Tracking down bypass popping

[mark2]

Does anyone have ideas for how to find and eliminate loud popping I get when bypassing or engaging the following?

It doesn't seem to be from the LED; I tried both mitigating and clipping it out altogether.

The volume knob amplifies it.



EDIT: Also does anyone fit the bill or know somebody with expertise on this topic who would consult on a different pedal? Preferably in the US since it'd probably be quickest to just ship you a prototype. I have another, much lower gain pedal that has a slight pop, but it's really problematic since it has a delay that repeats the noise any time you engage it. I can't even seem to mute it with an optocoupler and am considering hiring a somebody to analyze and fix.

[anotherjim]

I suspect the current pulses when operating the latching bypass relay is being "felt" in the audio path probably via the power supply routing. If you want someone to solve it, it might be worth showing the PCB routing.

[EBK]

The volume knob amplifies it.

Could you post some pics of the build?
My gut tells me you have some long/indirect ground paths that could be improved.

[mark2]

Here are some pics of the routing. The thicker traces are the +5V and +9V, and the plane is GND.  Gerber is here.

Let me know if you spot any obvious culprits. Thanks in advance.

[R.G.]

I like A.J.'s answer.
Relays have all the same issues as true bypass switching, and add the new issues of power supply glitches, ground glitches, and coil voltage coupling. Magnetic coupling of coil current pulses is theoretically possible, but I've never encountered it.

The current to and from the coil can cause voltage blips on the power supply and ground. Attack this by making a local power/ground domain for the relay only. Put a local decoupling cap of ~100uF to 220uF as physically and PCB-trace-wise close to the relay and driver transistors as you can put them, and isolating the capacitor from the rest of the power in the pedal with a "refill" resistor. This resistor/cap should be the place where PCB traces from the rest of the pedal touch the local relay power/ground traces. This forces all the current from the relay coil switching to be satisfied from the local cap, not the rest of the pedal power supply. The exact PCB traces to/from the relay coil domain matter - a lot.

The refill cap should really be as big as you can bring yourself to make it. The refill resistor is much bigger than you would use to decouple a general power voltage. It's fine if the relay cap voltage drops a lot when the coil pulse happens. I size them so the decoupling cap refills in about 50-100ms. Humans can't really operate a footswitch faster than 20 times a second, so the cap is filled back up and ready for the next human action. Using a slow refill and a local decoupling loop may fix your issue entirely.

Capacitive coupling of the coil voltage to the audio path is solved by using a better relay. The internal construction determines how much of the coil voltage edges couple to the contacts.

Your circuit will allow you to run your relay from a higher voltage than 3.3V, as you're using transistors to fire the coils. The lower the coil voltage on a given relay, the more current the coil pulls and the worse the power supply/ground glitches will be. Use a relay with a coil rated for the incoming pedal power voltage. If you have 9V, use it for the relay.

Second, relays vary a lot in how much the volts-size coil signal couple to the contacts. Your input signal runs right through the relay, and worse yet, the input path impedance is up in the megohm range, the worst possible combination for coupling coil glitches. Use relays specifically rated as low-signal relays. Some low-signal relays have an internal shield around the contacts to keep coil glitches out. Use them if it's economically feasible. If you can't use shielded-coil relays, try different brands. Different internal construction matters in how big a glitch you get from the coil.

[mark2]

Thanks for so many great pieces of advice. I'm still working through the suggestions (and will probably be ordering a handful of different PCBs to experiment with).

and isolating the capacitor from the rest of the power in the pedal with a "refill" resistor.

Can you point me in the right direction to learn more about the "refill" resistor? I don't think I'm understanding where to put it or what it's for.

Thanks again, and in advance.

[Rob Strand]

Thanks for so many great pieces of advice. I'm still working through the suggestions (and will probably be ordering a handful of different PCBs to experiment with).

and isolating the capacitor from the rest of the power in the pedal with a "refill" resistor.

Can you point me in the right direction to learn more about the "refill" resistor? I don't think I'm understanding where to put it or what it's for.

Thanks again, and in advance.

If you look at R20 it provides a ground path for C1.   C2 does not have a permanent ground path.  It needs it's own equivalent of  "R20".   I suspect this is causing the pops.   You can wire the resistor to the top side of R20 (instead of ground) if you like to prevent additional input loading.

[mark2]

So something like the following? And any tips on sizing those R13 and R14's? I chose 1M as arbitrarily large.

If you look at R20 it provides a ground path for C1.   C2 does not have a permanent ground path.  It needs it's own equivalent of  "R20".   I suspect this is causing the pops.   You can wire the resistor to the top side of R20 (instead of ground) if you like to prevent additional input loading.

That makes a lot of sense. But I did a quick test with a 1M from C2 to GND but no change in sound.

I'm shelving this, but have to solve a similar but much less severe problem on another pedal. But the other, while being quieter, is more pronounced since it has a delay that'll repeat the offending noise. Anyhow, I'm obviously trying to apply any lessons here to that one as well.

Any chance you , anyone else here or somebody you can recommend (preferably US based) are available to do an initial consultation on that other pedal (not yet shown)?  I'm thinking of
1. putting together the schematic and notes/vid on the symptoms, as well as a list of experiments I'm planning to try with a new batch of PCBs.
2. (soon, initial consultation) Having somebody with more experience validate/shoot-down those ideas, give other suggestions if needed, and give the whole schematic a once over for sanity. e.g. 1-2 hrs of consulting.
3. build a new one, or a handful as needed, then
4. (~3 weeks later) Bring that person back in, optionally sending some prototypes as needed to finish debugging. Size of that job TBD, but hopefully small.

I can also start a separate thread with more details.