How do I know a transformer is shorted?

[chipper]

Hello everyone! I have a shelved Rockford amp it's a punch 500a2. The problem is the mosfets for the power supply keeps shorting! I have replaced them probably 3 or 4 times already. I checked out all the solder connections on the mosfets no bridges I always make sure it's clean. Before it failed I can here a clicking sound in the transformer and the music would pause on each click the transformer would make. I have checked all the output fets to make sure there isn't a short in any and they all test fine. I don't want to waste anymore of my little stock of fets on this amp. I'd like to find out what's causing this if possible.


Oh yeah the fets are IRF3205 & the gate resistors are all 47 ohms.

[timwebb]

see if u hear any arcing on the inside . wiggle it to see if the current change.

[sdoom]

if you have access to an oscilloscope check the gate signals of the mosfets. When using both channels of the scope you can compare the two parts of a bank of mosfets driving the transformer (the primary site of the transformer consists of two coils, each coil is powered by a set (or bank) of mosfets. The signal should be clean square wave (a little "knotch" is ok).



When you compare both mosfet gate signals from each part of the bank you need to see that while one bank is in "ON" - state the other part of the bank needs to be in "OFF" state, if there is no dead time in between the mosfets are all open at the same time and thus causing a high current flow through the transformer.

If the drive signal looks real ugly there´s a good chance that the driving transistors are not working properly anymore.

[ultra]

@sdoom
sorry for being a bit off topic here, but the pic you posted, is it from a gate resistor ?

Which amp is it?

[sdoom]

yes and no . It´s a gate resistor signal but from an older type amplifier (Magnat) which uses a NE555 signal generator, and of course it shows just one gate signal, not both banks at the same time. Sorry for not having an exact pic of a TL494 / 3525 sort SMPS driving signal (but that could be googled as well). Just wanted to show that at the gates there must be a clean signal. The following shows both signals.


there you can see that one signal is "inverted" and there is a dead time between the "ON" states

[chipper]

Man only if I had a Oscope but I don't. I replaced the driver transistors from the get go. And I remember checking after installing them I had a 5.6v at each gate pad. I am suspecting the transformer because when it was last working I could hear a clicking type noise coming from it. Then one of two mosfets would get hot when that clicking noise occurred.

[ultra]

The overshoot is huge.

Reminds me of this pic, but this is the secondary of the trafo at full load

(30A input).

[sdoom]

Hmm, I only had it once that a transformer died and there it had been very obvious, the core itself cracked and a huge "V"-shaped part was missing out of the core material, as well as the transformerwires were burned,too.

You could take out the transformer secondaries from the board to have the transformer run only on primaries. If the mosfets still get hot it´s more likely that the drive signal is not correct. Did you replace the gate resistors as well? Make sure there are no bad solder joints at the mosfets and the driving transistors.

Mostly all smps in car amps produce a short sound during startup, so this could be considered "normal" , but if something goes wrong with the driving signal most transformers cause an ugly "scream" sound. If I am not mistaken that already happens even if there is no load connected to the outputs ?

[Perry Babin]

Measure the resistance between the gate legs of the two parallel power supply FETs (meter probe on the legs near the body of the FET, not the board). Do you read ~94 ohms between them?

Check both banks.

Does twisting the core when it's on make a difference?

Did you do anything to the audio section?

Do you have the transistors clamped and either a current limiter or a 10 amp fuse in the B+ line? If not, that will likely prevent you from blowing any more FETs until you find the problem.