Biasing SE class A, fixed bias

[Paul P]

In all the info on the net regarding biasing output tubes and don't believe
I've seen mentioned how to bias a tube if it class A and fixed bias. The
recommendation for push-pull is 70% of max dissipation. For class A it's
90%, but this presumes cathode bias which will limit the increase in
dissipation.

My amp is single-ended with parallel 6v6s with fixed bias. I've been biasing
the tubes to 10w dissipation (27ma for 384 plate voltage) which sounds fine
but if I play a bunch of strong full chords on my guitar the current goes up
to about 45ma or so. I suppose the plate voltage drops a bit at this point
though I haven't measured it. If not, the dissipation would be around 17w
which is more than the rating for the tubes.

Could I be stressing my output tubes too much ? I don't want to have to
think about what's happening to them while I'm playing.

Paul P

[Steve Conner]

Dissipation under load shouldn't be a problem: of that 17W input you calculated, some of it ought to be going to the speaker, not all of it is being burnt up on the power tube plates.

If the current draw increases under load, then it's biased too cold to be Class-A anyway. A true Class-A amp will keep the same current draw from idle to flat-out, or even decrease slightly when it's driven hard into clipping.

[Paul P]

If the current draw increases under load, then it's biased too cold
to be Class-A anyway. A true Class-A amp will keep the same current
draw from idle to flat-out, or even decrease slightly when it's driven
hard into clipping.

I understand that a usual class A will be running at lower plate voltages and
higher current, and the current draw shouldn't vary much or actually go down
under load as you say. So what happens when you run a tube at higher plate
voltages and lower current, do you end up with some sort of lop-sided class A ?
It can't be anything else but class A since it's conducting all the time, but
current goes up under load so it's not a 'true' class A. I'm pretty sure I could
make the current draw go up even more with a stronger input.

There must be a reason you don't see this configuration often, if at all.
Perhaps it's because of this unwanted current draw ? I'm waiting for
delivery of some suitable cathode resistors so next time I have my amp on
the bench I'll set it up to run properly with cathode biasing (I can switch
between the two) so I'll be able to have a 'normal' amp but I'm curious as to
what I have now, and how to bias it properly.

I don't plan on playing flat out all the time, far from it, so I probably don't
have much of a problem.

Paul P

[Synaptic Amps]

Are you running NOS tubes or something? If not, being overly concerned about the tubes of all things just doesn't make a lot of sense. It's the transformers you should be worried about overheating if anything and I really doubt that'll be an issue running two 6V6s at 27mA. Maintaining a careful watch, throw in a cheap pair of 6V6s and crank up the bias current until you see 'em just start to red plate - calculate your current then back it off a bit. That'll give you a sense for how much juice they can really handle. Don't be a fool like me - one time I had some tubes red plating while I got distracted by someone and thinking I'd left the amp on standby I came back 15 minutes later to find the tubes smelling mighty toasty. I burned the labels right off of those 6V6s but truth be told - they still work just fine. Don't be so scared to crank 'em is my point. 27mA shared between two 6V6s isn't nearly enough, you're starving them. Try 60-90 mA. The tubes'll be able to handle it just fine. It's your transformers you've got to watch out for especially if your filament line was only intended for a single 6V6. Feel your PT with your hands and make sure it's running warm but not hot. If it's running cool to the touch, warm it up.

[Paul P]

I should have provided more information. I'm using new production Tung-sol
6v6GTs. The 27ma I mentioned is for each tube, at idle, as was the 45ma
playing loud. My transformer is a beefy 180ma Heyboer from Allen which is
rated for 6l6s so it should be ok. I like the idea of bringing the tubes up to
redness to see how much that is. I'll wait till I have a couple of tubes I don't
care about too much before I try that.

It seems to me that running them at 17w dissipation for any length of time
wouldn't be very good for them. But if some of that is getting used up
elsewhere, like Steve Conner says, then I'm probably ok.

Paul P

[Synaptic Amps]

Sorry, I misinterpreted your info. I love those Tung-Sol 6V6s, it was a similar EH 6V6 in my red plate experiment above.

Even per tube 27ma sounds very much too low. My 6V6 datasheet lists a maximum plate current at 45ma. I think you're falling victim to a sort of false logic that since you've exceeding the maximum rated plate voltage for your 6V6 that you have to make up for it by limiting the current. Two wrongs don't make a right, right? There's a long standing tradition in guitar amps for exceeding the maximum plate voltage, you're only 10% over. If you exceed the maximum rated plate voltage by 10% then it'll follow as a natural consequence that you'll exceed the maximum rated plate disipation by at least the same margin. Indeed, increasing the plate voltage would have a tendancy to draw an increased plate current. Boosting the voltage and starving the current seems like a double whammy of sorts.

I understand you're probably concerned that if you raise the idle current from 27ma to say 35-40ma that you'll end up drawing even more than 45ma on your loudest chords but this isn't necessarily true. As you said, if you were truly biased class A your idle current should be set at or near max dissipation from the outset. Your amp is therefore, as you describe it more likely class AB.

If you ultimately find your amp is running your 6V6s too hot for you to be able to comfortably bias them where you want them then you might try taking a step up to the new Tung-Sol 5881s - my favorite current production tube. They only cost a little more but they'll be able to handle your slighly elevated voltages better and let you get all you can out of your amp.

[mooreamps]

You could leave the plates where they are at, and drop the screens down somewhere in the 300 volt range. Then put in a 470 ohm 3 watt bias resistor, and see if you get any better results from it. Also, install a 1 K 3 watt screen resistor for the 6V6.

-g

[Bruce / Mission Amps]

I understand you're probably concerned that if you raise the idle current from 27ma to say 35-40ma that you'll end up drawing even more than 45ma on your loudest chords but this isn't necessarily true. As you said, if you were truly biased class A your idle current should be set at or near max dissipation from the outset. Your amp is therefore, as you describe it more likely class AB.

I thought class AB is a designation for two tubes in push pull... I understand what your are trying to say but don't you think once the two parallel tubes are driven hard enough to be pinched off in the negative going signal... that is class B? Well, that's the actual definition of Class B isn't it?
What I think he has is simply, over biased Class A SE PA with two parallel 6V6s.
Also, I'm an old school guy and was taught that all that dissipation data is based on DC parameters, not the AC part.
I think with proper cooling, a good power tube can be pushed about 20%-30% past it's DC dissipation rating.
Of course most guitar amps chassis and builders are guilty of TERRIBLE cooling thought.

[Rob Mercure]

Seems like what this fellow wants to do is feed a good 400 Hz, or so, signal into the input with sufficient level to produce desired clipping levels on a scope and then to insure that the clipping on the scope is occuring symmetrically - no? This way he's Class A with as much dissipation as he can get w/o meltdown. Should be a pretty "squarish" wave. Then he can back down as much as he wishes.

[Paul P]

Actually I'm just looking to not burn up my output tubes

However I'm also interested in what is going on in my amp and why you don't
see high plate voltage (380+) along with fixed bias in many SE amps. It looks
to me like the answer is probably a combination of 1) the current can get out
of hand and exceed the power dissipation for the tube and 2) the output at
some point will be distorted asymetrically.

I have fixed bias in my amp for when and if decide to go to push-pull output
tubes. I've been planning on using cathode bias for SE operation but since
I don't have the proper resistors (470 ohms / tube isn't enough) I've been
using fixed bias, which seems to work well enough under less than maximum
conditions.

Paul P

[kg]

Actually I'm just looking to not burn up my output tubes

However I'm also interested in what is going on in my amp and why you don't
see high plate voltage (380+) along with fixed bias in many SE amps. It looks
to me like the answer is probably a combination of 1) the current can get out
of hand and exceed the power dissipation for the tube and 2) the output at
some point will be distorted asymetrically.

I have fixed bias in my amp for when and if decide to go to push-pull output
tubes. I've been planning on using cathode bias for SE operation but since
I don't have the proper resistors (470 ohms / tube isn't enough) I've been
using fixed bias, which seems to work well enough under less than maximum
conditions.

Paul P

you just don't see very many high plate voltage SE designs period, at least with typically used receiving tubes. that's because SE for audio use requires class A, and class A with high Vp means you're throwing off a lot of heat all the time. most receiving tubes will have a <30w Pa max.

steve hit the nail on the head... if the average plate current changes drastically (ie more than 5-10%) then that's a good indication you have asymmetrical clipping going on--BUT that also supposes you have a symmetrical signal going INTO the output stage. that will not likely be the case with anything other than a signal generator feeding right into the grids. what could very well be happening is that your DRIVER stage is clipping and the output stage is faithfully reproducing that lopsided signal. or the asymmetric clipping could be happening further upstream.

using cathode bias will do two things: add degeneration at DC which will help prevent thermal runaway, and drop some of that plate voltage down so that you're not running the tubes so hot. every watt you throw off with the Rk is one watt you don't place on the tubes.

basically you're SOL if you want to run a clean class A with 6v6s and nearly 400v across them. however some compromise could still sound sweet and not kill the tubes prematurely. as bruce said the big enemy is heat, and good ventilation or perhaps even forced air cooling will go a long way to keep that at bay.

hth
ken

[mooreamps]

Well, here's the thing. I believe a vacuum tube is more sensitive to an over-current condition, rather than in an over-voltage condition. I don't view a vaccum tube as an "electrical device". I view a vacuum tube as a mechanical device, with electrical properties. ..... So, I don't believe running 400 volts on the plate of a 6V6 is going to hurt it so long as one does not exceed the max plate dissipation.... However, to keep the plate current down such that you can run the tube at a proper bias, lets say for Class A, of -10 volts or so, requires to run the screen at around 280 to 290 volts. This means you have to drop the screen voltage down by 100 volts to keep the plate current in check. So if we say the max plate dissipation for a 6V6 is 14 watts, then max plate current at 400 volts works out to be 35 milliamperes. ...


Now as far as running a pair of them in push/pull ; either as class A or class A/B ; using either fixed bias, or self bias, or even combination bias, is solely up to the designer. These are some decisions I am considering right now on my next build. I'm really thinking about bringing up an 18 watt mixed mode power amp, with 2 EL-84's running class A/B, probably using fixed bias ; along with 2 6V6's running class A ; probably running self bias. It's very possible I may end up running all four tubes with fixed bias. I just haven't decided yet.


-g

[imaradiostar]

I some to recall reading in an old tube book about using plate loads to compensate for high voltages and prevent a tube from exceeding it max dissipation. Maybe this only applies to push-pull designs but sticking with the aforementioned logic you could use a tube well over its rated wattage just like guitar amps have done for years. Would the high plate impedance and a fixed bias setup allow you to take advantage of the higher B+ voltage and generate cleaner if not higher power?

As an example I have a pair of SE transformers with 9k:8 ratios. I haven't used them in anything because the ratios are a little high. I wonder if a 6v6 or 6005 tube with 420ish volts at the plate and a fixed bias arrangement would sound decent using this sort of setup? I calculate around 30ma per tube would be around 12-13 watts idle dissipation. This is less than many Fender amps run. I guess I could build it and see.

I have been wondering about this concept. Thanks for posting the thread Paul!

jamie

[Steve Conner]

So, I don't believe running 400 volts on the plate of a 6V6 is going to hurt it so long as one does not exceed the max plate dissipation.... However, to keep the plate current down such that you can run the tube at a proper bias, lets say for Class A, of -10 volts or so, requires to run the screen at around 280 to 290 volts.

This is a good point, and it's the reason why you won't see many SE amps with high plate voltage and fixed bias.

If the plate voltage is high, then the plate current must be kept low to save the tubes from overheating. If you keep the plate current down by applying a large negative grid bias, then the stage will clip asymmetrically, with the current increasing under overdrive conditions, which is exactly what the original poster noticed. It does this because there's a lot more headroom to push the control grid positive than negative, ie, the tube is running at much less than half of the maximum current it can give.

And yes, reducing the screen voltage would be one way to fix it. This wasn't done in commercial SE amps, because they were usually the cheapest practice amps, and wouldn't tolerate the extra cost of a gas regulator tube, dropper resistor or whatever.

However, if it sounds good, then who's to say that it needs fixing?

[imaradiostar]

And yes, reducing the screen voltage would be one way to fix it. This wasn't done in commercial SE amps, because they were usually the cheapest practice amps, and wouldn't tolerate the extra cost of a gas regulator tube, dropper resistor or whatever.

However, if it sounds good, then who's to say that it needs fixing?

I think this is a perfect application for the circuit Bruce mentioned elsewhere- use one set of diodes into a 40uf cap for the plate supply and a second set into a choke input filter for the screen and preamp supply. I think this would do the trick!

jamie