Mesa Subway Blues some hum/buzz 100Hz, out of ideas

[iefes]

yes, you're right. It is not grounded but one end of the voltage divider that provides the elevating DC is grounded. I've put a voltage divider across the B+ supply ("A" in the schematic). I've done this like is explained on Rob Robinettes page. So its: GND --- 33k ---- elevation DC ---- 220k ---- B+
I hope this helps to clarify

[Pedro Vecino]

It´s the input jack through where the ground is connected to the chassis? Exclusively?

[iefes]

Hi and thanks for your reply!
Yes, the Input Jack is the only connection to the chassis. But after reading your comment I have tried it with another connection, as was originally intended by Mesa, but this would not change the buzzing.

I have also cleaned and resoldered the socket of V2 without success.

I was desperate, so I again chopsticked many parts on the PCB and noticed a slight "pop" coming out of the speaker when softly hitting two of the coupling capacitors , one being the 0.02 after the effects loop and the other one being the coupling cap after the plate of V2b (0.047). Is this normal behaviour? I thought I'd order some new caps for these positions as there are still the original parts in there and see if it helps.

What do you guys think?

[Pedro Vecino]

I d´ont know exactly how did the test. But I would connect a wire from the negative of the first electrolytic (next to the power connector) to one of those screws to the chassis and then isolate the ground from the input jack to the chassis to see what happens. It would even remove the send / return jacks from chassis avoiding that same contact in case they are not isolated.

[iefes]

Hi Pedro, thanks for your suggestions. That's exactly what I did actually. The chassis connection is made with a wire just right next to the negative lead of the first filter cap. I've isolated the input jack from the chassis for the test. The FX-loop is completely removed from the amp at the moment, so no worries about this part.
Actually, Mesa used the chassis as the only Gnd connection between the preamp and the power supply. The grounds are connected from the input jack to the chassis and then from the chassis to the power supply. I have modified this by connecting the negative lead of the C+ filter cap with a ground-connection near V1. Additionally, I removed the chassis connection near the power supply. But this wouldn't help. In fact, all modifications I tried regarding the grounding scheme didn't change anything to the buzzing.

[Pedro Vecino]

It´s difficult for me to understand that the noise level is identical. The idea was to look for differences, although it may happen that is overlapped by another major noise of different origin.
In this case, I would disconnect anode capacitors in each triode. From the phase splitter to the input (in that order).
With high volume on the monitors gives me the sensation the possibility of filtering between filaments and cathode in some triode although I understand that it would be something exceptional. I would also try to disconnect the cathode resistors/capacitors in that same order. If this were the case, I think it would be more enlightening.

[iefes]

Pedro, thank you very much for your suggestions. I highly appreciate them!

I have ordered new parts and am now preparing the amp to put them in. During this I have pulled the coupling capacitor between the FX-Loop (in the schematic, the FX-Loop is not used in my amp right now) and the grid of V2b. The buzz is gone with this coupling cap pulled. So this means that the "bad" part is somewhere before this particular capacitor OR that the capacitor itself was the culprit, right? Do you think it's still necessary to try pulling the anode capacitor of V2b and the cap directly in front of the PI, which are both located further behind circuit-wise?

The parts will arrive tomorrow and then I will put them in successively. I will change all of the potentiometers (the volume pot is only barely working, most of the times it's on 10; I have tried cleaning it and everything) and maybe also the sockets of the preamp tubes. I have ordered new capacitors for some of the coupling positions.
And the PCB will get some cleaning with acetone at some places.

If putting in the coupling caps will not help I will definitely go the route and try pulling each part one by one for the anode capacitors and the cathode resistors+capacitors.

[Pedro Vecino]

If removing that capacitor the noise disappears forget to go towads the phase splitter. That would be the starting point to test it and go backwards (towards the input) if necessary using the same procedure.

[iefes]

Okay, so today I tried pulling the next anode capacitor (0.047u on anode of V2a). This resulted in unaffected hum/buzz. So I have now narrowed down the location where the buzz is "collected" to be between those capacitors. In the schematic there are these 3 parts that go to the Reverb circuit. They do not exist in my amp, as I have the non-Reverb version. Same for the Reverb pot, thats just a 100k resistor in my amp. Then theres this junction with the 680k and 500pF. These parts are enclosed by a blob of epoxy, you can trace them on the PCB-picture I provided above. In my amp there is not only one but two capacitors in parallel, but I cannot tell which they are and I cannot remove them as they are covered by 3cm of stiff epoxy.

The next coupling cap in the schematic is not in use in my amp as I have jumpered directly from there to the 0.02 cap on the grid of V2b.

So there are only very few parts in this area, most of them being resistors forming voltage dividers to lower the volume. I have also tried to jumper directly from the plate of V2a to the 0.02 cap on the grid of V2b but this resulted in loud hum. I suspect that was just because all those voltage dividers were jumpered, so the volume was just too high, amplifying all the noise etc.

I suppose that this strange 500pF cap that forms the loop between the two stages is causing the issues (or in my case those two caps in parallel). One of the heater traces runs directly below it, so maybe this thing is picking up the buzz.

Now I'm wondering what I should do. I could try to remove all this epoxy but risk damaging the PCB by doing so.
I don't know what would be the best approach to this issue. And I still don't know for sure that the issue is really caused by those capacitors.

So I think I got closer to a solution thanks to Pedro (!), but I'm still not there.

[Pedro Vecino]

If the 33K resistor, 15K resistor and 0.02 capacitor (those connected to the second point marked "x") are accessible, you could lift a pin in each one (those corresponding to that point) and connect them without touching the board . In this way you can check exactly what this strange loop circuit does and also remove from circulation the effect that the proximity of the heaters line can cause.

[Chuck H]

So... Is this an accurate representation of what is in your amp right now?

[iefes]

@Pedro:
Thanks. The parts unfortunately are somewhat far from another but I'll try it :-)

@Chuck:
Sorry for me not providing a correct schematic. When I find the time later I'll put a proper schematic together. Your schematic is almost correct, I don't have the bright-switch adapted and my input stage is made by a 10k grid stopper and a 150pF shunt capacitor (á la Merlin). And instead of the 500pF going to "X" I have 2 parts in parallel, supposedly 2 capacitors, maybe adding up to around 500 pF.

[iefes]

next update:

I lifted the parts at the junction of 15k/33k/0.02u like Pedro suggested and connected them with a wire floating above the PCB, so the strange loop capacitors were not connected anymore. BUT no better. Still the same buzz. One of the legs of each capacitor is still in the circuit (the one at the 680k/100k junction) but this shouldn't affect anything, right?

I'm starting to get really confused

[Pedro Vecino]

I would disconnect the resistor and capacitor from the V2A cathode to see what happens.
Once disconnected, it would also measure if there is any resistance between pin 3 and the heaters line.
I understand that you have tested in this socket different 12Ax7 ...

[Chuck H]

Maybe I'm impatient.?. So I redrew the schematic again. I wasn't so much concerned about what's happening with the input and the bright switch (because that doesn't seem to be where the problem is). I'm trying to get a bead on the circuit as it's wired near the problem circuit. Mostly, what goes into and comes out of the epoxy blob and how you are coupling from V2a to V2b?