Understanding Power Tube Bias

FractalAudio

Administrator
Fractal Audio Systems
Moderator
The Advanced menu of the amp block has a parameter called "Pwr Tube Bias". This sets the quiescent operating point of the virtual power tubes. What is the quiescent operating point you ask? It is the amount of idle current flowing through the virtual tubes when no signal is present.

Power tubes are basically nonlinear controlled current sources. They can only sink current (current flow in one direction only) so to get an alternating signal you have to have some amount of idle current then you modulate that.

The push-pull power amp was invented to increase efficiency by allowing lower idle currents. The name "push-pull" refers to the fact that one tube is responsible for mainly the positive portion of the waveform and the other for the negative. Some overlap of the responsibility is required since power tubes are nonlinear and without that overlap crossover distortion occurs.

Setting the idle current, or bias as it's commonly known, affects the resulting transfer function of the power amp. Too little bias and there can be excessive crossover distortion. So-called "Class A" amps bias the tubes quite high.

The following graph depicts the transfer function of the Axe-Fx II for five different bias settings: 0, 0.2, 0.4, 0.6 and 0.8. These are the red, green, blue, cyan and black traces, respectively.

bias_plot.jpg


Note the severe crossover distortion at zero bias. At a bias of 0.4 the transfer function is almost a perfectly straight line. At 0.8 the response gets softer. Amps like AC-30s run the bias around 0.7 which gives them the soft distortion characteristics. Note that this applies to power amp distortion which only occurs when the power amp is driven hard (Master Volume set high).

Some amps are intentionally biased cold to generate crossover distortion. Small amounts add an aggressive distortion. Some amps (i.e. Boogies) are intentionally biased cold to avoid having to set the bias and thereby reducing maintenance and warranty costs.

Negative feedback around the power amp (Damping) further linearizes the power amp. So the transfer functions depicted are only accurate when Damping is zero.

In conclusion, Power Tube Bias is a powerful parameter that can allow you to fine-tune the power amp distortion characteristics to your particular style.
 
You read my mind Cliff - I was thinking about asking about this parameter. Does the parameter relate one-to-one to bias current (i.e. .350 = 35ma)?

Thanks for the explaination.
 
So in something like a "hi end audio" tube power amp, would the typical approach be to set bias very high as well as negative feedback, or are there other factors involved?

I guess I'll have to try it out for myself...

Thx again for taking to time to explain these little things to the forum.
 
so cliff is it ok to plug my guitar in and start rockin or will it explode in my face ???? .....joking ,i'm just unfamiliar with so many in depth details ,happy YOU know those technical affairs ,thats what your responsable about and thats why we all trust you , thank you for sharing all that infos ,allthough i dont really dig them so much ,i just plug guitar tweak bass,tremble gain and off it goes ,the rest comes out from fingers and heart
wish you the best to all your team
keep on rockin in the ''free?'' world
 
Thanks for this- this is one of the best explanations for how power amps handle a signal that i've seen. Clear and concise!
 
Are there any sims in the AxeFx whose real amps are biased at .8 or higher?

What are most of the aggressive, high-gain amps biased at?
 
You read my mind Cliff - I was thinking about asking about this parameter. Does the parameter relate one-to-one to bias current (i.e. .350 = 35ma)?

Thanks for the explaination.

No, there is no relation between the calculated number in the axefx and the effective current flow in a circuit per one output tube! There were amps working on a higher plate voltage, set their operation point lower....otherwise they run too hot!
The transfer curve above is a theoretical attempt, mainly to explain you the difference on the transfer function between too low, almost ideal and higher bias point.

@ Philip34: Class A amps have higher values.
For a aggressive "explosive" sound on high gain amps, set the bias point in the axefx a bit lower than the default value.....
 
here is a "translated" picture of the above transfer curve. Unfortunately, I got no fancy app to draw sine waves - so this is handmade, hope you guys don't mind.....it looks very crappy, but anyway - you might get the idea.....

red is still zero, blue is 0.4

bias_plot.jpg


@ Cliff: Sorry for hijacking your picture...:numbness:
 
No, there is no relation between the calculated number in the axefx and the effective current flow in a circuit per one output tube! There were amps working on a higher plate voltage, set their operation point lower....otherwise they run too hot! The transfer curve above is a theoretical attempt, mainly to explain you the difference on the transfer function between too low, almost ideal and higher bias point. @ Philip34: Class A amps have higher values. For a aggressive "explosive" sound on high gain amps, set the bias point in the axefx a bit lower than the default value.....

I know class A amps run hot, but are there any that are biased at .8 or hotter in the AxeFx?
 
Is the virtual power amp driven by an "ideal" phase-inverter so both the positive/negative excursions of that stage are exactly gain matched and of equal amplitude, or does that depend on the particular amp sim/circuitry, etc.? Just curious if you've found any kind of desirable distortion/coloring from symmetric/asymmetric PI output waveforms driving the power amp.
 
Is the virtual power amp driven by an "ideal" phase-inverter so both the positive/negative excursions of that stage are exactly gain matched and of equal amplitude, or does that depend on the particular amp sim/circuitry, etc.? Just curious if you've found any kind of desirable distortion/coloring from symmetric/asymmetric PI output waveforms driving the power amp.

The PI model is asymmetric (usually). There is a slight imbalance between the gains as in the real amps. This adds a little warmth IMO.
 
I think these are great discussion topics. I wish you had more time to cover every aspect within the Axe for each amp and effect in detail. It's much better then posting about emotional issues :lol , again great stuff.
 
I appreciate these pearls of wisdom but a lot of them them I just can't wrap my head around.
To simulate an amp that is biased hotter do you turn up the power tube bias?
 
Back
Top Bottom