• 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 Bias for an EL34 ? Any recommendations / experience ?

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I love the idea of Menno and others to keep the core saturation as low as possible, especially as I am using Menno's toroid opt. I tried his autobias modules in his ul-40 as descibed, burned three of them over time and than changed over to fixed bias. My suspicion is that the module has not enough protection circuits build in whan a tube starts to misbehave and the russians sec el34 of recent production seem to fail often. or the 220k grid resistor in his schematics is simply too high as well.

Nevertheless, I would like to combine the concept of fixed bias with autobias to achieve no electrolytic im the signal path and the other goodies of fixed bias and the target of no core saturation.

Is this possible by combining fixed and cathode bias ? I know that someone can combine them, but will my target be achieved or not ?

Let's say: At 400v, the El34 shall be run at 50mA. The normal SEC El34 I plugged in need -37V grid voltage for that... So, if I give each an unbypassed cathode resistor of 100 Ohm, we will get 5 V, the other -32V come from the fixed neg. bias network. I will have a bit of local feedback, but not a lot of loss on gain, but more linearity. Not to much heat as well as 5v*0,050A=0,25W.

when the tube now operates and changes its behavior due to heat, ac signal swing, ht variation because of ac signal swing or what so ever, the auto-bias effect of the resistor should kick in and a range of 5 V is plenty for grid voltage.

The only thing which I am not sure about: In an ideal world, I would measure the current effect of both bias (fixed and cathode) per tube once, and during operations the cathode r. bias will regulate that current stays the same. for the whole tube.

Will it work like that ? Or will only the proportion of current the resistor has initiated with his 5V be regulated ? So, let say tube conducts 2mA more due to signal swing or heat...will the resistor bring it down by 2mA or only by 2mA*5/37 approx. ? No core saturation in this concept or still core saturations ?
Last edited:
I used mixed bias: fixed bias (i.e. done the old-fashioned way with a pot) and cathode bias (resistor bypassed with an electrolytic). My aim was to run the valves cooler as the bias shift with signal would be smaller.

You need to be clear what it is you wish to regulate: quiescent current or average current. Pure fixed bias regulates quiescent current and so is theoretically best. The typical bias servo regulates average current and so forces quiescent current down when a signal is present - theoretically bad but some people use it OK. Cathode resistor bias is a compromise: good enough for most purposes, simple and cheap.
I am sure that toroids can still tolerate an offset . I ran mains toroids as output transformers single ended using a secondary as tertiary cancellation winding and found even with a 30VA core there was about +/-3mA either way before saturation crept in . If you have unregulated HT then you also need unregulated grid bias supplies . Combining the two (fixed/cathode) has the worst of both worlds as you still retain the decoupling capacitor . Why not try a Blumlein 'garter' in the cathode circuit or just go for fixed bias ?

Well, currently I have fixed bias...

whatI try to achieve is the following:

Autobios improved

Menno claims that 0.3mA difference already has negative effects and that his circuit ensures a difference max. 0.05mA...and has some measurements there to proove his point.

...so basically an alternative way how to ensure that quiescent current stays always equally the same between the tubes, both from a DC and from an AC signal aplied point of view...and I would love to have a ruggidized solution which will not smoke up that easy like Menno's kit...
Last edited:
I used both, the 6040 and for monoblocks the 3070....and I can tell you, when it works (with autobias originally), it is unbelievable stunning. I have no electrolytics anywhere and use duelund cast for coupling, which made as well there a big impact, much bigger than in the speaker crossover. Damned expensive, but worth it.
That is what I did originally with the first one...the last one is just a black, fried chicken...nothing to repair there...

I have read here that you should not use a grid resistor for an EL34 which is more than 39k if you use fixed bias...Menno used 220k in his setup after the Module....maybe this is a cause or its simply the SEC tubes are faulty after a while...maybe I will give it a try with samller grid resistors, I run my fixed bias setup now with 39k as well...
I have read here that you should not use a grid resistor for an EL34 which is more than 39k if you use fixed bias...Menno used 220k in his setup after the Module....maybe this is a cause
It's not the grid leaks. The EL34 is rated for 500k grid leak resistance in fixed-bias, so 220k is quite safe. 39k is ridiculously low. Only a direct short to cathode would burn out that 10 ohm resistor, so it's either bad tubes or user error.
Have you considered using a cathode follower or source (mosfet) follower to derive the fixed bias voltage? It wouldn't specifically fix the problem of current balance under all condition but it might have other benefits related to that problem. For instance, you can't normally put an unlimited capacitance in a self biased cathode capacitor circuit because the bias voltage will then "ride" long passages of loud music and there will be a delay in the return to normal bias after the sequence is over because of the large capacitance.

But if you use a follower circuit to drive it my understanding is that the follower circuit's low impedance drive characteristics will drain any excess voltage at the grid from the cathode capacitor that is not DC related. You could then put an unlimited cathode cap in the cathode circuit. You could also set up the bias on the follower circuit so that it only partially provides all the negative bias that is needed and then use a resistor in the cathode circuit (which the cathode capacitor bypasses) to track any changing DC parameter of tubes going out of balance. You would also eliminate the blocking distortion caused by a straight fixed bias solution as an additional benefit.
Last edited:
Well...typically this happened after some months...the user did not do anything except listing...
please see this thread...
No information. All you said was "220k grid resistor...this really burned a lot of tubes over time".
FYI, EL34s are used with 220k (and more) fixed-bias resistors in thousands of amps worldwide, without problems. Whatever killed your bias modules, it wasn't this, so don't blame Menno without further investigation.

In the latest eh data sheets its speced for max 100k...
What datasheets? 220k will not kill an EL34; the bias servo has more than enough negative voltage to handle it, even if it was too high, which it isn't.
Last edited:
Merlin, I have not blamed Menno, but I used his module in his amp and schematic with the specs and I am sitting on 500 $ devices which simply burned out after a while. Others like triode dick had the same. Do you own one of these devices that you can share factual experience or why are you defending this stuff.

And dont get me wrong: I love the idea. I love the wonderful sound they bring, especially with Menno's OPT. It is absolutely outstanding and I bought three kits of him as I love his stuff ! I am just tired to get a nerve-breakdown after a couple of months each time and just want a reliable amp which has the same sound-quality as with his modules.

...and for the lazy ones not wanting to google stuff themselves, butlet other people do it for them:


So, it might be that the el34 from NOS like 220k, but it seems that recent russian production does not.
Last edited:
Just to be precise: I own three ul-40 and three autobias modules and the setup is like described in Menno's white paper linked above, so all the information is there. the amp, the setup etc. The only difference in the monoblocks where I used 3070opt instead of 6040. In the stereo amp the el34 runs currently at 6000 ohm load impedance with 400v and 48mA. In the monos with 350V and 58mA.

I changed the input part towards mono-bill II with 6N6 and mennos mec-50 gyrator for the input tubes, which is marvelous dead quiet...highly recommended...I eun the 6N6 with 12,5mA across 200k anode resistor. Some people even put 22ma through it as I learned.

Last edited:
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.