Half Power switch - best way?

[wizard333]

Assuming a 4 power tube amp, what's the consensus here on best implementation of half power:

1) Run a resistor in parallel with the speakers (Koch does this for example)

2) Turn off 2 tubes by:
a) Lifting Cathodes?
b) Lifting Screen Grids? I'd guess this is the touchier option due to the voltage on the screens vs. cathodes

3) Other? (no triode switch stuff please).

[mhuss]

I don't like the parallel resistor, it puts more stress on the tubes and will wear them out quicker.

Opening the cathode circuit is easy and safe (and keeps two tubes idling).

Adding a large-ish resistor in series with the screen grids also works (good for two tubes too).

[wizard333]

I don't like the parallel resistor, it puts more stress on the tubes and will wear them out quicker.

Opening the cathode circuit is easy and safe (and keeps two tubes idling).

Adding a large-ish resistor in series with the screen grids also works (good for two tubes too).

Why would the parallel resistor wear out the tubes quicker, assuming the impedance switch is set to half the speaker load and the resistor value matches speaker impedance (not the same thing, but close enough for this purpose)?

DO you know the calculation for the size of resistor to add to the screen grids to cut power ~50%?

Thanks!

[mhuss]

Running the two tubes at full tilt (or some subset thereof) will certainly wear them out more quickly then idling.

I typically don't bother changing the impedance setting because the mismatch reduces the power even further.

The parallel resistor would just be 16 ohms (or 8 or which ever tap you are parallel with) and probably a wattage rating similar to the amp (e.g., 100 watt amp, 100 watt resistor). Note that many newer high power resistors expect to mounted on an ambient temperature heat sink (not a hot tube amp chassis), and even the 'free air' ones need some air to meet their wattage rating.

I built an 8-ohm "1/4 power" box that has three 8-ohm 50 watt resistors (the speaker is the fourth, wired like a typical 4x12 cab). Still not a lot quieter...

[GainFreak]

Maybe it's a stupid idea but is it possible to just disconnect two of the tubes from the PI capacitors and temporarily connect them to an independent bias node. This way voltages and impedance won't change.

[Dave H]

Maybe it's a stupid idea but is it possible to just disconnect two of the tubes from the PI capacitors and temporarily connect them to an independent bias node. This way voltages and impedance won't change.

Voltages won't change but now you have two tubes driving the same impedance as four tubes were driving. You could set the impedance switch to half the speaker impedance to match for two tubes.

[The Dude]

Lots of correct methods. My preference is disconnecting cathodes with a couple of caveats. 1) Match speaker impedance depending on full or half power. 2) Make sure B+ doesn't go too high when 2 tubes are "disconnected"- usually not a problem.

[Steve A.]

With a twin reverb it was suggested to pull one pair of tubes and disconnect one speaker... that way there was no impedance mismatch.

Steve Ahola

[Enzo]

And again, these are guitar amps, no laboratory gear. having the impedance off by one "notch" won't hurt anything.

[g1]

And the mismatch will help further reduce power, so the switch will have a more audible effect.

[Steve A.]

I mentioned the impedance thing for the benefit of the EE's out there but the important thing was that disconnecting the second speaker really helped reduce the power of a twin reverb. This from an old AMPAGE post...

Steve

[The Dude]

And again, these are guitar amps, no laboratory gear. having the impedance off by one "notch" won't hurt anything.

Agree 100%. If only one speaker jack or no impedance switch, I wouldn't worry about it even a little bit. However, if there are different impedance jacks or a switch, why not do it "right"?

[Enzo]

Sure, but I see a half power switch on the panel, I might flip it. As a musician, how likely am I to also go in back and change the speaker connections? I bet my lunch money I won't. And I doubt someone will engineer their half power switch to not only cut power, but also reroute the output transformer wires.

Lose one speaker? Sure, in fact, simply doing that and nothing else should tame it a little.

[g1]

Back to the OP, my preference would be lifting cathodes, but under the 'other' category, have a look at the 94 Twin.
They switch down to '25W' mode by taking power off the mid point of the first totem pole series caps, while also changing the bias with the same switch.

[GainFreak]

Voltages won't change but now you have two tubes driving the same impedance as four tubes were driving. You could set the impedance switch to half the speaker impedance to match for two tubes.

The 2 "disconnected" tubes will be at idle (connected to the B+) just not driven so wouldn't the overall impedance stay the same? I'm sorry but I'm not very much into theory.