outside:
inductance: 32uH,
leakage inductance with all other windings shorted - 8uH
Is this ok?
I am going to wind some more coils differently, to see if I can improve on this ratio.

 

[schmitt trigger]

By all means improve the ratio,
you may want to sandwich the primary and secondary in layers to improve the coupling.
Also, 32 uH is a little on the low side for a transformer...are you sure you don't have unwanted air gaps?

 

[dynamag]

By all means improve the ratio,
you may want to sandwich the primary and secondary in layers to improve the coupling.
Also, 32 uH is a little on the low side for a transformer...are you sure you don't have unwanted air gaps?

Yes, you are right. I see that as the core is held tighter, the inductance value increases.
I obtained a value of 45uH on the same core. This is still low. The formula says I should get near 60uH, with an AL value of 3700 on the ETD44 without air gap.

What should I use to glue/fix the core halves together without any air gap? Presently I am tying the halves together as tightly as I can! This will need to change, I know, once the transformer is finalized.

About the ratio of leakage to prim. ind., I suspect my meter. So I am going to correct that situation first.