Variable Cathode Bypass Capacitor?

[MonkHouse93]

So Has anyone tried something like a variable cathode bypass capacitor before?

Only not with a variable cap, but with a potentiometer. . .



Rk=1500, use something huge like a 1MEG pot in parallel so that the parallel resistance doesn't change much, and then one way the cap fully bypasses the equivalent Rk value and then it can be gradually dialed out until it is shunted to ground.

Is this a silly idea?

I'm putting a 5F10 in a 5E3 cab/chassis and I have an extra pot not doing anything, and I noticed on the 5F10 the first half of the 12AX7 after the tone stack doesn't have the cap and the 5E3 does.

Better ideas?

[Enzo]

Actually it is a very old idea.

Before wiring it up just because you have a spare control, I suggest you tack solder the cap you intend to use across the existing cathode resistor to see if you even like the change. You can also just wire the 1M pot as a variable resistor and put it in series with the cap. The way you have it drawn MIGHT result in a scratchy control.

[gui_tarzan]

I think I understand why it might be scratchy, but why did some amps only have a cap on one side of the tube and not both?

[Enzo]

Same reason some amps don;t have them either place. It served the designer's needs.

What matters is circuits, not parts, whole systems, not little sections. And more gain is not always the proper tool to use. For example, a small amp with a 6" speaker is not helped by a big bass boost. Sometimes they worry about a small or low power speaker, so they don;t push it. Maybe they found the phase inverter would overdrive. Now that I look, the 5F10 uses a 6AT6 input triode, the tube has amplification factor of 70. The input 12AY7 of the 5E3 has amplification factor of 40. SO the added gain of the second stage compensates for lower signal level. But mainly, the 5E3 has no negative feedback, the Harvard does, and what do you know, it feeds back into the stream right where you want to put a cap. That cap would bypass your NFB to ground. And I suspect that is the whole story right there.

[gui_tarzan]

Oh, ok. That makes sense.

[MonkHouse93]

putting the pot in series with the cap will just cause the 1500 to remain the bias resistor value since the pot in series with the cap wont really have any DC current flow through it, but if it were 1MEG in series with the cap, that's not quite the same as totally shorting the cap to ground and just having the two parallel resistor is it? circuit analysis I'm not quite up for atm, but thank you for the idea. I haven't put this whole thing together yet but hopefully I'll get the majority of it tonight

[Enzo]

This will not be a precision lab circuit either way.

I wouldn;t expect a 1M in parallel with the 1500 ohms to make much difference in terms of resistance. In the parallel circuit, dialing the cap to ground just leaves a 1M and a 1.5k in parallel. And for all practical purpose, that remains a 1.5k. SHorting across the cap at the bottom end makes the cap disappear. In either circuit, having a large resistance in series with the cap, either entirely or as part of that parallel scheme, renders it largely ineffective, so the less resistance we get, the more influence the cap has. You might find in my series circuit, you can experiment with lower value pots without upsetting the 1.5k tube bias. You may find 50k is more practical for example. (Ignoring for now that it would kill the NFB).

[MonkHouse93]

I didn't notice it would kill the NFB. That's why I post here, Enzo and a lot of other guys are real smart and point things like that out to an amp noob like me

BUT! using the pot in the first configuration I posted, dialing the cap out of the circuit also puts the NFB back in the circut, so it's not completely killed. . .

So in the case of say the AA764, the NFB loop feeds back to the second preamp stage, but instead of at the cathode, there's 47ohm from ground, then the NFB loop, 1500, and the bypass cap all share a node, and then the other side of the 1500 and cap are at the cathode's potential. So if I wanted to keep the NFB in effect while having the cathode bypass cap in effect as well, I should use something like that configuation? with a relatively small resistor to ground so as to keep that node fairly close to ground potential?

EDIT: that 47 ohm between the NFB, 1500, cap node and ground isolates the feedback to that particular stage? as to where if the NFB were connected directly to ground there, it would become like a global feedback for the entire circuit?

[Enzo]

AA764 what? I see a Bronco with that date, seems close enough.

Well yeah, with the cap dialed out, the NFB is back because with the cap gone you are back to the stock circuit. So I guess you could have a control that dials from cap to NFB. But your premise was you wanted to put the cap on that cathode, and if you do, it kills the NFB.

Remember it is the circuit/system, not the parts. There is going to be a certain voltage on the speaker output, we feed a sample of that back into the amp out of phase. In the 5F10 there is a a 56k feedback resistor to the top of the 1500 ohm. That forms a voltage divider. About 2% feedback. Now the Bronco has a 2700 ohm feedback resistor to the top of that 47 ohm. Again, about 2% feedback. So to pull that off, you;d have to adjust both resistors, not just adding the 47 ohm.

If you connected the NFB to ground anywhere, it would simply be a resistor to ground parallel the speaker output. And that would have no effect on anything.

[MonkHouse93]

Okay I think I just about get it now. . . I didn't necessarily mean adding specifically a 47 ohm, but in order for the bypass cap not to kill NFB, I would have to use a configuration similar to that in the AA764 (was supposed to be a champ, but bronco is the same thing), where the NFB is at the bottom of the 1500 and bypass cap, and then a resistor between there and ground, rather than the cap shorting the NFB to ground?

[Enzo]

sure the Bronco circuit would work, but that is what experiments are for. Maybe you'd prefer the amp without the NFB anyway. Disconnect the NFB line and find out.

And NFB or no, try tacking the cap in just to see.