MOSFET driving with transformer

[Danko]

Hi!

H have built a half bridge SMPS, wich uses a gate driver transformer, to drive the MOSFETs. The SMPS is working, the efficiency is not so bad at (relative) low loads.
The problem is (if it is really a problem, and not a measurement error), that the Gate-Source waveform looks like this:
http://sziget.mine.nu/~danko/aramkor/dscn1463.jpg
I connected the 'scope directly to the low-side MOSFET's pins.
At the circled parts, I think the MOSFET is turning on. The Gate-driver IC was TPS2811 (1.5A peak current), and when I changed to MC34151 (1.5A peak current), the left circled part "disappeared", but the right part was the same. Changing the IC to MC34151 improved the efficiency about 5-7%.
Oh, another important thing: The scematic:
http://sziget.mine.nu/~danko/aramkor...-07_202558.png
The component values are not "top secret", I just turned them off, not to make the picture too "messy".

So, what do you think, SMPS gurus, why is there an undershot? Is that REALLY an undershot, or I just measure something wrong?

[Workhorse]

I think a capacitor in series with primary winding must be required....I have seen many gate drive transformer primaries coupled through capacitors....just to avoid DC offset....

[Danko]

Hi!

In the first version of schematic I included a capacitor, but I found a paper, wich said, that the capacitor can be omitted, becouse the small DC offset is making a very small current.
But i'm going to try that!

[Workhorse]

See this link and go to Page3 look at the gate drive transformer...

http://www.qscaudio.com/support/libr...ries/PL1.0.pdf

[Danko]

I tried to connect series with the primarz a 2.2uF WIMA foil capacitor, but no change on the 'scope.

Thanks Workhorse, the schematic!

[Eva]

Those spikes in red circles are due to transformer leakage inductance and due to the own voltage drop of the driver IC, that usually results in a 2V dead band between sourcing and sinking current.

I think that the best solution is to use turn-off buffer transistors after the transformer, directly at the gates, to prevent the turn-off gate current from flowing through the windings (see other threads). Also, you should consider a driver IC with a MOS output stage, like IR4427, that does not exhibit voltage dead bands unlike ICs with bipolar output stages. Schottky diodes from the outputs to the rails may help with bipolar driver ICs when driving gate transformers, though.

Concerning the coupling capacitor, its purpose is to prevent gate transformer saturation (and smoke) in case of slightly asymmetric drive. You should rather leave it in the circuit

[gearheaddruid]

I did not see the picture, but I had a problem in the past that turned out to be a scope probe problem. I tried a different probe and the "glitch" was gone. Just a suggestion.

[Nejc_car]

Hi,

I am also having trouble with gate trafo, but unlike Danko's mine doesent work. I'm using SG3525, min DT. Full-bridge configuration.

That's how the signal looks on gates, without any input voltave.
http://i206.photobucket.com/albums/b...rivesignal.jpg

But when I applyed 30V the lower fets started to warm up.

And here is the schematic I'm using.

http://i206.photobucket.com/albums/b...tedrivesch.jpg

Note: there are 4 sec, I just didn't know how to do it

Edit: ups, D15 is a ZD15V.

[Nejc_car]

Anyone??

[ifrythings]

I may be way off here, but isn't there usualy a snubber on the primary? or is that only for power transformers?