audio AB switch questions

[luisdent]

I'm planning to make an audio switch to compare source devices (phone/mp3/etc.) and was hoping I could get some help on the electric aspects of the design.

To give as much info as possible, I've already created a small AB switch in the past using an old JDC C5 amp chassis. I used really small 3.5 connectors and a tiny 3PDT switch. Grounded the jacks to the chassis. Worked great for a while, but ultimately the components are failing.

So... I wanted to make a version 2 with a few functional improvements, but I'm a little weak on the electrical comprehension. My goal is the following:

- more robust switches and jacks
- LED indicator(s)
- transparent audio quality
- passive audio design
- randomization funtion to make somewhat "double blind"

My idea so far is to use a 1590B alumimum chassis for the encolsure. I have a 3d printer, but I've read using a plastic enclosure can allow interference to get into the audio signal? The metal chassis prevents this more by acting like a faraday cage? It's also heavier which may make switching a bit easier.

In any case (pun intended), I was thinking of using a rotary switch like a 2P6T switch. That way I have separate pins for the LEDs and I could essentially spin the knob in a way to sort of randomize the position so I don't know what it is until after I make a guess at which time I could verify the LED status.

I would panel mount the switch and a few decent 3.5 jacks to the case. All simple enough, but here are the aspects i don't fully understand.

- What specifications does the switch need to be? If the audio signal is a low level phone/mp3, etc. meant for earphones/headphones, I assume I need to at least meet some minimum electrical specifications (amps/volts) but how do i determine the ideal specs and is it possible to go too far above the specs and cause an issue? For instance, would i use a 30V/1A switch? If so, and 10A is overkill would it sill work or cause undesired effects?

- What guage of wire is ideal? I assume for that signal level and short distances something like 22awg is fine, but is 20awg better? Any or no benefit? I understand wires have impedance/resistance and conductivity might differ, but i'm assuming those factors are infinitesimally small and irrelevant to this use case in terms of affecting the audio signal?

- Do I ground the switch and all 3.5 jacks to the chassis? Or is there a better method like grounding together and then one to the chassis? I assume that doesn't make any real difference? Do I leave any components (i.e. the switch) ungrounded for any reason?

- Is it possible to find a switch that makes very little to no noise when switching inputs? I've ready about relays and other things, but I'm not too familiar with those. I could also understand using something like arduino to program a randomized "switch", but I don't understand how it would control the physical signal path and not make any noise when doing so? I'm open to any suggestions. I would prefer it isn't insanely expensive or require a lot of battery/power, and I also don't want it to "require" power for the audio (aka, it still works if the LED power dies). The LEDs could run off a 9V battery or something, but not be needed for the audio portion.

- How do I wire the 9V battery/resistors for the LEDs without it interfering with the audio ground/signal quality?

For reference, here is my original AB switch: https://imgur.com/a/ab-switch-Q02JMCv

I just repurposed the shell of the c5 amp (i had extra "shells") and drilled extra holes and wired the components with basic wire. Very basic.

it works perfectly, but if I'm honest, I didn't "fully" understand how to spec the switch... I just found components that fit the case. So I think I just used the switch model that JDS used in their altoids amp.

Anyhow, that's all great, but the desire for something more silent, durable, and perhaps randomized has me confused, and I'd like to make I'm using components and specs to keep things as audibly transparent as possible.

[microwhat]

I'm going to assume you're dealing with low-impedance headphones and sources intended to drive them.

In my experience, headphone signals don't need any shielding, whether in cabling or enclosures.

If the switch lists a minimum switching voltage/current, its behavior is undefined if used below that minimum. Switch contacts intended for making and breaking larger currents use different materials that will survive the inevitable arcing, but those materials won't contact reliably at much lower voltages. If no minimum is given, then I would not expect any difficulty.

Use whatever gauge of wire is convenient. Your signals are probably in the range of 1mV-1V, and a few tens of mA. Larger gauge wire will only irritate you. For much longer runs (>>10'), cable construction (i.e. conductor twist rate, other details of construction that affect the cable's capacitance, etc) matters more than gauge.

It's generally best to ground/common all devices at a single point, but remember that bonding is more important than grounding. If all your ring/common/pin1/$commonContact are connected together with a wire, there's no need to also connect them with the chassis (it won't hurt if you do so anyway), or to a stake driven into the dirt in your back yard. Connecting the chassis of the rotary switch, for example, with the signal ground, won't do anything useful, but may allow ESD from the user's hand to enter something expensive.

Pops and clicks when switching audio signals can come from many sources, but often from DC offset coincident with the signal. I've never needed or attempted to mitigate this, so I can't say any more.

If you switch audio signals with different poles than DC for indicators, I would not expect the indicator power to interfere with the audio signals.

[luisdent]

My sources are anywhere between around 1-2V (portrait dap) to even 10V (probably my most powerful jds headphone amp). And the headphones range from as low as 8-32 ohms for iems to around 100-300 ohm max for some over ears.

As i review a lot of equipment it could be all over the place, but that's probably a good ballpark.

[Andy Chee]

This page describes A-B switch box design in detail

https://www.sound-au.com/absw.htm

https://www.sound-au.com/abx-tester.htm

[luisdent]

Nice. thanks. That looks a bit more involved than i was hoping. To be clear, the randomization aspect is probably the least important to me. Although it would be nice to have, I'd like to keep things simple and affordable. I'll look at those links again later when i have more time though.. thanks.

[microwhat]

I don't think it needs to be complicated. Audio frequency signals between an amplifier and headphones are relatively forgiving. I suggest you put some connectors and a switch into whatever box you have available, and see what it does. If it works, great. If it's noisy, you can cross that bridge when you get to it.

It's unlikely that whatever you build will be worse than the manual plugging of headphones into $source2 rather than $source1.

[Andy Chee]

The complexity stems from the "Double-blind" protocol.

Otherwise the switching scheme is electrically trivial.

[luisdent]

So would the rotary switch i mentioned work with no issues?

Regarding it being fine as long as there's no noise... To me, the point of the switch is to see if you can tell the difference between two audio sources. If I'm listening to very very small differences, I don't want to rely on the fact that it might be due to noise from my components. I want to know that technically, spec-wise, there's no reason the components should be affecting the audio quality.

Is there a way to properly spec the switch? From what I gather the wires aren't as critical, as long as they are sufficient for current flow. So the only component that really matters would be the switch. At what point are the specs not sufficient? And at what point are they overkill? I guess that's the main question...

Can someone recommend a value range for what sort of switch I should be looking for.?

[Analog Kid]

You're dealing with the amount of power to drive a set of headphones.
Worrying about the capacity of switches for such a device is absurd.
Just use whatever switches are suitable.
Same thing with wiring; thin wires are fine. No power-carrying capacity considerations needed.

(It would be different if you were switching speakers.)

[microwhat]

If you want to switch both left and right channels plus the indicator LED, then you'll need 3 poles. Otherwise, pretty much any rotary switch you want will work. I wouldn't suggest buying it from aliexpress.

[Analog Kid]

Yes.
I don't want to insult your intelligence, but here's the schematic for what you need:



Any decent 3PDT rotary switch will work (probably even cheap ones from ... who knows where?).

[themadhippy]

Any decent 3PDT rotary switch will work

To be on the safe side go with a break before make variant,not a make before break

[Analog Kid]

Any decent 3PDT rotary switch will work

To be on the safe side go with a break before make variant,not a make before break

Sure, if you have that choice. (Prevents both channels from being connected at once.)
But if you don't have that choice, doesn't really matter much. Nothing's going to get damaged (including your ears).

[TimFox]

The important thing to avoid "pops" when switching is to ensure there is no DC component to the signal voltages being switched:  use series capacitors driving shunt resistors if needed.