Orion 2500D trouble shooting- protect mode

[Scanlines]

Hi, I need a little repair advice;

Customer brought in two of these 2500D's
both go into protect mode as soon as they are powered up.

(I can see my ammeter go "thump, thump, thump.....)

None of the mosfet devices appear shorted or leaky

To make a long story short, I carefully cut and mended B+ and B- traces until I discovered that the power supply will run, the rectifiers are good and whole thing will power up correctly if the B+ line to the two, up-right, daughter boards is cut.
The preamp section is running +12V and -12V so this has lead me to suspect a problem on the 2 little driver boards or the digital bridge section.

The up right boards are difficult to get meter probes on components. that would entail removing them from the main board to really test. I definitely don't want to damage the main pcb as I don't have powered solder-vac stuff.

Has anyone else tread this path ?
Am I getting in over my head ?

Thanks,
Scan

[Perry Babin]

Is the problem possibly a DC offset problem that's causing the amp to go into protect mode?

If the output is swinging toward one of the rails and there isn't excessive current draw (without a load), I'd suspect the differential comparator (U305 - LM361) on the audio PWM board.

You don't need anything other than a good soldering iron and a good desoldering pump (Edsyn DS017) to remove the driver boards. If you still have trouble removing them, using ChipQuik solder makes them easier to remove.

[Scanlines]

No, I did not see any DC offset on either output, but I now see that the problem is a PCB and/or slodering issue.

I found just 4 of the FB31N20D's on one channel getting real warm.

If I flex the board just right, or smack it just right with a rubber mallet, the thing will kick out of protect and function normally.

It's real finicky. sometimes I can't get it to go back into protect unless I twist or pound just right and visa versa.

By the behavior, I suspect that the gates of the 4 hot FB's are going open and are being allowed to float into saturation.
but this is one of those physically intermittent gremlins that had me chasing it around all afternoon.

I will post my progress.

[Perry Babin]

Does it look like the driver boards have been removed previously? If so, there could be damaged vias where the driver boards are soldered into the main board.

The resistors standing off of the driver board and the 12v zeners can develop bad solder connections. Either can result in low drive voltage and will make the FETs run hot. Look at the driver voltage on the other 31N20s to see if they are all the same.

If you can make it go into and out of protection and can find the pins that have different voltages when in/out of protect, that could help with troubleshooting.

Which FETs are getting hot (Qxxx)?

[Scanlines]

No, it does not look like the driver boards have been removed.
They're absolutely clean.

At this point, I'm fairly sure that I found the problem anyway.

two windings on one of the secondary toroids are shorting together intermittently. I could hear the spark when I was bending the board which led me to it.

It looks like the transformer feeds rail voltage to the drains
(I think) of Q420, 21 and 22, which were the guys getting hot.

Not having a schem to work with I don't now off hand how exactly
the short was causing this.

http://members.aol.com/ta7205/2500D.jpg

I got a pic that points to the offending toroid. You can see Q420-Q422 all snug with their temporary heat sink I fashioned so that I wouldn't burn them to kingdom come.

I'm going to patch up this transformer now and run some tests.
I'll post if there's still a problem.
Thanks for your help.

Scan

[Perry Babin]

You can use the service manual for a JBL BP1200.1 as a guide to the audio circuits in this amp. This amp is essentially two 1200s.

Those are the output inductors. They are connected to the drains of the FETs. The FETs have their source legs connected to the rails.