Calculating cap and resistor size in a cat bias amp?

[Bluefinger]

Hi,

I have found this instruction for a switchable cathode/fixed bias setup http://music-electronics-forum.com/t16121/. That makes sense to me. I know how to bias a fixed bias amp and how to select resistor sizes but what about the cap and resistor in the cathode bias section? In a typical Tweed Pro/Super with 6L6 or 5881 tubes, what values should I use and why?

Thanks a lot in advance!!!!!!

[stokes]

Look at the schem for the Pro or Super,it shows a 250 ohm with a 25uf cap.Start there and then measure the idle current and adjust if needed.I like to use 50v caps for an added safety margin.

[R.G.]

You first split up the question into DC and AC parts.

First DC. The cap doesn't matter for this part. You pick the resistor so that at the DC current you want, the voltage across the resistor elevates the cathodes by a voltage which causes the DC-grounded grids to be just enough negative to cause the current to be the right value to cause the voltage.

Circular and confusing, yes? It is. The DC condition is an example of DC feedback setting its own conditions. There are two ways to figure out where to start. One is by looking at the characteristic curves for the tube you want, the other is to start anywhere and then change things toward what you want.

That implies you first know what you want. And that means you need to already know what current you want the tubes to sit at (that's the end-result you're trying to achieve in the first place). We know this by both characteristic curves and by long experimentation with biasing tubes. Then you need to know approximately what voltage causes that current.

Maybe the simplest thing is to measure the cathode current and bias voltage from an amp that's fixed bias and biased the way you want. When you have those two numbers, the resistor is then R = V/I by ohm's law. Even if it's not perfect, it will be close, and then the magic of feedback starts. The tubes and the feedback will try to move the tubes toward what you tried to get, even for varying tube properties.

The cap? It needs to be big enough that it acts like a short circuit at the lowest frequency you want to use. For guitar in standard tuning, Low E is 82Hz, and for bass in standard tuning, it's 41Hz. Odd tunings go lower. The beginner answer for the cap is to make it equal to the resistor at the lowest frequency, or C = 1/(2*pi*R*F) where F is the lowest frequency. That loses you half the advantage of the cap at that frequency. If you want to lose even less, have the response be substantially flat down there, you need the cap to be about 5x bigger than that to get the AC loss from rising cap impedance down to 1/10 of the resistor value.

[Bluefinger]

Thanks a lot for the replies!

R.G. ... I guess I'll have to reread that a couple of times until I understand everything you said but I love to learn new things.

[R.G.]

I really said the same thing as Stokes did, but used a lot more words.

[stokes]

Yeah R.G.,you answered the "why".I figure Leo and countless engineers since him have already done that,so I just cut to the chase.I've got a 500 ohm 25 watt wirewound pot I use to find the ideal idle when doing a cathode bias amp.The 500 ohm pot covers pretty much all the tube varieties,never did like math.

[Bluefinger]

I assume I should get the same idle readings on the cathode biased amp as on a fixed bias amp ... correct?

I usually like to bias those 5881s at roughly 65%. Would that be fine in the cathode mode as well?

All the old cathode biased dual 6L6 Fender schematics I have found use a 25µF cap and a 250 ohms resistor so I guess that's what I'll start with.

[R.G.]

I assume I should get the same idle readings on the cathode biased amp as on a fixed bias amp ... correct?

Yes and no.

On a fixed bias amp, the full power supply appears across the tubes because the grids are biased from a separate supply. On a cathode biased amp, the cathode bias voltage is subtracted from the power supply. You're getting about 40-50V less B+ for the tubes and OT to work with.

So yes, you can potentially bias it at the same cathode current, and...

I usually like to bias those 5881s at roughly 65%. Would that be fine in the cathode mode as well?

... yes, you still can bias it at 65% (presumably of dissipation maximum) and it will do OK. But this is not the same as operation at the full power supply voltage and output power will have gone down some because of the loss of B+.

Then there's the issue that cathode bias wanders around with power level in a Class AB amp. The whole point of Class AB is to reduce the no-signal power dissipation and allow more of the tube power dissipation limit to be used for putting out signal instead of static power burning. That necessarily means that cathode current is minimum at no signal, and goes up with output power in Class AB.

If you bias with cathode current through a resistor, then in class AB the average cathode bias goes up (that is, the negative bias voltage on the grids gets bigger) with large signals, and this makes for a bias shift toward cold on big signals.

This does not happen in class A biasing. That's why cathode bias is usually a mostly-class A fixture, as in the AC30 (yes, which is not completely class A; same principle applies).

The cathode capacitor is there to keep the instantaneous signal from being modulated by this effect. To stop bias shift from happening to any noticeable degree, the necessary cathode cap gets very big.

[jrfrond]

FWIW- in a cathode-biased amp design I've been building since 2004, which utilizes standard Fender-replacement 50W Bassman transformers, I use 250 ohms/25W (bolted to the chassis) and a 22uF/100V cap. With the close to 500V on the plates, I can freely swap 6550's (which is standard), EL34 and 6L6GC with no modification, and no thermal runaway, which is particularly important with EL34's. This produces 70W, 60W and 50WRMS respectively. BTW- the screen resistors are 1.5K.