• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

mixed mode bias

Has anyone ever tried using part cathode bias and part fixed bias on the same output stage?
My idea is to follow the Mullard data sheet on the EL34 in Triode mode and use an un bypassed cathode resistor on one tube, as a reference, and on the other tube use about 10% less cathode resistance and a op amp bias servo to make up the difference with "fixed" bias.
The goal is to be able to use a toroid for the OT.
My guess is it will be like this: Amplifier auto bias circuits: Alan Dower Blumlein's garter circuit
 
Why not just use the garter bias circuit in the first place? Note however that since this is a form of cathode bias you will need really good bypass caps...

Another alternative would be a CCS in each cathode bypassed by a big cap..

What you propose will work moderately well of course at the complication of an op-amp, a transistor or two and an additional power supply. The tube being servo'd to could be fixed biased, note that maintaining balance might be a problem if you approach class B operation due to the time constant of the integrator in your servo circuit.. You won't have this issue in pure class A PP as the plate current should be constant..
 
I want tighter balance than what is suggested as possible in the Garter bias.
Class A is not really an option for me.
I don't know how well a CCS will work in the cathode, I'd need to research that. But, a CCS and big, high quality cap are a lot more money than a quality op-amp and its power supply. Isn't there also a time constant problem with a big cathode bypass cap too? The version of the Mullard lash up I plan to use has un-bypassed cathode resistors with the tubes strapped as Triodes.
 
I want tighter balance than what is suggested as possible in the Garter bias.
Class A is not really an option for me.
I don't know how well a CCS will work in the cathode, I'd need to research that. But, a CCS and big, high quality cap are a lot more money than a quality op-amp and its power supply. Isn't there also a time constant problem with a big cathode bypass cap too? The version of the Mullard lash up I plan to use has un-bypassed cathode resistors with the tubes strapped as Triodes.

Single stroke have ruled out garter due to balance requirements, CCS ruled out because that is applicable only in class A, your proposed servo config because it will not track changes as the output stage enters class B due to the integrator time constant required to make sure it does not respond to audio signal currents in the reference path or the integrator feedback path. I would servo each tube independently which a very long time constant so that the servo cannot respond to signal currents in the cathode circuit.

An unbypassed cathode in an output stage is going to result in a rather high rp and very little voltage gain as it is local current feedback.

Most people here who use power OPTs for outputs tend to use tubes with a low rp, and CCS derived bias. Perhaps Fran will weigh in as he has a lot of experience with power toroids used as OPTs.

I recommend you simulate your design in LTSpice IV which is free and for which extensive tube libraries are available - it is Windows native, but runs really well under Wine in Linux as well.
 
Last edited:
I use part fixed, part cathode bias for UL EL34 - a modified version of the Mullard 5-10. My aim was to reduce the bias shift with signal, so the valves could start with lower quiescent current, but still have some protection from bias failure. The Mullard design has 470R cathode resistors. I planned to use 330R, but ended up with 390R as it was inconvenient to increase the negative bias supply. I guess that is the trouble with DIY: every build is a prototype!

To make adjustment easy, I used a bicolour LED in a bridge arrangement run from the cathode. Too much current=red; too little current=green. I just twiddle the bias pot until the LED goes out. I aimed for about 55mA cathode current; Mullard used 70mA.