250.32(B) and Handbook Exhibit 250.17

[JDBrown]

I'm stumped looking at Exhibit 250.17 from the Handbook (see below -- I'm pretty sure this is considered fair use).



Now, it was my understanding from 250.32(B) that a new feeder to a separate building needs to include an EGC. That EGC would be landed on the ground bus of the remote building's panel, which would also connect to the remote building's GEC, but would not be bonded to the neutral bus. The only exception to this is for an existing building with a feeder that has a Grounded Conductor but no EGC.

This seems to contradict what is shown in Exhibit 250.17 (above).

The connection from Building 1 to Building 2 looks correct (existing feeder with no EGC and no continuous metallic paths between the two buildings). The Grounded Conductor between Building 1 and Building 2 does double duty as an EGC, and is therefore bonded to the ground bus in Building 2.

The problem I see is with the feeder from Building 2 to Building 3. According to the language of the Code, there ought to be an EGC from the ground bus in Building 2 to the ground bus in Building 3. ... Right? I didn't think it was ever acceptable to use a local electrode for your only equipment grounding connection -- we talk about this all the time, how a ground fault in Building 3 in this scenario likely wouldn't be able to carry enough current to trip even a 15A breaker.

So, is the Handbook Exhibit wrong, or am I missing something?

* This comes from the 2008 Handbook, which is the most current one I have (and the current code cycle where I live). If this is a mistake that has already been corrected in later editions, I apologize for wasting your time with this post.

 

[Gregg Harris]

I'm stumped looking at Exhibit 250.17 from the Handbook (see below -- I'm pretty sure this is considered fair use).

View attachment 9264

Now, it was my understanding from 250.32(B) that a new feeder to a separate building needs to include an EGC. That EGC would be landed on the ground bus of the remote building's panel, which would also connect to the remote building's GEC, but would not be bonded to the neutral bus. The only exception to this is for an existing building with a feeder that has a Grounded Conductor but no EGC.

This seems to contradict what is shown in Exhibit 250.17 (above).

The connection from Building 1 to Building 2 looks correct (existing feeder with no EGC and no continuous metallic paths between the two buildings). The Grounded Conductor between Building 1 and Building 2 does double duty as an EGC, and is therefore bonded to the ground bus in Building 2.

The problem I see is with the feeder from Building 2 to Building 3. According to the language of the Code, there ought to be an EGC from the ground bus in Building 2 to the ground bus in Building 3. ... Right? I didn't think it was ever acceptable to use a local electrode for your only equipment grounding connection -- we talk about this all the time, how a ground fault in Building 3 in this scenario likely wouldn't be able to carry enough current to trip even a 15A breaker.

So, is the Handbook Exhibit wrong, or am I missing something?

* This comes from the 2008 Handbook, which is the most current one I have (and the current code cycle where I live). If this is a mistake that has already been corrected in later editions, I apologize for wasting your time with this post.

 

[ActionDave]

The illustration is not about EGC with the feeder. It is showing where to attach the grounding electrode conductor in each scenario.

 

[GoldDigger]

The illustration is not about EGC with the feeder. It is showing where to attach the grounding electrode conductor in each scenario.

But it still explicitly shows a ground to neutral bond wire in box #2. That is a problem no matter what the illustration is "about". And the explicit reference to the absence of a continuous metallic path between 1 and 2 seems to be there to indicate that there is neither a wire EGC nor a raceway EGC, so the bonding jumper is required.
I do not have a 2005 NEC, so I cannot say whether that installation would be compliant under it. It would not be for 2008 or later.

 

[ActionDave]

But it still explicitly shows a ground to neutral bond wire in box #2. That is a problem no matter what the illustration is "about". And the explicit reference to the absence of a continuous metallic path between 1 and 2 seems to be there to indicate that there is neither a wire EGC nor a raceway EGC, so the bonding jumper is required.
I do not have a 2005 NEC, so I cannot say whether that installation would be compliant under it. It would not be for 2008 or later.

Yes, it was legal prior to '05 to bond the neutral at a separate structure if there were no metallic connections between them. Note the note below the illustration.

 

[david luchini]

But it still explicitly shows a ground to neutral bond wire in box #2...I do not have a 2005 NEC, so I cannot say whether that installation would be compliant under it. It would not be for 2008 or later.

The neutral-ground bond for an existing premise wiring system (Building 2) was permitted in the 2008 Code. The new feeder to Building 3 would not allow the neutral-ground bond, as the picture shows. Yes, an EGC would have to be run with the new feeder.

 

[jumper]

Am I blind, tired or stupid are those service ungrounded/hot connections not correct or is this a 120v service?

 

[ActionDave]

Am I blind, tired or stupid are those service ungrounded/hot connections not correct or is this a 120v service?

That's not what's important. The drawing are primarily there to show where to connect your earth.

 

[jumper]

I'm pretty sure this is considered fair use.

Yes, and this is IMO and others that agree.

 

[jumper]

That's not what's important. The drawing are primarily there to show where to connect your earth.

Agreed.

 

[ActionDave]

Agreed.

Where is the emoticon for fist bump?

 

[JDBrown]

Thanks for the clarification, guys. I still think they ought to have shown the EGC from Building 2 to Building 3 -- for clarity and completeness if nothing else -- but at least it makes sense now why it's shown the way it is (illustrating the connection point for the GEC at each building).

And, for what it's worth, the 2011 NEC still allows the Grounded conductor to do double duty as an EGC for existing installations (e.g. Buildings 1 and 2). They re-worded it a bit, apparently to clarify that it's only acceptable for existing installations that were Code-compliant to begin with, but the gist is essentially the same:

NEC 250.32(B)
Exception: For installations made in compliance with previous editions of this Code that permitted such connection, the grounded conductor run with the supply to the building or structure shall be permitted to serve as the ground-fault return path if all of the following requirements continue to be met:

(1) An equipment grounding conductor is not run with the supply to the building or structure.
(2) There are no continuous metallic paths bonded to the grounding system in each building or structure involved.
(3) Ground-fault protection of equipment has not been installed on the supply side of the feeder(s).

If the grounded conductor is used for grounding in accordance with the provision of this exception, the size of the grounded conductor shall not be smaller than the larger of either of the following:

(1) That required by 220.61
(2) That required by 250.122