Combination cathode bias and fixed bias

[tubeswell]

In my new 5G9 build (which is adjustable fixed bias) I am running a pair of new sensor 'tung-sol' 6V6GTs.

Everything is stock , except for the adjustable fixed bias, which is to say that I have replaced the 56k resistor to ground with a 50k linear trim pot in series with a 39k resistor wired as a variable resistor.

On another post I've been discussing the different plate current levels.

I currently have the plates biased at about 27mA (I think) with a Plate voltage of 363 (giving about 9.8W per tube)

What I was wondering is can I have this circuit with say some chunky 100R cathode resistors AND have the adjustable bias too, and run the tube current level higher (say 31mA plus) to get the tubes close to max disspiation at idle? I was wondering whether this would enable me to get the 'best of both worlds' so to speak, as far as I think I undertsand that cathode biasing lets you run the tubes at max dissipation, and you get good distortion when cranked, versus the fixed bias - where current level goes up and down with output power (and in the 5G9 the trem circuit connects to the grid bias supply, so I would want to keep it there). I think this means that I wouldn't need to run the negative grid voltage as high, i.e. I'm hoping the tubes could run hotter without harm coming to them.

I set the amp up initially to run the tubes at 31mA with 371 plate volts (giving 11.5W dissipation at idle), but I now understand that this might severely shorten the tube life (so that is why I am asking these stupid questions).

Is this combination method of biasing typically done? and what are the things to watch-out for?

Or am I in La-la-land?

[BackwardsBoB]

So you want to increase the bias current, which moves your operating point closer to saturation, and you want to drop your plate voltage, which moves your operation point closer to saturation. Do you want to get your 6V6's really hot, reduce the headroom and get saturation distortion in your output stage? If so, you're on the right track.

If you've got a 100 Ohm cathode resistor and 31mA of bias current, it will shift your bias voltage 3.1V positive compared to a grounded cathode at the same current.

Don't you like the pretty little 6V6s? Assuming that this is a push-pull circuit with a bit of negative feedback, I really think you're better off running them at 18-25mA, cranking up the plate voltage beyond what their designers and God intended, and driving the fecal matter out of them. Then they will be happy little 6V6s. Oh yes. But that's just me.

[Enzo]

Tubes, I think you are confusing cathode bias and class A operation. Often class A amps are cathode biased, but that is not why they are class A.

[tubeswell]

Yep thanks for that reminder Enzo. Is it the Class A plate current characteristics that results in saturation distortion in the power tubes? Or can you get that with Class A/B?

I was hoping to get some nice distortion from the power tubes when I had the 5G9 cranked. When I was using it at practice for the first time last night I noticed that I couldn't get distortion when the amp was cranked, even tho' it was loud (this was after I had rebiased it down to 28mA), yet I did get fully cranked distortion earlier when I first got it going and the bias was around 31mA.

Tonight (thanks to MWJB's advice) I figured out that the tremolo footswitch should be switched to cut the trem completely when checking the bias, as I get a much more accurate reading that way. Dialing out the trem doesn't deal with the residual oscillation in the circuit. I have rebiased the amp to 22mA because this is where the plate and screen voltages seem to at least be relatively close, yet still be having the plate above the screen on both tubes (even if it is only by 1V!). At 20mA, the plate on one tube was below the screen by 2 or 3V. (Does that matter?)

[Rob Mercure]

Tubeswell,

You can even saturate your OT in Class C! Has to do with operating points and not output class. While some experimenters have made the scheme work well in general mixing cathode and fized bias hasn't got a great history. CBS engineers when they owned Fender added this feature to a few models - Bassman, Super, Bandmaster, etc., - and one of the most frequent "mods" is to rip this stuff out and return the circuit to pre-CSB fixed bias.

I do believe that K. O'Conner describes some methods of mixing in one of his volumes - you might want to check there.

Rob

[tubeswell]

Thanks for the info Rob

Given that it seems like this type of thing has been tried by CBS of all people and has been given teh thumbs down by subsequent modders, that I'm not sure that I'll bother following it up. Besides, I think I'm starting to get the hang of biasing the amp properly

Cheers

[hasserl]

I remember one of the members of the old Ampage web site tried this with a SF Bassman test mule, and he really liked the results. I think he was only using about a 25 ohm resistor, IIRC. This gave a few volts of Bias voltage, then he dialed in the rest with the adjustable fixed bias supply. He claimed this added some nice compression as a result of the bias shift from the cathode resistor. I've been wanting to give it a try sometime since then, just haven't gotten around to it.

Note, just because it became popular to rip out on old Fenders doesn't mean the idea is flawed, it could have just been Fender's implementation of it; or the poor performance could have been attributed to something else, but since this was a change everyone assumed it was responsible. Don't just go with the flow, give it a try for yourself and see what you think.

[stokes]

I tried it in a SE 6L6 amp I built and I like it.I ended up using a 75 ohm resistor.Originally I built it with fixed adjustable bias and tried an adjustable 500ohm rheostat for the cathode resistor and settled on the 75 ohm value.It gives it a bit more "singing" quality when pushed.At lower volumes it has little effect.

[tubeswell]

Okay so if I revisit this concept as hasserl and stokes seem to suggest (thanks guys), do I just fit the cathode resistor (making sure it is chunky enough to deal with the power dissipation) and then bias the tube with the adjustable bias until the tube is hitting something-percent of max dissipation at idle? What percent should that be?

If I try to work through this myself, I begin with an example of an amp that is straight cathode biased, and I would have 470R on each cathode or 250R in parallel (something like a tweed deluxe or 5E9A)

So if I go for (say) 1/2'n'half fixed/cathode bias, I would have a 250R resistor on each tube (or 120R for both tubes in parallel) and then use the bias adjustment pot and take measurements of the cathode voltage divided by the resistance to get the tube current. And would that mean (in this theoretical example) that I could run the tube at (say) somewhere between what I would run it at if it were either just cathode biased (i.e. max current at idle) or just fixed biased (i.e. upto 70% max dissipation at idle), being in other words halfway between 70% and 100% i.e. 85%?

Would the relative amounts of fixed/cathode biasing have a linear relationship to the (lets call it) 'safe' tube current? or not?

Presumably if the rk is say a 50R resistor on each 6V6, then I would have to set the adjustable bias colder than I would for the above example?

Is there a simple mathematical formula I could use to determine what the appropriate fixed bias level would be dependent on the relative extent to which the tube was cathode/fixed biased?

Can somebody please explain to me how a hybrid system works (in layman's terms) so I can work ths through?

I like the idea of being able to get more compression/sustain out of the 5G9. The tremolo on it sounds just awesome. All I want is more of a departure from 'clean' when I crank the vol pots.

Cheers

[hasserl]

Maybe I'm misreading your post, but I think you're confusing method of bias with operating range. JUst because you use cathode bias that does not mean the tube runs at max dissipation. In a Class A amp you can run the tube at max dissipation, because a Class A amp will not increase current draw across the tube when a signal is applied. But cathode bias does not equate Class A operation. I'll assume the amp in the example given is a Class A/B amp, just set the bias to achieve say 65 - 70% max dissipatioon. Use the plate to cathode voltage in your dissipation calculations. Also, use your ears when doing this, 70% dissipation isn't a target, it's a limit. If you like the sound and feel of the amp at 60% dissipation that is fine too.

[mooreamps]

Well, I think what one could do is, run the power tubes with self bias operating Class A, and then switch in the fixed bias to run the power tubes Class A/B.

-g

[tubeswell]

I'll assume the amp in the example given is a Class A/B amp, just set the bias to achieve say 65 - 70% max dissipatioon. Use the plate to cathode voltage in your dissipation calculations. Also, use your ears when doing this, 70% dissipation isn't a target, it's a limit. If you like the sound and feel of the amp at 60% dissipation that is fine too.

Ok thanks

[tubeswell]

Well, I think what one could do is, run the power tubes with self bias operating Class A, and then switch in the fixed bias to run the power tubes Class A/B.

-g

You mean switchable options right? Not running a combination of both types together?

With the 5G9, the tremolo comes off the bottom of the grid load resistors/bias voltage. So where would one run the trem from when switched to the cathode biased option? In my mind I saw a hybrid bias setup partly to overcome this issue, but with my limited imagination/experience I can't see where else to take the trem from?

[MWJB]

I think it might be worth taking a step back for a moment...a handful of mA (like the difference between 27 & 31mA) at the power tubes will affect tone, but it shouldn't be night and day with respect to whether you get reasonable distortion.

Have you tried a 12AX7 in V1? Fender used 12AY7 for improved fidelity & less distortion, but plenty of players may be perfectly happy with 12AX7.

[mooreamps]

You mean switchable options right? Not running a combination of both types together?

With the 5G9, the tremolo comes off the bottom of the grid load resistors/bias voltage. So where would one run the trem from when switched to the cathode biased option? In my mind I saw a hybrid bias setup partly to overcome this issue, but with my limited imagination/experience I can't see where else to take the trem from?

Yes, I mean running a combination of both types together. This is why it's called "combination bias". With repect to the 5G9, I'd have to see the print in order to make a comment on that.

-g