amp sounds fine with signal generator but not guitar

[thetragichero]

I've got an ampeg j-12-r (https://el34world.com/charts/Schematics ... _j12r_.pdf) I'm looking at for a friend and I'm scratching my head. with guitar it's making almost a bitcrusher sound and trailing out quietly. if i wait a few seconds and hit the strings hard again the volume will increase but still trail out crackly.

when i hook my signal generator with a sine wave output is clean and constant

before it came to me he recapped all electro caps and signal caps and replaced the power resistors (there's some lightening on the board where the dropping resistor between A and B nodes was so he replaced the resistor and it's now raised above the board for airflow). he did not replace the caps on the volume and tone controls. i thought the cap to ground on the time control might be contributing so i disconnected it from ground

all tubes test strong. i added grid resistors right at the tube sockets at points K and S because I've had similar problems fixed by that

I'm scratching my head here and any help would be greatly appreciated. i will admit I'm not familiar with the phase inverter on this

[xtian]

Measure resistance from all grids to ground, make sure you get 1M or lower resistance. Sounds to me like a broken grid-to-ground reference, which can choke and cut off the signal.

[stephen_w_keller]

I spent a fair bit of time chasing a "problem" in an amp that eventually turned out to be a bad pickup in my guitar. It was one of those "doh!" moments.

[Stevem]

Since you have a signal gen you need to go thru a logical process of elemination starting at the output stage and then working back up to the first gain stage after the input jack.

Put a .02 uf cap at 600 volts on the hot of your gen and start with hooking it up to the input of the P ( V3 pin 7 )I and then work on back.
This blocking cap will allow to inject signal on the tube plates also besides protecting you gen from a slip up in touching circuit sections.

Also do you have a resistive load and a scope to look at what the output wave looks like and also how much output voltage the amp is producing?
A signal gen set at 1k or so will not drive the amp like even a first position E chord will.

I would set my gen for 160 hz and then see / hear what the results are that take place.

It's not uncommon for these amps output transformers to go open on one end, as I have seen that many many times, so make sure your seeing good plate voltage on each output tube.

Maybe this was the reason that guy started to rebuild the amp in the first place?

[thetragichero]

just to wrap things up: grids-to-ground were fine, output transformer tested fine

since grids were fine i started checking cathodes... and that's when i found that the 150k cathode resistor on the phase inverter measured open!

works now, only issue is that the reverb won't turn all the way off. but that's a luxury problem!

edit: why is best practice to not use the pot casing as a grounding point? because you're relying on a mechanical connection instead of an electrical connection. jumper from the pot to a relevant grounding point and it works as it should

there's still quite a bit of hum/buzz but that's to be expected with a flat chassis and suboptimal grounding scheme. i don't think he wants to invest what it would take to try to get rid of it (and I'm skeptical even reworking the ground bus would eliminate it)

[R.G.]

edit: why is best practice to not use the pot casing as a grounding point? because you're relying on a mechanical connection instead of an electrical connection. jumper from the pot to a relevant grounding point and it works as it should

There are several different kinds of things we call "ground". They're all generally connected to a reference point (the "One True Ground" or OTG) but the currents through them and the stability/resistance of their connections to the OTG vary. In general, you want the metal enclosure to be a Faraday Cage, an all-enclosing metal shield layer to stop radiated junk. Using the chassis as a current carrier mixes up the shield nature of the chassis with returning currents to the OTG. It works mostly, enough to keep people doing it, but it also can result in add feedback and hum problems.

The casing of a pot is an enclosing metal shield to keep external junk (electrical AND mechanical) out of the insides of the pot. If you ground signal to the pot casing, you are relying on the pot casing's connection to the (presumably... ) grounded chassis, and you inherit all the problems with using the chassis as a ground conductor. The best practice is to use shields as shields and not current returns, so the pot case is a not-so-good place to ground signal.

There is a large amount of explanation about why chassis grounding isn't optimal, and how to think about ground conductors, as well as the fundamental rule that ground is a single point, not a conductor or network, in Guitar Amp Wiring Notes:
https://www.dropbox.com/scl/fi/vpijj3al ... haf4k&dl=0