Switching speakers on the same amp in real time

[T Wilcox]

I use a KHE 8x4 switcher for this
https://www.khe-audioelectronics.com/shop/acs-8x4
8 amp inputs and 4 cab outputs switched via MIdi foot controller
Works great
I am also able to run the cabs in pairs as well with it
Kahayan makes a similar switcher as well

[pjd3]

Whoa Jimminee crickats.

It has to be some serious tone monster that would rig up with one of those. VEry impressive what that system does.

I'll be an instant hero if I can just figure out a way to switch between 2 speakers when channels are switched. Its coming though. One fine day.

Thanks,

PJD3

[pjd3]

For starters The A/B make before break relay sounds worth a go, with each of the 2 speakers on a separate impedance matched tap-to-speaker. Like RG mentioned, there would never have to be a time when there wasn't some load or another grounded and/or connected to the OT. So even if the relay or switching went haywire there would be some saving load on the amp.

Its a simple straightforward idea and if it would work, great. I would then just be looking at various parts of the output section during switching to see if there were any transient spiking things going on.

Thanks guys,
PJD3

[R.G.]

Upon reading more comments...

The essence of not damaging a >>tube<< amp when switching loads is to never let the load go open. Solid state amps really need the opposite, never going shorted to ground, but that's another story.

I've been doing mostly microcontroller stuff for a decade or two now, and these things make it easy to do things that are difficult or expensive with just wires, switches, resistors and caps. A sneaky way to mess with switching issues is to use a microcontroller to sense when the "Scotty - switch me now!" signal and have it then turn on a MOSFET solid state relay for a fraction of a second. The MOSFET can switch on and off in a few ...microseconds... meaning it is fast enough to make a slow, bouncing mechanical switch look like a fancy make-before-break switch. It takes longer (another few microseconds) for the controller to sense the switch signal than the actual switching.

So ensuring that a no load situation never arises, a usable scenario is to sense the action of the mechanical switch changing, turn on a MOSFET output relay to ground, perhaps through a resistor, and let go of it when the mechanical switching is done, 50 milliseconds or so later. The controller would effectively be waiting around for the mechanical switch to quit bouncing. Or you could just load the output for a fixed time, maybe 100mS or so.

Hmmm. As long as you have the controller in there, sensing switches and controlling things, the next sneaky idea is to have the microcontroller sense the actual person-available switch and have it actually control the speaker loading relay, and also turn on the MOSFET speaker mute while the speaker relay is moving slowly to a new position. I have used exactly this in a commercial production product I helped design. It works, and is very, very quiet in its operation, as well as being so fast that you can't detect the short muting in the middle of the switch movement.

Next obvious step: Put in multiple relays and suitable input/output jacks and mode switches, and have the controller do all the stompswitch sensing, turning relays on and off, MOSFET muting, MIDI receive/send, blinking lights, displays, sirens, and pyrotechnics. All this sounds complicated, and it once was. But a $3 Arduino clone is capable enough, although I'd probably use a $0.75 PIC. The biggest challenges in an any-to-any head/speaker setup are, in order, (1) the exterior box appearance (2)user interface (3) financing, advertising, manufacturing, distribution, etc. ...
The design of the actual stuff inside is pretty short and direct.

[pjd3]

Thanks for that RG, with all the dinking around I've done with fun electronics plus some serious stuff in work (defrib and ventillators) I have never worked one on one with microcontrollers or PICs although I've been very interested to see what they are about and felt I've owned it to myself to get to know it. I see this would be an opportunity.

Earlier today however during lunch, I sat down and "designed" a logic circuit that would provide a MBB condition with 2 SPST relays, a 2 input NAND gate (four gates) and a Dual 555 timer. And 2 switching transistors for the relay coil source. I suspect It would work but, it wouldn't provide the Mosfet silencing like your tactic would, if, my dreamed up approach wasn't approaching optimum.

This is all great stuff to keep in mind and, I thank you all for such generous input. This thread has left me with alot more things to think about than I had originally anticipated.

Best,
PJD3

[R.G.]

Overlapping two different relays works, too. Relays bounce when they change state, just like switch contacts. Be prepared to make them overlap by maybe up to 200mS so that the "make" switch has quit bouncing before the "break" switch opens.

G'donya.

[pjd3]

Thats just what my circuit was meant to accomplish, there are 2 discreet SPST relays each ultimately switched by its own NAND gate where one of the 2 inputs are coming from its own 555 timer. I figured you start with a timing that maybe you can hear, and continue to decrease the time until its seamless, if it will let you, and you can also be looking at a couple of things on the scope incase something arises.

The other 2 NAND gates would have there inputs ties so that they become inverters, put them in series and tap off each one to create the "toggle" of sorts.

Thank you and Best, '
PJD3

[tony hunt]

Palmer made a 4 way speaker switcher for tube guitar amps. I have one, but forget the name.
I used it for trying out different speakers for myself on quite large amps running quads of old KT77s.
The basic principle was the output was muted for maybe a second between relays switching the speakers.
The unit is probably still around somewhere if you want me to take a closer look.

[pjd3]

Thank you for that offer Tony.

That sounds like it may be in line with RG's mention of having a big Mosfet across the secondary and go into a low impedance conduction during the switching of the relays. I will definitely call on you if I'm not having any luck with my little crazy idea.

Digikey just sent me a couple of SPST relays, a Hex inverter with open collector outputs, a NAND gate with Schmitt trigger inputs and a couple of 2N2222 transistors to drive the relay coils. I drew up a little circuit that will time the relays for a Make-Before-Break with the 2 relays and will see how this works. I'm mulling over where and how I would look for indications of any flyback or spike voltages during switch as I surely don't want to damage any part of my amp. I suppose I would start with a very low input voltage to see if there is any calamity. And of course examine the behavior of the circuit and relays with the scope before anything.

But, looking forward to see how to get this done.

Thank you very much for stopping by!
Best,

PJD3

[pjd3]

The circuit I concocted for timing up 2 SPST relays into an RC adjustable Make-before-Break action seemed to have worked very well. You can see using 2 channels on the scope that as the 2 relays are switched back and forth there is a window of time both ways by which they are both on simultaniously, as relegated by the time constant set with a cap/resistor pair.

I had the cap/resistor pair sand-witched between the output of an open collector inverter and a Schmitt trigger NAND gate and they did just what I asked them to do!

Also, riding on one of RG's ideas I ordered an Exclusive NOR gate and 100v/20A Mosfet that is to drive that power MOSFET to bring the whole OT secondary to a lowish impedance/resistance during the course of the Make-Before-Break relay action. Of course, I'm asking myself if that is redundant and my answer is, yes, it sort of is, but. So what. If it works, each system could be considered a back up for each other. The Make or Break relay action will keep a reasonably constant load on the OT but, there still will be a point where an inductive speaker will be dis-connected from the OT while energized. I have no idea at this point what the real life implication of that will be/are but, perhaps a MOSFET going low during all of that would give any spiking energy a place to go for a very brief window. I found a cool little simple circuit with the Exclusive OR gate that generates a pulse whether the input is going high or low. And like the relay circuit, you use a little RC pair to set the duration of the pulse that will put the MOSFET into conduction momentarily.

So, looking forward to the next step. Ho-pe this is a solution

Thanks for stopping by,
Phil Dl

[tony hunt]

Thank you for that offer Tony.

That sounds like it may be in line with RG's mention of having a big Mosfet across the secondary and go into a low impedance conduction during the switching of the relays. I will definitely call on you if I'm not having any luck with my little crazy idea.

Digikey just sent me a couple of SPST relays, a Hex inverter with open collector outputs, a NAND gate with Schmitt trigger inputs and a couple of 2N2222 transistors to drive the relay coils. I drew up a little circuit that will time the relays for a Make-Before-Break with the 2 relays and will see how this works. I'm mulling over where and how I would look for indications of any flyback or spike voltages during switch as I surely don't want to damage any part of my amp. I suppose I would start with a very low input voltage to see if there is any calamity. And of course examine the behavior of the circuit and relays with the scope before anything.

But, looking forward to see how to get this done.

Thank you very much for stopping by!
Best,

PJD3

My apologies for the delay, I hope this reply includes a PDF of the guts of the palmer unit.

2023-02-26 10.36.34.jpg

Cheers, Tony

[R.G.]

My apologies for the delay, I hope this reply includes a PDF of the guts of the palmer unit.
2023-02-26 10.36.34.jpg
Cheers, Tony

Very informative. Four relays, one for each output probably. Microcontroller (ATTiny) to manage footswitch bounce, lights, etc, and time the relays. No obvious shorting device, so the code on the microcontroller probably turns on the new one before turning off the old one, using the overlap method to never leave the amp output open. Very straightforward, probably works just fine.

[pjd3]

I can really feel my "greenness" on all things microcontroller on little systems like this.

I did have some great results with producing a stealthy MBB with logic gates and relays but it would be satisfying to make that happen with microcontrollers. Maybe its time to grab and Arduino and get down to it.

Just received a power Mosfet and XOR gate that will produce a positive pulse to switch the Mosfet on and off. That's next. I figure having both wouldn't hurt anything and it provides a backup situation incase one system breaks.

What I'd like to get under my belt now is how and where to measure a flyback or inductor induced spike on a tube amp. I suppose the EE's at work could help quite a bit with that but, they are often busy and I feel like I'm bothering them. But yeah, I may have to create a thread sometime focusing on that topic and see if we could gather some ideas or approaches on how to look at that say on a scope, and based on what you see, how to go about sizing up components to address the issue. But for now, its see what a Mosfet switch plus MBB can do for speaker switching.

thanks again all,
PJD3

[CraigGa]

I can really feel my "greenness" on all things microcontroller on little systems like this.

I was struggling to build myself a multifunction timer using ebay modules and somebody suggested using a Pi Pico.
I'd never heard of them or used a microcontroller but I had it coded and running in a couple of days.

You could get a Pico and a 4 channel relay board for £8 $10, it was a great learning experience too.

Craig

[pjd3]

That sounds like a very worthy investment Craig and one I ought to delve into.

Thanks for the info and the inspiration. Its time.

Just to mention, my particular intention was to eventually build another amp that has both a British and American preamp that can be foot-switched, and the idea was to have a British and American speaker switch with the channels to compliment the tone. I thought that might be a good point of flexibility when playing in bands that do a rock selection spanning 3-4 decades.

Best,

Phil