Protecting from two power supplies accidentally connecting

[wayfaerer]

TL;DR: If, through user error, two power supplies end up connected (9V to 9V, GND to GND), will that damage anything, and if so, can diodes prevent it?

Hello! I'm providing an expression input jack in a pedal--the jack will power a vactrol which in turn controls one of the pedal parameters. Here's what it looks like:



Between the jack and the LED is some active filtering, and long story short, the LED side of the vactrol needs both power and a control voltage--that's why it's labeled the "LED Circuit" instead of just a solitary LED.

I want to provide the option to fully isolate the Exp In and LED circuit from the audio circuit and main pedal power. This is what the DPDT is for. For example, if I attach a noisy digital MCU to the exp in jack, it can send its own power, ground, and control voltage to the LED circuit, in which case I'd flip the DPDT switch to disconnect the pedal's power supply from Exp In. The MCU's digital noise would then stay out of the audio circuit.

Assuming it needs protection, I want to protect against the case where I forget to flip the switch and the MCU's power supply ends up connected to the pedal's power supply. Is this the same thing as reverse polarity protection... i.e., should I just put a diode in series with the 9V (and same thing in the MCU wiring to prevent the main pedal's power supply from flowing into the MCU)? Of course, if the worst case scenario is just that the MCU's digital noise ends up in the audio signal, then I can forego the protection and just treat the noise as a reminder to flip the switch.

[Rob Strand]

It's not 100% clear from you pic the connections of the two supplies but I think I get the idea.

If, through user error, two power supplies end up connected (9V to 9V, GND to GND), will that damage anything, and if so, can diodes prevent it?

Protecting against *all* types of abuse can often get quite complicated.

One thing that needs considering is the LED circuit can (and should) be designed to be less dependent on the supply voltage.   For example you don't really want the LED to be driven hard from say an unregulated 9V supply (which could be 12V) and less on a regulated 9V supply.  That would make the "calibration" of the expression pedal dependent on the external supply voltage.   One way to solve this is to make pedal control the current through the LED.  The other would be some form of crude voltage regulation.   You don't anything complex, for example a zener in your existing circuit would work.  Maybe it already runs off 5V internally?

The main point here is if you can get the circuit to run at lower voltages then you can use diodes as protection.   In the simplest case:
- Add one diode so the anode on the effects +9V and the cathode connects to the LED circuit +V rail.
- The separate +9V adaptor for the "LED Circuit" connects to the anode of a second diode and the cathode connects to the LED circuit + rail. 
That mixes the supplies together and there is way no power can pass from one to the other.

Now if you wanted to simplify the isolation you could just put 10 to 100ohm, depending on the LED circuit current,  in series with with the ground and +V lines between the expression pedal and the Audio circuit.   

This should provide enough isolation between the two circuits and you don't need to use the DPDT switch.  The switchover from one power source to the other is automatic.  A "flaw" in this idea from a pure isolation perspective is if the effect is actually +12V and the LED circuit is 9V , the diodes will cause current to pass from the effect +9V instead of the LED power.    However the idea is the resistors keep the circuit isolated.  In fact with resistors and caps you probably don't need the separate PSU source.

If you start getting into problems with low +V rails you can use Schottky diodes and tweak the resistors in the PSU.   The LED and PSU ckts should have some caps on their +V rails.

If the LED circuit is pulling a lot of current then a different solution might be required.

[EDIT]
I should say this a simpler way:
- If you can't get the LED circuit to work off the effect supply noise free you shouldn't offer that option.
- The thing is I'm sure you can get it to work.
- And if you can get it to work there's no need to have the separate supply option.

[wayfaerer]

Thanks for your thoughtful reply.

I should say this a simpler way:
- If you can't get the LED circuit to work off the effect supply noise free you shouldn't offer that option.
- The thing is I'm sure you can get it to work.
- And if you can get it to work there's no need to have the separate supply option.

A few days ago I would have disagreed with this. I've been working on this circuit for months, and I've always had difficulty keeping the MCU's digital noise out of the audio (previously I was using digipots). I came to the conclusion that it wasn't possible. I made this vactrol circuit so that I could fully isolate them and be done with it.

After reading your post I decided to try powering the MCU through the pedal's power and, surprisingly, there was no noise. I guess the vactrol setup isolates the audio circuit from the MCU "enough"--the audio signal doesn't go anywhere near it. So I guess I don't need external power and a switch!

But I'm still interested in my original question for future reference. Just to clarify: was your suggestion to use diodes *or* 10-100ohm resistors, or both? And you think it would provide enough isolation that even a switch wouldn't be necessary?

Thanks!

[wayfaerer]

Forgive the quality, but I just wanted to provide this as a clarification to the original diagram.



But as stated above, I plan to remove the switch and, when using an MCU, just power it from the pedal's power rather than from a separate one.

[Rob Strand]

But I'm still interested in my original question for future reference. Just to clarify: was your suggestion to use diodes *or* 10-100ohm resistors, or both? And you think it would provide enough isolation that even a switch wouldn't be necessary?

The resistors and diodes are doing different things.  You can have either or both but for protection and isolation you need both the resistors and the diodes.

The purpose of the diodes is protection ie. to stop the power from one supply feeding into the other.   The side-effect is power is supplied from the rail with highest voltage and the other supply doesn't contribute; that's the "flaw" I mentioned.

https://4.bp.blogspot.com/-N7a81J8p9Fw/UxSz1n2y5WI/AAAAAAAAAcc/rna55boyzXQ/s1600/F.jpg

The purpose of the resistors to prevent two separate circuits interacting via a common power rail.   Caps on each circuit's power rail help further.  Glitches on any power rail have little chance of coupling back into other circuit.  Use small as possible resistors to prevent drops on the PSU rail.

Now if the resistors are deliberately made large enough they will cause enough voltage drop so the local supply comes from it's own external power source.   This sort of undoes the flaw.   The side effect is when running of a single supply source is the "other" circuit, which is supplied via the resistors, will see the voltage drops.

The thing to note is these are not nice clean general solutions.   They all have little side effects and it's upto you to make sure the circuits can deal with the side-effects.   If they can't the solutions aren't good enough.