Multiple Earth Grounds

[Rick Christopherson]

I doesn't?
The first part of Don's quote is talking about previous code allowing the neutral to be used as the grounding conductor, no differant then at a service.
The second part of Dons post is talking about the 2008 removal of this allowance and the requiring of a seperate EGC ran to other buildings, so this EGC would not be current carrying like the neutral would in Dons first part of his post, and he pointed out this EGC would most likly be smaller then a neutral that was allowed pre-2008.

Like Don, you are getting hung up on names and/or colors; as opposed to functions.

If you have a ground-to-neutral bond at both locations, then you have parallel conductors, and both of them will be current carrying conductors. Electrons are color blind. They don't care whether they are flowing down a green colored wire or one that you choose to label "EGC". Parallel paths (e.g. ground loops) completely negate the EGC between buildings. You have 2 grounded conductors, and 0 grounding conductors. They both carry "normal" current.

If you are going to carry the EGC from one building to the other, then only one building can have a G-N bond. The exception to this is a separately derived system using an isolation transformer. However, this leads to the next point....

By the requirements in the NEC you would still be required to bond across the transformer and still have the voltage problems in building A don't look for band aids to fix a real problem.

By this statement, you have wiped out the very nature of a separately derived system. Does this mean we can strike those words from the NEC?

You indicated that it was required to bond across a transformer. If you bond across an isolation transformer, then it is no longer a separately derived system. This negates the very purpose of the isolation transformer. To say that it is required to bond across a transformer, means that there is no such thing as a separately derived system, unless of course we are referring to off-grid generators.

An isolation transformer is in fact a valid solution (not a band-aid) however, as the OP already stated, not very cost effective.

 

[Rick Christopherson]

If indeed hurk is right and there is no main bonding jumper installed at the service there would likely be similar voltage issues even if the second building was supplied through a separately derived system. The EGC for the supply would still be required to be bonded to the secondary grounded conductor and this is where the stray voltage problem is originating.

You're mistaken. In post #12 I was the one that suggested that the symptoms indicated a faulty or lacking G-N bond in building A, and the possibility of the active G-N bond being in building B.

However, as I explained in my follow-up just above, that wasn't the purpose of the posting you quoted.

To repeat what I said earlier, the OP should first examine the G-N bond in the main building (separate from building B). This should be corrected regardless of the conditions in building B.

As for building B, I personally do not agree with the changes made to code over the past couple cycles on this matter, and I find it wrong that an inspector forced new code onto a grandfathered installation.

So ignoring the fact it would be a violation of current code (therefore I'm not recommending it) I would prefer to see building B treated as though it was a main service with its own grounding electrode system and G-N bond.

This would solve his problem, but you would have to get someone to sign off on it. The fact that it was originally done this way (and possibly permissible at that time), and an inspector wrongly forced it to be changed, might carry some weight.

If the service does not have properly installed main bonding jumper I would take some caution at just putting one in - there could be a ground fault on a phase conductor and installing the bonding jumper could really be enlightening. Apply a load between the neutral and ground and check what happens to voltage when doing so to get an idea of what is happening.

P.S. I agree with this approach before installing a G-N bond in building A.

 

[Buck Parrish]

To simpilfy it, you need a 4 wire from one building to the other and so on.
Isolate your grounded conductor , then connect your electrodes to your grounding conductor.
This would be easier then the new transformer system you mentioned.

 

[don_resqcapt19]

Rick,
You missed the point of my comment. You said that under previous codes you could have an installation where the grounding electrode system at the second building was isolated from the grounding electrode system at the first building.

It is unfortunate that the NEC eliminated the provisions for separate buildings having separate grounding electrode systems, because this sounds like the perfect situation where is would be needed (unless there are other ground paths between the two).

My comment says that the code has never permitted the grounding electrode systems of the two buildings to be isolated and that the only difference between the current code and the older codes is the color of the wire that bonds the two electrode system together. In the old codes it was the white conductor and in the current code it is the green conductor.

 

[don_resqcapt19]

By this statement, you have wiped out the very nature of a separately derived system. Does this mean we can strike those words from the NEC?

It is almost impossible to have an "isolated" system under the rules of the NEC. The transformer is the most common SDS, but the code rules require the interconnection of the supply and load equipment grounding conductors. This has been an issue with the NEC Article 100 definition of SDS and there is a change in the 2011 code to make it clear that a transformer is a SDS even with the primary and secondary EGCs connected together.

Separately Derived System. A premises wiring system whose power is derived from a source of electric energy or equipment other than a service. Such systems have no direct connection from circuit conductors of one system to
circuit conductors of another system, other than connections through the earth, metal enclosures, metallic raceways, or equipment grounding conductors.

The text in blue was added for the 2011 code.

 

[don_resqcapt19]

...
You indicated that it was required to bond across a transformer. If you bond across an isolation transformer, then it is no longer a separately derived system. This negates the very purpose of the isolation transformer. To say that it is required to bond across a transformer, means that there is no such thing as a separately derived system, unless of course we are referring to off-grid generators.

An isolation transformer is in fact a valid solution (not a band-aid) however, as the OP already stated, not very cost effective.

Please cite the code section that permits you to not to connect either the primary or secondary EGC to the case of the transformer. It is a very rare case where you can do this and therefore truly isolated systems are very rare in code compliant installations.

 

[don_resqcapt19]

...

To repeat what I said earlier, the OP should first examine the G-N bond in the main building (separate from building B). This should be corrected regardless of the conditions in building B.

Yes this is a serious issue that needs to be corrected.

As for building B, I personally do not agree with the changes made to code over the past couple cycles on this matter, and I find it wrong that an inspector forced new code onto a grandfathered installation.

Exactly what code changes are you talking about here?

So ignoring the fact it would be a violation of current code (therefore I'm not recommending it) I would prefer to see building B treated as though it was a main service with its own grounding electrode system and G-N bond.
This would solve his problem, but you would have to get someone to sign off on it. The fact that it was originally done this way (and possibly permissible at that time), and an inspector wrongly forced it to be changed, might carry some weight.

How would this change anything. The original poster was concerned about the grounding electrode systems in building 1 being connected to the one in building 2. This still happens when you put a second G-N bond at the second building. The neutral will be bonded to the grounding electrode system at building one and to the new grounding electrode system at building 2.

 

[Rick Christopherson]

How would this change anything. The original poster was concerned about the grounding electrode systems in building 1 being connected to the one in building 2. This still happens when you put a second G-N bond at the second building. The neutral will be bonded to the grounding electrode system at building one and to the new grounding electrode system at building 2.

I don't care that they are sharing a grounded conductor. What I do care about is that they share BOTH a grounded conductor and a grounding conductor, AND there is a problem with G-N bond in one or both buildings.

Please cite the code section that permits you to not to connect either the primary or secondary EGC to the case of the transformer. It is a very rare case where you can do this and therefore truly isolated systems are very rare in code compliant installations.

I never said the transformer would be isolated. I said the source and load would (could) be isolated. I have not found anything in code that mandates that the GES' of two buildings separated by an SDS share a common GES. The purpose of an SDS in this situation is to break one or both, the grounded conductor and/or the grounding conductor between buildings.

The Hurk said it was required to Bond Across the transformer. This implies that it is required to connect the grounded conductors together on both primary and secondary sides of the transformer. That may not be what he meant, but his choice of words, coupled with the statement that it wouldn't resolve the issue implied it. This is supported by the definition of a bonding jumper in 250.2

Bonding Jumper, System. The connection between the grounded circuit conductor and the equipment grounding conductor at a separately derived system.

There is a bonding jumper on both sides of the SDS, but that is not the same as saying you need to bond Across the transformer.

 

[don_resqcapt19]

I don't care that they are sharing a grounded conductor. What I do care about is that they share BOTH a grounded conductor and a grounding conductor, AND there is a problem with G-N bond in one or both buildings.

I don't understand what you are telling me.

I never said the transformer would be isolated. I said the source and load would (could) be isolated. I have not found anything in code that mandates that the GES' of two buildings separated by an SDS share a common GES. The purpose of an SDS in this situation is to break one or both, the grounded conductor and/or the grounding conductor between buildings.

There will still be a direct physical connection between the grounding and grounded conductors of both buildings and of the grounding electrode system of both buildings. This will happen because the EGC of both the primary and the secondary will be connected together at the transformer. The source and load will not be isolated.

 

[tom baker]

All receptacles should be ungrounded receptacles, or protected by a GFCI and marked as ungrounded.

I would like the code references on this. Perhaps from the OP we need when it was installed.

 

[Rick Christopherson]

There will still be a direct physical connection between the grounding and grounded conductors of both buildings and of the grounding electrode system of both buildings. This will happen because the EGC of both the primary and the secondary will be connected together at the transformer. The source and load will not be isolated.

First off, I think you need to stop using the terms "EGC" and "GEC" for a little while because I can see from your other posts that these terms are tripping you up a little.

Let's assume the transformer is physically located at the remote building B. How many conductors will you have at the primary side from the main building? You will have 2 ungrounded conductors and (arguably--see note below) 1 grounding conductor. (There will not be any grounded conductor--or "neutral" if you prefer.)

On the secondary side of the transformer, you will have 2 ungrounded conductors, 1 grounded conductor (via the secondary-side G-N bond), and a grounding conductor.

(I will use "Neutral" here for better clarity.)

Even though the Neutral of building A is grounded in building A, and the Neutral of building B is grounded in building B; there is only a single circuit path between these two points of bonding (the grounding conductor). However, a complete circuit requires 2 paths to form a circuit. A single conductor can share a voltage reference between ends, but by definition, it cannot sustain a current. You don't have a circuit. (Remember, this is NOT an autotransformer.)

Even though the grounding conductor is shared by both buildings, the grounded conductor(s) are not shared. They have been isolated by the transformer. The existence or absence of a G-N bond in one building will have no impact on the other building.

Now My Note: This isn't terribly important to the discussion, and that is why I added it as a note. Because the transformer is physically located at building B, it is required for the transformer to be both grounded (chassis ground) and bonded (CT bond) to building B's grounding system.

I don't find any requirement in code where you are forced to carry a grounding conductor from building A to the transformer of an SDS system in a separate building. You are required to bring one for a branch circuit, but the same language does not exist (as far as I know) for an SDS to a separate structure.

If the main and SDS are located within a single structure (the type typically described in code), then you do have to have a common ground because the structure itself is a common ground.

I'm an engineer, not an expert on code, so I could be wrong about this. However, do you have a citation that says a grounding conductor must be carried to the transformer from building A (assuming no other ground-path exists without one)?

 

[iwire]

First off, I think you need to stop using the terms "EGC" and "GEC" for a little while because I can see from your other posts that these terms are tripping you up a little.

Man that is just plain uncalled for, Don knows the difference between them and is not tripped up with the meaning or intended use of either.

 

[lauraj]

I would like the code references on this. Perhaps from the OP we need when it was installed.

He changed his scenario after my post, so it doesn't apply to this situation.

 

[don_resqcapt19]

First off, I think you need to stop using the terms "EGC" and "GEC" for a little while because I can see from your other posts that these terms are tripping you up a little.

First off you cannot have a discussion of the rules of the NEC without using the terms of the NEC. The terms as used in the NEC must be used and understood for the purposes of this discussion.

Let's assume the transformer is physically located at the remote building B. How many conductors will you have at the primary side from the main building? You will have 2 ungrounded conductors and (arguably--see note below0 1 grounding conductor. ...

There is no question that the rules of the NEC will require an Equipment Grounding Conductor to be run with the feeder conductors to the transformer. This can be any type of EGC that is permitted by 250.118.

On the secondary side of the transformer, you will have 2 ungrounded conductors, 1 grounded conductor (via the secondary-side G-N bond), and a grounding conductor.

(I will use "Neutral" here for better clarity.)

I don't see how the term neutral provides any clarity over the term "grounded conductor". However in most cases the terms can and are used interchangeably.

Even though the Neutral of building A is grounded in building A, and the Neutral of building B is grounded in building B; there is only a single circuit path between these two points of bonding (the grounding conductor). However, a complete circuit requires 2 paths to form a circuit. A single conductor can share a voltage reference between ends, but by definition, it cannot sustain a current. You don't have a circuit. (Remember, this is NOT an autotransformer.)

Even though the grounding conductor is shared by both buildings, the grounded conductor(s) are not shared. They have been isolated by the transformer. The existence or absence of a G-N bond in one building will have no impact on the other building.

I have never said that the grounded conductors are shared. I have only said that they are not isolated. They are connected together. There will be some grounded conductor current of each system flowing via the connections to the EGCs and via the earth. This is a parallel path for grounded conductor current and current flows on all available parallel paths. In most cases this will be a very small amount of current given the relatively high impedance of the path via the earth.

Now My Note:

This isn't terribly important to the discussion, and that is why I added it as a note. Because the transformer is physically located at building B, it is required for the transformer to be both grounded (chassis ground) and bonded (CT bond) to building B's grounding system.

I don't find any requirement in code where you are forced to carry a grounding conductor from building A to the transformer of an SDS system in a separate building. You are required to bring one for a branch circuit, but the same language does not exist (as far as I know) for an SDS to a separate structure.

You have to have a fault clearing path for the primary side of the transformer and the only way to do that is to run an EGC with the primary feeder.

If the main and SDS are located within a single structure (the type typically described in code), then you do have to have a common ground because the structure itself is a common ground.

You have a common ground because you have primary and secondary EGCs and the fact that both EGCs will be connected to the transformer case.

I'm an engineer, not an expert on code, so I could be wrong about this. However, do you have a citation that says a grounding conductor must be carried to the transformer from building A (assuming no other ground-path exists without one)?

250.4(A)(5)

 

[mivey]

Man that is just plain uncalled for, Don knows the difference between them and is not tripped up with the meaning or intended use of either.

Well I would not sweat it too much. I'm sure he wasn't being personal. Rick is an equal opportunity offender. :grin:

 

[PetrosA]

Am I misunderstanding something or does this not sound like a classic case of a high impedance neutral connection somewhere before the main? If the ground rod at Bldg. A is bonded to the neutral at the main as well as to the ground system of Bldg. B through the EGC between the buildings and if that fancy ground system at Bldg. B has seriously low impedance back to the utility transformer, couldn't that 70V reading be unbalanced neutral current returning to it's source through the ground system at Bldg. B? It would also explain why shutting certain loads in Bldg. A reduces the voltage. Has the neutral connection been checked yet?

 

[iwire]

Well I would not sweat it too much. I'm sure he wasn't being personal. Rick is an equal opportunity offender. :grin:

Me too ....

 

[PaulWDent]

Separate ground rod at outbuilding

Separate ground rod at outbuilding

It is very bad to have a separate ground rod at an outbuilding that is not part of the main panel ground system.. If lightning strikes close by, the potentials of the unconnected grounds will be vastly different and this will damage equipment.

On the otherhand, I don't like bonding neutral to the ground at the outbuilding when it is already bonded at the main panel. It is not an appropriate use of the neutral to bond two separate grounding electrodes. If the outbuilding is supplied by a GFCI or AFCI in the main panel, that would screw up the breaker operation.

The technically correct solution is to use an isolating transformer to feed the outbuilding. The neutral of the primary should be grounded only at the source, i.e. the main panel, and the neutral of the secondary should be grounded only at the outbuilding. However, there should also be a separate ground wire run from the main panel to the outbuilding to bond the main panel ground to the remote ground to avoid the lightning strike potential difference issue, which could put high stress between the transformer primary and secondary insulation.

I have bought 5KVA isolating transformers (outside rated) on E-bay for $125.

 

[kwired]

It is very bad to have a separate ground rod at an outbuilding that is not part of the main panel ground system.. If lightning strikes close by, the potentials of the unconnected grounds will be vastly different and this will damage equipment.

On the otherhand, I don't like bonding neutral to the ground at the outbuilding when it is already bonded at the main panel. It is not an appropriate use of the neutral to bond two separate grounding electrodes. If the outbuilding is supplied by a GFCI or AFCI in the main panel, that would screw up the breaker operation.

The technically correct solution is to use an isolating transformer to feed the outbuilding. The neutral of the primary should be grounded only at the source, i.e. the main panel, and the neutral of the secondary should be grounded only at the outbuilding. However, there should also be a separate ground wire run from the main panel to the outbuilding to bond the main panel ground to the remote ground to avoid the lightning strike potential difference issue, which could put high stress between the transformer primary and secondary insulation.

I have bought 5KVA isolating transformers (outside rated) on E-bay for $125.

And why is that better than a feeder with an EGC? You still have an EGC with either install. You do not bond neutral if there is also an EGC. If there is a fairly direct lightning incident, there is high voltage between many points both grounded and ungrounded.