Ground Lift Switch or...?

[Octavian]

Building a combo looper & A/B/Y amp selector pedal and wondering about the group loop. First, do I need a ground loop switch, or could I just use the plastic jacks that don't touch the chassis? Second, if I do, can I just build the switch into the A+B footswitch?

Also, for the switch, I was looking at this schematic: https://mwynwood.files.wordpress.com/2014/03/2w2g31u.png and was wondering if someone could explain how this works exactly. It would seem to me that the ground is connected between the jacks regardless of which position the switch is in?

[merlinb]

Building a combo looper & A/B/Y amp selector pedal and wondering about the group loop.

What ground loop?

It would seem to me that the ground is connected between the jacks regardless of which position the switch is in?

Correct.

[wavley]

Ground loops only need to be solved if they exist.  Your two amps and pedalboard may not hum from a loop, just try a simple y cable out of a buffered pedal and see if it hums.

If you do have a ground loop, you will have to use jacks that are isolated from the chassis in some way and you can either do a ground loop switch or something like RG's Hum Free ABY on Geofex.  I am a really big fan of the latter.  Currently I've actually separated my transformers from the rest of the circuit (partially to save space on my board because I play baritone and Bass VI so I used bigger Edcor transformers) the other advantage is that the buffered signal runs the longer length of cable and the transformer to amp length is just patch cable.  I'm thinking about switching my buffers to be balanced to that I run mic cable from my board to the transformers for those big stages with long runs of cable, plus mic cable is easily made longer or shorter by mating cables of different lengths and you don't even need an adapter like a 1/4" cable would.

[Octavian]

Ground loops only need to be solved if they exist.  Your two amps and pedalboard may not hum from a loop, just try a simple y cable out of a buffered pedal and see if it hums.

This pedal isn't for me, so I'd like to be able to kill any ground loop that *might* exist.

Is there any downside to integrating a ground loop switch into the A&B footswitch? And am I understanding correctly that (from the diagram I linked to) the switch in the position connecting the middle and bottom position is for when the A&B switch is off?

[wavley]

Ground loops only need to be solved if they exist.  Your two amps and pedalboard may not hum from a loop, just try a simple y cable out of a buffered pedal and see if it hums.

This pedal isn't for me, so I'd like to be able to kill any ground loop that *might* exist.

Is there any downside to integrating a ground loop switch into the A&B footswitch? And am I understanding correctly that (from the diagram I linked to) the switch in the position connecting the middle and bottom position is for when the A&B switch is off?

Not at all, if you're building it for someone else or planning on not always using the same amps that play nice together then it's a great idea to include a ground lift.  I'm a much bigger fan of properly lifting the ground of the signal by using a switch or transformer instead of doing the unsafe thing of lifting the safety ground on an amp (which you will find lots of stupid folks recommending to do on teh webz)

I would imagine that it's just a two way switch so one way it's open and the ground of jack 2 goes through the cap and resistors and the other it shorts them and connects all the grounds.

[R.G.]

This pedal isn't for me, so I'd like to be able to kill any ground loop that *might* exist.

If you ever find any way to kill any ground loop that *might* exist, please let me know how to do it.    

"Ground loops" is a term for a myriad of hum issues. Killing one doesn't necessarily kill the others. Especially when ABYing to more than one amp, the ABY box needs to make up for the internal AC leakage sins of either or both amplifier and any issue with ground versus neutral versus AC wire drop in the building's walls. As with so many things, you have to know all the ways something can fail before you can make things that fix the failures. If all the ways to get hum were clearly known, someone - me, if no one else!   - would have published a checklist of things to do to fix them all.

There are a couple of "right ways" that nearly always work, but they're not things that can be simply and easily done on typical guitar setups, and can't necessarily be done for a reasonable price. They all involve putting some kind of isolation between the "ground" potentials of the amps and pedals. Transformer-isolating signal into all but one amp usually works, but there are miscellaneous issues with how the secondary signal paths to the amps are themselves grounded, and prevention of hum pickup by the transformers themselves.

Differential/balanced cabling and wiring of all signal lines and using only low impedances is a great help, and that's why most studios use this for all the internal wiring. Standard guitars don't work well with balanced cabling and low impedance cable runs, unfortunately. And many of the side effects of buffering/balanced cable driving make the guitars work not so well with some vintage effects, so that's out. The ILPN (Internet League of Potty-mouth Naysayers) would shriek and scream the process to death if it didn't work perfectly with every effect every make, even if they also pottymouthed the effect.  

Is there any downside to integrating a ground loop switch into the A&B footswitch?

Cost, complexity, and coverage. It makes the thing more expensive and complicated, and may not cover all the hum issues. Any hum will be labeled a "ground loop" whether it is or not.

[wavley]

I was hoping RG would chime in on this, he's much better at explaining these things and thinking of the bigger picture.

Hum is like an onion, you keep peeling away layers of hum and you cry a little with each one.

I was so focused on your ground loop solution that I didn't think of the bigger picture of where hum might sneak in, which believe me I have a complicated rig and tracking down all sources of hum until you've killed them all is PAINFUL and time consuming (YEARS), but it also leads to A LOT of learning about signal flow and sources of hum... and there are a lot of sources.  My RG inspired transformer isolation is only the tip of the iceberg with finding and solving all of my hum issues... fact the loop was not between the amps, they could be connected to each other just fine, it was between my pedalboard and amps.  RG pointed out the advantages and disadvantages of balanced cabling which is why the only place I'm considering balanced cabling is as the last point in my signal chain, just replacing the unbalanced buffers with balanced drivers and the amps still just see the transformers like they always did.  Currently, I don't have a problem with my set up, the balanced solution I want is mainly because I like the length adaptability of XLR cables and there is already a transformer sitting on top of my amp that can be converted from an unbalanced input to a balanced.

[Octavian]

Yeah, I'm not too worried about getting every last bit of hum, just want to provide SOME recourse for hum at least!

Another question - so I've got this ABY going on - and everything I've read talks about the grounds from the two amps - but what about the ground from the pedal (assuming it has lights and whatnot). Should the DC ground just be connected to the LEDs, and not to the enclosure (and the grounds from one of the amp outs can get connected to the enclosure, I guess) or should both amp outs be isolated from the chassis? What should the input jack ground connect to in the former case?

[merlinb]

Should the DC ground just be connected to the LEDs, and not to the enclosure (and the grounds from one of the amp outs can get connected to the enclosure, I guess) or should both amp outs be isolated from the chassis? What should the input jack ground connect to in the former case?

The pedal enclosure should always be connected to audio ground. You can use non-isolated jacks if you want. Assuming the audio circuit is completely passive (just mechanical switches) then LEDs don't need to be connected to ground or the case or the audio circuit in any way.

[Octavian]

The pedal enclosure should always be connected to audio ground. You can use non-isolated jacks if you want. Assuming the audio circuit is completely passive (just mechanical switches) then LEDs don't need to be connected to ground or the case or the audio circuit in any way.

Could you clarify? Are you saying that the DC negative/ground gets connected to the LED negatives (because certainly they DO need to be connected to SOMETHING to work!), and that the chassis gets the ground from the incoming audio cable but the DC negative (ground) and the mono jack ground don't get connected as in other stompbox circuits?

Either way, is it better for the two output jacks to be connected to nothing, to the input ground, or...?

[R.G.]

There are multiple purposed that "ground" serves.

The metal box outside the circuits needs to be attached to ground to drain away any radiated interference that it intercepts.

The input jack sleeve needs to be attached to ground to serve as a reference voltage for the input of the circuit it serves.

The output jack sleeve needs to be attached to ground to serve as a reference voltage for the input of whatever is attached to the jack externally, and as a return path for any current that the remote input needs.

The power supply "ground"  is a return point for all of the "used electricity" that the circuit has used in doing whatever it does. It also serves as a return point for the "used electricity" that the LEDs and whatever else in the box have eaten.

All of these "grounds" must be connected by conductors. But real conductors have resistance, so they generate a voltage by V = I * R proportional to any current flow. Real wires are not "shorts", they are resistors, albeit low value resistors. You get into trouble with grounds by either leaving off a wire to something that should be grounded, or by connecting multiple sources of ground current to the same wire. This mixes the currents, and produces a mixed voltage, and if one of the things connected to the "ground" wire is a reference input with a high gain, you get ugly grounding problems.

You have to know and provide for the currents that flow in every ground wire to do it right. Most of the time, pedals are so small and the currents so low that just tying everything together works OK-ish  - until it doesn't.

Now to your questions.
1. DC negative is not ground. Ground is an abstraction, a theoretical concept involving designating one point in the circuit as the only true, real point that is zero volts. All other voltages are measured with respect to this point.
2. The external box must be connected to ground by some kind of conductor to drain away intercepted EM radiation.
3. The input and output jacks must be connected to ground; these wires are reference grounds, and in a perfect world will carry no currents other than the signal currents.
4. In a box which does not contain any active circuitry, only switches, jacks, and LEDs. The LED wires and power supply do not need to be connected to ground because they do not need to be referenced to the signal ground. In this special case, the LEDs and power supply may be left floating, not connected to the shield ground or signal reference grounds at all, and they will still work. This is not the best thing to do, but will work.
5. In small signal audio work (and this is that), the currents are so small that you often get away with just using metal-sleeve jacks for input and output, and letting the sleeves connect to each other through the metal exterior box. Any circuitry and power supply can be connected to either the input jack sleeve contact or the output sleeve contact. Neither one of these is ideal, but generally you get away with it in pedals.
6. For jacks requiring ground isolation, you must - well, isolate them. The sleeve of the jack should not be directly connected to the reference ground of the box. In practice this often means using a jack with a non-conducting bushing, and then explicitly connecting the sleeve contact of the jack where it needs to go in the box. Such a jack has to have its sleeve contact connect somewhere, and this is often a transformer secondary. It *may* be the output of some isolating amplifier circuitry.
7. No jack sleeve contact may be connected to nothing. They all need to connect to something.

As for what is best; there are an infinite number of grounding setups that work OK. The only way to know ahead of time that your grounding setup will work is to either use star grounding, or to know the currents in each "ground" and provide for those currents to get to the One True Ground, the place of zero volts that you designated as ground.

Grounding practice in large systems is a topic that often leads companies to hire grounding-wizard consultants.

In your case,
- the enclosure must be connected to ground
- using an input jack with a metal sleeve will do this, and will be OK in the small setup you're describing
- if your setup uses DC power for LEDs only, you do not need to attach *either* the positive or negative side of the power supply to the chassis or signal ground. But you can attach one side, but not both, and you can connect it through a resistance. All these work. Be aware that your DC input jack will almost certainly be connected through a daisy chain connector to the ground of other pedals, and worry about what that does.

[merlinb]

Are you saying that the DC negative/ground gets connected to the LED negatives (because certainly they DO need to be connected to SOMETHING to work!), and that the chassis gets the ground from the incoming audio cable but the DC negative (ground) and the mono jack ground don't get connected as in other stompbox circuits?

Either way, is it better for the two output jacks to be connected to nothing, to the input ground, or...?

All three 1/4 inch audio jacks must have their grounds connected together. The metal box/enclosure must also be connected to this ground. In a pinch you can use non-insulating jack sockets, so they make their ground connections through the metal box itself (until they work loose...). You can think of the box as a cube-shaped segment of the shield inside your guitar cable.

The LEDs have nothing to do with audio. They just light up. The incoming DC does not need to be connected to the audio ground or metal box. It just goes to the LEDs and back again. Like RG said, the DC negative supply is not 'ground'.  It can be connected to the box if you want to, but pretty much the only reason you would do this is if you're stuck with a non-insulated DC jack socket.

[Octavian]

Thank you both, especially R.G. for that thorough explanation! The incorrect use of terms in some of the circuits I've been studying (pedals mostly) definitely got me confused, but I think I've got a much clearer understanding of the grounding issues now.

I'd still be curious to know how the 470k resistor and 47nf cap work in the drawing I posted , if either of you wouldn't mind shedding a little more light!

[R.G.]

I'd still be curious to know how the 470k resistor and 47nf cap work in the drawing I posted , if either of you wouldn't mind shedding a little more light!

It's an RF ground. The 470K is a high enough resistance that this is not a "ground" in the normal sense of the word, as small voltages at low frequencies (like 60Hz or 120 Hz) see it as a high resistance, and minimal current flows. The cap however is a "short" at radio frequencies since the cap has a very low impedance to frequencies well above the audio range. So the ground lift switch alternates between a metal-contact ground path and a path that is high resistance at low frequencies, but conducts RF interference well.

[wavley]

Are you saying that the DC negative/ground gets connected to the LED negatives (because certainly they DO need to be connected to SOMETHING to work!), and that the chassis gets the ground from the incoming audio cable but the DC negative (ground) and the mono jack ground don't get connected as in other stompbox circuits?

Either way, is it better for the two output jacks to be connected to nothing, to the input ground, or...?

All three 1/4 inch audio jacks must have their grounds connected together. The metal box/enclosure must also be connected to this ground. In a pinch you can use non-insulating jack sockets, so they make their ground connections through the metal box itself (until they work loose...). You can think of the box as a cube-shaped segment of the shield inside your guitar cable.

The LEDs have nothing to do with audio. They just light up. The incoming DC does not need to be connected to the audio ground or metal box. It just goes to the LEDs and back again. Like RG said, the DC negative supply is not 'ground'.  It can be connected to the box if you want to, but pretty much the only reason you would do this is if you're stuck with a non-insulated DC jack socket.

Let's not confuse the guy, directly connecting all three grounds together defeats the purpose of his ground lift switch, at the very least that jack has to be isolated from the enclosure so that when switched on the ground is referenced through the 470k/47nf combo in the event that the two amps have one of the things we call a ground loop *see RG's post about ground loops

I'd still be curious to know how the 470k resistor and 47nf cap work in the drawing I posted , if either of you wouldn't mind shedding a little more light!

It's an RF ground. The 470K is a high enough resistance that this is not a "ground" in the normal sense of the word, as small voltages at low frequencies (like 60Hz or 120 Hz) see it as a high resistance, and minimal current flows. The cap however is a "short" at radio frequencies since the cap has a very low impedance to frequencies well above the audio range. So the ground lift switch alternates between a metal-contact ground path and a path that is high resistance at low frequencies, but conducts RF interference well.

[Octavian]

I'd still be curious to know how the 470k resistor and 47nf cap work in the drawing I posted , if either of you wouldn't mind shedding a little more light!

It's an RF ground. The 470K is a high enough resistance that this is not a "ground" in the normal sense of the word, as small voltages at low frequencies (like 60Hz or 120 Hz) see it as a high resistance, and minimal current flows. The cap however is a "short" at radio frequencies since the cap has a very low impedance to frequencies well above the audio range. So the ground lift switch alternates between a metal-contact ground path and a path that is high resistance at low frequencies, but conducts RF interference well.

Interesting - thanks for the explanation. I think that might explain some weirdness that was happening - I managed to get together enough stuff to test the pedal. Everything was pretty ok - very slight hiss with the ground lift off. With the ground lift on there was a very loud buzzing. Thinking back, I might have tested it without screwing on the bottom, so all of the radio frequencies bouncing around wouldn't have been stopped by the enclosure and the pedal might actually have acted as an antenna for them. Does that thinking make sense?

[merlinb]

If it helps the OP to visualise what's happening, this is basically what the ground lift switch is doing in the layout he posted. It breaks the connection to one of the output grounds.

[Octavian]

Thanks merlinb - that makes sense.