Its noisy with or without the source plugged in. I just tried it using a dB meter, and here's the values I noted:
Citation 12 with Polk RT5 monitors, no source : 52dB unweighted at 1 meter, 48dB A-weighted
Citation 12 with Polk RT5 monitors, Carver CT-3 preamp connected by 3 meter AXG Sapphire interconnects, preamp OFF : 48dB unweighted at 1 meter, 39dB A-weighted
The noise is nowhere near as bad as yesterday, when I had the vintage Pioneer CS88 speakers in service. Larger cone area, more sensitive than the Polks I'm sure. Unfortunately I didn't remember to do what you asked about connecting only one channel. I'll have to try that when I get back into the workshop.
Since I've replaced every passive component on the board with the exception of the sand cast resistors I think the only thing that could be causing this buzz would be the transistors, which are basically impossible to find now.
Citation 12 with Polk RT5 monitors, no source : 52dB unweighted at 1 meter, 48dB A-weighted
Citation 12 with Polk RT5 monitors, Carver CT-3 preamp connected by 3 meter AXG Sapphire interconnects, preamp OFF : 48dB unweighted at 1 meter, 39dB A-weighted
The noise is nowhere near as bad as yesterday, when I had the vintage Pioneer CS88 speakers in service. Larger cone area, more sensitive than the Polks I'm sure. Unfortunately I didn't remember to do what you asked about connecting only one channel. I'll have to try that when I get back into the workshop.
Since I've replaced every passive component on the board with the exception of the sand cast resistors I think the only thing that could be causing this buzz would be the transistors, which are basically impossible to find now.
I believe I found the issue.
While I had the amp plugged in I flipped it over and started touching wires to see if there was a loose connection. I found that touching a pair of white wires caused the noise to diminish quite a bit. I traced them out to the top of the chassis and they went to the output transistor heat sink on the far left side of the chassis. I noticed that transistor was not an RCA, it was a Sony.
I decided to switch that transistor with its RCA cousin on the right side outboard heat sink and see if the noise transferred. Thats when I discovered the mica insulator underneath the Sony was basically dry. It looked like a faint smear of something clear was on there but that could have been anything from vegetable oil to horse sweat. It sure wasn't thermal paste.
I have a healthy supply of fresh TO-3 mica insulators from Wakefield-Vette so I smeared up a couple and set them in place. I also made sure to switch the transistor positions just to follow through with my first theory.
End result: noise is greatly diminished. The right channel is basically silent now, and the left channel is approaching bearable. I still have to uproot the inboard transistors and have at them with fresh insulators and paste. When I have all that done I'm going to try the tests you recommended.
Ray, thank you very much for your advice. I ended up ordering the Bourns 3386S-1-502LF pots, which I received and installed yesterday. They slipped right into place and I was able to solder all the leads in without any trouble, btw. They've turned out to be ideal for this application.
While I previously said that I'd verified the passives multiple times, now I have to walk that one back.
After installing the new trimmers in the bias circuit I found the exact same problem as before - the bias values were about 10 times what they should be. Instead of 40mA +/- 10mA, I was seeing values around 400mA or higher. The closest I could get it was in the low 100mA area. Something else had to be going on here.
I tested the values of every single resistor on the card, and I found one resistor, R721, that was off. It was supposed to be 3.3 ohm, instead it was 21ohm. I couldn't believe I'd gotten it mixed up with another resistor, but then again all Dales look alike until you get them on a DMM.
Since I usually order multiples in each value on a project, I looked through my remaining stock of resistors from this amp and couldn't find any more 3.3 ohms parts. I then started thinking "wait, where did I even get a 21 ohm Dale? I didn't order any for this job. I've never used a 21 ohm value in any project, ever, Dale or not".
Then I realized what happened. Since Dales are ridiculously expensive in some values, I only order two 3.3 ohm parts. And since they weren't on a tape that means they were hand-selected out of bulk and the picker got it wrong. So somewhere out there there's a guy who ordered 21ohm resistors who just put a 3.3 ohm part in his project. Hopefully that's a non-destructive error just like this one was.
All those diodes are standard Silicon type diodes, nothing special.
CR701 and 702 are way over sized at 2A, a 1A diode would work fine.
I'd use a 1N4004, yes higher voltage rating but why not.
CR5 and CR6 are mounted off board, to the heat sink I take it?
Are they in a metal can? I've not seen how these are mounted but the metal
can provides better thermal conductivity to the heat sink/output devices. They
are forward biased so their reverse rating doesn't even matter. I would not
replace them. But if you are determined to do so, the old single diode metal
can device is the 1N3754 which is not
made anymore: https://www.radiomuseum.org/tubes/tube_1n3754.html
You'd have to use 2 and figure out how to mechanically mount them.
The modern way to do this is to use a (pair) medium power device like a BD139 or
anything similar with some exposed metal, but in these devices the metal is
connected to the collector of the device so you'd have to insulate them. The
newer devices of this type employ a very thin layer of the plastic package as
an insulator - it has very poor thermal conductivity. One way to compensate is
if the original diode was thermally coupled to the heat sink, mount the newer
device with the plastic side in thermal contact with the case of the output
power devices, taking one insulator out of the path. Again, I would not change
them unless you can prove that they are bad.
CR1-4 protect the output transistors from back EMF from an inductive speaker
load, also Si type and not critical. But 3A would be better there, 1N5404 is fine.
CR701 and 702 are way over sized at 2A, a 1A diode would work fine.
I'd use a 1N4004, yes higher voltage rating but why not.
CR5 and CR6 are mounted off board, to the heat sink I take it?
Are they in a metal can? I've not seen how these are mounted but the metal
can provides better thermal conductivity to the heat sink/output devices. They
are forward biased so their reverse rating doesn't even matter. I would not
replace them. But if you are determined to do so, the old single diode metal
can device is the 1N3754 which is not
made anymore: https://www.radiomuseum.org/tubes/tube_1n3754.html
You'd have to use 2 and figure out how to mechanically mount them.
The modern way to do this is to use a (pair) medium power device like a BD139 or
anything similar with some exposed metal, but in these devices the metal is
connected to the collector of the device so you'd have to insulate them. The
newer devices of this type employ a very thin layer of the plastic package as
an insulator - it has very poor thermal conductivity. One way to compensate is
if the original diode was thermally coupled to the heat sink, mount the newer
device with the plastic side in thermal contact with the case of the output
power devices, taking one insulator out of the path. Again, I would not change
them unless you can prove that they are bad.
CR1-4 protect the output transistors from back EMF from an inductive speaker
load, also Si type and not critical. But 3A would be better there, 1N5404 is fine.
The OP direct messaged me about intermittent noise and this was my response in case
it might help anyone else:
connect some speakers that you don't care about
turn the volume down on the preamp, and tap parts that you have not replaced
with the plastic end of a BIC pen or any other plastic probe. This is a good way
to find cold solder joints and intermittent components.
I built my Citation 12 from scratch so I've only read about this but there is talk
of how the plug in boards do not always make good contact so certainly clean
them and check connectors for tension. This is probably the very first thing to
check.
it might help anyone else:
connect some speakers that you don't care about
turn the volume down on the preamp, and tap parts that you have not replaced
with the plastic end of a BIC pen or any other plastic probe. This is a good way
to find cold solder joints and intermittent components.
I built my Citation 12 from scratch so I've only read about this but there is talk
of how the plug in boards do not always make good contact so certainly clean
them and check connectors for tension. This is probably the very first thing to
check.
All those diodes are standard Silicon type diodes, nothing special.
CR701 and 702 are way over sized at 2A, a 1A diode would work fine.
I'd use a 1N4004, yes higher voltage rating but why not.
CR5 and CR6 are mounted off board, to the heat sink I take it?
Are they in a metal can?
After tracing the circuit out I finally figured out that CR5/CR6 are indeed mounted on the heat sink as you suggested. They're glass package DO-41 or smaller. That picture is from another amp somewhere else on the planet and it has the same diode as the one sitting on my desk right now, so I'd say this is a stock diode. Its certainly not a metal can and I don't see why this was represented as a double diode in the schematic. That thing looks weak compared to the other diodes.
I gotta figure this thing out. The left channel has disappeared into a bad hum, with intermittent crackling like the channel is trying to come back. I took a fat highlight marker and used that to prod and tap every part on the board and the chassis, and nothing changes. Since I'm getting a hum I'm going to dismount the outputs on that side and repack them with more thermal paste, maybe all the parts switching thinned out the paste and its finding a way to make a connection.
I've applied the thermal paste. Still nothing. I replaced the left channel differential pair with fresh 2N5087 and now the thing works and the channel is quiet. I've also replaced the 2N5232 at Q708. I can't get the bias adjusted down to the recommended 40mA. I had it there before this all started. The lowest value I can get is around .170mA. I replaced the bias adjustment pot with a new 5K Bourns, and there's no difference.
Even better, now I can't get any bias value at all to show up on the right channel. That thing is reading 0.00 no matter where I turn the pot. But both channels are working just fine. I'm starting to think its time to get a new Fluke. WTH?
Even better, now I can't get any bias value at all to show up on the right channel. That thing is reading 0.00 no matter where I turn the pot. But both channels are working just fine. I'm starting to think its time to get a new Fluke. WTH?