Question about independent bias

[toml]

First, I would like to say thanks to all that contribute here. I am just getting started in amp building and have learned a tremendous amount from reading and studying the threads here.

So, here is my first question. I am working on a D'Lite 44 kit. It is now up and running and I am in my tweaking stage.

I would like to use some old 6L6 tubes I have that are unmatched. So, I figured that I could modify the bias circuit for independent control of the 2 output tubes.

I found this link that describes how one person accomplished this:
http://people.cornell.edu/pages/mt24/Amp/mm/bias.html

As shown in the above link, the first option is to split the two 220k grid leak resistors. Replace the 10k bias pot and series 10k resistor to ground with:
a first 25k bias pot and series 15k resistor to ground with the wiper of the first 25k bias pot to the first 220k grid leak resistor;
a second 25k bias pot and series 15k resistor to ground with the wiper of the second 25k bias pot to the second 220k grid leak resistor.

The motivation for the change in values is to preserve the load.

My first question, is how important is it to preserve the load in the bias circuit?
A traditional screwdriver adjust bias pot at 25k is not easy to come by but 10k screwdriver bias pots are. So, if the voltage drop is the key, couldn’t I simply duplicate the 10k bias pot and series 10k resistor to ground for each of the 220k grid leak resistors?

The author also suggests that it's probably better to use a pot to set the bias level, and a balance pot to distribute it to the tubes.

In this regard, what I envision is keeping the current single bias pot circuit. The modification would comprise splitting the two 220k grid leak resistors, connecting the first 220k grid leak resistor to lug 1 of a linear taper pot, connecting the second 220k grid leak resistor to lug 3 of the linear taper pot, and running the bias pot wiper to the new “balance” wiper.

Has anyone implemented either of the above designs? Can someone recommend approaches/motivations/ specific values?

For example, my first guess is maybe to replace the grid leak resistors with 200k and use a 50k “balance” pot. That seems like a pretty simple change to get the benefit of being able to use unmatched tubes.

I apologize if my description is unclear. I do not have a schematic of second option, although you can see the first option at the above link.

Thanks in advance for all your help.

Best regards

Tom

[Structo]

I did exactly this a couple weeks ago and I followed that schematic that you linked.
I found some NTE 25K pots with short slotted shafts, with lock nuts.
Then I used the 15K tail resistor.
I mounted the second pot right next to the existing hole in the chassis and it was a tight fit!

You have to preserve the total load like it was in the stock wiring,ie, 20K.
So with the two bias pots and tail resistors of 40K in parallel you get the same 20K.
And yes, split the two 220K resistors.

The reason you want to preserve the load is to keep the same adjustment range.
If I recall mine will adjust from around 18ma to well over 100ma.
Be sure to back the pots all the way down before you fire it up.
I didn't and had them set about half way and the current was over 100ma!

His solution seems to work well and it was my desire to be able to balance unmatched NOS tubes instead of paying the piper for matched sets that are very expensive.

Not sure about using the stock bias pot setup with a balance pot, as I didn't do it that way but I suppose it would work.

[toml]

Structo:

Thanks for the feedback. I think I may give this a try.

Tom

[Bob-I]

I kinda disagree with your solution. Bias only adjusts the idle current, not the working state. If the tubes are really not matched you can't get a matched output by adjusting the bias as the tubes will amplify differently.

Typically seperate bias adjustments are used to balance an output xformer with unequal DC resistance on the 2 sides of the windings. If the AC impedance matches you'll have some xover distortion if the idle is not matched.

JMO

[toml]

Bob:

Thanks for your advice. I guess I was thinking specifically of idle current. To provide an example, with a particular pair of 6L6 tubes in my DLite 44, I measured my plate voltage at 397v. So I dialed up the bias to about 53 ma for V4 at the corresponding bias test points. When I measured the idle current for V5, it was sitting over 60ma. So, I figured that was a bit too hot. I backed down the bias until V5 sat at 52ma and the amp sounded okay, but now V4 was also lowered. That got me thinking that a balance or independent control would allow me to dial the idle currents a little closer together, even if not a perfect match.

I understand that the 70% bias is just a target range and nothing is set in stone, so I will experiment more with bias settings.

I will also do some more studying with regard to biasing. Any other thoughts or suggestions are appreciated.

Tom

[novosibir]

I kinda disagree with your solution. Bias only adjusts the idle current, not the working state. If the tubes are really not matched you can't get a matched output by adjusting the bias as the tubes will amplify differently.

I agree!

If two tubes are way off in bias, then they usually are also way off in amplification (transconductance).

Larry

[novosibir]

I'm not a fan of dual bias, but for those who want to try it, see the attached schem, what actually was designed for Marshall type amps.

Larry

[Firestorm]

Some tubes (all of the "KTs," for example) SPECIFY separately adjustable bias supplies, so it's not entirely a mismatch fix. As Bob-I noted, separate bias controls can help achieve better DC balance in the OT, but you'll only be adjusting current, not transconductance.
Of course, it has been suggested that matching for current and deliberately mismatching for transconductance can provide interesting (maybe even pleasing) results. It can reduce the phase cancellation of even-order harmonics and, if it introduces a little cross-over distortion, add some odd-order hamonics, as well.
That said, I wouldn't try any of this with tubes that were wildly different to begin with. As with any kind of tweak, there's a point where it will start to sound like cr*p.

[Structo]

The reason I went the way I did is because it simply duplicates what's already there.
You guys are talking a little over my head in regards to the DC balance of the OT but I have installed it as in this schematic and I must say it works well and I don't notice and tone degradation.
Of course I am not running wildly mismatched tubes either.
But, I have found that even the tubes sold as matched pairs can often time vary by 5ma between the two.
With this bias circuit (b), I don't have that.

[LooseChange]

Tom,
In example b, you can join the two pots together and halve that 15k to ground. 've got that setup in all my DIY amps. Works great.

['67_Plexi]

IIf I recall mine will adjust from around 18ma to well over 100ma.

Tom, thats an awful lot of range. I would set a range to limit around 45-50mA.

[dimitris]

Toml wrote :

couldn’t I simply duplicate the 10k bias pot and series 10k resistor to ground for each of the 220k grid leak resistors?

I was qoing to to do that thing.Will this work or it is necessary to change the values of the bias pots?
Thanks
Dimitris

[Structo]

IIf I recall mine will adjust from around 18ma to well over 100ma.

Tom, thats an awful lot of range. I would set a range to limit around 45-50mA.

Well I wanted to preserve the ability to run 6V6 tubes if I wanted.
Currently I have the JJ 6L6 tubes biased at 35ma.

When I had talked this mod over with Moss he told me that I should aim for a total resistance that mimicked the original.

[odourboy]

I kinda disagree with your solution. Bias only adjusts the idle current, not the working state. If the tubes are really not matched you can't get a matched output by adjusting the bias as the tubes will amplify differently.

I agree!

If two tubes are way off in bias, then they usually are also way off in amplification (transconductance).

Larry

2nd that. Could you put a trimmer in the circuit to compensate for the differences in the gain for unmatched tubes? Anyone ever seen this or tried it? (Just curious)

[toml]

Thanks to everyone for their respnses.
Earlier in this thread, Bob-I said

Typically seperate bias adjustments are used to balance an output xformer with unequal DC resistance on the 2 sides of the windings. If the AC impedance matches you'll have some xover distortion if the idle is not matched.

Can someone please explain this, maybe by example? Or better yet, direct me to some suggested reading on bias supplies in general?

Thanks
Tom