JCM900 noise issue – this is a weird one...

[SoulFetish]

JCM900 50W head comes in with a strange issue: (schematic below)
Marshall-JCM900-Dualrev-50W-4500-Schematic.pdf

Customer complains that the amp is humming/buzzing in standby, and when the amp is fully on, the noise is louder and continues to get louder.
Noise in "Standby", we asked. So, we plugged it in and confirmed it's just as the customer says. (see audio clip)

JCM900_Noise_Issue_6-8-19•US•.mp3 - The audio you first hear is the amp with just the power switch in the on position. When the audio changes and the noise gets louder, is when the amp standby switch is turned to the on position. Then the amp is put back in standby for a few moments and then turned completely off.

This is a strange one indeed. The amp theoretically should pass no audio in standby. But, there you have it. I pulled the two existing tubes and tried some bench output tubes and got the same symptoms. To compound the weirdness, I pulled output tubes altogether and still got the same noise in standby! WTF.
The quick version of troubleshooting was, the HT voltages all measured okay. Voltages at the tube pins seemed fine as well, with both grids measuring a bias of around -41V and change.
I was also getting +/- 16V on the opamp chips where you would expect them as well.
(One other thing to note, was I was not able to pass signal to the output injecting a sine wave into the input, or in through the effects return in failure mode)
So, it was time to do some signal tracing through from the input through the output stage. At first, I was getting a noise and distorted signal (not my injected signal) at pin 5 of the closest chip to the input. I mistakenly thought it was IC1, but it wasn't. In any case, I went to probe signal on the plate of V1 and accidentally shorted two pins together, which was evident by the sound and visible spark. Happened to a few other people before... maybe.
I powered it down, and went to check the fuses assuming I probably blew one of them. Now, here's another really weird part of the story. The 900 has a fuse for the output tubes on the back panel, not traditional HT fuse found on many other Marshalls. When I pulled the fuse anyway to check it, it came apart and the end was stuck in the fuse holder cap. The mains fuse was fine, so I replaced the output tubes fuse with a new 500mA fuse it specified. When I powered it back up and went back to signal tracing, I probed nice clean sine waves through all stages into the Output stage. I turned to my boss and said/questioned "no way that fuse was the issue, right??" With his original tubes back in, took it off the dummy load and put it into the speaker, and no noise. Testing it with the guitar through both channels, and the amp was showing no sign of the original issue. We set it aside, for further testing and I moved on to replacing an OT in an old SVT.
But, what the Hell?

Any thoughts?

[g1]

Hard to say if the fuse was intact, blown, or some resistive amount in between. Was the tube fail LED lit up?
That fuse disconnects the cathodes, but there is still the led in series with the 100K offering some small current path for the power tubes.
So now it's quiet in standby?

[Randall]

I'm thinking fuse, but I don't know why. Interested in hearing what your boss had to say to your question.

[SoulFetish]

Hard to say if the fuse was intact, blown, or some resistive amount in between.

I know. If I were to guess, I'd say some resistive amount. But I can't say for sure.

Was the tube fail LED lit up?
That fuse disconnects the cathodes, but there is still the led in series with the 100K offering some small current path for the power tubes.

No. At least I'm pretty sure it wasn't. I definitely didn't notice the red LED, but I initially thought that fuse was the HT fuse. The 900s are probably the Marshall model I'm least familiar with, and I had it chassis exposed/knobs facing me while I was checking voltages and tracing signal.

So now it's quiet in standby?

Yes. But I would stress the word "now". A little more detail:
After I was able to trace a clean signal, I powered down to check the audio through a speaker (after replacing the fuse as stated above). When I first powered back up into the speaker, I could hear a bit of that noise. However, It quickly went away after taking it out of standby and when returning and switching in and out of standby, the noise never returned.
This makes me think that maybe there might be more than one issue causing what seems like one problem, and why it was set aside for further testing.

[alexradium]

What was the noise,was it a specific frequency?
This circuit is a bit different from what you find in most amps,when you are in standby,the B+ is not zero Volts and the bias is -80V,around that,so the combination of those with some possible defective component could lead to noise in the amp.
Check out the bias circuit.

[SoulFetish]

What was the noise,was it a specific frequency?
This circuit is a bit different from what you find in most amps,when you are in standby,the B+ is not zero Volts and the bias is -80V,around that,so the combination of those with some possible defective component could lead to noise in the amp.
Check out the bias circuit.

*See link for clip:
https://music-electronics-forum.com/...0&d=1560042614

[alexradium]

filter caps

[Pedro Vecino]

Could it be related to a problem around D6, D7, R36 and C16? (The group of components between the star earth and chassis)

[SoulFetish]

Could it be related to a problem around D6, D7, R36 and C16? (The group of components between the star earth and chassis)

Hmmm... to me, it does sound like a power supply grounding noise. So, this would make a lot of sense. When I get back on Tuesday, I'm looking into this. I have a hunch you might be onto something, Pedro.

[nickb]

If you don't know what the cause was it's not really fixed. It wasn't the fuse for sure.

With the standby off there are two circuits energized, the bias and the heater. The heater current to ground will be small so I choose bias, notably half wave rectified as is the noise when in standby. If I'm looking the right schematic, the speaker jacks connect to GN7 which is shared with the bias return and then connected to the star point. The other side of the output transformer goes the the star point. Therefore I'd suspect that wire & its connections. That would also tie in with a 100Hz noise when the standby switch is closed.

[Chuck H]

re: The bias supply and/or the filament supply: I'm not sure how modulation in those circuits could be amplified if the tubes aren't conducting So there's THAT to consider. The SS amplifiers aren't powered via the standby, but anything they are doing should also require the power amp out of standby to be amplified.

My thinking is that something has to be modulating on the speaker output to be making noise. Without conducting amplifiers (amp in standby) that could indicate some elevation from 0V on the speaker outputs allowing power supply noises from charging caps (probably bad caps) to appear on the speaker. The schematic is awful, but it looks like the ground for the power amp board is connected via a flying lead other than the filament chassis connection. And I can't tell what else that chassis connection is doing because of the bad schematic. But I'd be using a scope to look for an analogous noise on any active circuits when the amp is in standby. Even if the amp isn't making noise in standby now. Like Nick said, if the problem hasn't been identified then it isn't fixed. I suspect bad caps (bias supply or power filters) and possibly a dirty or faulted ground somewhere.

EDIT: Ok... I just listened to the clip and the actual noise sounds like an oscillator to me. I might have suspected ripple overtones because of the frequency, except that the frequency seems to shift with voltage, so that wouldn't be ripple which should remain constant. In any case, same as I said above... I think there must be some bad filtering caps and possibly an iffy ground connection that you may have inadvertently improved (though not fixed) while tinkering.

[nickb]

re: The bias supply and/or the filament supply: I'm not sure how modulation in those circuits could be amplified if the tubes aren't conducting So there's THAT to consider. The SS amplifiers aren't powered via the standby, but anything they are doing should also require the power amp out of standby to be amplified.

No, not amplified, but simply passing through the speaker ( and OPT secondary) by means of the bad ground connection mentioned.

[SoulFetish]

If you don't know what the cause was it's not really fixed. It wasn't the fuse for sure.

I wasn't seeing how it could've been the fuse. Since it was the only change that I made to any part of the amp, and it normalized right after, I didn't know what the heck was going on.

With the standby off there are two circuits energized, the bias and the heater. The heater current to ground will be small so I choose bias, notably half wave rectified as is the noise when in standby. If I'm looking the right schematic, the speaker jacks connect to GN7 which is shared with the bias return and then connected to the star point. The other side of the output transformer goes the the star point. Therefore I'd suspect that wire & its connections.

Yeah, this is helpful and gives me some direction in where to start looking for a broken connection. Thanks, N.

That would also tie in with a 100Hz noise when the standby switch is closed.

120Hz, but who's counting.

[Helmholtz]

Since it was the only change that I made to any part of the amp

Well you also caused a HT short. This may have produced an arc across a bad contact and might have (temporarily) restored it by spot welding (even inside a cap).

[SoulFetish]

nickb for the win!
This was an easy one to spot once I pulled the back panel board.