Which direction should banded (foil case) side of signal caps go in an amp circuit?

[mikepukmel]

It is not a *potential* issue at all, but an *impedance* one.
The outer foil is a shield by definition, a metallic foil wrapping the rest of the capacitor, and as we already know, shields to do their magic, must be grounded.
Ever saw a guitar cable with center conductor grounded and shield carrying signal?

If not actually grounded, best (or least bad) is "the point with lowest impedance to ground"
Imagine a 100k resistor tied to a plate, and a 1M resistor tied to next grid ... which one has lowest impedance to ground?

That said, you are shielding just a part of a path which may be much longer.
If plate gos to a grid 1 inch away, fine; if signal travels through 4 inches of wire or more, wonīt do much.

As of paint influencing thermal behaviour of a capacitor, IF the capacitor dissipates power (what a resistor does) the thin layer of paint will increase thermal resistance and resistor will work (a little) hotter, but capacitors do not dissipate significant power, so .......
Capacitors take whatever ambient temperature surrounding them, whatever it is; a paint layer in theory will make that process infinitesimally slower .... unmeasurable.

Even if the capacitor is wrapped in styrofoam, it will work at ambient temperature, it will just take a little longer to reach it (suppose you just turned your amp on).

That's where I get lost (the impedance to ground arg). OK one of the many places I get lost These caps do some really cool stuff: one side has AC and high voltage DC, the other side, hopefully just the audio signal. So, impedance implies we're talking about AC. So, there's a shield tied to one end of the cap, to one lead. Suppose its the cap right after V1. On the V1 side, there are two paths to ground, through the 100k resistor, but that leads back to the dropping resistors and caps., and the other path is through the plate, down through the tube through the cathode and its resistor to ground. The other side of the cap goes through a whole bunch of stuff before the signal gets to ground. I can't quite get why these 3 paths are different enough to make a difference in the audio signal if you flip the cap around.

[g1]

The supply is basically AC ground. Keep that in mind. That's why you don't see the AC signal on the supply rail when you scope it. If you do, one of the filter (decoupling) caps is probably bad.

[Dave H]

Suppose its the cap right after V1. On the V1 side, there are two paths to ground, through the 100k resistor, but that leads back to the dropping resistors and caps., and the other path is through the plate, down through the tube through the cathode and its resistor to ground. The other side of the cap goes through a whole bunch of stuff before the signal gets to ground. I can't quite get why these 3 paths are different enough to make a difference in the audio signal if you flip the cap around.

Have a look at the Tone Stack Calculator circuit for the Fender plate driven tone stack. It has Zsrc i.e. the impedance looking into the plate as 38k. If it's just driving a 1M vol pot (say) the impedance to ground looking the other way is the 1M.

[mikepukmel]

Have a look at the Tone Stack Calculator circuit for the Fender plate driven tone stack. It has Zsrc i.e. the impedance looking into the plate as 38k. If it's just driving a 1M vol pot (say) the impedance to ground looking the other way is the 1M.

OK thanks, then that is why a few amp builders flip these caps around depending on the position in the circuit.

[J M Fahey]

Yes, the supply rail *is* Audio ground.
You have a big honkinī10 to 47 uF capacitor to ground there.

Forget supply rail resistors, the local capacitor is already a short to ground at all Audio frequencies so plate load does not even "see" them.

If it werenīt so, signal from one stage would reach the earlier one and oscillate or at least cause instability ... exactly the symptom when one such decoupling cap opens, or dries up and has high ESR.

Or has severely reduced capacitance, so it still does its job at mid-high frequencies (sort of) but not at lower ones ... thatīs how motorboating appears.

As of where the "38k" used in TSC comes from, itīs the result of 100k load resistor to rail "Audio ground" in parallel with triodeīs own "internal impedance", again to its own cathode which is grounded or very near ground, and which is usually between 60 and 70k.

Now if it were a pentode stage driving that Tone Stack, "generator" (the tube driving the TS) impedance would be straight the load resistor (say, 100k) because a pentode internal impedance is very high, usually >500k so it gets out of the picture for all practical purposes.

TSC, as well as most simulators, call all audio sources "generators" with a given voltage output and an equivalent series resistor to study the tone stack all by itself, with no need to argue about tubes/transistors/Op Amps/pickups or whatever is driving it.

[mikepukmel]

Thanks Juan, Im able to understand most of what you said pretty well. Some of this stuff is sinking in!! So, re the original question, its complex enough, inside the amp, capacitor to capacitor, that if you had them all lined up "the right way", whatever that is, listened to the amp, then pulled them all out and put them back in randomly, you probably would not hear a difference. Or might, if you knew what to look for?

[J M Fahey]

You will not hear a difference in "sound" , they work the same both ways, and if the Audio path is short , say capacitor goes from pin 1 (plate 1) in a 12AX7 to pin 7 (grid 2), you donīt really have much of an antenna there anyway.

Now in the other extreme, when going to a relatively distant panel control / Jack or simply from somewhat separated Pre to Power, or if too near to another high gain stage, it will help.

Not a drastic difference but every bit of hum you can save here and there makes the grand total less annoying.

[bob p]

That looks like the coolest places in the whole world. I can imagine what my wife would say if she saw that lab!

"I can't stand to live in this house -- I want a divorce."

Dare I say it, looking at that picture of his Man Cave, I'm inclined to think that Mr. C is a bachelor.

Here's a walk through tour of his lab. He's either a bachelor or he's got a very understanding wife. His home lab has got to be one of the greatest estrogen-free zones on the planet.

As great as his lab is, I think he needs a bigger workbench. His workplace is way too small.

[DrGonz78]

As great as his lab is, I think he needs a bigger workbench. His workplace is way too small.

I think he thought the same thing... lol!!

[bob p]

Well that removes any doubt from my mind. He's a bachelor.

He's opened up a second lab that's packed to the gills with old gear, including old broken test equipment that needs to be fixed so he can use it to fix old broken gear... and he says that the second lab doesn't even show all the stuff that he has that needs to be fixed... which means that there's yet another room full of stuff in his house... I don't think any wife would but up with that, I think he's a bachelor.

BTW, which piece is the receiver from a B-29? When he mentioned that, he pointed at a Heathkit O-10 oscilloscope. I'm thinking the B-29 receiver was sitting it it's right.