Ringing in a Multiple output Flyback converter

[tocotukan]

Hi all!

I have some problems with ringing in my flyback converter when the MOSFET turns of, see picture.

View attachment PWM_14.bmp
View attachment RAMP_31.bmp
View attachment Stability.bmp

I'm using an UC1825 equal to the UC3825, running in Current-Mode.

Theories of failure is oscillation in the Voltage Feedback or the Current Feedback, or something else?
I have measured the loop stability as attached and RAMP voltage for the UC1825.

Could someone help?

 

[mtwieg]

What do the two waveforms represent? Drain voltage and oscillator ramp? If the first one is you drain voltage, then the ringing on turn off actually looks very well damped, and should be fine. Are you talking about those bursts of very high frequency signal? Because those aren't happening on turn off, but rather just after turn on, which is odd.

 

[tocotukan]

Yes, RAMP is the oscillator ramp and the PWM is the gate voltage at the MOSFET, its the ring just after the turn off I am wondering about. The drain-source voltage over the MOSFET looks like this:
View attachment VDS_61.bmp

Basically it is the spikes on the oscillator ramp thats worries me. Compared: View attachment E00.bmp

 

[FvM]

I would like to see the circuit schematics, particularly the gate driver circuit. The oscillation shape suggests a parasitic MOSFET oscillation.

 

[mtwieg]

The gate drive waveform doesn't look like it's oscillating. The burst of ringing happens after the gate has been taken to 0V, so it shouldn't be able to oscillate there. But the drain waveforms do look strange. I see what looks like high frequency ringing as the voltage approaches the supply voltage, and also the slow slew rate after that point is odd. Are you using some kind of snubbing circuit on the drain?

 

[tocotukan]

This is how the input basically looks like: View attachment basic.bmp

 

[FvM]

Ringing usually shows an exponential decay, not a burst. But I agree that the gate circuit doesn't seem to allow oscillations. The previously posted "PWM_14" waveform is the controller output, not the gate voltage I think. It doesn't exactly tell when the gate voltages falls to zero.

 

[tocotukan]

Yes, the PWM_14 waveform is the controller output! I have added some snubber network between the "supply" and the drain of the MOSFET, which only resulted in lower ringing after the turn-off spike happens!

 

[vimalkhanna]

The output /input windings need to have lower winding capacitances for the underdamped flux restoration .You need to to put leadless ceramic caps across the secondary winding to compensate .As there shall be power losses , snubber resistances shalll be required .as the switch OFF slopes shall lower .
It is best to rewind the transformer again using less turns/layer to reduce caps intercoupling between turns .of the transformer .

 

[mtwieg]

Okay so the high frequency stuff showing up on the ramp waveform is almost certainly EMI being generated from some switching event somewhere in your converter. Probably a diode switching somewhere, or your ground bouncing. Could you take a waveform showing the noise on either the ramp or gate waveforms, at the same time as the drain waveform? Use a fast time base of 500ns/div or less. The output side rectifier could also be source. There should be some obvious correlation between an event on the drain or output which is synced with the noise bursts.

Also, why is your switch only working at half the ramp frequency?

 

[FvM]

I agree, that e.g. the ramp waveform seems to show picked up interferences or a bad probe connection. The collection of waveforms doesn't reveal the primary source of switching noise. We would need to see the complete circuit and a layout photo for more precise guesses. I would try to locate the interference source with a probe coil swept around the ciccuit.

Also, why is your switch only working at half the ramp frequency?

Obviously, only one output of the 3825 push-pull controller is connected to a switch transistor.

 

[mtwieg]

Right, forgot that the 3825 is push pull.

 

[tocotukan]

Well I am using the UC1825 which is both a flyback, forward, push-pull and half-bridge controller.

 

[Orson Cart]

I suspect one of the output diodes - probe around with a 2nd probe until you find a switching event that is in phase with the noise on the ramp. It looks like a ringing causing a reverse recovery on a diode some where...primary snubber circuit? perhaps...