Swirling Fizzy Tone on Decaying Notes

[syscokid]

I'm still messing with my 6G3 clone project that I started a few months ago. I’m just now sure if the issue is normal for this type of amp, or there’s something mischievous going on. The slow swirling fizzy tone is happening on both channels and when the amp is into overdrive territory.

I built this 6G3 clone without the tremolo circuit. Therefore, only half of V2’s 12AX7 is used. I also split the V1 cathode, and gave the Normal channel a Plexier flavor.

Test sample…
LP, neck humbucker, cable, Bright 1 channel, Volume and Tone at 3:00 o’clock:

My 6G3’s Fuzzy Tone.m4a

[syscokid]

Fender_deluxe_6g3.pdf

[xtian]

Lead dress, parasitic oscillation. Photos?

[syscokid]

… Photos?

IMG_2147.jpeg

IMG_2148.jpeg

IMG_2149.jpeg

IMG_2150.jpeg

IMG_2151.jpeg

IMG_2152.jpeg

What I tried so far:
Swapped out all six tubes.
Swapped all filter caps including bias filter cap.
Checked and measured all resistors and all other caps for proper values.
Have experimented with two different output trannies: 1) Modulus Amps Tweed Deluxe OT 7.2K primary. 2) Pacific Audio Magnetics Fender Brownface Deluxe OT 8K primary.

Other component facts:
JJ6V6S power tubes biased at 60%.
Modulus Amps Deluxe PT produces 355-0-355 vac (loaded) at 120 vdc. B+ was originally 437v, but recently I installed a zener diode with MOSFET that dropped the B+ to 380v. Power tube plates are at 372v.
I’m also using a Staco variac while testing the amp.

[syscokid]

Almost forgot…
The OT and speaker jacks are grounded at the speaker jack’s mounting holes. The rest of the circuit is grounded at one point to the chassis at the input jacks.

[nuke]

Since you omitted the tremolo, might try connecting the unused plate/grid/cathode pins to ground.

[pdf64]

Have you verified that the feedback loop is negative? eg opening the loop increases gain.
Granted it would be unlucky to get it wrong with 2 different OTs.
It's probably best to leave the heater of the unused triode disconnected.

[jabguit]

different speaker?

[Stevem]

Several things could be making for your issue.

First off your Daisy chaining of all the power supply filters to ground is an open invitation to interaction between gain stages.

Your first two filter nodes ( output plate and screen ) due to there high current draw should be split off from that chain you have and be brought over to a separate ground stud / bolt near the PT .

The Transformers V+ ac center tap should go right to the negative of the first filter ( output tube plate ) and then to ground .

As mentioned your issue sounds like an oscillating condition going on that only can be heard when a note being played tapers off.

Long grid wires are an invite to this condition.
Especially the V8 grid wires with large signals on them like from the PI outputs to the grids of the output tubes .

Because if this possible condition I would try moving your PI output coupling caps right to the output tube sockets.

The best way to wire up any amp is to have your grid wires as short as possible.

So it’s best to make your plate wires longer which will allow your grid wires to be shorter.

Also as posted it could be something as simple as your speaker making this issue which is something I have heard many times .

Take your headphones and plug them into the amp output if you do not have another speaker to audition the amp thru, do you still hear the problem taking place then?

Another simple thing to try is to open up the circuit and eliminate / bypass most of it.
Take the output of your first gain stage and run it right to the PI input grid after you have disconnected its normal input.

Do you still hear the issue going on then?

[syscokid]

different speaker?

Yes! I forgot to add this to the list of things I tried.

Since you omitted the tremolo, might try connecting the unused plate/grid/cathode pins to ground.

This should be easy to try, thanks.

Have you verified that the feedback loop is negative? eg opening the loop increases gain.
Granted it would be unlucky to get it wrong with 2 different OTs.
It's probably best to leave the heater of the unused triode disconnected.

Feedback loop: I haven’t verified, but I assumed it was OK because the amp is not producing that crazy squeal when the primaries are backwards. I will give it a quick try though.
Heater of the unused triode disconnected: That would be pin 5 of V2?

[R.G.]

+1 for Steve. Yeah, parasitic oscillation is the best guess based on the description.

As Stevem says...

First off your Daisy chaining of all the power supply filters to ground is an open invitation to interaction between gain stages.

Your first two filter nodes ( output plate and screen ) due to there high current draw should be split off from that chain you have and be brought over to a separate ground stud / bolt near the PT .

The Transformers V+ ac center tap should go right to the negative of the first filter ( output tube plate ) and then to ground .

+1 for this. PT CT goes only to the negative of the first filter cap - the One True Ground. For the highest hum/oscillation immunity, the cathodes of the power tubes go here on their own, separate ground wire(s), and so does the bypass filter cap for the screens (which are in fact, a low-gain signal input as well as the normal grid)

As mentioned your issue sounds like an oscillating condition going on that only can be heard when a note being played tapers off.

This kind of thing is very common on solid state stuff with oscillation/wiring issues. Very often it's seen as "bottom side fuzzies", a buzz of RF only on the negative going output signal, a result of the often lower frequency response of PNP output devices. The oscillation can only sustain itself for a very small region of signal levels.

Almost forgot…
The OT and speaker jacks are grounded at the speaker jack’s mounting holes. The rest of the circuit is grounded at one point to the chassis at the input jacks.

There is a dark-horse issue lurking here. If the OT and speaker jacks contact the chassis at one place, and the rest of the circuit, including the PI, is grounded at the input jacks, the chassis is conducting any signal current for feedback from the output to the PI. Any ground currents on the existing ground buss get mixed into the feedback to the PI, and also to the input stage. This is unusual, but does happen - I have these scars...

[syscokid]

First off your Daisy chaining of all the power supply filters to ground is an open invitation to interaction between gain stages.

Your first two filter nodes ( output plate and screen ) due to there high current draw should be split off from that chain you have and be brought over to a separate ground stud / bolt near the PT .

The Transformers V+ ac center tap should go right to the negative of the first filter ( output tube plate ) and then to ground .

As mentioned your issue sounds like an oscillating condition going on that only can be heard when a note being played tapers off.

Long grid wires are an invite to this condition.
Especially the V8 grid wires with large signals on them like from the PI outputs to the grids of the output tubes .

Because if this possible condition I would try moving your PI output coupling caps right to the output tube sockets.

The best way to wire up any amp is to have your grid wires as short as possible.

So it’s best to make your plate wires longer which will allow your grid wires to be shorter.

Daisy chaining of all the power supply filters to ground is an open invitation to interaction between gain stages: Ah yes… I’ve been warned about the grounding layout I’ve been following, from another member at another forum months ago when discussing the possibility of this project. It will also be easy to try your suggestion, but first I’ll give the previous suggestions a try.

As mentioned your issue sounds like an oscillating condition going on that only can be heard when a note being played tapers off: Exactly! Which is why sometimes I think I’m overly obsessing over what might be this circuits natural traits. When I compare this amp with other internet video demos of true 6G3’s, my clone sounds very similar. But I’ve yet to find a demo where the guitar player is using a guitar though a neck humbucker into a very overdriven 6G3 and letting those fat bassy notes ring out.

Long grid wires are an invite to this condition…: True, but I didn’t think that these grid wires were overly long. On the other hand, I don’t remember twisting these wires either.

The first version of this build was set up where the power tubes were cathode biased. Then after some contemplation, I decided to return the circuit to its glory with a fixed bias, but with an added trim pot for adjustability. That’s why you see that “afterthought” bias circuit board over the main board in the pics.

I seriously need to have a new circuit board created and make it all look pretty and functional too!

[syscokid]

Almost forgot…
The OT and speaker jacks are grounded at the speaker jack’s mounting holes. The rest of the circuit is grounded at one point to the chassis at the input jacks.

There is a dark-horse issue lurking here. If the OT and speaker jacks contact the chassis at one place, and the rest of the circuit, including the PI, is grounded at the input jacks, the chassis is conducting any signal current for feedback from the output to the PI. Any ground currents on the existing ground buss get mixed into the feedback to the PI, and also to the input stage. This is unusual, but does happen - I have these scars...

Isolate the speaker jacks from the chassis and run a ground to the PI filtering?

[R.G.]

Isolate the speaker jacks from the chassis and run a ground to the PI filtering?

That is one way. We know that output jacks to chassis and chassis ground to PI isn't 100% an oscillation creator, because many amps do this, and get away with it. But an amp that is manifesting with demons and gremlins might have this issue. This is discussed on pages 20, 21, and 22 of Guitar Amp Wiring Notes, the first link below. I don't know this is the case here, because you don't have any obvious high currents through the chassis, but best practice is always to have one and only one wire connecting the chassis to the negative of the first filter cap, or to the input jack only, and nowhere else.

I actually think that the grounding bus for filter caps and maybe the ground path for the output cathodes are more likely; so are the long grid wires Steve brought up. Grid wires are the most sensitive things in the amp, and they are also prone to capacitive pickup. Could be either ... or both, conspiring against you. It would be easy for you to exclude the output jack wiring by temporarily removing it from the chassis and running an explicit signal grounding wire back to the "ground" you have closest to the PI stage. If nothing changes, you can pretty much exclude the output jack grounding and put it back on the chassis.

Hiding down at the bottom of a lot of amp troubles is the fact that there can only be one >>point<< that is truly ground, the point of absolutely zero volts for the whole unit. Wires, the chassis, bus bars, etc. are all low value resistors, so any current through them makes the "ground" at the far end of the wire, etc. be not the same voltage as the One True Ground by V= I*R. We get into trouble when we think that any old wire or plane connected to the single-point ground is OK for a reference for a high gain input, and there are signal and hum currents running through the same wires.

Here's the full version of Guitar Amp Wiring Notes:
https://www.dropbox.com/scl/fi/vpijj3al ... 9kpkh&dl=0

And here's the TLDR version:
https://www.dropbox.com/scl/fi/soyg9csg ... wnv38&dl=0

[syscokid]

Could shielded grid wires for the output tubes counteract any possible oscillation?

Yesterday I tried some of the suggestions posted on this thread, but the results are still the same:
Grounded the unused pins of V2.
Disconnected heater at pin 5 of V2.
Verified feedback loop is negative.
Separated speaker jacks from chassis and grounded them to the PI ground node.
Grouped and grounded the HT CT, heater HT, main and screen filter, power tube cathodes, and bias circuit, to one of the PT bolts. Grouped and grounded the PI circuit at a separate grounding point.
Replaced the original PI to power tube grid wires with a pair of equal length and twisted set of wires.

Later today I’ll try Stevem’s suggestion of:
Another simple thing to try is to open up the circuit and eliminate / bypass most of it.
Take the output of your first gain stage and run it right to the PI input grid after you have disconnected its normal input.