Biasing questions

[12AX7]

To date i have always just set bias to about 50% according to one of the online calculators. But i read Aiken's info on biasing yesterday and it has me wondering about several things. The first being my not very well spec'd OT for a 6V6 pair. I read one thing that i coudn't understand well but seemed  to be was suggesting that the tubes could redplate very easily or at least much easier with a OT with a low primary impedance. The other question involves tone at stage volume because i can't test my amp at home and obviously can sit there biasing and testing various bias points at a gig. So i wanted to ask about that and how biasing at say 50% might sound compared to 70% and whether one will sound more complex harmonically, or smoother, etc. I can't tell at the home volumes i'm limited to.

So first Q is where would you bias my amp? It's 2X6V6 JJ's with a marshall style PI and the OT is a custom made 4K (which is low for 6V6 i think?) 30 watt heyboer with only 8 and 16 ohm taps. So as i understand it, the ideal load would be a 16 ohm speaker off the 8 ohm tap but my speaker is 8 ohm so i'm runnning that off the 8 ohm tap. I have it biased right now at a cool 50%/17.5mA (calc'd using robrobinette's calc) with 400 volts on the plate. Given that info, where would you bias it for best tone at stage volume and safety, or is it possible to be safe with that OT and speaker? I have a NYE party to do tonite and only gig rarely so i haven't used this amp out but once before and i couldn't determine how the output sounded because i had a preamp tone design issue at the time that has since been rectified. So the question is what bias setting would you shoot for with this setup to insure it's safe and for good tone? I could simply ask about tone and how it;s affected biasing somewhere between at 50 to 70% but after reading Aiken's info it seems i have to take my OT into consideration for both tone and safety due to it's less than optimal spec.

I wouldn't have asked all this if not for having read Aikens info. But after that it has me a bit worried that with my OT and bias setting i'm either compromising tone or risking the amp/tubes or both and i have a few hours to make any changes before i have to load up the car. So any advice as to what you'd do with this setup would be great.

[sluckey]

I'd set the bias to 70%.

[12AX7]

Even with that OT, and why? (just curious)

[sluckey]

I think 70% is a good safe compromise between too cool and too hot. If you are concerned about the impedance mismatch, use a proper speaker load.

[12AX7]

I would if i could but all i have right now is that one speaker. I tried again to see if i could hear a difference between 24 and 17mA and i do think it sounds a tad better at 24 but it's subtle at home volume. Just wondering how it will translate to stage volume. But the OT and load are the biggest concern so no knowing if the OT/load is an issue with 70% bias i guess i'll just leave it at 17 to be safe.

[sluckey]

I am not concerned with even a 100% impedance mismatch. I'm also not playing in any situation that would require driving an amp hard. If I was, I'd use a proper speaker load. That's just the tech side of me talking.

[12AX7]

Well, i don't think i'll be driving it real hard. I usually get to stage volume around 10:00 on the master. But then after about noon it just gets more driven and no louder.

[HotBluePlates]

... where would you bias my amp? It's 2X6V6 JJ's with a marshall style PI and the OT is a custom made 4K (which is low for 6V6 i think?) 30 watt heyboer with only 8 and 16 ohm taps. So as i understand it, the ideal load would be a 16 ohm speaker off the 8 ohm tap but my speaker is 8 ohm so i'm runnning that off the 8 ohm tap. ...

4kΩ and 30w sounds typical for a pair of 6L6's (or 4x 6V6's) at typical guitar amp voltages.  It probably wouldn't be optimum for maximum clean output power from a pair of 6V6's, but also likely does no harm.

I'd set the bias to 70%.

Even with that OT, and why? (just curious)

The below recently occurred to me; I've never read anyone explain the "70% rule" as I will below, but it might be out there somewhere.

Class A is 100% dissipation at idle, and presumes the supply voltage and OT load are such that the output tube(s) will swing from idle current to 2x idle current, back to idle and down to (almost) zero current and back to idle.  At no point in the maximum clean output signal cycle does the tube shut off.

Class B (theoretically) idles at 0% dissipation, just turned off.  Since the tube is cooling (off) exactly half the max output signal cycle, much bigger plate voltage and current swings are permissible, for much more power output.  This requires the supply voltage and load to be sized accordingly.  It also assumes the two half-cycles generated by the push-pull output stage can be stitched together perfectly to create a unified output.

Class AB is literally anything between Class A (100%) and Class B (0%).  The designer can tweak in the bias to be exactly where they (think they) want, based on their supply voltage, OT load, and expected plate dissipation throughout the signal cycle.  Where they think they want it is necessarily based on some assumed ideal tube.

In the real-world, that last statement on Class B is tough to achieve.  Class AB which is operating very cool (closer to Class B than Class A) may sound rough due to crossover distortion and incomplete "stitching together" of the signal.  Class AB conditions (high supply voltage, very low OT load) can overheat the tubes if they're idling too close to Class A.

Look at a "Unit Circle" (a foundation for discussing sine waves).  The midpoint between a sine wave at 0 degrees (zero, off, or reference point) and its positive peak at 90 degrees is at 45 degrees.  The distance in the Y-direction hasn't rising to 50% of the peak, but is actually at 70.7%, 1/√2, √2/2 or sin45-degrees (all basically the same number).  This is also the percentage used to convert an sine wave's peak to its RMS value to express what value of d.c. would have an equivalent heating effect (same amount of power over time).

The "new idea" (I'm probably just now recognizing the "old idea") is that idling at ~70% is the same as "halfway between Class A and Class B, when the output stage is driven by sine waves" (as it often is in testing).

So just try 70% dissipation as an idle point. Ass PRR says, if it sounds too rough with small signals, idle a little hotter; if it tends to redplate when pushed hard, idle a little cooler.

[HotBluePlates]

... how biasing at say 50% might sound compared to 70% and whether one will sound more complex harmonically, or smoother, etc. ...

As noted above, if you idle too cool (at/near Class B or 0% dissipation), soft sounds will be harshly distorted.  The amp will actually get cleaner when fed larger signals.

Below where the tubes redplate during use, there's a broad range where the amp sounds essentially "the same" regardless of where you bias it.

If the output stage receives a peak signal equal to/bigger than the (absolute) value of the bias voltage, the tubes won't be able to stay clean (unless your amp circuit can drive gracefully into grid current, and most guitar amps cannot/don't try).  You idle hotter by reducing the bias voltage, implying that the output stage will now distort sooner than with cooler bias.  Depending on a lot of factors, this might mean the amp output less clean power before hitting its limit.

But usually folks say only they notice "fatter sound" or "earlier grind" etc.  Really what they're noticing is the amp distorts somewhat sooner than it did with cooler bias.  "Harmonic complexity" is somewhat a misnomer in a push-pull stage, because even harmonics are cancelled by a push-pull stage (unless it's "run wrong" to unbalance the output stage).  Changing idle bias might have a slight impact on how quickly the amp tends to break into distortion, though the supply voltage and load generally influence that much more.

I don't know if ~24mA would sound any better than the ~17mA you have now.  Try it & see.  Shouldn't hurt anything.

[PRR]

> if it tends to redplate when pushed hard, idle a little cooler.

Actually, plate dissipation falls at maximum output, and idle current probably have very little effect (unless idle is way hot to start).

Below is my *guess* what happens. If you idle near 50% Pdiss, dissipation drops for small outputs but changes little at high outputs. THD likewise unchanged at high output but increasing at low output. This is the GE 6L6GC data, and GE was pushing at Hi-Fi applications, so low-level THD was of concern, more than idle dissipation. Also note GE was suggesting idle at 87% of rated Pdiss.

> "Unit Circle"

Yes, a good tool. But in an amplifier we also want the change of Gm with current, the zone where one tube may "cut-off", how sharp it actually cuts-off (the change-of-change of Gm with current), and how B+ balances with Pdiss.

I think after looking at a lot of amps (plotting the gain curve; see Kuehnel) you will find that "cleanest" performance is only possible with fairly low current, so we can run rich current without melting the plates. It has been suggested that the original 5F6a Bassman "sound" is in-part because at that B+ you can't run the 5881 hot enough safely to smooth the A-B crossover transition, adding "growl" to the decay.

Also that 70% is always safe (for rational loading), 100% is not much smoother and is flirting with trouble, 50% is for most designs audibly rough (depending how you play).