160v/8a

[Fob]

Hi all,
I have a success with my 160V/8A but in full bridge topology (no more flyback above 200W!). But it still blows with 7A load. I see nice waveforms on gates with slope below 350nS, no rings on primary and fully stable work without noise or heat with 6A load.
Any comment will be grateful.

[areza]

very nice and interesting care to show some pictures , perhaps a full schematics,

[megajocke]

Which part of the circuit blows up?

What does the waveforms on T1 secondary look like? I suspect you might need snubbers over the diodes to lessen the spike generated after diodes turn off because of reverse recovery in combination with leakage inductance.

Are you sure L1 is not saturating?

It could also be cross conduction. Which types of FETs do you use? Is it IRFP460, IRFP460N or something else? Regular IRFP460 have pretty high reverse transfer capacitance and will need very stiff turn off and soft turn on drive or negative gate bias when off if you don't want cross conduction.

[areza]

Sorry didn’t quite catch the word “blowing up” , 960 stable watts not bad, still like to see it

[Tahmid]

Hi fob,

Your circuit seems ok, but you could try the following:

1. For quicker MOSFET gate discharge, add an anti-parallel signal diode across R2 (maybe a 1N4148).

2. At the 270-300v input, you should add an inductor of about 1mH before the 470uF capacitors to limit peak inrush current.

3. Increase the output inductor L1, from 150uH to about 400-500uH.

4. Since your switching frequency is 98kHz, the internal freewheel diode of the MOSFETs may be too slow and thus, cannot protect the MOSFETs from transients, while switching off. That's why, you can add anti-parallel ultrafast recovery diodes across each MOSFET of the bridge.

5. Since your switching frequency is high (98 kHz), you can remove R6 and R7 as your transistors are working in emitter-follower configuration. This will quicken the turn-off of the transistor.

Try these and observe and inform regarding the result. I think it will improve.

Thanks.

[Fob]

Thanks all you,
Sorry for my English (Winglish) and all misunderstand.
The fired device is IRFP460, the source pin on high side FET disappears with glory at 7A after couple of minutes. And I don’t have output snubber and this cause some ringing on secondary side
Thank you Tahmid, tomorrow I will post the scope of waves. I did not pay any attention to reverse diodes in mosfets and I will limit the inrush current.

[megajocke]

I'd just move the emitter of Q1 to the other side of R2 to give better turn off. If shoot through is a problem R2 can then be increased giving slower turn on and thus less parasitic turn on off other side fets.

You could make (or steal from a pc supply) a current transformer and measure current through transistors to see if you have cross conduction. In series with one of the upper transistor drains is probably a good place.

I'd replace those 22µF electrolytics in the driver with film caps, they shouldn't need to be this big. 22µF electrolytics have high ESR, it can be over 10 ohms!

"4. Since your switching frequency is 98kHz, the internal freewheel diode of the MOSFETs may be too slow and thus, cannot protect the MOSFETs from transients, while switching off. That's why, you can add anti-parallel ultrafast recovery diodes across each MOSFET of the bridge."

The diodes have slow turn off, the turn on with any diode is pretty much instantaneous so this shouldn't be a problem. The problem with MOSFET body diodes is when you try to force them off after they have been turned on. This can cause latchup and kaboom... But it should only be a problem in phase shifted ZVT bridges, ZVS and ZCS converters. (and in totally different applications like synchronous rectifiers and class-D amps)

[Fob]

I changed the mosfets to STW20NM50 with twice faster diodes. Also I put RC across the output diodes 470pF+50 Ohm. Here is the result with 5A load hard pwm (without feedback).

[Fob]

These are the current through diodes, GS front & end and rectified pulse before inductor. All these WVForms are with 5A load.

[edfed]

you can directlly plug the driver primary widing on sg3525 outputs.
Q6,7,8 & 9 are not needed at all.

D1, R1, Q1, R2, R3 are not needed too.

but R4 is.

and you need protection diodes on the SG3525 outputs to supply rails.

indeed, this is one of hte best way to drive a full bridge SMPS, and the one i use.

maybe you need a RC snubber just before the self. after the rectifier diodes