spike problem !

[Andrew Eckhardt]

The turn off edge of the visible transistor doesn't look very fast (164659d1269901278-spike-problem-5.jpg), but when you look at the turn on edge something Really gets kicking even before the gate drive voltage rises very far. In fact it looks like the gate voltage is getting a little help from the gate drain capacitance. Maybe it would be better if your driver were able to hold the gate low better until turn-on, like a FET output stage could.

[Andrew Eckhardt]

I'm also a little concerned about how the drain signal looks Totally different than in your previous shots. Maybe you can explain that.

[Andrew Eckhardt]

see ya

[mgm2000ro]

ok ! Because I was tired, I put wrong the diode reverse on the gate mosfet!
I ajust dead tine at 40%!
what can try again .. what to change?

[orfeus]

It seems an IC current problem to me. Do this. Put a resistor of 2k in series with a diode, both in parallel with a 100n capacitor and put the parts in series in the line between IC and totempole.

[rajudp]

check the transformer design, what is the calculation you used for transformer and what is the flux you taken.

[savu]

Also try and wind the cores so that the magnet wire is evenly spread on all it's surface.
I think that you use to bigger cores for your windings or vice-versa.
put the snubber at primary sides of trafo and secondary side, not on transistor, keep high current paths and signal paths as short as possible, put the totem pole drivers (emitter followers) as close to the mosfet transistor as possible, drive the fets with 12V or more (up to 18V)
put 100uf or more low esr capacitor near the sg as possible (you have very ugly waveform at the gate), try to use litzwire for the winding (if you cant find or afford it i will help you by sending you homemade or teach you how to make it yourself).

Best regards,

Savu Silviu

[Psychobiker]

This is, I am going to wager, the problem caused by leakage inductance in that transformer. When the transistor switches off, all the current that was flowing during the on-time has to go somewhere! And it can't go across the MOSFET due to its huge impedance, so it'll head back on through the transformer and that ringing you see is due to the 'parasitic inductance' of the transformer's primary before the current can 'fly out' through the diode bridge on the other side of the transformer. An RC snubber should aid that, or a Zener Clamp. PM me if you need a breakdown. I'm an advocate of the zener, mind...

L

[Perry Babin]

Most snubbers go across the transformer windings, not across the FET.

Are the primary and secondary windings would on top of each other? If not, poor coupling could be causing the problem.

[Psychobiker]

Snubbers/clamps are a real tradeoff - and are best across the transformer. But - something as simple as a <1nF high-voltage ceramic cap across the drain and source of the FET can work too. Sometimes it's just easier to overspec the MOSFET than find components with enough clout to eat up large spikes without heating up and effecting the overall efficiency.

Also - did you split-primary the transformer? Much better!