Some questions on a mid60s Vox AC50

[Robert M. Martinelli]

The fixed term just means that the bias voltage is set by potential divider (usually resistive), the variable term is used to describe circuits in which the bias voltage can be easily adjusted on a pot. But they're essentially the same thing.
The other type of bias is cathode bias, where the grid is held to DC ground, with the cathode current generating a bias voltage which kind of balances itself.

Hi Peter,
"Fixed", "cathode" and "automatic" bias are only different words to describe the same thing. The bias voltage is obtained by means of a bias resistor located between the cathode and ground, resulting in a voltage divider in which the first leg of the divider is the tube itself, and the second leg is the bias resistor. The cathode is lifted from the GND potential thus making the control grid negative with respect to the cathode.

This system is simpler, and it' s almost universally used in gain and PI stages, where grid voltages and cathode currents are minuscule, but, when it comes to output stages, it has two drawbacks - firstly it needs high power resistors, secondly it' s applicable with good results only with tubes with reasonably low currents, where biases in the region of -10 to -12 VDC are needed ( EL84/6BQ5 ).

With tubes needing bias voltages from -35 to -45 VDC ( EL34s/6CA7s - 6L6s/5881 and the like ) and high currents this system becomes inefficient as you would have to drop all the above voltage on the resistor at a high current, so the resistor's power rating would grow accordingly; Vak voltage would decrease as well ( because you "lifted" the cathode potential away from GND ), resulting in the tube being supplied at an "apparently" lower voltage and decreasing the stage's output power.

That is where "adjustable" or "variable" bias comes to help :
"Adjustable" or "variable" bias is obtained by means of an external source, which generates a negative voltage to be applied on the control grid. This is then supplied to the control grids via trimmers acting as voltage dividers ( only they work in the negative region ). This way the cathodes can be connected directly to ground, ( and no power gets wasted across the cathode resistor ) and the output stage can release its full power.

To complement and further prove all of the above, if one would like to build a "variable" bias network with a cathode potentiometer in place of the fixed resistor on a tube high power output stage, the pot should be able to withstand many watt of dissipated power - that would be expensive, less efficient and unreliable, so no one does that.

Hope this helps

Best regards

Bob