Bias: Grid-Voltage

[HotBluePlates]

Guitar amp enthusiasts fixate on "idle plate current" when discussing tube bias.  However, the tube's grid-volts (or more correctly, "grid-to-cathode voltage") is just as important, and sometimes even more informative.

Let's consider a case using just the conditions printed by the tube manufacturer in a data sheet (click the link for the full-size image):



   - Class AB Push-Pull
   - Plate & Screen Volts:  250v
   - Grid Volts:  -15v   (a note indicates this push-pull section is fully-driven at 15v peak input)
   - Idle Plate Current:  35mA per tube
   - Gm:  3.75mA/volt
   - Plate Load:  10kΩ plate-to-plate

   - Max Signal Power Output: 10 watts

       * Note:  the "Max Sig Plate Current" as-documented in the sheet requires some explaining; ignore it for now.


We take the tube manufacturer at their word that this power section is "Class AB."  That means one side is cut off at peak output power, so the load seen by the remaining tube is "Plate-to-Plate Impedance / 4":  10kΩ / 4 = 2500Ω.

Max drive signal is when the peak AC volts equals the grid-bias voltage:  15v peak into a grid biased at -15v.

Gm has been rendered in the European manner:  3.75mA per volt.  Peak plate current beyond the idle plate current is:
     15v peak x 3.75mA/volt = 56.25mA peak additional plate current.

Total peak plate current = Idle plate current + Peak additional plate current = 35mA + 56.25mA = 91.25mA

Tube plate current is pulled through the OT primary impedance, and creates a voltage drop across that impedance:
    Peak Current x OT Impedance = 0.09125A x 2500Ω = ~228 volts

    The power supply must have enough voltage to support the peak volts dropped across the OT primary, and still leave some volts across the tube.
    250v - 228v peak = 22v ----> absent curves, let's assume the tube can actually pull its plate this low at 250v plate & screen.

Peak Power = Peak Volts x Peak Current = 228v peak x 0.09125A peak = 20.82 watt peak.

RMS Power (of a Sine Wave) = Peak Power / 2 = 20.82 watts peak / 2 = 10.4 watts RMS ----> what the sheet claimed minus some rounding.


So What?

   - Idle Plate Current is not the end-all, be-all of biasing.

   - The output section is designed to deliver a certain power output, and the designer juggles tube plate current (when fully driven), supply volts, and OT primary impedance as the main factors to consider.

   - Class A output stages typically swing plate current from "Idle Current" to "2x Idle Current" to "Almost Zero Current."  When everything else is properly juggled, max output power occurs when the tubes are idled at 100% plate dissipation (and therefore the largest feasible Idle Current).

   - Class AB and Class B amps have a Peak Plate Current that is more than "2x Idle Current" (our example had a peak plate current ~2.6 times idle plate current).  Sometimes the peak is 5-6x idle plate current, as in the Super Reverb.

   - When the Peak Plate Current is many-times the Idle Plate Current, cooler bias is needed to keep the tube from overheating.  Exactly what "Idle Plate Current" that requires changes depending on how-many-times-more the Peak Plate Current is.

   - Individual tubes won't deliver the same idle plate current with a given plate/screen/grid voltage, nor will they exhibit the same Gm.  However, when you actually measure hundreds of tubes, you find they land around an Average Value, which then gets printed in the tube data sheets.

   - The variation between "tube data sheet values" and an actual tube means the actual tube  may deliver a few-tenths-of-a-watt more/less than what was calculated/claimed.  Since hearing is logarithmic, "double-loud" requires "10x the power," and 1/2dB is about as small a change as we can easily hear, it would take a change of about +/-1.5 watts for tube characteristic variance to even make an audible difference.

[shooter]

Idle Plate Current is not the end-all, be-all of biasing.

as a tech;
I set idle bias, verify the system is working properly, NO signal, do the napkin math with numbers measured, re-adjust bias as necessary.


I monitor (PA tube(s) via 1 ohm resistors cathode to ground), all knobs at "5", INPUT signal ~~ 100mVac rms (adjust to stay Class A) for 5 minutes.


More math using stored AVG value of 1 ohmers, during the 5 minutes and Note MAX/Min
Adjust bias accordingly, ie; If the AVG math says my bias is HOT, I adjust cooler.


I let system cool down for 15 minutes then bring it up fully operational, speaker, guitar, pedals, ...monitor 1 ohmers, and play for 20 minutes (simulating the owners "gene")
math again, bias adjust again.

[HotBluePlates]

It's worth pointing out that one could set the bias voltage per the schematic, then measure the resulting plate current with average-performing tubes.  Thereafter, one has a known-current-measurement applicable to that amp.


Mesa uses non-adjustable fixed-bias, and sells tubes that "meet Mesa's specs."  Some folks get suspicious that Mesa is just trying to find an angle to make more money.  But Mesa will tell you, "We design around an 'average tube' and screen the tubes to verify they indeed perform like the data sheet says."

Mesa's argument then continues:  "When the tube performs the way the data sheet says, the one grid-bias-voltage chosen at the design stage does indeed deliver the expected results (including idle plate current) in the amp."

[shooter]

the one grid-bias-voltage chosen at the design stage does indeed deliver the expected results

from a tech/electronic standpoint that works, then you add "guitarist" to the formula and it all goes sideways 


It sounds to weak
I want more growl
I don't care about tube life, I need my sound

[pdf64]

It's worth pointing out that one could set the bias voltage per the schematic, then measure the resulting plate current with average-performing tubes.  Thereafter, one has a known-current-measurement applicable to that amp. ...

It's worth noting that some schematics should perhaps be treated with caution, as their noted voltages can be somewhat different to many of the amps 'as built', eg 6G3 :)