Noise from pedals in a patchbay when not in signal path

[James Perrett]

If your iMac is providing the ground then it is likely to be via a fairly tortuous path which could easily cause problems. You ought to try a much more direct path to ground.

[Sam Spoons]

I don't know if iMacs have the same characteristic as MBPs but it's probably worth checking that the earth path between the earth pin in the mac and the earth of the chassis or USB lead is low resistance. MBP's PSUs have a resistance to earth of around 1kΩ, (but only if you use the 'optional' extension lead, if you use the included plug adapter there is no connection to earth).

[Echos]

If your iMac is providing the ground then it is likely to be via a fairly tortuous path which could easily cause problems. You ought to try a much more direct path to ground.

I can lift one of the speaker grounds, but then I have two paths to ground, one via the speaker and one via the iMac which is breaking the single path to ground rule isn't it? Unless I buy a USB isolator which seem to be incredibly expensive for a decent one. At least that way I would be able to provide a single path to ground.

Do you have any other suggestions?

[Echos]

A lot is going to come down to specifics of exactly how you've got your patchbay wired and the path the cables are taking, which is very hard to describe. I don't know if you could draw it out for us at all and post a photo (you have to host elsewhere and link to it).

Here you go: I've drawn up the connections. (Sorry. Adding images inline appears to be broken).

[Echos]

I don't know if iMacs have the same characteristic as MBPs but it's probably worth checking that the earth path between the earth pin in the mac and the earth of the chassis or USB lead is low resistance. MBP's PSUs have a resistance to earth of around 1kΩ, (but only if you use the 'optional' extension lead, if you use the included plug adapter there is no connection to earth).

Thanks for the suggestion, but I've checked the Mac's plug and it's definitely got a metal earth pin. Can't believe they don't include an earth on the laptop power supply. That's terrible.

[Hugh Robjohns]

I can lift one of the speaker grounds, but then I have two paths to ground...

There's nothing wrong with having 100 ground paths, provided any circulating ground currents don't get into the audio ground reference anywhere.

So it's really all about the detail of the engineering of the specific kit and how it's connected. These things can't usually be fixed with a simplistic approach.

It's late here now so I'll take a look at you diagram tomorrow to see if anything obvious is apparent.

[James Perrett]

If your iMac is providing the ground then it is likely to be via a fairly tortuous path which could easily cause problems. You ought to try a much more direct path to ground.

I can lift one of the speaker grounds, but then I have two paths to ground, one via the speaker and one via the iMac which is breaking the single path to ground rule isn't it?

There is no hard and fast single path to ground rule. It is a convenient suggestion when people encounter ground loops but in a well designed system you can have as many ground connections as you need.

The problems arise when you include a poorly designed piece of kit that relies on there being no currents flowing in the ground lines of the signal cables. This kit will be susceptible to ground loops. If everything is properly grounded through its power supply then there should be fewer problems because any ground currents should flow through the mains connectors rather than through the signal cables.

The ground connection on the Mac is likely to go through various components which may not affect the DC resistance but may increase the impedance at audio frequencies. So the ground on the Mac may actually make things worse. (NB I don't own a Mac these days so I'm making this assumption based on my experience with other computers.)

Hang on - I've just seen your diagram and noticed a power conditioner. Take it out - they're not normally needed and can cause more problems than they purport to fix. Take out the Radial box too - it isn't needed and is only going to confuse things.

[Sam Spoons]

I don't know if iMacs have the same characteristic as MBPs but it's probably worth checking that the earth path between the earth pin in the mac and the earth of the chassis or USB lead is low resistance. MBP's PSUs have a resistance to earth of around 1kΩ, (but only if you use the 'optional' extension lead, if you use the included plug adapter there is no connection to earth).

Thanks for the suggestion, but I've checked the Mac's plug and it's definitely got a metal earth pin. Can't believe they don't include an earth on the laptop power supply. That's terrible.

Lots of laptops don't have an earth connection*, the MBP does but only if you use the 'extension cable' and even then that is a high resistance path. Both the adapter and the extension lead have a metal earth pin so it's existence won't tell you if the it is actually connected or whether it is low resistance, you'll need to get the multimeter out I'm afraid.

* Laptops invariably require a low voltage DC supply and, as such, don't require a safety earth.The PSU is usually a class II device (double insulated) so does not require a safety earth either, if they are class I they will have a three pin connector including a safety earth but may or may not connect that through to the laptop.

[ef37a]

I would tie all the pedal jack sleeves together but not return that link to another ground. That way every pedal is grounded whether on or off.
I assume the 'all off' condition is not required?

Dave.

[Echos]

If your iMac is providing the ground then it is likely to be via a fairly tortuous path which could easily cause problems. You ought to try a much more direct path to ground.

I can lift one of the speaker grounds, but then I have two paths to ground, one via the speaker and one via the iMac which is breaking the single path to ground rule isn't it?

There is no hard and fast single path to ground rule. It is a convenient suggestion when people encounter ground loops but in a well designed system you can have as many ground connections as you need.

The problems arise when you include a poorly designed piece of kit that relies on there being no currents flowing in the ground lines of the signal cables. This kit will be susceptible to ground loops. If everything is properly grounded through its power supply then there should be fewer problems because any ground currents should flow through the mains connectors rather than through the signal cables.

The ground connection on the Mac is likely to go through various components which may not affect the DC resistance but may increase the impedance at audio frequencies. So the ground on the Mac may actually make things worse. (NB I don't own a Mac these days so I'm making this assumption based on my experience with other computers.)

Hang on - I've just seen your diagram and noticed a power conditioner. Take it out - they're not normally needed and can cause more problems than they purport to fix. Take out the Radial box too - it isn't needed and is only going to confuse things.

Thanks. The problem with pedals is they all seem to share ground between audio and power cables. I picked five of my pedals, a Boss, a Strymon, an EHX, a Wampler and an Eventide. I took the pedal and plugged a power cable in (with the other end of the power cable connected to nothing), then plugged a jack into either the in or out jack and again connected it to nothing, I then did a continuity test with a multimeter and in all cases there was a clear signal path between the power ground and the jack ground. This bring the case it's hard to see any way of eliminating ground loops because as soon as there is a connection between two pedals via a TR cable, there is a ground loop running from the power supply to the first pedal via the power cable, along the TR cable between the two pedals and back down the other pedal's power cable. This is surely multiplied for every pedal added to the signal chain. It would be a different matter if the signal path was isolated from the power supply, but if there is an inherent ground route via all power supply cables and via the audio cables that run between them I can't see any way of eliminating it.

The reason I have the Radial before the speakers is to ensure there aren't multiple paths to ground. If I remove it there will be three - via each speaker and via the iMac. Although given what I've said above about the pedals, there are already many paths to ground via each pedal's power cable. All connected by the signal path.

[James Perrett]

I think you are far too fixated on eliminating multiple paths to ground when you may find that the more paths to ground there are, the quieter the system.

The main thing to avoid is to have half of your setup plugged into one wall socket and the other half plugged into a wall socket on the other side of the room. Keeping everything plugged into the one wall socket should be fine with no extra precautions.

[Sam Spoons]

Have you tried it though? Hardly ant wall wart PSUs have a safety earth...

[Echos]

I think you are far too fixated on eliminating multiple paths to ground when you may find that the more paths to ground there are, the quieter the system.

The main thing to avoid is to have half of your setup plugged into one wall socket and the other half plugged into a wall socket on the other side of the room. Keeping everything plugged into the one wall socket should be fine with no extra precautions.

I probably am. But I'm really struggling to conceptualise the problem. As I said in my original post, it's taken me a lot of time to get this far, and that was based on what seemed to common advice to eliminate ground loops by eliminating multiple paths to ground. I'm honestly completely confused now. Thanks for trying to help.

[forumuser931182]

I think I would be tempted to try a simple non insulated patch bay and make that the main audio earth point. If you have a preamp or mixer available you could also bypass the audio interface and Mac and feed the patchbay straight into the studio monitors as a test.

[Hugh Robjohns]

... in all cases there was a clear signal path between the power ground and the jack ground. This bring the case it's hard to see any way of eliminating ground loops ...

This is an inherent problem with unbalanced connections where the transferred signal reference is also the power supply ground. As a result, any power supply noise is automatically injected into the audio signal.

I took a quick look at your diagram and I see you have some digital fx pedals sharing a power source.

I've come across problems before where digital noise from one pedal gets passed to others via shared power supplies.

It might be worth experimenting with the original individual supplies (if you have them) or with batteries (where possible) to see if that helps.

My other thought is that -- as someone suggested above -- your application might actually work better with a patchbay where all the grounds are tied together. Some have an internal option for that, but I'm not sure if the Samson does.

[Wonks]

The Voodoo Labs 3 power supply should have individually galvanically isolated outputs, so there shouldn't be any connection between the 9v negative and the PSU's mains ground pin. If there is, it's faulty.

How are the pedals mounted? Within a rack on a rack shelf?? If the PSU casing is touching the same metal rack and the metal case of one of more of the connected FX units are also directly in contact with the metal shelf, then that could create a path to the PSU power ground pin. I know some people like to screw the FX units down rather than use Velcro or 3M Dual Lock, but if the PSU and the FX casings are connected to the same metalwork, then you've created a ground loop that you definitely want to avoid as you've negated the PSU's isolated output feature.

If the above is the case, then it's probably easiest to mount the PSU so it's insulated from the metalwork.

For the non-patched FX units, you could try unplugging any without true bypass, and put the rest in bypass. This (I hope) should stop noise generated within the pedal from being sent up the connection leads to the patchbay and make things quieter.

[Echos]

The Voodoo Labs 3 power supply should have individually galvanically isolated outputs, so there shouldn't be any connection between the 9v negative and the PSU's mains ground pin. If there is, it's faulty.

Thanks for drawing attention to this. I tested with a multimeter and there doesn't appear to be any continuity between any of the wall plug pins and either center or sleeve on the 9V power cables plugged into the Pedal Power 3 Plus. So this means there is no route to ground back down the power cable from the pedals doesn't it? so they'll have to ground via either the rack case, or the audio signal path?

How are the pedals mounted? Within a rack on a rack shelf?

The pedals are sitting on a wooden frame. Picture here

The power supply is currently completely separate and on a wooden shelf, only connected to the pedals via the power cables. The floor between the rack and shelves is wooden.

[Wonks]

The path back to ground for the pedals (and the guitar) should ideally just be via the signal cable.

Your wooden frame for the pedals is fine.

[ef37a]

I did wonder Wonks if the pedal power supply was isolated and it seems it is. Paradoxically this could be the reason for the noise? It seems to me to be a lack of a ground that is the problem here, not a lack?

Dave.

[James Perrett]

It would be interesting to know whether there is any capacitive coupling to ground through the pedal power supply. Or any capacitive coupling to the other pedal power supplies. If there is then again, it could cause problems. It might have been better if the power supply had offered ground lift as an option rather than mandating that all grounds should be lifted.

[Echos]

The path back to ground for the pedals (and the guitar) should ideally just be via the signal cable.

Your wooden frame for the pedals is fine.

Thanks. I've bitten the bullet and ordered:

- a USB isolator
- a Neutrik NYS-SPP patchbay that has configurable grounds.

My plan is to:

- Isolate the USB connection between the Audient iD44 and my iMac
- Only lift the ground on the right speaker connection with th Audient iD44
- Configure the new patchbay so that all the grounds are bussed

Theoretically this will provide a unified ground for all pedals so long as they are plugged in to the patchbay, but regardless of whether they are patched into the audio signal path via the left and right TS cables, to the Audient iD44 and to the left speaker and via its power cable.

I also plan to try powering digital and analogue pedals with different power supplies.

If none of that works I give up.

[Wonks]

You want to have only one ground path, but the audio interface needs to be grounded somehow. So if you sue the USB isolator, you still need one ground path. Check you have one with a multimeter (if your leads are long enough).

The Samson S Patch manual indicates that the sleeve connections are connected to ground. I don't know if this means that all the sleeves are connected together, or only for each set of inputs and outputs.

If they are all connected together, and also to the rack unit metalwork, it may not be a good idea to have the ground running off to the water pipe (though I expect you had the noise before you did that, so it's probably not the main issue).

The real issue is that you don't have any excess noise when only the used pedals are plugged in to the patch bay, so the presence of the cables in the patch bay from units that aren't in the signal path seems to be introducing noise into the signal path. Which to me indicates that the lack of ground reference of those unconnected pedals allows noise form the pedals to be transmitted from those cables.

What happens if you leave the unconnected FX cables in the patchbay, but remove the power from those FX units? We know that there is no change in any ground connections when you do that, but if it is noise from the unpatched pedals that's the culprit, then that should stop if they aren't powered up.

If that's the cause, then you may have to make up some jacks that are linked back to a ground connection, so at least the output of each unused pedal has a sleeve connection to ground, to stop the cables acting as transmitters. You could link the T and S in these jacks together to make doubly sure.

[ef37a]

The path back to ground for the pedals (and the guitar) should ideally just be via the signal cable.

Your wooden frame for the pedals is fine.

Thanks. I've bitten the bullet and ordered:

- a USB isolator
- a Neutrik NYS-SPP patchbay that has configurable grounds.

My plan is to:

- Isolate the USB connection between the Audient iD44 and my iMac
- Only lift the ground on the right speaker connection with th Audient iD44
- Configure the new patchbay so that all the grounds are bussed

Theoretically this will provide a unified ground for all pedals so long as they are plugged in to the patchbay, but regardless of whether they are patched into the audio signal path via the left and right TS cables, to the Audient iD44 and to the left speaker and via its power cable.

I also plan to try powering digital and analogue pedals with different power supplies.

If none of that works I give up.

Am I chopped liver here? It will take you but a few minutes to bond all the pedal jack cases together (but no further. )

Dave.

[Echos]

You want to have only one ground path, but the audio interface needs to be grounded somehow. So if you sue the USB isolator, you still need one ground path. Check you have one with a multimeter (if your leads are long enough).

The interface would be grounded through the TRS cable running to the left speaker wouldn't it?

The Samson S Patch manual indicates that the sleeve connections are connected to ground. I don't know if this means that all the sleeves are connected together, or only for each set of inputs and outputs.

I've tested with a multimeter. There is no shared ground between jacks (other than front-to-back pairs). So each jack doesn't share a ground with the jack's next to it or above/below it. Unless a cable joins the hem together. There is also no connection between jack grounds and the chassis of the patchbay.

The real issue is that you don't have any excess noise when only the used pedals are plugged in to the patch bay, so the presence of the cables in the patch bay from units that aren't in the signal path seems to be introducing noise into the signal path. Which to me indicates that the lack of ground reference of those unconnected pedals allows noise form the pedals to be transmitted from those cables.

Agreed. This is why I am think a patchbay with bussed grounds might be the answer. That way all pedals connected to the patchbay will share the bussed ground, regardless of whether they are in the audio path. Even if a pedal is only connected to the back and isn't patched in at the front, it's still on the same bus as anything patched in at the front.

What happens if you leave the unconnected FX cables in the patchbay, but remove the power from those FX units? We know that there is no change in any ground connections when you do that, but if it is noise from the unpatched pedals that's the culprit, then that should stop if they aren't powered up.

The noise stops. It clearly reduces slightly for each unplugged pedal until the last pedal is unplugged, at which point it is gone.

If that's the cause, then you may have to make up some jacks that are linked back to a ground connection, so at least the output of each unused pedal has a sleeve connection to ground, to stop the cables acting as transmitters. You could link the T and S in these jacks together to make doubly sure.

Wouldn't a bussed ground shared amongst all jack's achieve the same thing?

[James Perrett]

You want to have only one ground path, but the audio interface needs to be grounded somehow.

I think we need to get away from this one single ground path idea here. Echos is obsessed by it and it is doing them no favours. Everything in the system needs a good solid ground.

Echos also needs to lose all the power conditioning and isolating gadgets. The system is far more complex than it needs to be thanks to this obsession with a single ground.

Once the system has been stripped back to the basics we can then see if there are any problems.

[Echos]

You want to have only one ground path, but the audio interface needs to be grounded somehow.

I think we need to get away from this one single ground path idea here. Echos is obsessed by it and it is doing them no favours. Everything in the system needs a good solid ground.

Echos also needs to lose all the power conditioning and isolating gadgets. The system is far more complex than it needs to be thanks to this obsession with a single ground.

Once the system has been stripped back to the basics we can then see if there are any problems.

I tried replacing the power conditioning strip with a standard strip and there is a lot more noise (which is why I got the power conditioning trip in the first place). Removing the Radial doesn't make any difference to the noise so I'll leave it out for the moment.

[Wonks]

You want to have only one ground path, but the audio interface needs to be grounded somehow.

I think we need to get away from this one single ground path idea here. Echos is obsessed by it and it is doing them no favours. Everything in the system needs a good solid ground.

Echos also needs to lose all the power conditioning and isolating gadgets. The system is far more complex than it needs to be thanks to this obsession with a single ground.

Once the system has been stripped back to the basics we can then see if there are any problems.

I disagree here. The whole point of spending £200 on a PSU with isolated outputs is to get away from having more than one ground path between the pedals, otherwise you might as well use a £10 PSU and daisy chain the lot. There are some instances where multiple grounds work well, but not for guitar pedals in my experience, especially when you've got quite a few digital pedals.

And remember that there is no issue with the guitar signal chain if no unused cables are patched in to the patch bay. Unless there is some common grounding going on in the patch bay that is linking all the jack sleeves together, there there should be no physical/electrical connection between the unpatched FX and the patched FX.

[Hugh Robjohns]

I'm going back to the beginning because in threads like these it's so easy to get unintentionally misleading information and go down blind rabbit holes in ever decreasing circles.

I'm confident that any noise I hear is coming from the character of the pedals (fuzz, compressor etc) rather than ground loop etc, but this only holds true as long as pedals are in the signal path. If I have pedals that are patched in to the back, powered, but not patched into the signal path (via the front jacks), I get noise.

This is almost certainly because the power supply has isolated and 'floating' outputs, and some of your pedals are digital. The combination means that all manner of interference is being radiated around the unbalanced audio cabling and generating unwanted 'noise'.

This noise increases for each pedal that is powered and not in the signal path. If I patch one of those pedals into the signal path, the noise is reduced. Unpatch it and the noise returns.

When patched the pedals pick up a solid ground from the signal path and that provides effective screening against the radiated interference.

So a patchpanel that allows all connected grounds to be tied to a solid earth would probably help, a lot. Most panels keep each circuit ground separate to reduce the risk of unwanted ground loops... and while that makes sense in many installations, it's not actually helpful in yours.

1. Ferrite chokes on usb cable from computer to audio interface.

No problem with that, as long as it doesn't effect USB reliability.

2. Radial Stagebug SB-6 isolator in between audio interface and speakers, with both grounds lifted, so that the (powered) audio interface becomes the only root to ground for the cables entering the interface (the power supply for the pedals is isolated).

Looking at your schematic, there is no ground for the interface in this arrangement. The guitar isn't grounded. The effects pedals aren't grounded. The interface could be grounded via the USB and computer... except that the computer probably isn't grounded either. (I can't remember what you've done about the iMac power supply).

The speakers possibly are grounded, and since you're using balanced connections between the interface and speakers (which reduces the risk of ground noise getting into the audio signal path), I'd try removing the SB6 to ensure the interface (and iMac) are grounded via that route alone.

3. Swapped out my MXR power supply with a Voodoo Labs Pedal Power 3 Plus which is fully isolated and seemed less noisy then the MXR.

Fair enough... but be aware that its isolated outputs mean no grounding via the power line for any effects pedal. That is probably helpful, but only if the fx pedals are grounded via another route!

4. Fully isolated the patchbay from the rack using nylon screw collars (and confirmed this with a multimeter)

Probably not helpful in this situation.

Everything runs into the same power-conditioned strip and into a wall socket (240v, 3-pin as I'm in the UK).

The 'power conditioning' of these things is negligible and can actually make the grounding impedance worse rather than better. However, they do have the benefit of encouraging a star power distribution which is a good thing as it minimises the risk of ground loops.

So in the current configuration, the only route to ground for the pedals is through the signal path, into the audio interface and out through its power cable ground.

As per your diagram, there was no ground path as the interface psu was a double-insulated type with a plastic earth pin. There was no ground in the audio signal path anywhere!

My advice would be to use a patchbay in which the screens of all connected inputs can be tied together and, ultimately, to a solid earthing ground. That should keep your pedals quiet and stop them radiating noise when not in use.

I'd also remove the stagebug so that the interface acquires a ground from the mains-powered speakers (assuming they are themselves class-1 and grounded types). That ground will be passed to the iMac via the USB connection. The chokes may well be unnecessary, but you can experiment with removing them once everything else is sorted.

There is still a possibility that the digital pedals are generating interference which is spreading via the shared power supply (even though they are isolated outputs -- the isolation might not be as good as claimed!). If, after using a grounding patch bay, the noise remains, I would try using individual power supplies on the digital pedals and keep power leads as far apart as possible. Chokes might help on those too!

Hope that helps.

Resolving this kind of problem with unbalanced cabling and multiple sources and destinations can be very tiresome and time consuming, and made even harder by non-real-time remote diagnosis and where the tests and results aren't may inadvertently not be exactly as requested, giving unintentionally misleading results.

But these things can usually be sorted out eventually... so stick with it!

[Wonks]

The downside of grounding all the patch bay grounds (which may not be that easy to do) is that the connected pedals may now become noisier. pedal noise through ground loops is fickle and some combinations may be fine and others really noisy.

Are you really getting any sonic benefit from disconnecting the unused pedals rather than simply turning the pedals on and off as required? With either a clean buffer as the first pedal or a simple boost pedal (say an Xotix EP) 'always on' pedal that drives the rest of the pedal patches, then the extra length of the patch cables won't be loading the guitar for those pedals in bypass.

You've already got a buffer pedal before the stereo input effects, though I'm not sure if it provides much benefit as a signal splitter, so you could always try that as the first pedal in the line.

If you do tie the patch bay grounds together, then I'd make up some special patch leads with the signal cable connected to the tip at both ends, but the shield only connected to the sleeve at one end - doesn't matter which - so the shield still provides screening of the signal cable. I'd also just tie the screens together on just the input or output connections, to minimise the number of ground paths.

[resistorman]

You really should get rid of the patchbay. In my experience, when I try to set up all my gear for maximum flexibility I spend most of my time obsessively arranging the tools and not making music. Lately, I set up a simplified rig to work on a song and dive deep into that setup, it's surprising how much more you can get from your tools when you focus on fewer.