Different ground potentials causing ground loops

[PhaseShift]

I've heard of different ground potentials at opposite ends of a shielded cable causing ground loops.

What causes these different ground potentials?

How are these ground potentials measured? Are they measured from ground grid to remote earth?

 

[gar]

100713-2232 EST

PhaseShift:

I am not sure what is a definition of a "ground loop". I do not have time to discuss it at the moment.

Different potentials do not determine the existence of a ground loop.


Most practical materials have resistance, some low and some high. If a current flows thru a finite resistance, then there is a voltage drop across the resistance. The earth has currents flowing thru it and resistance, therefore voltage differences.


Take two screwdrivers and insert them in your backyard, about 12 feet apart. Connect an AC DVM across them. You will probably read from 10 MV to several volts depending upon your soil and location.

This experiment created a loop. If there was a high intensity AC magnetic field encompassing the loop, then an error voltage would be induced into you ground voltage measurement. I have not encountered any significant errors from this kind of one turn loop where I have made ground voltage measurements.

.

 

[hurk27]

most common form of what most call as a ground loop is when we hear the nice 60hz hum in a bands loud speakers, but also occurs in video equipment, computer equipment (older network stuff) and a few other things.
the most common cause is because of neutral current on the equipment grounding conductors between two points.
Here is one example:
a band has set up there PA system and amps on a stage in one corner of a club, they install there mixing board across the room, the stage has a panel on it, but is only wired with three conductors (PVC pipe) neutrals and grounding is bonded together at this stage panel, the mixer is on a receptacle from another panel, because the neutral to the stage is also the grounding conductor the amps, are grounded to this current carrying conductor, so is the snake shielding, and since the pathway has current on the grounding, and current takes all paths, some of this current will also be on the snake shielding, which the amplifiers will love to make louder.

this will be also the same setup where audio or video cables run between buildings that have differant services.

the other is just where a neutral has been grounded some where in the grounding path between the equipment, intentional or not?

for almost all audio equipment you can just put in 1:1 isolating transformers for the impedance your using I.E. XLR to XLR, ?" to ?" or even RCA to RCA, When I did sound I always kept these in my bag of tricks, and the same for video, but for the older computer networks see Gars above post, PM him

 

[hurk27]

I should add that if there is current on the grounding, in most cases there is a code violation somewhere, but not always, having two or more services feeding the same building or group of buildings will cause the same effect and be totally code compliant, but in either case it should be checked out to see which applys.

 

[PhaseShift]

I have heard that you do not want to ground both sides of a shielded cable on an analog cable because this could create a ground loop on the shield. I believe this ground loop current on the shield is caused by different voltage potentials at ground on both sides of the shield?

 

[gar]

100714-2253

Change the word sides to ends.

Later I will discuss what you are calling a ground loop.

.

 

[zbang]

I have heard that you do not want to ground both sides of a shielded cable on an analog cable because this could create a ground loop on the shield.

The key word is 'shield'. The shield's purpose is to keep external fields from affecting the surrounded conductors, not to carry any signal-related voltage/current. It can to that with only one end connected. (There are a few situations where the shield is intentionally floated to a certain voltage away from 'ground', but they're very special cases.)

 

[hurk27]

I have heard that you do not want to ground both sides of a shielded cable on an analog cable because this could create a ground loop on the shield. I believe this ground loop current on the shield is caused by different voltage potentials at ground on both sides of the shield?

This can be true if the shield is not part of the intended signal pathway, but in such as audio and video cables in most cases it is, except XLR, which is a balanced pathway.
Leaving a shield open ended when it is a signal pathway for one will make the signal circuit incomplete, (no return path) also it will cause much more unwanted signals induced into the pathway, take a common audio cable on the back of you stereo and pull it out to the point only the center is still connected, it will cause a nice hum, you might even hear your local AM radio station in the mess of all the signals.

to sum this up:
If the shield is part of the circuit carrying the wanted signal, then it must remain bonded on both ends, if the shield is not part of the signal circuit (as in a balanced signal circuit) then you can leave it un connected at one end, but I have always been taught to leave the shield open on the source end away from the receiving equipment.

SO if the shield is part of the signal pathway you only method of removing a ground loop is to either find the cause of the ground loop or use some form of isolation, weather it be 1:1 transformers, or optical coupling as used in data, or running some form of balanced signal pathway such as Ethernet cable system.

 

[PhaseShift]

This can be true if the shield is not part of the intended signal pathway, but in such as audio and video cables in most cases it is, except XLR, which is a balanced pathway.
Leaving a shield open ended when it is a signal pathway for one will make the signal circuit incomplete, (no return path) also it will cause much more unwanted signals induced into the pathway, take a common audio cable on the back of you stereo and pull it out to the point only the center is still connected, it will cause a nice hum, you might even hear your local AM radio station in the mess of all the signals.

to sum this up:
If the shield is part of the circuit carrying the wanted signal, then it must remain bonded on both ends, if the shield is not part of the signal circuit (as in a balanced signal circuit) then you can leave it un connected at one end, but I have always been taught to leave the shield open on the source end away from the receiving equipment.

SO if the shield is part of the signal pathway you only method of removing a ground loop is to either find the cause of the ground loop or use some form of isolation, weather it be 1:1 transformers, or optical coupling as used in data, or running some form of balanced signal pathway such as Ethernet cable system.

The cases I'm most familiar with are analog signals in industrial facilities. Analog signals are shielded to avoid induction from other cables however usually are only shielded at one end to avoid ground loops. I always thought that these ground loops were created by difference in potential at the ends of the shield?

 

[zbang]

I always thought that these ground loops were created by difference in potential at the ends of the shield?

I suppose you could think of it that way- if there isn't a potential difference, there won't be a current flow, and it's the current over the shield that causes the problems (except for electrostatic charges, but I'm ignoring them). Likewise, if you have two electrically connected points separated by more than a few feet, there will be a potential difference between them, albeit very small if the distance is short.

IIRC, the Audio Cyclopedia (Tremaine) has a chapter on signal grounding techniques for large installations, but I don't have my copy with me to check.

A recent audio magazine had a chart of shielding connections and how they reduced noise. I'll see if I can find it.