Ground loops

[lile001]

OK, a dumb question. I know that ground loops, or grounding an electrical system at more than one point, cause trouble. It's a good general rule that applies to 480 volts or 24 volts. I've fixed electronic systems by identifying a ground loop and removing it. I can understand how the code requires a building neutral to be connected to ground at only one point, that's fine. I have busted the I.T. department for wanting their own ground rod at the Server closet, not connected to the building ground and erroneously thinking this would be less noisy.

I was trying to explain WHY these things cause trouble the other day, and came up short. (pun intended) Why do we so scrupulously try to avoid ground loops? What is the mechanism involved? What sort of trouble do they cause? Do they cause extra trouble when lightning is in the system?

 

[petersonra]

Re: Ground loops

if they conenct the ground rod to the EGC, I don't have an issue with what they are doing. I don't think it will do anything useful, but it won't hurt anything.

If they want to connect it to the neutral at that point, IF they have an SDS, I have no issue with that either (although code would seem to require a connection to building steel as well).

Most of the time a UPS is an SDS anyway, so there is a neutral-gnd bond at that point.

 

[tom baker]

Re: Ground loops

Research shows an isolated ground rod intruduces more noise into the isolated equipment. The reason why is there is voltage difference between the two different ground systems and this causes a current. If the system does not work tied into the building electrical ground system then its an equipment problem

 

[lile001]

Re: Ground loops

OK, isolated grounds are bad on data processing equipment because the ground may be at a different potential than the building ground, resulting in a current, thus a voltage where you don't want one, maybe noise where it isn't wanted. Good explanation.

How about in an electrical system in general? Why (besides its a code violation) shouldn't a guy drive a ground rod at either end of a 100 foot long building, at the service and on the end of a cable to a big motor load, say? What unwanted effects does this cause?

 

[petersonra]

Re: Ground loops

Is this guy bonding the two rods together?

 

[physis]

Re: Ground loops

If the guestion is what problems are caused when the grounded conductor is grounded in more than one location the answer is that the ground path that is only intended to carry fault current for a short time becomes a parallel path for neutral current. The grounding path can be made from things like conduit couplers and connectors who's resistance might not be as low as the grounded conductor connections. Points of higher resistance will heat up when current flows through them. If there's enough current and the right amount of resistance you can have a fire hazard.

And if the grounded conductor ever becomes disconnected it might not even be known and you would have shock hazards and fire hazards.

 

[physis]

Re: Ground loops

That wasn't the question was it.

If the question is why can't you jsut drive a ground rod at the service and one at the equipment and forget about the grounding conductor, the answer is that the earth doesn't have low enough impedance to return fault current to the panel and the overcurrent device wont open. If there's a ground fault everything will stay energized and there will be shock hazards.

 

[lile001]

Re: Ground loops

Of course this is hypothetical, But let's say you drove your ground rod at the end of a long feeder, and then bonded that ground rod back to the ground rod at the service entrance via the normal equipment grounding conductor, creating a loop. This is specifically not allowed, but what kind of problems does it cause? I'd like to be able to explain the theory behind this to a novice.

 

[iwire]

Re: Ground loops

Originally posted by lile001:
Of course this is hypothetical, But let's say you drove your ground rod at the end of a long feeder, and then bonded that ground rod back to the ground rod at the service entrance via the normal equipment grounding conductor, creating a loop. This is specifically not allowed, but what kind of problems does it cause? I'd like to be able to explain the theory behind this to a novice.

Actually it is allowed and if the feeder runs to another structure it is required by 250.32(A)

 

[physis]

Re: Ground loops

Tom is right. Using sort of slang terms, you'll be bleeding off any voltage difference at that piece of earth putting current on the EGC. That current is essentially noise. Even if it's a pure in phase sine wave. But it wont be.

 

[iwire]

Re: Ground loops

Originally posted by physis:
Tom is right. Using sort of slang terms, you'll be bleeding off any voltage difference at that piece of earth putting current on the EGC.

I agree, but the NEC is not concerned with that, in fact it embraces it in some locations.

1)You can always add unnecessary electrodes.

2)Separate structures require additional electrodes which along with being at different potentials due to their spacings may actually have part of the neutral current flowing through them.

 

[dereckbc]

Re: Ground loops

OK, I am going to stink things up a bit. Ground loops are your best friend and worst enemy. They are impossible to eliminate in any commercial application. Case in point condiut, cable tray, etc installed attached to concrete or steel beems.

The noise created is call common mode noise. Chew on that for a while and come up with some questions.

For now I will tell you we in the telephone buisness rely on ground loops. i.e ground rings with multiple entry points. It's all in the design and how you use them.

 

[physis]

Re: Ground loops

For now I will tell you we in the telephone buisness rely on ground loops. i.e ground rings with multiple entry points. It's all in the design and how you use them.

I'm chewing on that instead. Now I don't see how I can get out of asking you to elaborate.

[ July 18, 2005, 05:22 PM: Message edited by: physis ]