Hi everyone,
I would like to ask for your tips and hints regarding the design of amp selector switch which is compatible with tube amps (but not limited to).
what I mean is a switch selecting between two or more amps into a single pair of speakers. Sharing my thoughts so far:
I would appreciate sharing any constructive feedback.
Aviv.
I would like to ask for your tips and hints regarding the design of amp selector switch which is compatible with tube amps (but not limited to).
what I mean is a switch selecting between two or more amps into a single pair of speakers. Sharing my thoughts so far:
- Assuming intended only for isolated amplifiers, is it necessary to switch both the positive and the negative outputs? i.e. is there a resaon why I should not tie all negative speakers output terminals of the two (or more) amps together permanently (still maintaing separation between left and right negatives terminals)? this is obviously a curcial question, as if both +/- outputs have to be switched, for every stereo amp I need to switch four poles.
- For a generic tube amp, a common advice is that the amp output should not see an open load. Following this advice, I see two options- A) have a pilot load (100 Ohm resistor?) connected permanently to each amp output (input of the switching box). B) Use a shorting type selector (Make Before Break) with a neutral position between every two amps- an idea which was described here by Chris. I have to admit that I like the idea, but again supporting 2/3 amps, left+right, +/- switching, I have to find a selector switch with 8/12 poles and 3/5 positions. shorting type, which is quite hard to get (or very expansive).
- Switch selector vs. relays- While shorting type selectors are not very common, with a uC controlled matrix of relays I can achieve any switching logic I want. I can have the uC switch the dummy load in before breaking the connection, I can have the dummy load permanently connected, or not connected at all. in the worst case, for every amp, two double pole ralys are required plus one additional dual pole relay for a full control over the dummy load. Quite std. relays, not hard to get. simple uC programing and custom PCB are not a big deal. The downsides- I have to design the PCB and the switching box requires power for the control gear.
I would appreciate sharing any constructive feedback.
Aviv.
Don't operate a tube amp without a load on the output transformer is about the only advice I could give. Long cable/IR activation to switch plus a random function may be useful to switch - recording the time means you're blind to which is being played.
Also something to help set the playback level would make it faster.
Also something to help set the playback level would make it faster.
For two amplifiers, use a Very Well Rated DPDT switch (Double Pole, Double Throw).
It is robust, can handle several Amperes of current, and 250VAC, etc.
For more than two amplifiers, probably a rotary switch (but rotary switches usually do not work with the kind of current in many hi fi / stereo systems.
A true 8 Ohm speaker, and true 8 Watts rms takes 1 Amp rms (1.414 Amps peak).
What about a 60 Watt amplifier? Into 8 Ohms = 3.87 Amps peak . . . what rotary switch?
Stay away from Make Before Break Switches.
Connecting all grounds together is likely to cause ground loops. Hum.
In general, isolated amplifiers are Not Safe, Especially Vacuum Tube Amplifiers.
Safety First!
Prevent the "Surviving Spouse Syndrome".
If any amplifiers are powered up without a load (including some solid state amplifiers), Your Mileage May Vary.
Relays . . . Perhaps
How can you always keep track of all of these things, to prevent a mishap?
Do not switch when the amplifiers are powered on.
Or, Your Mileage May Vary.
If you are trying to do A / B testing, with two amplifiers, it will be very complex.
The signal Inputs and the signal outputs need to be switched simultaneously.
Lots of switch contacts.
Depending on the amplifiers in question, Your Mileage May Vary.
At VSAC 2008 and an A / B demonstration, I used identical speakers, and only the signals to the amplifiers inputs were switched; the speakers were always connected; simultaneous synchronization was not necessary.
It is robust, can handle several Amperes of current, and 250VAC, etc.
For more than two amplifiers, probably a rotary switch (but rotary switches usually do not work with the kind of current in many hi fi / stereo systems.
A true 8 Ohm speaker, and true 8 Watts rms takes 1 Amp rms (1.414 Amps peak).
What about a 60 Watt amplifier? Into 8 Ohms = 3.87 Amps peak . . . what rotary switch?
Stay away from Make Before Break Switches.
Connecting all grounds together is likely to cause ground loops. Hum.
In general, isolated amplifiers are Not Safe, Especially Vacuum Tube Amplifiers.
Safety First!
Prevent the "Surviving Spouse Syndrome".
If any amplifiers are powered up without a load (including some solid state amplifiers), Your Mileage May Vary.
Relays . . . Perhaps
How can you always keep track of all of these things, to prevent a mishap?
Do not switch when the amplifiers are powered on.
Or, Your Mileage May Vary.
If you are trying to do A / B testing, with two amplifiers, it will be very complex.
The signal Inputs and the signal outputs need to be switched simultaneously.
Lots of switch contacts.
Depending on the amplifiers in question, Your Mileage May Vary.
At VSAC 2008 and an A / B demonstration, I used identical speakers, and only the signals to the amplifiers inputs were switched; the speakers were always connected; simultaneous synchronization was not necessary.
Last edited:
Attached is a schematic of amplifier selector with dummy load for safely switching tube or solid state amplifiers. The selector is a Grayhill 44A45-02-03S three position shorting style rotary switch, one per channel. Center position is for "speaker off" to prevent both amplifier outputs shorted together.
The reason of using shorting style (make-before-break) rotary switch is to ensure both amplifiers will be connected to a load at all times, either to the speaker or a dummy load.
44A45-02-2-03S Grayhill | Mouser
Relay with make-before-break contacts is hard to find, I utilize a make-before-break rotary switch to control relays for speaker and dummy load.
The reason of using shorting style (make-before-break) rotary switch is to ensure both amplifiers will be connected to a load at all times, either to the speaker or a dummy load.
44A45-02-2-03S Grayhill | Mouser
Relay with make-before-break contacts is hard to find, I utilize a make-before-break rotary switch to control relays for speaker and dummy load.
Relay controlled A-B-C amplifier selector:
It utilize a SP5T make-before-break rotary switch to control relays for speaker and dummy load connections. In order to ensure load (speaker or resistor) is always connected to the amplifiers, the circuit need an extra switching position in between A-B and B-C.
It utilize a SP5T make-before-break rotary switch to control relays for speaker and dummy load connections. In order to ensure load (speaker or resistor) is always connected to the amplifiers, the circuit need an extra switching position in between A-B and B-C.
This is'nt good enough !
The output should be connected to the resistor before the speaker is disconnected , a moment
with both speaker and resistor as load does not harm a tube amp, but a moment with no load could.
The trick is the make-before-break rotary switch, when the switch move from "A" position to next position, the spst relay will engage the dummy load resistor to "amp 1" before the dpdt relay break the connection of the speaker.
When the switch move from that position to "B" position, the dpdt relay will engage to the speaker before the dummy load resistor disconnect from "amp 2".
The Grayhill 44A45-02-03S three position shorting style rotary switch current rating is 5A.
Grayhill 44A45-02-2-03S
Semi-con output amplifiers will really, really not like being connected together. They may or may not survive it, but it's a gamble. The issue is that each will try mightily to impose its own idea of the common output voltage at any particular moment, supplying as much current as each can to accomplish its duty. One will lose the battle. So, make-before-break has an inherent limitation, if intended to be used generally.
Valve output amplifiers, if indifferently designed, can burst oscillate if unloaded. The oscillation rapidly drives output anode voltages to dangerously high levels, risking the transformers' insulators' ratings. So, for the most general use, all tested amplifiers must be terminated with at least a Zobel to maintain stability. Something like 10R + 0u1F should do.
I'd strongly recommend switching all four poles of 2-channel amplifiers, if only to remove ground loop issue confounders.
All good fortune,
Chris
Valve output amplifiers, if indifferently designed, can burst oscillate if unloaded. The oscillation rapidly drives output anode voltages to dangerously high levels, risking the transformers' insulators' ratings. So, for the most general use, all tested amplifiers must be terminated with at least a Zobel to maintain stability. Something like 10R + 0u1F should do.
I'd strongly recommend switching all four poles of 2-channel amplifiers, if only to remove ground loop issue confounders.
All good fortune,
Chris
Hi so many choices! I would be tempted to use the microcontroller solution as one could remote control the switching from a phone app, IR remote control or LAN. I would also switch both conductors as others have suggested, a fail safe "load" relay could be fitted using the normally-closed contacts to connect a separate dummy load to each amplifier in the event of a power failure to the micro/relay PCB. One could also incorporate a RF oscillation detector, power meter and clipping detector into the same box, after all the micro is already there. BTW relay switching could accommodate high power amplifiers.
Have fun.
Ken K
Have fun.
Ken K
I tried using a heavy duty SW for switching speakers, it lasted about a week, I now use banana jacks and "U" patching links. To make SWing between tranny and homemade valve amps with different speaker connectors I use little patch cables made with hefty wire, bare wire soldered only at the very tip to stop fraying to a 4mm banana socket. For switching source's I terminated all my sources like PC, phone,TT etc into 1/4" stereo jack male plug, I had some 1/4" stereo sockets to 2 x RCA from a closed hifi shop sale. This is a little fiddly reaching behind the gear to swop over but the big plugs etc help, bish bash, bosh, job done. No micro controller's,switches or anything complicated to go wrong. Remember what the wise man said KISS. If you wanted ease of access build a patch panel, it's how the pro's do it in the recording studio.
If your in the UK I have some 1/4" stereo sockets to RCA spare if you want them, or they are available online.
Andy.
If your in the UK I have some 1/4" stereo sockets to RCA spare if you want them, or they are available online.
Andy.
Thank you. I remember going down a similar sort of route and learning my lesson when I built my dummy load. I started with complicated switching to select between 2,4,6, 8 & 16 ohm loads, 100:1 & 10:1 attenuation with variable attenuation, numerous BNC with switching out the ying yang: this nightmare didn't last long, I ended up with two 400w 8r loads or a gurt big tunnel heatsink out of a welder, with a fan, to SW to a speaker I used the above 3 x banana sockets with removable U links, a stereo fixed 10:1 attenuator with 3 BNC's for each "channel". I rarely use the stereo functionality, for other loads I just stick a resistor in a cup of water.
It's easy to get carried away and build something complex,but then you find odd stuff happening when listening and testing, with a complex doodad it's harder to fault find and easier to get disheartened.
Andy.
Last edited:
My friend wants a switcher can safely switching tube or solid-state amplifiers to one pair of speaker, so I built two identical speaker level amplifier switchers and one line level switcher for him.
I put a master/slave switch on the amplifier switchers and use DB9 cable connect them together, so he can easily use that for mono blocks or bi-wire/bi-amp system. If both amplifier switcher set to "slave", the line level switcher will becomes the "master" switcher and switch the source/preamp to the amplifier input.
The line level switcher will mute the input signal while switching for additional safety, it also shorted the input of the amplifier that is connect to the dummy load, so the dummy load can use a lower wattage (5Watt) resistor.
I put a master/slave switch on the amplifier switchers and use DB9 cable connect them together, so he can easily use that for mono blocks or bi-wire/bi-amp system. If both amplifier switcher set to "slave", the line level switcher will becomes the "master" switcher and switch the source/preamp to the amplifier input.
The line level switcher will mute the input signal while switching for additional safety, it also shorted the input of the amplifier that is connect to the dummy load, so the dummy load can use a lower wattage (5Watt) resistor.
Great discussion goes here 
@chrisng - I referenced you in my first post. Thanks for sharing your relay based design as well.
After some thought, I tend to go for the uC controlled relays based design, as it gives much flexibility when it comes to switching sequencing. I made a first design draft, would appreciate your feedback. One comment regarding this design- as for now I'm considering A/B only, I chose switching the speaker terminals using DPTD relays in order to save two relays. There are some drawbacks to this design, but it can easily be changed to pure single pole per amp contact per channel in case more than two amps have to be supported. The control logic should work more or less like this:
Current stare: Active amp is connected to speaker out, its dummy load is disconnected. Non-active amp is either connected to it's dummy load or open (depending on the switch position).
Transition: connect dummy load to the current active amp (unless its dummy load switch is set to off) -> switch the main relays -> disconenct the dummy load from new active amp (in case it was connected).
@chrisng - I referenced you in my first post. Thanks for sharing your relay based design as well.
After some thought, I tend to go for the uC controlled relays based design, as it gives much flexibility when it comes to switching sequencing. I made a first design draft, would appreciate your feedback. One comment regarding this design- as for now I'm considering A/B only, I chose switching the speaker terminals using DPTD relays in order to save two relays. There are some drawbacks to this design, but it can easily be changed to pure single pole per amp contact per channel in case more than two amps have to be supported. The control logic should work more or less like this:
Current stare: Active amp is connected to speaker out, its dummy load is disconnected. Non-active amp is either connected to it's dummy load or open (depending on the switch position).
Transition: connect dummy load to the current active amp (unless its dummy load switch is set to off) -> switch the main relays -> disconenct the dummy load from new active amp (in case it was connected).
Last edited: