Power Tube Biasing and Headroom

[Structo]

Hi guys,

I've been researching the relationship between the bias current relationship of power tubes (6L6) and headroom.

There are a few conflicting opinions out there and it seems that the Hi Fi crowd has a bit of a different opinion verses the guitar amp crowd.

What I have been able to glean is that a power tube has a sort of bell curve in it's bias range.
According to what I have read that the highest headroom is in the middle of that curve which I guess would be in the 50% range.

[img:217:254]http://site.tubedepot.com/BiasPointFig4.gif[/img]

So take a 6L6GC that has a 30 watt dissipation rating.
In my amp I have a plate voltage around 440v.

Agreeing that the highest I should bias these tubes is 70% of the dissipation and at the other end at 50% I get a range of around 34ma to 47ma.

Would it be safe to assume that the highest headroom before breakup would be in the middle of those two extremes?

It seems the D amps prefer a bias setting on the cool side of things.
I usually have my amp biased around 34ma.
Would increasing that bias to say 40ma give it more headroom?

Or are there other factors I am not understanding that should be considered?

[dave g]

I think "headroom" is a oft misused/misunderstood term. Strictly speaking, it is the magnitude of an input signal that an amplifier may amplify without transitioning into a nonlinear region of operation. A lot of times, people think of it more as "how loud can the amp get before it distorts." That's a much more complicated situation, so I'll try to explain a bit here:

With any sort of active device, be it a BJT, MOSFET, or tube, you generally have a "linear" region of operation sandwiched in between two "nonlinear" regions of operation.

In the case of a vacuum tube in class A operation, the linear region of operation is bounded by Vg - Vk = 0 (we will naively assume this, but in actuality it is a bit more complicated) on one side, and on the other side by Vg = -Vcutoff where -Vcutoff is the grid voltage at which point no current flows from plate to cathode (which will depend on the load line).

So ideally to get the maximum amount of headroom, you would bias exactly in between the two nonlinear regions. If you didn't you would hit a "wall" on one side before the other, causing you to clip asymmetrically. If you want maximum headroom, you should be hitting both "walls" simultaneously. Of course, optimal headroom means symmetrical clipping, which means near total cancellation of even order harmonic distortion.

In the case of a class AB1 push-pull amplifier, it's not so simple because at any point in the input cycle, one tube should be conducting and the other should be (close to) cutoff. This means that you can theoretically swing a voltage (almost) twice what you could in class A operation, and therefore put out ~4x the power (not that simple, though...). The colder you bias, the less "overlap" there will be between the pair of output tubes, so the voltage swing you can accommodate is larger (This doesn't mean more power, necessarily, because the lower the current goes the less power is dissipated). So by biasing colder, you can increase your "input headroom", but at some point you start getting crossover distortion. The hotter you bias, the less likely you are to run into crossover distortion, but the less headroom you will have because you're biasing closer to class A operation.

It should be clear, then, that there is an "optimal" point (or region) where you have the ability to swing a large voltage, are pumping enough current to produce a lot of power, and are still preserving enough overlap between the two opposing sides of the waveform to ensure that there is no crossover distortion. As it turns out, this usually ends up being around 70% of the tube's rated plate dissipation. You can play around in the neighborhood of this value; biasing colder will give you more headroom depending upon your definition of the term, but it may not seem like it since your output power may go down. Biasing hotter may give you a bit more power, but will also decrease your headroom (again, in the strict definition of the word). So in either case, you may or may not really notice much in terms of gained/lost headroom.

I say just bias at 70% to start, and then try different bias voltages that put you between ~60 to ~80% and go with what sounds best (and doesn't leave you redplating, of course).

[Structo]

Thanks David, that is a good explanation.

[Andy Le Blanc]

Headroom is a qualitative term, it represents apparent volume before audible distortion.
That's why a 100w amp can be favored over a 25w amp, the 25w could be a much better amp,
but the 100w will seem to present with more clean volume,
even if the % distortion is very large your perception of "apparent clean headroom"
has nothing to do with the bias condition.


70% of what? Isn't it about the relationship of Po to %thd?

Are you biasing to a % plate dissipation at the zero signal condition?...
or at a full power condition?

What other assumptions have been made?

Voltage conditions? Effective plate load? Feedback?

[Ears]

Andy I think dave is right on the nail.
Headroom is the allowable input range before nominal distortion, nothing to to with actual volume but a lot to do with bias condition.
No reason 1W amps can't have higher headroom than 100W amps.
In class AB Push pull headroom will decrease as bias is made hotter from the bias condition that results in class B operation at one extreme right through to the class A condition where the bias point allows roughly and equidistant plate current swing from cut-off to grid current draw.
My 2c

Edit - A good and simple way to see this is to use Thompson's method to create some composite curves for different bias conditions. Also edited as I noticed later that I had completey reversed description of effect in hasty post.

[vibratoking]

I think dave and Andy both have valid points. Dave's strict definition is correct from a theoretical design standpoint. Biasing the output to equal swings allows for the greatest headroom electronically. This is how I think of headroom as a designer.

I think Andy is referring to headroom as a guitarist. As a guitarist, I only care about headroom from the standpoint of how loud can I play clean tones without audibly distorting. Such as trying to be heard above that hammer-fisted drummer without killing my clean tone. Maybe this is known by another name for which I am not familiar.

For me there are two levels of headroom. A high wattage amp with no electronic headroom won't work for me in a gigging situation and a low wattage amp with a lot of headroom won't work for me either. I need both to have what I consider useable headroom.

[Ears]

Well I guess if it's one of those discussions on headroom (including that which is called headroom by some but actually isn't headroom), sort of discussions...



But Structo's OP confuses a couple of things in my view. He's seems to be seeking a relationship between Bell curves of tube performance (manufacturer's tolerances etc) and circuit headroom. Headroom is determined by circuit parameters as much as tube tolerances.

[Jana]

Pffft... shorter people have more headroom.

[Andy Le Blanc]

Ok.... Here's a real question concerning headroom.

The room is crap, the drummer is too loud, you cant hear the vocals....

What do you do?

The first temptation is to bring up vocals right?...... wrong......

That's the first step towards loosing the mix, your baby sitting the room,
you have started a process where you will push the gain at the wrong place
in your gain structure, creating distortion and reducing your headroom.

What do you do?

Headroom is reference of transient quality and clarity of the program.

The right thing to do (unless you've screwed your self into a corner) is....
to bring the whole mix down to -10, reduce the input gain on your amps
driving FOH, and use your compressor to drive the system until your
bury the stick swinging bonehead, meeting the venue demands for volume
distortion free.

If you want a guitar amp to do that, get a possessor. Other wise use a small amp a good mic,
and leave the headroom to guy who's job it is to worry about it.

[Bob-I]

and leave the headroom to guy who's job it is to worry about it.

Well Put Andy. It's also important to me as a guitarist to tweak my amp to sound good out front, not the sound coming out of the speakers. That's the biggest advantage of in-ears, you hear the mike, not the speakers and can tweak the controls to sound good, making the sound man's job easy.

I've found that since I'm using the in-ears, I play 1) much cleaner and 2) much lower in volume.

[Andy Le Blanc]

Monitors are your friends, and if your sound man is on his nut, they will be
set up so you can hear what you want to hear...... period,

You're happy, the sound guy is happy, the FOH is good, the audience is happy,
and the even venue owner/promoter might have a smile.

[vibratoking]

Ok.... Here's a real question concerning headroom.

The room is crap, the drummer is too loud, you cant hear the vocals....

What do you do?

The first temptation is to bring up vocals right?...... wrong......

That's the first step towards loosing the mix, your baby sitting the room,
you have started a process where you will push the gain at the wrong place
in your gain structure, creating distortion and reducing your headroom.

What do you do?

Headroom is reference of transient quality and clarity of the program.

The right thing to do (unless you've screwed your self into a corner) is....
to bring the whole mix down to -10, reduce the input gain on your amps
driving FOH, and use your compressor to drive the system until your
bury the stick swinging bonehead, meeting the venue demands for volume
distortion free.

If you want a guitar amp to do that, get a possessor. Other wise use a small amp a good mic,
and leave the headroom to guy who's job it is to worry about it.

All true. It is always better to turn down. There are rooms where that is not possible. If the drummer is killing the room without the aid of the PA then the sound man can do nothing. FWIW, I wasn't suggesting a volume war, I was making a point regarding electronic issues vs audio issues and clean tones/headroom.

Well Put Andy. It's also important to me as a guitarist to tweak my amp to sound good out front, not the sound coming out of the speakers. That's the biggest advantage of in-ears, you hear the mike, not the speakers and can tweak the controls to sound good, making the sound man's job easy.

I've found that since I'm using the in-ears, I play 1) much cleaner and 2) much lower in volume.

Ears are great. I wear them too. In my opinion, you cannot listen to a feed from a board and assume that what you hear in your ears is what the audience hears. It simply is not true for many reasons - room ambience, close mic'd only, bone conduction, Fletcher Munson, the speakers...Too many things are fooling you and the list goes on and on. If you are listening solely to a close mic'd feed then how do you know what it sounds like out front? You are just adjusting your amp to make it sound good in your ears, which has little to do with what the audience hears. The best technique that I have come up with is to put a good condensor out in the audience and monitor that in one channel of my in ears along with my vocals. It allows me to hear what my guitar sounds like out front and also allows me to monitor what the soundman is putting out. I am the only vocalist in my band. The biggest benefit of ears for me is that I can have a personal mix with the vocal level that I like. Then I can 1) sing much cleaner and 2) sing more relaxed, on pitch and with much lower volume. I don't use ears for guitar volume reasons...I don't play that loud.

[Andy Le Blanc]

Your all used to the fender model right....

There's three very clear design choices in a fender that increase the "headroom".

The inverter, the plate to plate loading, and the feedback.
All in combination to reduce THD and increase the usable bandwidth at volume.

This is why it has become such a widely used format, and why every manufacture
is forced to market features and not good tone.

Somebody please post test results to support or invalidate the argument of
the original post.

You should see a result in the usable response of the amp.

Ambient mikes are very useful....but you'll never know what the venue
demands really are.... that's the the sound engineers job.
FOH is a juggling act between the artist, the venue, the audience and the promoters demands,
The performance monitors are never ment to sound exactly like the FOH.
They're there so the artist can hear what THEY need to hear in that work area, Its your job.

YOU ARE NOT THE AUDIENCE, the venue owner makes the final
presentation choices, not you. That's why sound guys get grumpy.
And if you cant work with your drummer, hire one you can,
The sound guy can LET YOU SUCK all on your own.

[dave g]

Sorry, headroom is not a qualitative term. It has a very precise definition.

Like they have done with similar terms ("gain" to refer to distortion rather than a dimensionless scalar constant representing the magnitude of an amplifier's output signal with respect to that of the input, "tremolo arm" to refer to the vibrato arm, etc...), guitarists have warped it into yet another musical misnomer.

Like I said before, the "guitarist's slang" definition of headroom refers to the maximum perceived volume level of an amplifier before an appreciable amount of distortion is generated.

You won't be able to come up with any sort of meaningful "test results" to bring quantitative meaning to this definition. I can't pretend to be an expert in psychoacoustics, but then again I don't think there's anybody on the planet who completely understands psychoacoustics. Take an amplifier with a TMB tonestack and turn the bass all the way down while kicking the treble all the way up; by the slang definition, the amp will have a lot more "headroom" then if you turned the bass up instead because we tend to hear highs as more "piercing" and "louder" than low frequencies transmitting the same amount of energy.

You're right, the feedback and output tube loading in a push-pull tube amp have a lot to do with perceived "headroom". It's mostly due to the fact that speakers are wildly reactive loads, and because of this often cause transient spikes in the output waveform (e.g., overshoot). These spikes can drive the output tubes into nonlinearity even if the amplitude of the output waveform is less than the maximum swing the output tubes are capable of. Adding negative feedback primarily acts to dampen these transients, which results in more perceived "headroom".

[Andy Le Blanc]

I think we all have been waxing poetic over the wrong points.

The power side of an amp is supposed to be designed around a known maximum
signal voltage for the inverter or driver. So that if you know what the max
undistorted signal is from your pre amp then you base your design assumptions
around that, and not the other way around.

A guitar amp is a tone source and not a hi fi amp, unless you design with that intent.

This places the headroom requirements at the designers digression.

So.... after choosing a ratio of feedback whats the max undistorted signal
can you derive from a signal applied to the inverter?
If say you get 75v rms at less than whatever distortion figure you choose
than you set your bias condition accordingly, say -37.5 so the max p-p
signal never pushes the grids of the power positive. Every thing else is dirt.
The head room is yours to design.