Ground loop?

[mauiboy81]

So I just put together and amp with parts I had laying around. It is basically a matchless nighthawk / ceriatone creme brûlée / ac15 type deal. Ef86>12ax7>2xel84. Controls are Volume/ bright switch / 6 way rotary tone switch/ brilliance control (cut control wired in reverse) /lar mar type master volume (dual 250k pot w/ 2.2meg resistors).

I fired it up and it’s quiet and sounds great but there is what sounds like a loud ground loop when the the brilliance control and preamp volume are all the way down. If the master volume is all the way up it goes away but if I back it off at all it comes back. If I turn either the brilliance or preamp volume up at all it goes away. It only makes the noise on those setting.

I did not include the voltage divider between the volume wiper and pi input like the creme brûlée has so the wiper connects directly to the .01 pi input cap.... trying to eliminate the noise I did try adding a 220k resistor in series between the volume wiper and pi input like the trainwreck rocket and the noise is gone completely.

So as it stands it’s quiet but I’m trying to determine why it’s got the noise without the 220k into the pi. Without that resistor and when on the noisy setting if I touch the chassis around the pi tube I get a scratchy static sound. Also if I touch the pi input wire in a certain spot the ground loop noise goes away.

Hopefully this all makes sense and I’m not just rambling. If anyone has any insight please let me know.

Thanks!

[Colossal]

Can you post a high res gut shot of your amp? It would be helpful to see how you laid out your layout and where and how you made your ground connections in order to see if anything stands out.

[Stevem]

It sounds like the amp is somewhat oscillating and your adding that resistor has squelched it.
This is likely due to layout and or build issues.

That rotary tone selection switch is likely acting like a big antenna in a small chassis, so one of the first things I would try would be to bypass that with a length of shielded wire that is grounded at only one end, and that ground should be landed at the up stream tubes cathode resistor ground.

Try playing with that resistor value by lowering it until the issue returns and then go like 10k higher in value.

[mauiboy81]

I tried tying the grounds of each stage together but separate from the other stages and then ran a wire from each stage back to the respective filter cap. The only connection to chassis is by the input. Input and speaker jacks are isolated.

Noise is only in the pi section on those settings in the original post. If the amp has any volume at all the noise is gone.

[Colossal]

Well, that's a good looking build (nice and neat). It looks like all of your grounds tie back to their respective cap and then that ground bus terminates in one place, at the input. I usually build amps just as you have done it with excellent results. To troubleshoot, I would be inclined to separate the preamp ground and the power amp and PI ground. Take the OT, screen, and PI caps and ground them over near the PT (not using the PT bolt though) and then ground the preamp with its B+ cap right at the input, as you already have.

Quick question: Is this joint soldered? Hard to tell....

Pilot.png

[xtian]

The first pot next to the input jack, is that the Volume control? That should be grounded at the input jack.

[mauiboy81]

Thanks! There are a couple things in there that are just tacked in temporarily. The pilot light is alone on a separate 6.3v winding from the heaters. The pt had two 6.3v windings so I used both. I need to remove the light fixture to put a faceplate on so I will clean it up once I get it on there.

Grounds are:
OT CT to first reservoir filter cap
El84 cathodes, OT secondary, and master volume to screen supply filter cap.
PI and pre volume to pi supply cap.
Ef86 and input to the final filter cap.

Edit*
I grounded the volume at the PI because the wiper feeds directly into the phase inverter. Do you think I would be better off grounding it at the input?

[mauiboy81]

So the noise only happens when:
Pre volume is at zero so the pi input is grounded.
Reverse cut control is all the way down so there is no resistance in series with the .0047 cap.
Master volume is not maxed out.

If I change any of the settings from there it goes away completely. So in actual operation it is fine.

[Stevem]

Some VOMs have a frequency / Hz test setting, does yours?
If so test what frequency shows up when the noise level is the highest and report back with it.

[Colossal]

Thanks! There are a couple things in there that are just tacked in temporarily. The pilot light is alone on a separate 6.3v winding from the heaters. The pt had two 6.3v windings so I used both. I need to remove the light fixture to put a faceplate on so I will clean it up once I get it on there.

Grounds are:
OT CT to first reservoir filter cap
El84 cathodes, OT secondary, and master volume to screen supply filter cap.
PI and pre volume to pi supply cap.
Ef86 and input to the final filter cap.

Edit*
I grounded the volume at the PI because the wiper feeds directly into the phase inverter. Do you think I would be better off grounding it at the input?

Yes, I would definitely ground the Volume pot to the preamp ground at the input

[ChopSauce]

So we're back to "fighting" the theory according to which the pot should be grounded next to its related filter cap &/or the cathode ground for the next stage, and everything on the same ground bus/star/whatever - example given: http://www.valvewizard.co.uk/Grounding.pdf -




(just curious, as I've grounded my last build as you recommend)

[pdf64]

I too suspect that the issue may be oscillation, with the hum just being consequential from that. How about if the signal polarity of the output stage is flipped, eg at power valve control grids or anodes?

I think that shorting the balanced signals from an LTPs anodes tends not to be a good idea (ie even via a tone or snubber cap), as rather than signal cancellation, a common mode signal can result.

[mauiboy81]

Success! So this morning I removed the 220k to the pi and moved the volume ground wire to the input stage ground point and now it doesn’t hum in that setting. It’s quiet as a mouse. Thanks for your input!!!

So I looked back at merlins grounding page that was linked and I thought it confirmed my original ground scheme... grounding anything after the coupling cap to the following stages cathode as it is part of that stages grid leak circuit...

Does this not apply when it comes before the phase inverter?

So if it were a tone stack that feeds directly to the pi like in an ac30 top boost type circuit would you ground that back to the cathode followers cathode?

Edit*
On my last build from a while back I grounded the top boost tone stack to the pi. It’s quiet but is it better practice to move the tone stack ground to to cathode follower cathode ground point?

Couple Pics