DC potential between two circuit grounds?

[lion]

I've just found that I can measure a DC potential between ground/chassis of 2 different pedals.
Intially noticed it when I had a couple of my own builds connected. Stumped by what I found, I went on and checked a couple of pro/commercial pedals - same thing (to my relief).

Hooking up a number of random circuits I measure a range of  2 - 8 millivolts, depending on between which GND's I measure - unit A-B, unit B-C, unit D-A etc. Some of the units are fed from the same supply, others have their own/internal supply.

I'm confused. Why is this - and is it anything to be concerned about?

Erik

[GibsonGM]

Hmmm, yeah, I can see where you would be like "WTH?".   

I believe you see this because, due to resistance in the ground path, we DO actually have some VERY small currents flowing there....in theory, no, but in reality, yes.   This is why we try to do as clean a job with the ground bus as we can (star grounding).   You could picture this as a very low-value resistance in series with your ground.   

The mV you are seeing aren't significant, and our meters also don't resolve too well at THAT level of precision (unless you paid $2,000 for it?  LOL)

When you get into several 100ths of volts, that would seem to be when you might explore some more....  .3V,  .5V could indicate something like a leaky cap (always possible,  and also a disclaimer to what I wrote above, but not 'an issue' yet if just a few mV).

I wouldn't be concerned about it unless there's something else with it, like an increase in the noise floor or something.  My 2 cents. 

Edit:  Other things - if powered by more than 1 battery, you also now have multiple power supply internal resistances affecting things. But I think most of the (small) difference on the ground would simply be due to resistance.

[EBK]

My first thought is: Is your meter showing 0 mV for any of your measurements, or are they all measuring 2-8mV?

[lion]

Thanks gentlemen, much appreciated.

I believe you see this because, due to resistance in the ground path, we DO actually have some VERY small currents flowing there....in theory, no, but in reality, yes.   This is why we try to do as clean a job with the ground bus as we can (star grounding).   You could picture this as a very low-value resistance in series with your ground.   

Yes, OK. So this is a real life example of issues related to the very complex subject of GROUNDING.

My first thought is: Is your meter showing 0 mV for any of your measurements, or are they all measuring 2-8mV?

My meter is a low(er) budget DMM, so I guess you are right that the precision may be so so at that low mV levels. No, the meter is not showing 0 mV for any of my GND-GND measurements - but I read 1.2, 1.9, 2.6 and 4.6 mV etc.

Interestingly I found that removing one particular pedal from the chain lowers all mentioned DC readings by 50+ % - leaving more of them below 1mV. Surely insignificant.

I'll look into what's up with the pedal causing higher readings all over. Any clues as to what to look for?

Erik

[antonis]

Despite of the fact that any individual pedal has it's own grounding design which is added to (or subtracted from) the loop chain, you firstly make sure that your DMM shows "absolute" zero on DC mV reading with probes on air..

If it takes more than a few seconds to stabilize (or it never does it..) your readings in the scale of a few mV are worthless..

[anotherjim]

It's also worth measuring for AC volts. DC can be a result of AC being partly rectified due to oxidised  contact in the path.
https://en.wikipedia.org/wiki/Metal_rectifier
This is AC supply - not signal - in origin, and is of course the hum heard in a ground loop problem.
The AC can be higher voltage and easier to measure.

[lion]

With probes on air - took 3-4 secs to stabilize at 00.0 mV.

Measuring for AC my meter showed zero all over.

Erik

[GibsonGM]

Not that you can accurately measure it with our DIY-level meters (or even detect it at all), there likely IS some junk in the ground path.  And it's likely not causing any problems like noise or anything in your circuits.   MANY things play into why there could be some (very small) currents flowing there, most beyond our ability to truly control. 

So, what do you 'do about it'?   Just follow good layout and ground practice and enjoy your builds!   

[PRR]

Your ground wires are not superconductors.

Say 1 Ohm resistance counting the wire and both plugs. Not unlikely. Say 4mA or 8mA current from shared supply, or just general slosh. That's 4mV-8mV.

The "real" problem is your modern meter is too good. Back in the days of 1.5V meters we never knew about mV.

Similar case recently. Friend found dozens of volts on the idle wire of a 3-way 120V stair-light switch. Back in passive-meter (or buzzer!) days the meter would have loaded it down to insignificant. With a modern meter he was reading about 80% of the 60VAC electric field around the un-connected wire.

[EBK]

The "real" problem is your modern meter is too good. Back in the days of 1.5V meters we never knew about mV.

Similar case recently. Friend found dozens of volts on the idle wire of a 3-way 120V stair-light switch. Back in passive-meter (or buzzer!) days the meter would have loaded it down to insignificant. With a modern meter he was reading about 80% of the 60VAC electric field around the un-connected wire.

But how do the vintage multimeters "sound", and are germanium reproductions available? 

[Phoenix]

Similar case recently. Friend found dozens of volts on the idle wire of a 3-way 120V stair-light switch. Back in passive-meter (or buzzer!) days the meter would have loaded it down to insignificant. With a modern meter he was reading about 80% of the 60VAC electric field around the un-connected wire.

Ah! But now our fancy expensive modern electrician-grade multimeters have special "features" like low impedance mode to resolve ghost voltage problems. And we have the privilege of paying extra for it...
Full circle...

[GibsonGM]

The "real" problem is your modern meter is too good. Back in the days of 1.5V meters we never knew about mV.

Similar case recently. Friend found dozens of volts on the idle wire of a 3-way 120V stair-light switch. Back in passive-meter (or buzzer!) days the meter would have loaded it down to insignificant. With a modern meter he was reading about 80% of the 60VAC electric field around the un-connected wire.

Yeah, that can happen, umm, LOL...sometimes the only way to know is to load the circuit with a reasonable amount of resistance, maybe 10k (to assure very little current could flow, if it were actually there), turn the power back on, and test again...and see "0V", as that ghost voltage sagged right down even on that tiny load.   Actually, I rarely see ZERO volts...I see mV, which is the same thing per what we've been discussing.

No good playing with R's on a ground line, but knowing  how GOOD we have it with modern DMM's is definitely part of this!  Like Paul said - my rocket scientist father would never have seen that back in 1966, while working on guidance for moon shots...

[sajy_ho]

It's kind of like the situation that your amp is not connected to the Earth ground and everytime you touch the strings you hear a click, because earth is actually a CT for mains voltage and when your amp ground has a potential dc offset from the earth, your body connects the earth to your amp ground and that causes a discharge and click/pop sound...

[anotherjim]

There were problems when tech's had their old meters replaced by fancy DMM's. Especially in fields like telecoms and railway where there are long cable runs. Stray voltage checks for safety (don't want false feeds especially in railway stuff) are done with both ends of a cable disconnected. A lot of non-faults were found with the DMM. Solution was a specially made shunt resistor. Similar in form to the plug in current shunts with stackable jacks, but a 150k resistor in it.


A cheap DMM, you could probably fit a small push button that places a resistor across the V and Com jacks of the meter?

[lion]

So, what do you 'do about it'?   Just follow good layout and ground practice and enjoy your builds! 

I do - and very much so, but for me it's also a learning process. Everytime I run into a new problem/challenge/"discovery" I learn something new - often with the help of members of this board willing to share their knowledge and expertice. Which I appreciate very much!

Story is I stumbled into this particular issue when I was trying to locate a minor noise problem I encounter with a combination of 4 units - preamp/delay/compressor/amp sim - going to a power amp. The noise is a wavy high pitch whine - not a MAJOR problem, but annoying nevertheless. When moving the units around, in a first attemp to envestigate the problem, one enclosure happend to touch another and I heard a relatively loud crack/pop out of the amp.

Your ground wires are not superconductors.

Say 1 Ohm resistance counting the wire and both plugs. Not unlikely. Say 4mA or 8mA current from shared supply, or just general slosh. That's 4mV-8mV.

Yes, I see – but.......
Discovering that one particular unit happens to increase the ground DC potential (small as it is) between all 4 units, I have looked closer at the offending pedal. To my surprise, it seems that in that pedal the DC negative is lifted above ground by 235 ohms???

Pedal in question is a (DEAD CHEAP) Joyo AC Tone - but it is proved to be a near 95% identical clone of the Tech21 Liverpool. And - sure enough - the Liverpool schem shows the same (to me) oddity:



More learning needed! Looks to me that if the Tech21 is fed from same supply as other pedals (like in my case), good grounding practise seems to be sacrified. I'd guess there's a good reason - only I can't see it.

Erik

[anotherjim]

Paradoxically, the R in the 0v feed is to reduce ground hum. It makes the audio cable screen the preferred signal return path between devices instead of the power supply loop. Path of least resistance stuff. 
235R seems on the high side to me but that pedal maybe takes very little current.

Long ago, the first time I had 2 externally powered pedals (a chorus and DS-1 - yeh I was Kurt before there was Kurt!), I got terrible ground hum despite the 9v supply was a non-earthed double-insulated transformer job.  Urgent solution was to cut the 0v wire off the DS-1 DC jack. That fixed it.
So I removed the 0v feed, the DS-1 still got power return from the path via the Chorus chassis and audio cable screens.
Admittedly, I struggle to rationalise the benefit of doing this. Isn't it bad to send power return current mixed with signal return? In practice, it seems better that way than allow signal return current to go via the power supply.

Since then, if I make a multi output DC splitter, I put 10R in series with all the 0v feeds. Not had any problems doing that.

[PRR]

> in that pedal the DC negative is lifted above ground by 235 ohms???

That seems odd today. Perhaps R.G. (who diagnoses such problems as a side-job with a power supply company) will give us some of his insight.

My thought is: this pedal was NOT intended for "shared supply" operation. Alone on its own isolated output, the 235r is just a minor drop; or rather the supply sits "elevated" but nobody cares because it floats alone. On a SHARED supply things is different.

The TWO * 470r is double odd. This suggests they expected large currents. Even if we assume two 1/8W parts, they planned for like 7V! Which could be their unregulated wart soaring to ~~17V in some situation?

> makes the audio cable screen the preferred signal return path

Yeahbut.... 10 Ohms would be ample for that.

[PRR]

This is about Wall Power circuits, not pedal-stuff.

> load the circuit with a reasonable amount of resistance, maybe 10k

Yes, I have heard that value.

> problems when techs had their old meters replaced by fancy DMM's. ...telecoms and railway where there are long cable runs. Stray voltage checks for safety ... A lot of non-faults were found with the DMM. Solution was a specially made shunt resistor ... a 150k resistor in it.

Also a good value.

Either will cause a "significant" drop of "ghost" voltage. Both are fair approximations of "human body". (~~100+K dry, ~~1K when working wet at the sump-pump.) More realistic than the 10Meg of a DMM. If the voltage does NOT drop much with loads like these, the "ghost" has enough current and power to bite you bad.

They make "electrician's meters" with such loading built-in, by default or at button-press.

The 150K is probably sized so it "could" be left across a live line long-term without needing a huge resistor. Note that if the resistor is _ever_ burned-out, it will give false negatives! (Say "safe" when really dangerous.) If 10K is left on 125V it is 1.56 Watts, a fairly large resistor; at 240V it is near 6 Watts and a real lump. 150K 1/2W can be left across 270V for days at a time. The 10K in "120V" work is "OK" at 1/2W if the user understands not to leave it in with >70V across it. In this case the "ghost voltage" was just ~~50V DMM-load and dropped way off with the 10K.

[lion]

I learned something new again - and probably also have a couple of things needing further study and/or some experimenting.

Thanks.

Erik