Bias question

[daz]

If you have 2 tubes and when you bias one to 70% and the other is off considerably, do you just bias the hotter one to 70% or do you bias it a bit over so the other one isn't too cold? I calculated 70% on my amp at 40 ma, and when i set the bias to 40 ma, the other tube was about 31-32. Would you bias the one thats around 31 up to say 35 which would put the other at about 44, or keep the higher one at 40? I've had tubes off that much before but i never bothered with it and just always biased the hotter one to 70% and lived with that. i guess it doesn't really matter unless they are off more than that, but i'm just wondering what you all do. I think the imbalance is probably a good thing tonally unless they are too far off and one is much too cold. But i'm unsure about how far off they can be before the colder tube starts causing harshness.

[Enzo]

Just never bothered with it is the operative idea for me. Bias is just not a critical setting. It is a ball park setting. If I had to make up a rule, I would average the two tubes, and set their average at my target. SO one would be a little hotter and the other a little cooler than my ideal.

A valuable experiment is to swap the tubes. If tube A is 40ma, and tube B is 32ma, swap places. if the two tubes keep their current draws in the new sockets, then the tubes have differences. But if the one socket just seems to run hotter than the other, then you are looking at either the differences in OT winding resistance, or ther is some component out of spec in the tube grid circuits.

On many OTs, the two halves of the primary winding are not of the same resistance. SO the DC current through the tubes will differ. But the signal is transformed by the number of turns, not the DC resistance, of the transformer.

[daz]

Bias is just not a critical setting. It is a ball park setting.

I like that. So many act as tho it's so crucial but to be honest i have never noticed much if any difference unless they are way way off. I'm not tech, but i do have ears and from my experience they gotta get pretty darn cold B4 things start sounding worse.

[BackwardsBoB]

Most push-pull amps count on power supply noise on the output transformer being cancelled by equal AC current fluctuations at the transformer inputs. That's why you'll find old amps without a bias adjustment that have a hum balance adjustment that balances the bias current through both tubes.

You're probably only a pot, resistor and capacitor away from having independant bias current adjustments, and when tube bias requirements drift, you'll be able to use the tubes longer without hum.

That said, there's something to be said for asymmetric distortion on your amp's output, and you can have great fun mixing tube types within a pair, especially in a blues amp, with high harmonic distortion with minimal clipping. Again, independant bias adjustment is a great asset.

[daz]

I can envision how to do it, but to be sure can you point me at a schematic with independent adjustment?

[Enzo]

Actually it would be equal amplification on the two sides. You could have two tubes of different gain and they might bias up even on current but still be way unbalanced.

[Jazz P Bass]

Tube bias

For HiFi I would want the tubes "matched"
For guitar amps, most push pull phase inverters are not "balanced" either.
They try to be. (ie: 100K & 80K resistors)
Nor is the output transformer.
So what's with all these "matched tube" salesman.
Bias takes the tubes from Class B to a little bit of Class A.
This in effect rids the output signal of crossover distortion.
That is all.

[daz]

I'm not concerned with tubes being mismatched....i'm concerned with tubes being in completely different universes. like 15 MA apart or even more. I have had some drift so far that biasing one as high as i should leaves the other one so low it's nuts. That can't be good for tone.

[defaced]

I can envision how to do it, but to be sure can you point me at a schematic with independent adjustment?

Here's an old dual fixed bias schematic I did, but if you're working with the build you typically talk about, I don't know if it'll be much help. By breaking the junction of the two parallel tube using a single bias can you take the dual bias to quad bias. Also attached is a schematic PaulP and Enzo came up with a while ago for independent dual bias.

[daz]

Thanks. Not sure that will work for me tho.

[Enzo]

Why not?

[MWJB]

For HiFi I would want the tubes "matched"
For guitar amps, most push pull phase inverters are not "balanced" either.
They try to be. (ie: 100K & 80K resistors)
Nor is the output transformer.
So what's with all these "matched tube" salesman.
Bias takes the tubes from Class B to a little bit of Class A.
This in effect rids the output signal of crossover distortion.
That is all.

Well, firstly, some folks do like the sound of tubes mismatched on current (often described as warmer, grindier), there's no "right" or "wrong" as long a you like the sound and you're not damaging anything. But, guitar amp PIs (since the mid 50's anyway) are usually pretty well balanced, at least balanced well enough to not sound like they are malfunctioning anyway. As long as everything else is working properly/as intended, a mismatch in plate current will have the biggest sonic effect. +/- 15%-20% might be a good rule of thumb, if a vendor sells you a matched pair that exceed this, I'd expect a refund/replacement (it's what you have supposedly paid a premium for after all). Typically EL34/6L6 are "matched" to within 5mA (@30-35mA idle). Certainly by 10mA mismatch at idle, fidelity will be noticably reduced. Be sure to read the tubes in all sockets to determine whether any mismatch is inherant in the circuit, rather than just relating to a specific tube.

However I stress again, if you fit a pair of tubes that idle warm enough to not sound thin, not so hot as to damage anything and you like the sound (matched or not) then you don't have a problem. If you can hear a mismatch (when A/B'd with a matched pair) then you know what to do about it.

Several manufacturers even have controls on their amps to creatively mismatch a PP output. So, if possible, I'd recommend ear testing degrees of match/mismatch & going with your ears.

Output transformers may be a little mismatched as regard DCR (rather than AC impedance) which may have a knock on on plate voltage & current, despite this AC voltages may still be closely matched.

Personally I prefer a closely matched pair. Many brands of current production power tubes are pretty inexpensive, it's worth buying a few "matched" pairs, or quads, so you can cherry pick...if you turn over a lot of power tubes, you will usually be able to recycle moderately mismatched pairs over time.

[daz]

Why not?

I was looking at it wrong. I have a hard time visualizing schematics that are drawn very differently. However, there is one thing i'm unsure about. All the bias circuits i've used have a cap where the 50k pot is i that drawing. Is it ok to leave that out? Also, why do all the bias circuits i've copied use 10uf caps while this one uses 50uf? Sorry for my ignorance?

[MWJB]

"All the bias circuits i've used have a cap where the 50k pot is i that drawing. Is it ok to leave that out?" The load resistors (50K pot in series with a 15K fixed resistor) ARE bypassed by a 50uf cap. The load resistor (be it a pot, or fixed) is normally connected directly to ground.

"Also, why do all the bias circuits i've copied use 10uf caps while this one uses 50uf?" Tweed Fender derived amps (inc. classic Marshalls) use one 8-10uf cap to ground before the dropping resistor, then a 2nd cap in parallel with the load resistor to ground...the smaller value caps charge up quicker, stabilising the bias supply quicker. Larger caps can take a little longer to settle after turn on. In reality a lot of guys will use a value that they have a lot of in stock. E.g. the 50uf 70v caps in tolex Fenders are normally replaced with 100uf 100v as everybody carries these. 22uf would probably work just as well in the schem above.

Voltage rating should be 100v min.

You just need a value that eliminates hum being introduced via the bias supply, but doesn't leave the power tubes underbiased for long periods of time.

[daz]

got it, thanks much ! definately going to do this once i get done with a few other things i'm in the process of trying.