half power switch and impedence matching question

[kwijabo]

all - i saw this type of half power switch in a gut shot on the interwebs and am thinking about putting in my next build. couple questions:

1. does the output impedance of the power-amp change? so if you did this switch and moved to half-power, would you have to change the output impedance selector of the transformer like you do when you physically remove two output tubes?
2. when you do this, do the no-input tubes still draw current? my guess is yes they still draw, because when you turn on an amp with no input your pre-amp tubes still draw and cause a drop across the dropping resistors. so this switch's benefit is that if you wanted to move to two output tubes you wouldn't have to change any dropping resistors.
3. which output pair should share the cathode bypass - as drawn (V5/V6 and V4/V7)? or should V4/V5 and V6/V7 be paired? i couldn't see it clearly enough on the picture.

any other comments/suggestions?

[martin manning]

all - i saw this type of half power switch in a gut shot on the interwebs and am thinking about putting in my next build. couple questions:
1. does the output impedance of the power-amp change? so if you did this switch and moved to half-power, would you have to change the output impedance selector of the transformer like you do when you physically remove two output tubes?

Yes, if you want to keep the load on the pair still in play the same.

2. when you do this, do the no-input tubes still draw current? my guess is yes they still draw, because when you turn on an amp with no input your pre-amp tubes still draw and cause a drop across the dropping resistors. so this switch's benefit is that if you wanted to move to two output tubes you wouldn't have to change any dropping resistors.

Yes, and that is a good point, pulling two tubes or switching two of them off would raise the idle B+ a bit, and either way it won't sag as much with only two tubes working.

3. which output pair should share the cathode bypass - as drawn (V5/V6 and V4/V7)? or should V4/V5 and V6/V7 be paired? i couldn't see it clearly enough on the picture.

In the schematic they are separated so that the active pair has the same cathode resistor and bypass cap either way.

FWIW, I think most people who try a half-power switch don't find it all that useful. It's only -3dB.

[Reeltarded]

Totally agree..

Half power is a waste of time and money, unless it's something about the tone you really must have!

I need to make a saddle blanket attenuator video. Stylish and effective.

I rivet pairs of surplus wool Army blankets together, and throw a saddle blanket over those. I can play a 250 seater with 3x Super Leads and no deaf audience. Feels good too! Thump!

[epis]

I wouldn't agree with Martin's answer to first question.There is no need to change output impedance.V4 and V7 are still present in circuit, just idling.

[Alexo]

I wouldn't agree with Martin's answer to first question.There is no need to change output impedance.V4 and V7 are still present in circuit, just idling.

Load is moot when they are idling. It's the load they are driving while active that counts, and you really should double it in this case (which is a pain in the tush).

I've done it both ways - actually more effective when you do not change the load because not only do you go to half the output tubes, you introduce an impedance mismatch that also knocks a few db off and you get something more like 1/4 power, but with all the issues that go along with an impedance mismatch (not actually that big a deal if the mismatch errs on the low side like this, but the tone will change).

If you do it properly and switch the output Z, I have to agree that it's a waste of time - difference is so slight you almost think the switch isn't even hooked up.

BTW, the 220K grid-to-ground resistors are superfluous when the tubes are out of circuit, their grids can just go straight to ground (or bias).

[martin manning]

I wouldn't agree with Martin's answer to first question.There is no need to change output impedance.V4 and V7 are still present in circuit, just idling.

Idle current is DC... Only the active pair of tubes generates a reactive load, so they will see half of the load impedance as with all four working.

[selloutrr]

Not to high jack the thread.. how would you go about correcting the impedance while running only two tubes so that the transformer would act the same as it would if all 4 tubes were operating in the circuit? If it's an impedance issue I would assume you could make up the difference with resistance? (transformer, choke, resistor) that could be switched on (into the circuit) when the extra pair of tubes are removed... ?

by measuring the amount of resistance when all tubes are in circuit and when only two are in circuit. Then make up the difference.

I could be wrong... I've been drinking bottled water, It's a rather sobering experience, but not so great for the imagination

[Zippy]

Looks like a multi-tap OT - change the secondary load to restore the impedance balance.

[martin manning]

Not to high jack the thread.. how would you go about correcting the impedance while running only two tubes so that the transformer would act the same as it would if all 4 tubes were operating in the circuit? If it's an impedance issue I would assume you could make up the difference with resistance? (transformer, choke, resistor) that could be switched on (into the circuit) when the extra pair of tubes are removed... ?

Just run a speaker impedance that is twice the nominal value, or move the impedance selector to half of the speaker load you're running.

You could (and maybe this is what you meant) switch in a series resistor equal to the speaker load when two tubes are turned off. I don't know what that would sound like, but it would be a 1/4 power switch!

[kwijabo]

thanks for the replies. so yes, the idea is when you pull two tubes you use the impedance selector to change it, which i'll have in anyway.

to clear this up for myself, the way i read it is, for example, with 4 tubes running an 8ohm speaker and the impedance selector is set to 8ohms. i remove two tubes from the circuit. my options are 1.) run it with a 16ohm speaker, or 2.) use 4 ohms on the impedance selector.

so really the only benefit of the switch is only to keep the B+ voltages the same without having to physically remove tubes from sockets. i'm at the point in this build where i already have 4xEL84 sockets drilled and heaters connected. after hearing about the value (or lack thereof) of a switch, i think what i'll do is what this site (and IMO good engineering practices) typically recommends - voice it correctly during the build then hardwire in how you like it.

again i really appreciate the help. this is one area thats always been cloudy for me.

[kwijabo]

ok after looking at carvin's schematics - they use the same transfromer for the Vintage 50 (4xEL84's) and Vintage 16 (2xEL84's) and on the schematic for the Vintage16 it says "use 4ohm tap for 8ohm speaker. 4 tube output transformer with 2 tubes." so what martin and zippy said is correct about how to match it.

secondly, the statement that initially confused me was from an interview in Tonequest about the amp that i saw this switch in and it says, "Yes, and
I played the game with the output transformer and split the difference between the 8K and the 4K ohms so it works very well with 4K and with 2K. That’s the problem when you go to half power. Like in a Twin, if you pull two tubes you should actually increase the speaker impedance load by half."

i'm sure it was a misspeak and they know what they're talking about - but i like, how do increase something by half?? you can increase something 50%, or 1 and 1/2 times, but you can't increase something by half, so i wasn't which way to change the impedance.

also, it looks like what they did is pick an OT that has primary impedance somewhere between the 4k and 8k. i have a 4k primary OT so the switch isn't necessarily applicable for me without other adjustments.

[Firestorm]

If you do what martin et al. describe (disconnecting the drive signal to a pair of tubes and doubling the load), you wind up with an amp that sounds exactly the same but at -3dB, which you likely cannot hear. It might be more fun/useful/toneful to omit the switch and replace it with a dual-gang 250K pot so you can dial one pair of tubes up or down and run into the stock impedance. As long as the variable tubes aren't dialed out completely, they will contribute something to the combined transconductance of each fixed/variable pair so you should hear a change in tone, rather than a change only in output power. If the drive signal is big enough to push the tubes into cutoff, you have the ability to dial the pot to the point where two of the tubes stay in the Class A region even when the others have transitioned to "Class B" operation.

Or it might be fun to wire a triode/pentode switch to one pair to have that sonic capability.