Relays that are "audiophile" grade

[MakeMineVinyl]

I might add a third relay just for audio ground side, so it toggles between both inputs. Here I can just one of these Omron relays, as a minimal chance of crosstalk is not needed for this signal. Don't want the risk of groundloops.
As for pops, my amp has a speaker protection circuit in it (Neurochrome Guardian-86), so that *should* keep the speakers safe.

I'll post the PCB drawing once I have that figured out. Thanks for all the help!!

The pop problem can arise if for some reason both relays are 'open' to signal ground for an instant when switching. If it were me, I'd probably implement a jumper arrangement where I could select via jumpers whether the signal grounds are all connected together, or go through the relays. That way, you would have both bases covered.

I'd just leave the +12V 'GND' isolated. You might even consider placing a 0.1uF cap across the relay coil in parallel to the diode to suppress potential pops introduced by the coil energizing. You could always place a footprint there for a capacitor and only stuff it if you get a pop from the relay coil.

This stuff can really start playing mind games with you with all the details and possible ways things could go wrong, even with a 'simple' circuit.

 

[Count Arthur]

Go old school:

Elma 4 pole 6 way switch, 04-2264 | HIFICollective

Elma 4 pole 6 way switch, 04-2264

www.hificollective.co.uk

 

[pma]

Go old school:

Elma 4 pole 6 way switch, 04-2264 | HIFICollective

Elma 4 pole 6 way switch, 04-2264

www.hificollective.co.uk

 

[MakeMineVinyl]

Go old school:

Elma 4 pole 6 way switch, 04-2264 | HIFICollective

Elma 4 pole 6 way switch, 04-2264

www.hificollective.co.uk

That would be the ideal, but defeats the purpose of having a switch which is automatically activated by a trigger voltage.

 

[JayGilb]

I might add a third relay just for audio ground side, so it toggles between both inputs. Here I can just one of these Omron relays, as a minimal chance of crosstalk is not needed for this signal. Don't want the risk of groundloops.
As for pops, my amp has a speaker protection circuit in it (Neurochrome Guardian-86), so that *should* keep the speakers safe.

I'll post the PCB drawing once I have that figured out. Thanks for all the help!!

Why don't you add a pair of jumper pins to the power pin of the ground switching relay in case you're getting a pop from switching the ground. If you are getting a pop, just remove the power jumper and leave the ground relay not energized.

Edit: I see that MakeMineVinyl covered this idea.

 

[JayGilb]

That would be the ideal, but defeats the purpose of having a switch which is automatically activated by a trigger voltage.

Just add a 12vdc solenoid and a self-reversing ratchet mechanism to flip the switch back and forth.

 

[MakeMineVinyl]

Just add a 12vdc solenoid and a self-reversing ratchet mechanism to flip the switch back and forth.

I'm sure I could find an excuse to place a vacuum tube in there somewhere.

 

[Yasuo]

What about an optical sensor acting as a relay? That way you can isolate the circuit.

 

[starfly]

Updated schematic below. Added a jumper that will allow me to have the - (negative/ground?) audio signal either go to the output, or for it to just be grounded. Should I do the same for the pins 2 and 4 on the output?

 

[MakeMineVinyl]

Updated schematic below. Added a jumper that will allow me to have the - (negative/ground?) audio signal either go to the output, or for it to just be grounded. Should I do the same for the pins 2 and 4 on the output?

View attachment 157156

In either position of the grounds relay, the ground has to follow through to the output (J4) otherwise with the jumpers connected to 1&2, the output doesn't have a ground throughput at all. The ground relay should just determine whether the grounds from input 1 and 2 are tied together or not. The jumper could just determine whether that relay is active or not; the relay would just short the grounds from 1 and 2 together when activated, and not short them together when it is not active (or vice versa). If I wasn't so busy, I'd draw it out, but I can't.

 

[starfly]

In either position of the grounds relay, the ground has to follow through to the output (J4) otherwise with the jumpers connected to 1&2, the output doesn't have a ground throughput at all. The ground relay should just determine whether the grounds from input 1 and 2 are tied together or not. The jumper could just determine whether that relay is active or not; the relay would just short the grounds from 1 and 2 together when activated, and not short them together when it is not active (or vice versa). If I wasn't so busy, I'd draw it out, but I can't.

So, like this?

 

[MakeMineVinyl]

So, like this?

View attachment 157163

Realistically, the input grounds for L and R for each input can be connected together since they are common to each component, and that would involve only two contacts on the ground relay. That is not such a big deal though.

I would have the jumpers directly connect the input grounds directly to the output ground, which essentially connects the grounds together and bypasses the ground switching. That way you can have either hard connected grounds if there are no ground loop problems (and avoid possible ground switching transients), or have fully discrete and isolated ground switching. If you tie the input L and R grounds together, that would be 2 jumpers, and if not, it would be 4 jumpers.

 

[starfly]

Realistically, the input grounds for L and R for each input can be connected together since they are common to each component, and that would involve only two contacts on the ground relay. That is not such a big deal though.

I would have the jumpers directly connect the input grounds directly to the output ground, which essentially connects the grounds together and bypasses the ground switching. That way you can have either hard connected grounds if there are no ground loop problems (and avoid possible ground switching transients), or have fully discrete and isolated ground switching. If you tie the input L and R grounds together, that would be 2 jumpers, and if not, it would be 4 jumpers.

Thanks. I think I see what you're saying, and I changed the grounds relay to a SPDT instead of a DPDT:

 

[MakeMineVinyl]

Thanks. I think I see what you're saying, and I changed the grounds relay to a SPDT instead of a DPDT:

View attachment 157170

Yep, that should do it. When you lay out the PCB, make the "GNDS" net connected to the ground plane pour - don't attach it to anything else.

 

[starfly]

Ok, version 1 of the PCB. This is a dual layer PCB, with the green hatches outline representing the bottom layer fill (connected to the GNDS net), and the red hatches representing the top layer fill (connected to the +12V net).

Added top and bottom layer fills so it's easier to see.

 

[antcollinet]

Those are spider web width traces. I'd be making them much wider.

Gaps between traces to be about the same thickness as the traces you currently have.

 

[starfly]

Those are spider web width traces. I'd be making them much wider.

Gaps between traces to be about the same thickness as the traces you currently have.

0.25mm isn't wide enough? What should it be, 1mm?

 

[MakeMineVinyl]

0.25mm isn't wide enough? What should it be, 1mm?

My default trace width is 0.035 inch, which is about .9mm.

 

[starfly]

My default trace width is 0.035 inch, which is about .9mm.

Because of the bottom layer pour being for GNDS, and the top layer pour being for +12V, I don't actually have to connect those pins with traces, correct?

 

[MakeMineVinyl]

Because of the bottom layer pour being for GNDS, and the top layer pour being for +12V, I don't actually have to connect those pins with traces, correct?

I would just make both pours the GND net. There's no reason to make a power plane for the +12V since its just powering 3 relays. There could also be a potential problem if the +12V power supply is noisy, it could induce noise into the audio traces if it surrounds all of those traces. Remember if there are landlocked areas which won't pour, use vias from the top to the bottom layer so this doesn't happen.