Bias Excursion

FractalAudio

Administrator
Fractal Audio Systems
Moderator
I talked about Cathode Squish in the previous column. A similar phenomenon in tube amps is bias excursion. In this case I'm referring to bias excursion in the power amp.

Bias excursion occurs because the power tube grids forward conduct when the grid voltage is slightly greater than the cathode voltage. Now this isn't a problem by itself. However almost all tube amp designs use a capacitor coupled grid circuit. The phase inverter is coupled to the power tube grids via a capacitor.

When the grid voltage exceeds the cathode voltage, which is typically zero volts in a fixed-bias topology, the grid will become forward-biased and look like a low resistance. This clamps the grid side of the coupling capacitor. This occurs when the phase inverter signal is large and swings toward the B+ supply. When the phase inverter signal swings the opposite direction the grid stops conducting and the capacitor is no longer clamped. However there is now excess charge on the capacitor. During the time the capacitor was clamped charge was building up on the phase inverter side.

When the grid comes out of conduction that charge effectively reduces the power tube bias. For example, a typical 6L6 is biased around -50V. The clamping action would then push the bias voltage even more negative, say -75V. In some designs the the bias voltage can be reduced by nearly 100%! Since the bias voltage is shifted the phenomenon is referred to as "bias excursion".

Like cathode squish, bias excursion pushes the power amp from Class-AB operation towards Class-B operation. As we know Class-B operation has lots of crossover distortion. Now this may seem bad but, in fact, there are positive attributes associated with bias excursion. When designed correctly bias excursion can actually help an amp sound more "open". This happens because as the bias shifts the gain of the power tubes decreases. This in turn prevents the power tube plates from saturating as easily. However too much bias excursion leads to what is referred to as "blocking distortion" which can make an amp sound farty and generally unpleasant. Blocking distortion occurs when the bias shifts so much that the tubes are basically shut off for a period of time. If the capacitor charges up rapidly but bleeds off slowly, combined with lots of excursion, this leads to blocking distortion.

There are three associated parameters in the Axe-Fx II that allow you to alter the bias excursion behavior: Bias Excursion, Excursion Time and Recovery Time. Bias Excursion controls the amount of bias excursion. The higher the value the more the bias shifts when the virtual power tubes are overdriven. Excursion Time controls how rapidly the coupling capacitor charges when the virtual power tube grids are conducting. Recovery Time adjusts how quickly the excess charge bleeds off when the virtual grids are not conducting.

Preamp tubes also exhibit bias excursion and too much of it can cause blocking distortion. Like power tube bias excursion, though, a little bit can help. The trick is getting the right amount.
 
Excursion Time and Recovery Time are not accessible on hardware. Axe-Edit only?
 
Excursion Time and Recovery Time are not accessible on hardware. Axe-Edit only?

I believe so. They are accessible from the debug build. I'll have to check if they are accessible via Axe-Edit.
 
Very cool! So in sticking with the most realistic settings for Bias Excursion, Excursion Time and Recovery Time, would that be to leave them as default?


 
There are three associated parameters in the Axe-Fx II that allow you to alter the bias excursion behavior: Bias Excursion, Excursion Time and Recovery Time. Bias Excursion controls the amount of bias excursion. The higher the value the more the bias shifts when the virtual power tubes are overdriven. Excursion Time controls how rapidly the coupling capacitor charges when the virtual power tube grids are conducting. Recovery Time adjusts how quickly the excess charge bleeds off when the virtual grids are not conducting.

Would you be able to use these controls to simulate a directly driven power tube grid circuit (no capacitor/clamping/recovery) where the grids are driven positive without clamping? I've always been curious to hear how that would sound...some tube books/sites say it can be interesting. Another informative writeup, thanks.
 
Very cool! So in sticking with the most realistic settings for Bias Excursion, Excursion Time and Recovery Time, would that be to leave them as default?
Sticking with the default values would leave them closest to the original.
 
I talked about Cathode Squish in the previous column. A similar phenomenon in tube amps is bias excursion. In this case I'm referring to bias excursion in the power amp.

Bias excursion occurs because the power tube grids forward conduct when the grid voltage is slightly greater than the cathode voltage. Now this isn't a problem by itself. However almost all tube amp designs use a capacitor coupled grid circuit. The phase inverter is coupled to the power tube grids via a capacitor.

When the grid voltage exceeds the cathode voltage, which is typically zero volts in a fixed-bias topology, the grid will become forward-biased and look like a low resistance. This clamps the grid side of the coupling
capacitor. This occurs when the phase inverter signal is large and swings toward the B+ supply. When the phase inverter signal swings the opposite direction the grid stops conducting and the capacitor is no longer clamped. However there is now excess charge on the capacitor. During the time the capacitor was clamped charge was building up on the phase inverter side.

When the grid comes out of conduction that charge effectively reduces the power tube bias. For example, a typical 6L6 is biased around -50V. The clamping action would then push the bias voltage even more negative, say -75V. In some designs the the bias voltage can be reduced by nearly 100%! Since the bias voltage is shifted the phenomenon is referred to as "bias excursion".

Like cathode squish, bias excursion pushes the power amp from Class-AB operation towards Class-B operation. As we know Class-B operation has lots of crossover distortion. Now this may seem bad but, in fact, there are positive attributes associated with bias excursion. When designed correctly bias excursion can actually help an amp sound more "open". This happens because as the bias shifts the gain of the power tubes decreases. This in turn prevents the power tube plates from saturating as easily. However too much bias excursion leads to what is referred to as "blocking distortion" which can make an amp sound farty and generally unpleasant. Blocking distortion occurs when the bias shifts so much that the tubes are basically shut off for a period of time. If the capacitor charges up rapidly but bleeds off slowly, combined with lots of excursion, this leads to blocking distortion.

There are three associated parameters in the Axe-Fx II that allow you to alter the bias excursion behavior: Bias Excursion, Excursion Time and Recovery Time. Bias Excursion controls the amount of bias excursion. The higher the value the more the bias shifts when the virtual power tubes are overdriven. Excursion Time controls how rapidly the coupling capacitor charges when the virtual power tube grids are conducting. Recovery Time adjusts how quickly the excess charge bleeds off when the virtual grids are not conducting.

Preamp tubes also exhibit bias excursion and too much of it can cause blocking distortion. Like power tube bias excursion, though, a little bit can help. The trick is getting the right amount.

That's what I thought.
 
The trick is getting the right amount.
Understatement of the week!

Bias excursion is in front panel Amp adv parameters

Excursion time and Excursion recovery used to be there but were removed a while back

I haven't looked in tech notes, but I'd be curious how CF threshold, time, comp, and ratio influence tone/feel. Some of those exist on both dynamics tab and the adv tab.
 
I talked about Cathode Squish in the previous column. A similar phenomenon in tube amps is bias excursion. In this case I'm referring to bias excursion in the power amp.

Bias excursion occurs because the power tube grids forward conduct when the grid voltage is slightly greater than the cathode voltage. Now this isn't a problem by itself. However almost all tube amp designs use a capacitor coupled grid circuit. The phase inverter is coupled to the power tube grids via a capacitor.

When the grid voltage exceeds the cathode voltage, which is typically zero volts in a fixed-bias topology, the grid will become forward-biased and look like a low resistance. This clamps the grid side of the coupling capacitor. This occurs when the phase inverter signal is large and swings toward the B+ supply. When the phase inverter signal swings the opposite direction the grid stops conducting and the capacitor is no longer clamped. However there is now excess charge on the capacitor. During the time the capacitor was clamped charge was building up on the phase inverter side.

When the grid comes out of conduction that charge effectively reduces the power tube bias. For example, a typical 6L6 is biased around -50V. The clamping action would then push the bias voltage even more negative, say -75V. In some designs the the bias voltage can be reduced by nearly 100%! Since the bias voltage is shifted the phenomenon is referred to as "bias excursion".

Like cathode squish, bias excursion pushes the power amp from Class-AB operation towards Class-B operation. As we know Class-B operation has lots of crossover distortion. Now this may seem bad but, in fact, there are positive attributes associated with bias excursion. When designed correctly bias excursion can actually help an amp sound more "open". This happens because as the bias shifts the gain of the power tubes decreases. This in turn prevents the power tube plates from saturating as easily. However too much bias excursion leads to what is referred to as "blocking distortion" which can make an amp sound farty and generally unpleasant. Blocking distortion occurs when the bias shifts so much that the tubes are basically shut off for a period of time. If the capacitor charges up rapidly but bleeds off slowly, combined with lots of excursion, this leads to blocking distortion.

There are three associated parameters in the Axe-Fx II that allow you to alter the bias excursion behavior: Bias Excursion, Excursion Time and Recovery Time. Bias Excursion controls the amount of bias excursion. The higher the value the more the bias shifts when the virtual power tubes are overdriven. Excursion Time controls how rapidly the coupling capacitor charges when the virtual power tube grids are conducting. Recovery Time adjusts how quickly the excess charge bleeds off when the virtual grids are not conducting.

Preamp tubes also exhibit bias excursion and too much of it can cause blocking distortion. Like power tube bias excursion, though, a little bit can help. The trick is getting the right amount.

Not to put too fine of a point on it, and since this doesn't occur in a vacuum - in your research: How much does NFB affect this?
 
Fascinating. I really enjoy these threads because I enjoy understanding what's happening when I crank said parameter.

Keep these coming, Cliff.
 
3rd paragraph- just made me want to say something like 'yes, that is the widely accepted view' to insinuate that it was an oversimplification, to cover up the fact that I hadn't the slightest clue what all those big words meant and wanted to cry.
 
Not to put too fine of a point on it, and since this doesn't occur in a vacuum - in your research: How much does NFB affect this?

The question should be, what other parameter might interact with bias excursion? I would say, turn Power Amp low cut to higher values when you have lower bias excursion amount and lower recovery time for detailed simulation, as said 18 months ago....
http://forum.fractalaudio.com/axe-f...-version-12-04-public-beta-45.html#post984525
 
The question should be, what other parameter might interact with bias excursion? I would say, turn Power Amp low cut to higher values when you have lower bias excursion amount and lower recovery time for detailed simulation, as said 18 months ago....
http://forum.fractalaudio.com/axe-f...-version-12-04-public-beta-45.html#post984525

Thank you for the helpful reply.

My interest was a theoretical one. A zero-NFB state is different than a non-zero-NFB state. No biggie.

EDIT:

On one of the amplifiers I built, I think I counted about 41 circuits that could be construed to be 'unique' (certainly not 'independent'). That is a lot of interaction or cats to herd if you prefer analogies.
 
Last edited:
Back
Top Bottom