Half power switch question

[markmalin]

Hi all,


I'm going to use this circuit as a half power switch in an AC30 style build.  In this example, in the "High" setting the bias resistance is 50 Ohms, and in Low it's 100 Ohms.  If I understand correctly, the current draw for 4 tubes is twice the amount as 2 tubes, so halving the resistance on the "High" vs "Low" setting would keep the bias voltage about the same.  Why is it not necessary to also switch bypass cap values since the resistance changes?  (does this act like an RC filter?) 


Thanks guys.  I'd like to understand this better.


Humbly,
Mark.

[robrob]

In the "Hi" position all 4 power tubes share the 50 ohm resistor and pull all their current through it which is equivalent to 200 ohms for each tube.

In the "Low Power" position the the two right side power tubes' cathode is completely disconnected so they won't function. The bias resistance is doubled to 100 ohms since only two tubes are now used which gives the same equivalent of 200 ohms per tube.

The cap value doesn't need to be switched because the 25uF is enough for a full frequency AC signal bypass gain boost no matter the cathode resistance.

[markmalin]

Thanks.  That makes sense.




What would the switch have to be rated for Voltage and Current wise??

[shooter]

you could use about any switch, I believe.   If memory serves me, my quad ran about 50mA per tube with a around 6v bias on the cathode R's.  So .2A  and 6v, double and you're still way under most switches max rating

[markmalin]

you could use about any switch, I believe.   If memory serves me, my quad ran about 50mA per tube with a around 6v bias on the cathode R's.  So .2A  and 6v, double and you're still way under most switches max rating

Thanks.  That's good to know.  I've got a spare 250V switch which will do the trick :)  Overkill, but at least I know it's not going to get hot.

[kagliostro]

If you go half power that way, you must connect the speaker to the OT as to double the reflected impedance

If you don't want to swap the OT connection of the speaker, you must obtain 1/2 of the power shorting to ground two of the power tube grids



K

[markmalin]

I thought the impedance mismatch at half power wouldn't be that big of an issue ??  Granted, not having to deal with the impedance mismatch is kind of attractive ;)


So K, going your route, the two inner tubes are still in the circuit so still drawing current through the cathode bias resistor, and no need to change the value when running 1/2 power?  Hoffman's current circuit uses a 50 Ohm/10W...(see attached)


Also - Doug's circuit uses 1.5K resistors at the grids rather than the 8.2k's in your drawing.  Would I need to change those to 8.2?


I really appreciate the help guys.  Truly.

[kagliostro]

The schematic I posted was posted some time ago by our friend Geezer

you must not use the resistor value shown there, I used it only as an example

As to have the 1/2 power option, you must only shunt two grids to ground

K

[markmalin]

Cool.  Thanks, K.  If da Geezer posted it, it's got to be good :)


Seriously, thanks for the input.  As always, you guys are a huge help.

Mark.

The schematic I posted was posted some time ago by our friend Geezer

you must not use the resistor value shown there, I used it only as an example

As to have the 1/2 power option, you must only shunt two grids to ground

K

[Bieworm]

If you go half power that way, you must connect the speaker to the OT as to double the reflected impedance

If you don't want to swap the OT connection of the speaker, you must obtain 1/2 of the power shorting to ground two of the power tube grids



K

Quick question, maybe stupid.. but just to be sure..
When you shut off the grids from the 2 tubes, will this affect the bias? Or is the bias R-C in this case maintained for the quad of EL84's?

thanks!

[pdf64]

If you go half power that way, you must connect the speaker to the OT as to double the reflected impedance

If you don't want to swap the OT connection of the speaker, you must obtain 1/2 of the power shorting to ground two of the power tube grids



K

That rational seems wrong headed to me.
With only a single push pull pair of EL84 operational, the loadline for that pair moves to the right of the knee, anode dissipation under load will increase, max audio power output will decrease.
That an EL84 is sitting there in parallel with each, drawing DC but unresponsive to signal is immaterial.

[Merlin]

Quick question, maybe stupid.. but just to be sure..
When you shut off the grids from the 2 tubes, will this affect the bias? Or is the bias R-C in this case maintained for the quad of EL84's?

The bias is maintained. The circuit only switches off the AC signal going to the grids, but the DC conditions are unchanged.

With only a single push pull pair of EL84 operational, the loadline for that pair moves to the right of the knee, anode dissipation under load will increase, max audio power output will decrease.

True, but this is generally a safe direction to go in. Anode dissipation will increase only a little (or maybe not at all depending on details), but screen dissipation will decrease a lot, which makes for a happier tube.