Power tube cathode bypass cap - what does it do?

[atmars]

I build quite a few cathode biased amps and I was wondering about the bypass cap on the power tubes. What is it supposed to do and what does it actually do in the opinion of the forum members. It doesn't add gain like preamp bypass caps. I have had them on a/b switches from time to time and difference between cap/no cap is quite subtle. Why are these standard issue at 22 (or so) uf? What phenomenon is at work?

[sluckey]

I wasn't aware that 22µF was a standard. Here's some info from the Aiken site...

http://www.aikenamps.com/cathcap.pdf

[atmars]

I guess I didn't mean 22uf was standard, but I see common full frequency bypass cap values in this position ((22uf, 25uf, 50uf) maybe only in the vintage amps I design around?) usually not small (.68uf) or very large (1000uf) although I have heard of large values being used. Any first hand experiences?

Oh, also, I haven't finished the article yet, but I am interested specifically in cathode caps in AB as opposed to class A referred to in the article.

Thanks

[dude]

There was a big discussion over at 18 watt.com (year or so ago) about the size of that cap in parallel with the biasing resistor. I've never seem any sizes lower then 100uf there with usually 500uf as standard in these amps.

I have used up to 1000uf here with good results (some knowledgeable guys there even recommend 2000uf's), el84's in that amp lean on the bright side and a big cap will tighten up the bottom end making the amp sound more like fixed biased. 

al

[HotBluePlates]

100uF and higher (sometimes much higher) is a common tactic in hi-fi amps. The goal there is probably to minimize the effect of shifting bias as a momentary peak pushes a class A or rich class AB amp into a short overload. The idea is that a cap would keep the bias very steady, with a bigger cap doing a better job of that.

I kind of wonder about all that, and should probably test the idea. It seems to me that a momentary big peak would cause the tube to draw more current and the voltage across the cathode resistor to rise. That would charge the cap, keeping the slightly higher bias voltage longer, which slightly reduces the tube sensitivity.

I'm guessing that no cap makes this effect happen near-instantly, and the result is local negative feedback. A small cap reduces the tendancy, but allows the bias voltage to rise and hold longer than no-cap. A huge cap could probably store enough energy to supply the momentary extra peak current, so that bias doesn't change as much as with a small cap. But these are guesses, and I'd want to measure to see if that really happens.

[PRR]

In 1-tube single-ended, the gain about doubles when cap is effective (over 50uFd for guitar for most tubes). This can be a cheap place to shave bass.

In push-pul...... up to a point it does nothing, after that it gets too complicated. You can switch-in a handful of caps and see what it does in YOUR build, far faster than thinking.

[JayB]

In 1-tube single-ended, the gain about doubles when cap is effective (over 50uFd for guitar for most tubes). This can be a cheap place to shave bass.

In push-pul...... up to a point it does nothing, after that it gets too complicated. You can switch-in a handful of caps and see what it does in YOUR build, far faster than thinking.

Or tame an overly bright el84.

[atmars]

In push-pul...... up to a point it does nothing

Any idea why Leo included them (25uf) on his pp tweed amps?

[GroundhogKen]

Any idea why Leo included them (25uf) on his pp tweed amps?

Probably to improve transient response.  I don't think Leo expected guitarists to turn the knobs up to 12 and wail all night long as a steady-state condition.


Ken