Adjustable bias range?

[hcorneli]

I realize real-world bias is a bit hard to predict, given low bias currents and varying voltage supplies. But I'm trying to design an idealized bias adjust system. For those who don't know, Doug has a great page on bias circuits, and shows how you can use anything from a 10K to a 50K pot.

Let's say you want to bias an amp for a single power tube type -- say 6V6s -- and let's assume your PT has a 50V bias tap. Merlin notes a 10K pot is ideal, and that the other bias resistors (such as bias divider, bias tail) should be small-ish too, with the tail resistor (his 'idiot resistor' or ground shunt) possibly the size of the pot. Then he notes the max negative (most negative) voltage you'd need would be the screen voltage divided by the triode mu (10 for the 6V6), which would take the tube to near cutoff. So for 6V6s with screen voltage 400 you'd want max negative of -40.

If I'm right, the math is like this (skip if you want). The voltage divider formula is V(out)=V(source)*R2/(R1+R2) where here R1 is the divider and R2 is the tail -- in this case, the tail resistor + the pot. Note with a 10K pot R2(max) will be the tail + 10K (max pot setting dumps least bias voltage to ground).

The result? A big R1 (bias divider) gets you a narrow adjustment range, as little as ±3V. In some Fender layouts, if you center this range on the schematic bias voltage, you get a max negative voltage far less negative than the 40 Merlin wants. Of course, if you just swap in a small R1, you can get the bias adjustment range up to say 12V, with the upper limit at or near 40, but at the cost of making bias adjustment much less precise (let's assume you use a single-turn pot).

Yes, I know many people just run 25K or 50K and get a wider range, and I know you can use 10-turn pots, and yes, many people want to be able to bias for two or more tube types.

But for a single tube type (say 6V6) *what is a useful, real-world range of bias adjustment*? 3V? 12V? or???

[shooter]

the top of tube volts has something to say, 400vdc at plate wants way less current than 100vdc at plate.  does it still want a small, medium, or large range 

[PRR]

> A big R1 (bias divider) gets you a narrow adjustment range, as little as ±3V.

I don't follow. I may need a picture.

"OFF!" is likely Vg2/Mu as you say. In OVER-volted PA and guitar amps you may want a little more to safely bring-up the amplifier very-cool. A typical warm bias is 0.6*Vg2/Mu (see RCA manual). Tubes vary. But I'd think 0.55 to 1.1 times Vg2/Mu is a general target. Note that this is 2:1. If the variable is wired *as a potentiometer divider*, and liking 10k, then 10k pot and 10k stopper from a point that gives 1.1*Vg2/Mu .

When Leo was getting crates of tubes more consistent than +/-20% Fender did use small or no range. Heck, the vast majority of fixed-bias tube amps I have owned had no adjustment. The major exception was a Fisher which had 6 or 8 internal trims for EVERYthing.

[sluckey]

But for a single tube type (say 6V6) *what is a useful, real-world range of bias adjustment*?

-30 to -45 will cover most all 6V6 circuits.

[hcorneli]

> A big R1 (bias divider) gets you a narrow adjustment range, as little as ±3V.

I don't follow. I may need a picture.

"OFF!" is likely Vg2/Mu as you say. In OVER-volted PA and guitar amps you may want a little more to safely bring-up the amplifier very-cool. A typical warm bias is 0.6*Vg2/Mu (see RCA manual). Tubes vary. But I'd think 0.55 to 1.1 times Vg2/Mu is a general target. Note that this is 2:1. If the variable is wired *as a potentiometer divider*, and liking 10k, then 10k pot and 10k stopper from a point that gives 1.1*Vg2/Mu .

When Leo was getting crates of tubes more consistent than +/-20% Fender did use small or no range. Heck, the vast majority of fixed-bias tube amps I have owned had no adjustment. The major exception was a Fisher which had 6 or 8 internal trims for EVERYthing.

Thank you, good points. I remind myself the bias adjust range on a lot of vintage Fenders was 0V. 


I understand what you say about needing a picture. As Merlin says, there are a huge array of bias adjust schematics out there. I'm picturing something like this, basically inserting a pot into the topology of a 5F6a or (my current project) a 5G9. If I'm right, and if I understand linking an image file, the scheme would be:

____

[sluckey]

For the 5G9 I would replace the 56K with a 50K pot and a 22K resistor. That will allow you to swing above and below the stock bias voltage.

[hcorneli]

And if I'm doing the math and Excel right, here are bias range calculations on a 5G9. I just insert a 10K pot with Fender's 'big' R1 and R2 values (82K and 56K) in the first column, and with Merlin's 'small' values in the second column. (I assume 50V rectifies to -70V, and note Fender's fixed schematic bias voltage at the grid leak junction was -28V.) I bold the resistor values...


(I know pictures are better than words, but I still feel like my words may be confusing my pictures). 

[hcorneli]

But for a single tube type (say 6V6) *what is a useful, real-world range of bias adjustment*?

-30 to -45 will cover most all 6V6 circuits.

Thank you, that's very helpful. I'll work on trying to get a span like that. FWIW, I'm trying to build a circuit with a lower plate / screen voltage based again on Fender's nominal 5G9 voltages (~370 at the screens) but with modern wall voltage and all the other variables, it'll be better to have a few too many negative volts than too few.


And I hear you about a 50K pot. I have a 25K in my stash. Let's see, if I'm doing it right, here's a 50K pot, 22K tail, and 47K divider.

[PRR]

ALL the pix on the Merlin page you linked show potentiometer connection:


Yet you turned to the Rheostat connection:


The Rheostat gets especially awkward if there is no large excess voltage and resistance in the bias supply. As when we use a 50VAC tap. It appears to work dandy when we drop bias from a 300VAC winding through a large impedance as in some Marshalls.

Yes, the potentiometer connection really should have an added 12-cent part to bypass the wiper when it fails. Merlin's R3.

While you can design a smaller range for the tubes in your hand, I figure tubes may be replaced and bias parts should not need re-fixing for a tube change. (Unless, like the places Sluckey and I used to work, there is a dude on salary for just this purpose; and he needs to *look* busy.) So for 450V on G2:

[hcorneli]

ALL the pix on the Merlin page you linked show potentiometer connection:


Yet you turned to the Rheostat connection:


The Rheostat gets especially awkward if there is no large excess voltage and resistance in the bias supply. As when we use a 50VAC tap. It appears to work dandy when we drop bias from a 300VAC winding through a large impedance as in some Marshalls.

Yes, the potentiometer connection really should have an added 12-cent part to bypass the wiper when it fails. Merlin's R3.

While you can design a smaller range for the tubes in your hand, I figure tubes may be replaced and bias parts should not need re-fixing for a tube change. (Unless, like the places Sluckey and I used to work, there is a dude on salary for just this purpose; and he needs to *look* busy.) So for 450V on G2:

Aha, I noticed that Merlin fed the grid leaks off the wiper and I didn't -- as you say, potentiometer v. rheostat. In fact, I know Doug's bias schemes also do it that way; I followed them to build my PR and 6G2. Not understanding the implications, I borrowed this topology from a bias-adjust scheme I'd seen for a 5F6a. But I didn't know the bias voltage source made a difference.


Thank you. This is super helpful. I'll go back to work on my layout, and when I follow Merlin, I'll try to follow him better. 

[sluckey]

I prefer the way you drew it.

[hcorneli]

I prefer the way you drew it.

Sorry, you mean my original schematic? I barely understand scheme thinking, bear with me...

[hcorneli]

Trying to use more pictures -- I'm not good at either schemes or terminology -- I'm gonna try to post a pic of my layout...

[hcorneli]

Here's my version with a 10KL pot and a 51K tail to just 'split the difference' around Fender's 82K and 56K. I'm not putting this up to suggest I have the right values, but to make sure my schematic reflects reality.

[hcorneli]

I prefer the way you drew it.

Sorry, I'm still confused, but I've been reading more on rheostats vs. potentiometers, and potentiometers in bias feeds following a "bias range resistor." I gotta assume you mean you like the way PRR drew it, which was 100X more likely in the first place. 

I'm gonna see if I can't fit a plan like that into the existing eyelets on Doug's stock 5G9 board -- skipping the doghouse gives me some extra holes in the upper left.

[sluckey]

I prefer the way hcorneli drew it. Look at the schematic in this pdf...

     http://sluckeyamps.com/misc/bandmaster.pdf

[hcorneli]

I prefer the way hcorneli drew it. Look at the schematic in this pdf...

     http://sluckeyamps.com/misc/bandmaster.pdf

Thanks. Very helpful. That does look nicely Fender-y.  


Thanks to all. I really appreciate the help, and I'm learning a lot.

[sluckey]

The amp is fendery but the bias circuit is straight from Marshall.

[hcorneli]

Yeah, I recall Doug's Fender examples using potentiometer wiring with their bias range resistors -- as did my PR and 6G2. Merlin: "There seem to be as many different bias adjustment circuits found on schematics as there are different amplifiers." 

[hcorneli]

A little follow-up and a big thank you to all. I built it like this and it works perfectly. Nice swing from about -21V to -42V, and the 'likely 6V6 target' (to get roughly 17-22 mA at the cathodes) is roughly half way up the dial. Plus the range is not too wide to make adjustment fiddly; you can easily turn it up a mA at a time. Oh, and with inboard filters instead of a doghouse, it all fits on Doug's standard 5G9 board with a couple eyelets to spare.

[HotBluePlates]

... Plus the range is not too wide to make adjustment fiddly; you can easily turn it up a mA at a time. ...

When the wide range does get fiddly, a 10-turn pot solves the issue nicely.  I have a tube tester where I replaced 4 controls that set plate, screen & bias voltages for the tube under test with 10-turn pots.  The knobs spin easily, and it's simple to dial-in the precise value desired.