Marshall 3210 lead mosfet 100

[pontiacpete]

This amp is doing something weird. When it came to me it was motor boating heavily. There were a couple of badly burnt resistors R46 and 47. The main filter caps were oozing liquid. I replaced almost everything in the area in this picture. The board has some slight burning which I ground out and jumped over some the area with wire.
The weird thing is that when powering up the +/-15vdc and the +/-34v are not quite making it to spec. The -15v comes up slowly to about -12v and the +15 comes up immediately. The -/+34 is off by about 5 to 7v. Two of the caps C40 and C39 get really hot. I’m wondering if there’s a problem with board and that maybe all of the supply voltage components should be moved.
Any thoughts on this?

[R.G.]

Caps getting hot usually means either they're old/damaged and leaky, or that they're being asked to pull too much current through them and the current heats them by their ESR.
C39 and C40 are on the positive side of the power supply, so it could well be that the positive side of the amp is oscillating. That's unusual, but possible, especially with MOSFET outputs. You didn't say whether you replaced the rectifier diodes. A leaky diode or two can overheat a first filter cap.

Have you used your oscilloscope to see what is happening on the output jack and also how much ripple voltage is across C40 and C38? The size of the ripple voltage is directly proportional to the amount of current being pulled out of these capacitors.

It's odd that C39 is getting hot, as it's presumably only supplying the preamp stuff ( I don't have the schemo of the amp at hand).
Motorboating is a low-pass version of oscillation, and usually happens because of insufficient decoupling caps between the several stages of the amp.

With no more information to go on, I would first reflow all the solder joints to eliminate high resistance joints. It is (again) unusual but not impossible that decoupling caps are not making good contact with ground and the rails. The check the ripple voltage and look for high frequency oscillation. It is (again) unusual, but possible that the output MOSFETs could be oscillating all by themselves at 100-500MHz where simple o'scopes can't see.

[pontiacpete]

I’ll try to attach it

[R.G.]

OK, try attaching the base of TR6 to signal ground through a 22 to 100uF cap. This "shorts" any signal to ground before the power amp. If the caps still get hot, the power amp is doing the evil on its own. If not, the preamp is telling it to scream.

Oh, wait. Belay that. TR6 is fed from an opamp output through a cap. You'd have to lift one end of the cap. Capacitor shunting won't work well.

The objective is to separate the two.

[Stevem]

I have had burnt boards where I thought I had ground out all the new carbon resistors from the smoke show, but I had missed one spot and that was bleeding voltage to a grounded board standoff right near by.

As posted test the filters for leakage if you can and test the recto bridge with it unhooked.

[nuke]

Divide and conquer, remove R46 and R47, lift the 34V sides of R44 and R45.

Apply, with an external power supply, about +20v to the lifted leg of R45 and see that the +15V regulates properly.

Apply, with an external power supply, about -20v to the lifted leg of R44 and see that that -15V regulates properly.

That proves out the zener diodes and the preamp section. The effects loop send should be working, except for reverb, as the driver circuit uses the +/-34v. But you should hear the spring "clash" when you bump it. Just plug the send into another amp and check it out.

If something doesn't pan out, then correct the 15V supplies and/or loads.

Then move on to the 34v rails.

Step at a time.

And if you're human like, me, good chance the last guy who worked on it missed something.

[pontiacpete]

Here are pictures of two different pairs of output transistors, original and another pair I have but not sure if they are working.

IMG_2083.jpeg

[pontiacpete]

When playing through the return jack signal is nice a clean and clear. The send to another amp not so good

[R.G.]

When playing through the return jack signal is nice a clean and clear. The send to another amp not so good

That's very telling. Everything after the return jack is likely to be OK. A problem (or "the" problem) is likely before the send jack.

I don't know for certain, but the scope traces look like one of the opamps in the preamp are having problems with their inputs being forced outside their "common mode voltage range", causing that blip of phase reversal in the middle of the waveform. Maybe. It's a plausible guess anyway.

[LOUDthud]

A second vote for opamp being operated with excessive input. IC1 and IC3 being TLO72 are suspect.