Half Power switch - best way?

[Enzo]

Nope. They are not conducting current, so they amount to pieces of wire connected to an empty glass bottle.

R = V/I and current (I) would be zero. So R = V/0, which is infinite. SO put infinite impedance in parallel with whatever the other tube impedance might be, and they calculate out to whatever that other tube is. A not conducting tube might as well be the air.

[Dave H]

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.

It's the primary impedance which stays the same as it is set by the turns ratio and load impedance which are unchanged. If the primary impedance was 2k (say) for four tubes then it is still 2k but now it is only being driven by two tubes. The other two tubes are only conducting idle current (DC) not signal current (AC) so look like a high impedance to the signal. You have two tubes driving 2k which is a (theoretical) mismatch, which doesn't really matter but this is the 'Theory' forum...

[GainFreak]

OK, I get it now.

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.

The Sunn T reissue has similar solution.

Another thing I can think of is a "fixed power scaling" but it will be somewhat more complicated. Lowering the screens voltage by 100V and bias accordingly (assuming 470V on plates) should get you 50W of power.

[Chuck H]

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.
[ATTACH=CONFIG]39371[/ATTACH]

Glad YOU got the likes for this one. I suggested totem tapping for voltage reduction once and took a beating for it on some technical grounds I couldn't understand at the time. I've resurrected the idea a couple of times with more amicable responses though. Since the premise is voltage reduction I suppose totem tapping isn't dissimilar from power scaling. I think the 100W/25W switch would definitely have more of an instant gratification effect. IMHE the volume difference between 100W/50W (depending on the amp) can range from:

"Oh. That doesn't really do much, does it.?."

to

"I believe something changed because I know what the switch does, but I don't actually hear it."

A 100/50/25 switch would probably be the best case for adjusting to different venues.

[pdf64]

I like plenty of interaction between amp and speaker, so prefer little / no negative feedback on the power amp, speakers wired in series; a load resistor (of similar resistance to the nominal impedance of the speaker load) in parallel to the speakers would almost certainly interfere with that interaction, so I don't fancy it. It will likely reduce bass and treble, lessen the max available presence boost.

My feeling is that the facility of switching an amp so as to ~halve its rated power output is, in general, pretty pointless. This is due to the key benefit from the lower power mode being that the amp overdrives at a lower, more useful level. However, the overdrive response of lower power modes are generally perceived as being less musically pleasing than the full power mode.
This may be because overdriving a classic guitar amp involves not only the output devices 'hitting the rails' but also the power supplies sagging as the demand increases, and perhaps significantly the screen grid demand % increase may be greater than the % increase in plate demand.
Cutting power output by deactivating output devices etc tends to reduce power supply sag / increase operating point voltages under all conditions, which may change the amp's tone.
So the end result may be a lower power mode that has a harsher tone and stiffer response than the full power mode, which is the opposite of what I want.

As a simple way forward to reduce power a little, whilst increasing compression and softening the tone, I suggest just switching in a HT sag resistor, eg 20 ohms 25 watts.

I've got a Fender 75, which has the same 1/4 power switching arrangement of that Twin.
For the above reason (super stiff HT resulting in minimal sag / compression under overdrive) I never found the 1/4 power option to be much use.

[Hardtailed]

I've often wished my amp had less power, I've never wished it had more. So I'd rather just not build an amp that requires a half-power switch.
Why have a quad of 6L6s when you can have a pair of 6V6s, heck 6K6GTs even (I've found them to be the perfect spot volume wise, just wish they'd start making them again so we never run out).

[R.G.]

The power amp I liked best for tone qualities was a single-ended 6AQ5 circuit salvaged from a Magnatone home hifi "furniture" setup from the early 60s. About 2-5W. It went into the PA, of course.

[Mike Sulzer]

I like plenty of interaction between amp and speaker, so prefer little / no negative feedback on the power amp, speakers wired in series; a load resistor (of similar resistance to the nominal impedance of the speaker load) in parallel to the speakers would almost certainly interfere with that interaction, so I don't fancy it. It will likely reduce bass and treble, lessen the max available presence boost.

I do not think there is a difference between identical speakers in series or parallel (with the correct tap on the Xformer in each case of course). If the impedance of a speaker is Z(f), nominally 8 ohms, then it is 2*Z(f) in series or .5*Z(f) in parallel. If the source impedance of the amp is Zs(f) on the 8 ohm tap, then it is 2*Zs(f) on the 16 ohm tap, and .5*Zs(f) on the four ohm tap. A voltage applied to the primary of the Xformer results in the same current through the speakers in both cases.

[J M Fahey]

No absolute need to use 6V6 if they are becoming hard to find, you may very well build a 15W amplifier with a couple overabbundant 6L6, as lon g as you design for it:
* 15W appropriate PT, so it sags like any real world transformer when reaching max power ... while a subloaded 100W PT will be real stiff, as noticed above.
* proper 15W OT, for the same reasons.
Who cares it's labelled "for 6V6/EL84/whatever" , that's just to help the undereducated warehouse attendant to pick and pack the proper one for delivery, of course "15W 6V6" loads will provide close to that even with 6L6 , if in doubt just draw load lines on a 6L6 graph and check.
You *might* get 20W because the tube is sort of idling ... still in the ballpark.
* you might add cathode bias instead of a fixed one, it will certainly let it compress m ore.
* as a bonus: any 6L6, even cheap ones, will last forever under such light load.
* or you can bias them to *real* class A if that rocks your boat ... now you have leftover dissipation available
I mean, bias them to **6V6** 100% dissipation, which is what that circuit might demand .

6L6 will whisper in your ear: "hey Daddy ... is that all you have for me???? "

[Dave H]

I do not think there is a difference between identical speakers in series or parallel

Do you think there would be a difference with zero source impedance? In parallel both speakers see zero source impedance but in series each speaker sees the other speaker as its source impedance.

[Chuck H]

Do you think there would be a difference with zero source impedance? In parallel both speakers see zero source impedance but in series each speaker sees the other speaker as its source impedance.

And series vs. parallel inductance has also been discussed here and there in books and forums as it relates to bass response. I've never even ear tested that one for myself. I don't imagine it would be significant but some reports suggest it is.

[mozz]

No one has suggested a fixed ppimv switch. Would be just 2 resistors right?

[Enzo]

Just a nit pick on Dave H. The two off tubes would not be conducting the DC idle current. With no screen voltage or an open cathode lead, they don;t conduct at all.

[Dave H]

Just a nit pick on Dave H. The two off tubes would not be conducting the DC idle current. With no screen voltage or an open cathode lead, they don;t conduct at all.

Pick away Enzo

I was replying to GainFreak's post #5 where he is suggesting switching the inputs of one pair of tubes from the PI to the raw bias supply end of the grid leak resistors so they would still be conducting bias current but not signal current.

[Mike Sulzer]

Do you think there would be a difference with zero source impedance? In parallel both speakers see zero source impedance but in series each speaker sees the other speaker as its source impedance.

Ohm's law tells us that if you have zero source impedance and apply a voltage V to one speaker you get current i(f) if the impedance is Z(f). If you put two speakers in parallel each has i(f) through it. If you put two speakers in series and drive them with 2*V, each has current i(f) through it and voltage V across it. Speakers must be identical.

You have to be careful thinking about the damping of a speaker by a low source impedance. The audio industry has traditionally over sold the size of the effect.