Buzz that goes away when input grounded.

[wicked1]

I just hooked my new amps up to some efficient speakers and one channel buzzes. (didn't notice it on less efficient speakers)
It goes away when I ground the input. it also goes away if I squeeze or hold the input cap!

But, I can't find a problem.. I've checked the solder joints. I've unscrewed the rca jack from the panel, so I'm sure it's not creating a ground loop w/ the chassis.

Any ideas of where I should look next, or what causes this sort of issue?
(basically, I'm looking around the input for issues, and have resoldered all the joints. My limited experience tells me the issue is likely around there, because moving the input and touching the cap changes or stops the buzz... but if I should be looking elsewhere for the cause, let me know. Also, these are two monoblocks, constructed exactly the same, and only one buzzes, which makes this difficult to track down for me.. I know I constructed them well enough, since one is ok)

thank you, as always!!

 

[peufeu]

Please show a picture of the internals of both amps.

 

[wicked1]

Ok, here they are. The one on the left buzzes.
And I guess I lied.. They aren't identical, they are mirror images, which did change my heater wire routing.

But, I don't think it's the heater I'm hearing. And now, on my workbench, I have moved the heater wiring (the black and blue twisted) so it is looped well above the PCB, away from everything else. And the inputs are loose, and far away from the heater wires. Moving all the wires and resoldering everything hasn't changed the buzz much.

I also made another mistake in my last post, I dont have input caps.. It's the coupling cap I can squeeze and it stops the buzz! So, grounding the input or squeezing the first stage coupling cap will stop the buzz. (In the picture the coupling caps are the black caps towards the top of the PCB). I'll try replacing the cap, and check the tube socket for corrosion or any other issues.

 

[mr ed]

Check your interconnects for an open ground.

 

[wa2ise]

It's the coupling cap I can squeeze and it stops the buzz! So, grounding the input or squeezing the first stage coupling cap will stop the buzz. (In the picture the coupling caps are the black caps towards the top of the PCB).

The cap may be acting like an antenna and picking up some buzz that way. Maybe if you wrap some aluminum or copper foil around the round part of the cap, and have a wire in contact with the foil, and wrap some tape over it and connect the wire to ground may intercept the buzz and drain it to ground, like a shield?

Maybe lining the wood near the board with foil and grounding that foil may also help. And a bottom metal plate may also help, and also keep wayward fingers out of high voltage if someone picks up the amp to move it.

 

[Merlinb]

I just hooked my new amps up to some efficient speakers and one channel buzzes. (didn't notice it on less efficient speakers)
It goes away when I ground the input. it also goes away if I squeeze or hold the input cap!

Normally I would suggest it is buzz being picked up from the heater via grid-heater capacitance. But since just touching the input cap reduces it, it might be worth reversing the input cap so its output foil matches the quiet channel (presumably).

 

[wicked1]

OK, thanks everyone.
It's basically what Merlinb says.. I had the capacitor labels matching direction, but noticed one end of the cap had a slight convex shape, and that was reversed on the quiet side. I flipped the coupling cap and it's quiet!
I'm surprised!! But, it made a dramatic difference.

(All the other amps I've built are in the single digit watts, and this one is about 80 watts. I suppose other amps could have this problem and I don't even know, because it's not amplified enough to hear. I found a problem in my preamp, too, that I never noticed w/ the lower powered amps.)

 

[DF96]

Strange noises which appear/disappear when a component is touched can be a symptom of RF parasitic oscillation. As a further check, monitor DC conditions too. An oscillating stage is very likely to shift its operating point.