Mixed signals on same ground/wire

[TheWinterSnow]

I have a feeling that I am over thinking this, but if I have an audio (guitar) signal that has a ground wire that is sharing 15mA DC current from an LED have noticeable audio/noise issues given the LED/current limiter noise and the voltage drop across the ground wire given the current and thus acting as an antenna to any noise that may be received on the wire?

Would it be worth my time to add an extra ground wire just for the LED leaving the original for just the guitar's earth reference?

[GibsonGM]

A pic might be helpful, winter...there's no 'guarantee' that any wire will or will not act like an antenna / have coupling problems.   If it's easy to try it with the LED in, and then out, you may be able to just do it by ear.

The best way to run grounds IS to 'star ground' - audio path has its own connection to ground, and ancillary things like your LED would, too. But in a very simple stomp circuit, it may not be worth much to try to do so.  I've always just run LED grounds back to the PCB, where the "-" (or what ground you have) goes in to the board...easy 'nuff...

**EDIT**
I guess I should also say - if you're having long ground runs, long enough to have a non-trivial resistance that allow much of any current to flow...you need to rethink the ground scheme...but again, what is so complex as to make you run into that?  Not a typical stomp, I'd think.  Tube amp, maybe...

[R.G.]

What he said.

One other bit of finesse. You can separate the signals in frequency as well as separating them on wires. The guitar signal will presumably be amplified quite a bit. So any other signal confounded with it will be amplified too. But if you slow down the other signal so it's frequency components are not amplified with the guitar signal, you avoid a lot of the issues.

That's why one solution to switching transients like an LED going on/off is to slow down the transitions. It's only the audio band components of the sudden on/off that you can hear, and probably that the guitar amplifier will respond to. So if you make the LED current ramp up and down slowly enough that it eliminates the audio components, there is no audible interference.

As always, the cost of complexity waits in the wings. Sometimes another wire is cheaper.

[TheWinterSnow]

@GibsonGM:  My designs are based on the fact that I have a true bypass PCB that connects to both the input jack, the output jack, and runs the ground and board IO to the main PCB.  On the switching PCB where the 3PDT switch is located I have an LED that indicated whether the effect is on or not.  I am supplying power from the main board back to the switching PCB (the LED), but the guitar's ground is therefor sharing current with the LED.  My worry is that what small amounts of DC offset would render the ground plane of the switch PCB as an antenna and render noise onto the guitar's input by the time it reaches the main PCB and into the first semiconductor?

RG's post reminded me also that there would be a DC pop when you go from no DC offset when the LED is off to a DC offset when the LED is on...I think I should run them on two separate ground wires to the main PCB.

[GibsonGM]

The best of all situations WOULD be that you have a ground going directly from your LED to the power supply ground, yes.   Even that may not kill a pop, tho.   Like R.G. talked about, you may need to do the 'cap trick' to 'slow down' the current pulse to the LED.  Star grounding is good, tho, and I see zero down side to it. 

How much you're amplifying your signal will also affect the *POP*.  High-gain circuits can be nearly impossible to tame.  I have found that we inevitably live with some *POP*, but that when on stage, it is not audible - depending of course on what genre you're playing in, how you use the switch.  I've never had any issues with a pop being "embarrassing" or anything...

I think you'll find that if you connect LED ground (from switch to gnd...) right to the PCB, or to the battery ground, you'll see little difference at all, if your leads aren't long.   Try it!  Best way to see what's up.   

You won't get an "antenna" per se, but I know what you're saying - you could induce a current, perhaps.  If the circuit IS high-gain, use a shielded wire for your input - issue gone for 99.9% of cases.   I'd use shielded cable for high-gain, anyway, as a matter of principle; cuts down on noise...