PP Tube Biasing, how good is good enough?

[TheGimp]

I did some searches but have yet to find a good explanation of how well matched the bias current has to be in a Push Pull output stage.

I've been reading the Baby Huey thread (up to page 25) and have seen statements that using 1% resistors with current sources (LM317 and others) is good enough.

It seems to me that we are trying to limit imbalance to prevent core saturation and loss of low frequencies, yes, no?

Worst case analysis of the LM317 as a current source shows that just using the room temp spread in ref values and 1% tolerances on resistors ends up with spread of 10% in current.

In reality I suspect that most LM317 current sources are in the 1-2% range but worst case it is possible to see a 10% spread.

Is there any general guideline for setting the balance in bias current for PP amps? Does it matter if it is a guitar amp vs an amp for listening to wider bandwidth music?

Using a matched set of Groove Tubes EL34s, I got the Traynor YVM-1
(Fixed bias, -40V) to run at 26.6mA and 26.9mA (I swapped the tube positions and chose the configuration that gave the lowest difference of about 1.1%).

Using a pair of Mesa Boogie EL34s I got 26.2 and 26.0 by swapping the position to get the lowest offset (0.76%).

A pair of Sovtek (not matched) tubes yielded 41.4mA and 43.0mA(3.7%). A side question is why are the Sovtek running 50% hotter than the GTs and MBs?

All three sets of tubes sounded very similar (shutting down and swapping tubes prevents ideal comparisons).

So the question is how much mismatch is acceptable in PP amps.

I'm sure it is power related, so can it be defined as a simple percentage, or is it more complex?

[Poindexter]

Boy, are you going to get a lot of conflicting answers to this good question!

There are people who think that if there's a tenth of a milliamp imbalance in the quiescent draw of the two tubes, Ultimate Purity of the Sound will be compromised.

There are people (like me) who, with their brass ears, can hear no degradation with 1-2% or so imbalance in the pair.

There are those who will purposefully imbalance the pair a few percent, to move the smallest details of the signal off the zero-crossing point of the transformer, which is somewhat nonlinear in that region.

Why not just listen to your various pairs and hear what you can hear, rather than seeking advice from a bunch of webgeeks like me, who demonstrably don't know shirt?

Aloha,

Poinz
AudioTropic

BTW, the brand-to-brand variance in operating point is quite normal. I find that all the Russian tubes take 2-3 volts more negative bias for the same current draw as old stock stuff.

[Sch3mat1c]

Ten Percent Rule.

In general, if the tube is what it's labeled as, it's within 10% of its spec, so you don't really have to worry much. If you want to be a damnedfool about it, you can tweak it to arbitrarily low values (which, being damnedfools, we all enjoy doing). I don't think even a toroidial OPT would notice 1% imbalance.

Tim

[Miles Prower]

Originally Posted by TheGimp

I did some searches but have yet to find a good explanation of how well matched the bias current has to be in a Push Pull output stage.

It seems to me that we are trying to limit imbalance to prevent core saturation and loss of low frequencies, yes, no?

Yes. The primary of a PP OPT responds to differential current only. Any DC offset will lead to core magnetization. To improve the tolerance to DC offset, PP OPTs are often stacked differently from power xfmrs, groups from 4 to a dozen lams stacked together to increase the effective air gap that all E-I type xfmrs have.

With toroids, there is no air gap at all, and these are much less tolerant of DC offset.

Originally Posted by TheGimp

Is there any general guideline for setting the balance in bias current for PP amps? Does it matter if it is a guitar amp vs an amp for listening to wider bandwidth music?

Not really. Core saturation effects sound nasty. Since all my designs are fixed bias, balance is NBD. Just set both halves of the PP pair to the same current. Touch up as necessary when checking the bias once a month or so.

Originally Posted by TheGimp

Using a matched set of Groove Tubes EL34s, I got the Traynor YVM-1
(Fixed bias, -40V) to run at 26.6mA and 26.9mA (I swapped the tube positions and chose the configuration that gave the lowest difference of about 1.1%).

More than good enough.

Originally Posted by TheGimp

So the question is how much mismatch is acceptable in PP amps.

What kind of mismatch? You have DC characteristics, and AC characteristics. Which ones do you want balanced? Also, keep in mind that nothing will stay in balance forever.

I once saw an article somewhere I long forgot where someone once did a test by connecting a 45 to one phase and a 2A3 to the opposite phase as a PP pair. The conclusion was that it really made much less of a difference than you'd expect. Being fanatical about "balance" is audiophoolery.

Some guitar players also use deliberate mismatches (6V6 on one side / 6L6 on the other; or pulling out one half the finals from a PPP design on one phase) to change the tone.