Separate Ground Rod at Every Transformer

[lile001]

I have a fight with a chief electrican of a facility, and maybe also with the AHJ. This facility has installed a separate ground rod at every 480:208 volt transformer in a large facility (250,000 square feet). We inspected all the transformers, and most of them have a ground rod, no bond from the ground bus of the transformer to the primary ground bus or the grounding electrode system for the building, and no connection to building steel or water pipe. All of the ground rods are driven through a hole in the basement slab, into presumably dry soil - useless. Some transformer secondaries are not grounded or bonded at all, and they do admit these are noncompliant (and dangerous!). As I was photographing these ground rods, the electrician was saying that the AHJ instructed him to do it this way - a separate ground rod at each transformer. I was trying to explain to the electrician the problems that could be caused by this, he wants me to quote chapter and verse from the code where this isn't allowed. All of these transformer secondary grounds should be bonded together and to the main building system ground.

The problem is the Code doesn't come out and say this isn't allowed in plain english, it implies it several places. I'm trying to zero in on the key paragraphs that make the best case. I'm citing the 2008 NEC.

250.30(A)(7) doesn't help: Exception #1 allows separately derived systems to use any of the electrodes identified in 250.52 if building steel and water pipes are not available. It says nothing about bonding this darn ground rod to the main building ground. The AHJ is likely to hang his hat on this paragraph and say that's that.


250.4(A) (1) and (2) give us no ammunition, since they talk about "connecting to earth".

250.4(A)(3) and (5) require cases to be bonded, and (5) specifically says the earth is not an effective ground fault current path. However they could argue that there is no current between a separately derived system and the service entrance if there is a fault on the separately derived system.

250.50 requires all ground rods "that are present" to be bonded together to form an effective ground fault current path. That comes close.

250.58 requires a "Common Grounding Electrode" but does not specifically mention separately derived systems.

I wish there was an article that says "multiple separately derived systems shall be connected to a single common grounding electrode that is bonded to the building service grounding electrode system" but there is no such paragraph. 250.30(A)(4) doesn't say this.

What is worse, this is a hospital. Small differences in potential between two 120V systems could have deadly consequences. An amp of ground current, due to a nearby substation, could produce tens of volts between two ground rods, enough to stop a patient's heart if two machines are plugged into different separately derived systems and then connected to the patient.

 

[zbang]

As I understand it, you can install as many grounding electrodes as you want, as long as you bond everything to code.

 

[lile001]

As I understand it, you can install as many grounding electrodes as you want, as long as you bond everything to code.

Right! I don't care how many ground rods they drive, but I'd like to be able to point to a paragraph that says they all need a wire sized per 250.66 based on the secondary conductors that bond them all back to a common point. Or a water pipe or steel. Which paragraph says that most irrefutably?

Or - once they drive the ground rod, is this bonding wire required to be sized per 250.122 based on the primary overcurrent protection?

 

[don_resqcapt19]

250.30(A)(7) requires a connection to the building steel and or the metal water piping if those are available. Only if those are not available are you permitted to use a ground rod as the grounding electrode for a SDS.

 

[brian john]

I have a fight with a chief electrican of a facility, and maybe also with the AHJ. This facility has installed a separate ground rod at every 480:208 volt transformer in a large facility (250,000 square feet). We inspected all the transformers, and most of them have a ground rod, no bond from the ground bus of the transformer to the primary ground bus or the grounding electrode system for the building, and no connection to building steel or water pipe. All of the ground rods are driven through a hole in the basement slab, into presumably dry soil - useless.

How do you know they are useless? If you brought a statement like that to an meeting, you'd better have some back up.

 

[ptonsparky]

Egads!!

"Some transformer secondaries are not grounded or bonded at all, and they do admit these are noncompliant (and dangerous!)."

The neutrals are floating? Who...what the....


Good luck on this one.

 

[jim dungar]

250.50 requires all ground rods "that are present" to be bonded together to form an effective ground fault current path. That comes close.

Why do you think this does not cover your situation?

All of the building steel is bonded together, so from a grounding standpoint the transformers' secondary fault paths are not separate systems. Current flows on all available paths.

 

[petersonra]

250.50 requires all ground rods "that are present" to be bonded together to form an effective ground fault current path.

Does it really say that? Really?

 

[iwire]

Does it really say that? Really?

Thankfully no.

III. Grounding Electrode System and Grounding
Electrode Conductor

250.50 Grounding Electrode System. All grounding electrodes
as described in 250.52(A)(1) through (A)(7) that are
present at each building or structure served shall be bonded
together to form the grounding electrode system. Where
none of these grounding electrodes exist, one or more of
the grounding electrodes specified in 250.52(A)(4) through
(A)(8) shall be installed and used.

Exception: Concrete-encased electrodes of existing buildings
or structures shall not be required to be part of the grounding
electrode system where the steel reinforcing bars or rods are
not accessible for use without disturbing the concrete.

 

[SG-1]

As I understand this...

As I understand this...

The bonding you desire should exist, but not the way you want to do it.

Each SDS is feed from the same source. An EGC sized per 250.122 comes with the phase conductors to the transformer primary. It might be a metalic raceway. It is bonded to the transformer case & frame.

The secondary side of the SDS has the grounded conductor bonded to the case & frame. This would establish the connection to the "primary GES" that you desire.

As for an individual ground rod, can they show that it measures less than 25 ohms ? ( This is just for show because if it measured 2000 & driving a second rod brings it down to 1000 the code is satisified.) The GEC to a rod only is only required to be a 6 AWG CU (250.66(A).

Can you take voltage measurements between the different SDSs grounded conductors or rods to prove there is a voltage difference ?

Would that get their attention ?

 

[ptonsparky]

There is, or should be, so much conduit hung from building steel from all different SDS that I doubt you could get a VD from the two rods the farthest apart. Not counting all the EG pulled everywhere.

 

[Smart $]

....

250.50 requires all ground rods "that are present" to be bonded together to form an effective ground fault current path. That comes close.

....

Does it really say that? Really?

Thankfully no.

III. Grounding Electrode System and Grounding
Electrode Conductor

250.50 Grounding Electrode System. All grounding electrodes
as described in 250.52(A)(1) through (A)(7) that are
present at each building or structure served shall be bonded
together to form the grounding electrode system. Where
none of these grounding electrodes exist, one or more of
the grounding electrodes specified in 250.52(A)(4) through
(A)(8) shall be installed and used.

Exception: Concrete-encased electrodes of existing buildings
or structures shall not be required to be part of the grounding
electrode system where the steel reinforcing bars or rods are
not accessible for use without disturbing the concrete.

While it technically doesn't say what lile001 said, it does say what it says... and it says all 250.52(A)(1 through 7) electrodes present must be bonded to form the grounding electrode system. That includes ground rods.

Drive a rod into the ground qualifies it as present.

 

[lile001]

How do you know they are useless? If you brought a statement like that to an meeting, you'd better have some back up.

Reisistance of Ground rods are heavily dependant on soil moisture. The drier the soil, the higher the resistance. A slab below a football-field sized building is unlikely to have a lot of soil moisture, compared to the soil under a grass field next door. I could probably look up a few measurements of ground rod resistance in dry soils if need be to back it up. "Useless" is actually the term I use for all ground rods - they ought to be a last resort. There are a half dozen other ways of ensuring that the building electrical system and the earth are near the same potential that work better.

 

[lile001]

There is, or should be, so much conduit hung from building steel from all different SDS that I doubt you could get a VD from the two rods the farthest apart. Not counting all the EG pulled everywhere.

Wish we had some building steel in this building. Then this would be no problem. It has a concrete frame.

 

[lile001]

Egads!!

"Some transformer secondaries are not grounded or bonded at all, and they do admit these are noncompliant (and dangerous!)."

The neutrals are floating? Who...what the....


Good luck on this one.

Yeah - the only thing that keeps the neutral from floating up to some random high voltage is random, unintentional connections to ground on that panel, like maybe they anchored some conduit into a concrete beam, or happened to clamp onto the same unistrut as a water pipe. I don't think anyone is arguing this is acceptable. This is how you get a shock from a piece of conduit.

 

[lile001]

While it technically doesn't say what lile001 said, it does say what it says... and it says all 250.52(A)(1 through 7) electrodes present must be bonded to form the grounding electrode system. That includes ground rods.

Drive a rod into the ground qualifies it as present.

This, and the paragraph (250.122? It is Sunday morning and I don't have my code book handy) that says water pipe SHALL be bonded to the secondary look like the best things to make an argument with. This hammerhead AHJ thinks that soil is a perfect conductor, any two ground rods are bonded together like siamese twins. 'Tain't so, and Mr. Mike Holt has an excellent article showing a 120V fault can produce an 82 volt difference between two ground rods a foot apart. Enough juice to wake someone up, if they happen to be plugged into a diagnostic machine and touching a grounded bedrail!

 

[Smart $]

...and the paragraph... that says water pipe SHALL be bonded to the secondary look like the best things to make an argument with. ...!

That'd be 250.30(A)(7), but not limited to a metal underground water [supply] pipe. A structural metal grounding electrode is also possible. The nearest of the two must be used. There is no exception to using one of these electrodes, if either is available. The term available is not to be construed as meaning practical. Available in this context means it exists at the building or structure, period.

 

[petersonra]

That'd be 250.30(A)(7), but not limited to a metal underground water [supply] pipe. A structural metal grounding electrode is also possible. The nearest of the two must be used. There is no exception to using one of these electrodes, if either is available. The term available is not to be construed as meaning practical. Available in this context means it exists at the building or structure, period.

However, nothing says you have to run a separate conductor to make this connection to the required GE. If the EGC is a conductor (as opposed to a raceway) and it is of the appropriate size all the way to the GE it has to be bonded to, there is no need to run another conductor. Just use the EGC.

 

[iwire]

While it technically doesn't say what lile001 said, it does say what it says... and it says all 250.52(A)(1 through 7) electrodes present must be bonded to form the grounding electrode system. That includes ground rods.

Drive a rod into the ground qualifies it as present.

I agree with you but that was not the point of my post at all, Bob had highlighted

ground fault current path

which is what I was answering.

 

[Smart $]

I agree with you but that was not the point of my post at all, Bob had highlighted



which is what I was answering.

I know, that's why I said it didn't say what lile said (also)...

...but reading your reply from this end, it seemed to have a slight connotation that the bonding wasn't necessary and my reply was intended to clarify, not rebuke yours.