Grounding on relays

[BillP]

As a personal project, I seek to build a loop switcher with relays driven by a micro-controller.

I have played around with arduino for a while ( arduino being the software guy's POLR into hardware) and have bread boarded out some buttons, a relay driver, and midi using off the shelf demo circuits. This looks like it is working just fine.

The part that I am confused about is how to ground the thing.

My instinct is that tying the micro controller ground to the audio signal path ground might be a source of noise and problems.
RG's geofx articles on relays have me convinced that relays are electrically noisy on the solenoid side.

My initial plan was:
1.  tie all the audio jack grounds to the chassis at a single point (star ground, yes?)
2. use an isolated power jack
3. tie the microcontroller board ground to the power jack only, leaving it isolated from the audio signal path and vice versa.

I have googled all available looper schematics, and read a half dozen articles on grounding in pedals and still haven't been able to determine how to handle grounds, other than to be sure to space the returns far enough away from the sends that you don't induce feedback loops there.

Does anyone have some advice to a pedal newbie?

Thanks,

Bill

[PRR]

Try it isolated (separate grounds).

You may be over-worrying the "noise" (click/pop) issue; but if it happens, then try tying the grounds together.

Do NOT route your audio lines right-on the relay coils.

[ReCoC]

Came here exactly for the same reason
Same case, building a looper that will be switching relays via a uC and am also worried about the noise.
It might be switching up to 10 relays at the same time. So if you found your answer, please share it

Do NOT route your audio lines right-on the relay coils.

What do you mean by that? What is the other option?

1.  tie all the audio jack grounds to the chassis at a single point (star ground, yes?)
2. use an isolated power jack
3. tie the microcontroller board ground to the power jack only, leaving it isolated from the audio signal path and vice versa.

But if the audio jack grounds are tied to the chassis and the chassis is not really grounded, then any EMI noise will have nowhere to go.
Isn't that right? Or am I missing something?

[JFace]

But if the audio jack grounds are tied to the chassis and the chassis is not really grounded, then any EMI noise will have nowhere to go.
Isn't that right? Or am I missing something?

If the audio jack grounds are connected to chassis, then the chassis is considered grounded, and the circuit is shielded.

I have tested microcontroller ground schemes, and from what I could gather, it is best to power them from a separate, isolated power supply. When using a microcontroller inside of a pedal, even with the grounds coming together only at the DC jack, there was still audible DC pops coming from the surge current to toggle the relay position. I attempted to isolate the grounds using inductors, diodes, resistors, etc, but ultimately the audio circuitry was too sensitive to these current spikes. This isn't much of a problem in a dedicated relay bypass strip since it would get it's own power supply anyway, but when it's within a pedal, I haven't been able to use microcontrollers effectively (try other logic devices). These spikes may be "absorbed" by other pedals/amps and may be acceptable to some users. I plug directly into PA systems, so these little pops become big problems when amplified in the mains.

[Beo]

I've been playing around with this as well. Using a 5v regulator and a 5v latching relay provides some power isolation, but at some point the digital ground has to meet up with common ground, usually at the power jack where analog ground also connects. I've read about the idea of using a small resistor inline for the digital ground, but I haven't seen that as doing much in my testing.

The other question I've struggled with is with the use of an optotransistor for muting. I've seen relay bypass circuits where the opto ground is just connected to digital ground. To me it seems like this should be a true audio ground, isolated from digital and connecting directly to the effects circuit. The only place digital and analog ground should meet is at the power jack.

[JFace]

The other question I've struggled with is with the use of an optotransistor for muting. I've seen relay bypass circuits where the opto ground is just connected to digital ground. To me it seems like this should be a true audio ground, isolated from digital and connecting directly to the effects circuit. The only place digital and analog ground should meet is at the power jack.

I've tested this as well. Figuring out the build-out resistor for controlling the timing of when the opto mutes is tricky. I used a trimmer pot to get it perfect, and then the next day it wasn't right anymore. I've since given up on microcontroller bypassing, and use logic gates to control the states.