Amp Switching?

[tbryanh]

What is the best way to switch between one of two or more amps to the same speaker cabinet?

Example:
1. There are 3 amplifiers and 1 speaker cabinet.

2. One amplifier is played at a time, but the musician switches between amps during the same song.

3. Each amplifier is connected to a dummy load when not in use. This allows the volumes of all the amps to be set to the desired levels even when they are not connected to the cabinet.

4. The amps are protected during the switching process, maybe by a high value resistor connected across the output of each amp.

What part numbers and circuit diagrams are needed; e.g. what are the part numbers for the relays, and how are they connected, etc?

What problems might there be?

[Enzo]

OK, it can be done. There may even be commercial equipment that does it already on the market. If there is, I am not aware of it though.

To just do it is feasible, but to do it right, safely, and reliably offers many problems to solve.

I guess we are assuming the same input signal has already been distributed to all of them? Or is that also to be switched? That has some bearing too.

But...

The switching has to be absolutely seamless. The switching between speaker and load at the output of any amp can't have any space. That brief moment where nothing is connected between relay positions is enough to glitch an output transformer if it happens at a peak point on a signal waveform. I guess maybe you could put a minimal - not minimum impedance, but minimal as in least amount necessary - load permanently across an amp out and switch the speaker in and out of parallel. Even so, some amps might not like abrupt load changes with signal.

The speaker circuits ought to be kept isolated one from another. The potential for ground loops is there, plus any phase differences between amps could cause contention in some cases.

And consider each amp on a case by case basis. Mixing tube and solid state can cause trouble potentially. Not all solid state amps use ground as the negative speaker connection. Some elevate that over a small value resistor to use as NFB sensing. Any number of SS amps run the speaker load in the return of the PT CT. That might or might not get in the way - Haven;t thought enough about that detail. And even tube amps don;t always ground the speaker line anywhere. All this is more in favor of complete isolation of speaker lines.

[tbryanh]

Maybe the dummy load could be switched in before the cabinet is switched out, and the dummy load switched out after the cabinet is switched in.

Example:
1. Amp-1 is connected to the cabinet, Amp-2 is connected to Dummy-Load-2, and Amp-3 is connected to Dummy-Load-3.

2. Dummy-Load-1 is not connected to Amp-1.

2. The musician hits a stomp switch to choose Amp-2.

3. The following sequence of events occurs.

A. Dummy-Load-1 is connected to Amp-1.
B. The cabinet is disconnected from Amp-1.
C. The cabinet is connected to Amp-2.
D. Dummy-Load-2 is disconnected from Amp-2.

[tbryanh]

“I guess we are assuming the same input signal has already been distributed to all of them?”

Yes. At least for now.

“The switching has to be absolutely seamless.”

What about relay contact bounce?

“plus any phase differences between amps could cause contention in some cases.”

Not sure what is meant here.

“Mixing tube and solid state can cause trouble potentially.”

It would be just tube amps only. At least for now.

“The speaker circuits ought to be kept isolated one from another. The potential for ground loops is there . . . All this is more in favor of complete isolation of speaker lines.“

I have heard of isolation transformers for speaker output jacks. Maybe they are needed here.

[Enzo]

A. Dummy-Load-1 is connected to Amp-1.
B. The cabinet is disconnected from Amp-1.
C. The cabinet is connected to Amp-2.
D. Dummy-Load-2 is disconnected from Amp-2.
__________________

That is an involved bit of sequencing to achieve with relays. And I worry about pulling it off without causing an undesirable pause in sound as it is done. Search for discussion of several Peavey amps where owners complain about a "delay" when channel switching - a momentary mute of the audio.

What about relay contact bounce?

I wasn;t even thinking of that. I think bounce is probably faster than any audio problems. What I am concerned with here is the timing. Though fast, relays are going to be "break before make" switches. There will be that gap in circuit continuity while the contacts transfer. You might find or could special order I suppose some make before break contact form relays, but then you'd have to be EXTRA cautious that you never wind up with two amps coupled into the speaker at the same time, even for a millisecond.

Phase difference? What if on one amp the output is out of phase with the other, and at a given instant, the output is a positive waveform peak on one? That means the output on the other amp is a negative peak. I may be overthinking this, but that means as you shirft loads, your switch from one polarity peak to the opposite. I am concerned the inductance of the load might not like it and spit a current pulse back at the amps.

I am not aware of speaker isolation transformers.

[Robert M. Martinelli]

As already noted the main problem is that the switching has to be seamless. This, together with the timing problem, makes for a difficult solution if faced with a "conventional" ( hardware ) basis.

But.....

Everything can be solved pretty easily using a "micro PLC". These devices can be freely programmed, have enough "intelligence" and program memory to handle small logics, can be expanded to deal with a bigger number of signals if needed, have scan times of a few milliseconds, and are pretty cheap. By properly programming one of there units ( which are very easy to program, using a ladder-based language ) all the issues of amp/signal seamless switching can be dealt with and eliminated.

HTH

Best regards

Bob

[tbryanh]

PLC looks good.

[octal]

I have an idea. If the amps don't use NFB, you could float both leads of the output transformers from the chassis. Wire all three amps in series, then have the relays SHORT the secondaries of output transformers of the amps not in use. Worst case, none of the amps will be looking at a complete open circuit load and contact bounce will be less of an issue.

This MIGHT cause ground loops or oscillation, but it might also work fine.

Regardless of how you do your switching, you could also put clamping diodes on the primaries of the OPTs to clamp any voltage spikes during switching. Check out the schematics of some newer Fenders to see how to wire them.

Nathan

[tbryanh]

"If the amps don't use NFB, you could float both leads of the output transformers from the chassis. Wire all three amps in series, then have the relays SHORT the secondaries of output transformers of the amps not in use."

This is a good idea. NFB amps would probably be needed, so isolation transformers might be needed.

"Regardless of how you do your switching, you could also put clamping diodes on the primaries of the OPTs to clamp any voltage spikes during switching."

This sounds like a good idea.

[Steve Conner]

How about using make-before-break switching, so that Amp A and Amp B are both connected in parallel to the speaker for a few milliseconds. To my mind, that's less dangerous than an open circuit.

Or you could have a separate switched dummy load for each amp, and a common speaker changeover switch, and operate them so they cover each other:

Start position:
A to speaker, B to dummy load

Dummy load switched onto A:
A to (dummy load in parallel with speaker), B to dummy load

Speaker changed over from A to B:
A to dummy load, B to dummy load || speaker

Dummy load removed from B:
A to dummy load, B to speaker

To avoid ground loops, you'd use double pole relays and switch the grounds too. You can generalize this to any number of amps.

[Enzo]

Yay. And the switching system winds up more complicated than the amps it serves.

[tbryanh]

How much time might the relay contacts need to stop bouncing?

Motor contactor relays might be used. Do their contacts bounce?

[Enzo]

ALL switch contacts bounce. That happens on a very small time scale. Bounce does not concern me. I am concerned by the time between once side opening and the other side closing.

Steve suggested make before break switching, which would certainly eliminate my gap concern, but then the two amp outputs would be connected together during the make before break period. I don;t like that either.

[Gregg]

The switching has to be absolutely seamless.

It's easy to do with the help of an MCU.
The sequence is: Press amp1/2 button - amp1 input to ground - amp1 output to dummy load - amp2 output to speaker - amp2 input off ground.
Use industrial grade relay for speaker/dummy load switching and FET, LDR or relay for grounding the input.

Commercially available units are available as well:

http://www.tonebone.com/

[olddawg]

People have been fussing about this forever, and on just about any amp board you ever happen to look on. For the life of me I don't understand why anyone would do it in the firstplace. Every person I know of that has used an amp output switching scheme has damaged their amps. IMHO it is foolishness. The easiest off the shelf way to do essentially the same thing would be to use 2 stereo cabs for individual loads and A/B/Y the amp inputs. That would even leave you room for even a fourth amp if you wanted. Or... you could make a 3 load cab with 3 or 6 speakers if you wanted to just use three amps and all of the individual loads.