Shared vs. individual cathode bias resistors

[Gaz]

What are the pro/cons of having a shared vs. individual cathode bias resistor in a push-pull out put section?

I understand that if a tube fails, just that one can be changed, but are there any other advantages? Are there any tonal differences?

[statorvane]

Supposedly there are tonal differences. I read where Dennis Cornell's small 20 watt Plexi Jrs. (Clones of Marshall's 20 watt Lead and Bass amps) used separate cathode bias on the EL84s to get a bigger sound - whatever that means. I suppose it sounded more like a bigger Marshall, than Marshall's 20 watt amps.

I don't know of any electrical pros/cons. Possibly a little bigger layout footprint and slightly increased parts count with the separate bias resistors/caps. Cornell managed to shoehorn that into a similar sized chassis - probably not worth pointing out.

[Alexo]

One pro is that you don't need a matched set of power tubes to avoid one of them overheating/redplating - although a matched set is still desirable if you want a 'properly' balanced signal.

In theory, I don't think there is any tonal difference. In reality - maybe - but until we see a double-blind trial with control groups and a large sample size, it's more or less conjecture.

[Milkmansound]

think about it this way - you get one cathode bypass cap per tube if you split them up. That may make the amp sound "bigger" than 2 tubes sharing one resistor and bypass cap.

interesting idea. I might have to try it for myself!

[Colossal]

I have often wondered by KF built the Rocket with all four output tubes running off one cathode resistor and bypass cap but the Liverpool output section is divided into two tubes each with their own cathode R and C. I built my test Liverpool with the single cathode R/C and my ear is tuned for its sound (which is killer). I will split them into separate Rs and Cs and report back (not sure if I have the right values on hand though so will have to look). The amp already sounds huge (it's actually biased for about 45W of output) so it will be interesting to see if there is an audible difference with the split cathode bias.

[katopan]

A lot of people use a large shared cathode cap (~1000uF) to get more of a "fixed bias" sort of feel. But even with the big cap, under sustained signal the cathode voltage still comes up a fair way and cools off the bias under load. Seperate resistors and two large caps give a feel even closer to fixed bias. Each side gets a bit less than half the signal time (when the valve is cutoff) to recover and you'll find that the cathode voltage won't creep up anywhere near like for a shared arrangement.

Of course there's other options for more of a fixed bias feel like using a zener and smaller resistor in series, or clamping the amount the normal cathode resistor can rise in voltage with a zener across it in parallel. But yes, a split setup will give you a different feel.

[pdf64]

katopan, sorry to disagree but my findings are at odds with most everything in your post.
I've put individual 270 ohm cathode resistors for each el84 on my ac30, each bypassed with 220uF.
The cathodes idle at ~11V (ie +/-0.2V), at the max clean sine they rise to ~12.5V, at max square ~18V.
If I strap them all together the results are just the same; there's none of the 'fixed bias' effect you mention, when comparing 'shared' to 'individual'.
Tonally and feel-wise I can't tell the difference between 'individual' and 'shared' (strapped together).

Regards the cathode bypass cap value, the 220uF caps charge and discharge really quickly, there's no visible ramp up / down on cathode voltage between idle, max square and back to idle (11 - 18 - 11V).
A big cathode bypass cap (equiv to 1000uF in my case) does not hold / fix the cathode voltage, the cathode voltage still 'squishes' (rises and falls in the same way, as evidenced above) in the same way.
Maybe some fancy test gear could show a few milli-seconds more rise / fall time with 1000uF than 22uF but I haven't found anything of significance using regular ears / scope / DMM.
Regards the actual value / frequency break point of cathode resistor / bypass cap, be aware that the pre amp cathode equations don't work as expected on push-pull power amps with shared cathode resistor, because the inverse signals from the phase splitter appear at the cathodes and mostly cancel out, so the cathode bypass cap doesn't do much. Except that as the signal level increases and the tubes cut off for some of the wave, then the cancellation stops and the cathode bypass cap comes into play.
So listening tests at high signal levels are the key thing, there's no simple frequency breakpoint.
With individual cathode resistor / bypass cap, it's a lot simpler.
Pete.

[Gaz]

Really interesting, thanks!

[overtone]

Thanks for this post Pete.

edited: the rest was veering off topic about individual Cks rather than Rks.

Best, tony

[9pins]

There is a real issue of degeneration and how to use it.
A shared bias resistor reduces harmonic derivatives in the push/pull circuit.
You might just be hearing an increase in odd order distortion vs even order, or simply less feedback.
Might be useful when you implement global feedback.
should be able to measure the difference with a power test over a dummy load

[katopan]

No worries Pete. It's not the first time (the testing waveforms and suggestions about the historically 'hot' topic of the Express 3rd stage I posted went no where fast).

You won't see much difference with a test sine wave because what ever cap arrangement just charges up under this fairly static test load. But I measured a difference while testing with a real guitar signal input. It's the way the voltage moves dynamically that changes with a seperate cathode resistor and cap. You won't see it with a sine wave. Also the difference in feel with a 'big' cathode cap is well discussed in many other forums (AX84, 18Watt, Wattkins).

Anyway, this was in one particular amp and I haven't gone back to try it out in my other amps or really test out how much difference it makes under different scenerios. If I was going for more of a fixed bias feel in a cathode biased amp I would use partial zener/resistor biasing or zener clamping which are both much more effective. But I still think there's a feel difference in seperate cathode R/C's.

Totally agree with your statement about the freq equations not holding true for shared cathode cap.

Cheers, Craig.

[joCCe]

I'm thinking about splitting the cathodes on my rocket build. Would two 120 ohms 10W be enough? I plan to use the 260-0-260 taps on the pt.

Thanks
Joakim

[joCCe]

Did some more searching and a very useful thread from 2009 popped up.

http://ampgarage.com/forum/viewtopic.ph ... sc&start=0

Seems like 5W is more than enough. I think I have room for a couple of 7W - 10W, so I'll grab the largest one I can fit in there to keep them cooler.

[Phazor]

What about multi cathode resistors so if you pull 2 of 4 output tubes,(to lower output level) the resistance now goes up. 4 @200 ohm each in parallel = 50 ohm. Pull 2 tubes of 4 output and the ohms go to 100 ohm or 2 in parallel.

With 1 cathode shared, the 50 ohm load is constant. pull 2 tubes and current load doubles on remaining 2 tubes .

This ignores any good or band affect of running 2 tubes in a quad amp like an AC30/ Rocket ect.

[pdf64]

My view is that the real benefit from that is increased reliability; if a tube fails for some reason and stops conducting, the other tubes retain their normal operating point. With a shared resistor the others would run hotter than ever, because all tubes have to be functional for correct bias.