Main Bonding jumper w/multiple service enclosures

[c_picard]

Hi folks,

This is a follow up question related to an open topic way over in the PV Forum.

What are the available avenues for requesting a formal interpretation of an article in the NEC?

I am currently assembling some published articles and relevant code sections to forward to the DCA in NJ, with the intent of clarifying the PV industry best-practice (I believe the only code-compliant method) of bonding the grounded conductor at the PV system disconnect, when using a supply side connection.

All of the published material I can find, as well as the feedback I am getting from my contacts in the industry, indicate that this enclosure should be treated as a service disconnect, which is installed to disconnect an explicitly permitted "set" of service entrance conductors. The DCA's stance is that since PV is not a "service", no sections of article 230 would apply, and the grounded conductor should not be bonded in this enclosure.

Your thoughts?

I humbly ask the moderator to allow this post here, as I am soliciting the response of a broader knowledge base. I'm preaching to the choir in the PV chamber! I'm happy to provide code references, but I'd rather not lead the witness.

Thanks

 

[GoldDigger]

All of the published material I can find, as well as the feedback I am getting from my contacts in the industry, indicate that this enclosure should be treated as a service disconnect, which is installed to disconnect an explicitly permitted "set" of service entrance conductors. The DCA's stance is that since PV is not a "service", no sections of article 230 would apply, and the grounded conductor should not be bonded in this enclosure.

Applying common sense (not always the best guidance) the supply side connection is the service, regardless of which way the power happens to be flowing, and as such the disconnect is service disconnect with the POCO side being the service side. If the inverter grounded conductor (often but not always a neutral) is not bonded on the load (PV) side of that disconnect, it is not effectively bonded at all.

 

[c_picard]

Agreed. This may be why I am having such a hard time illustrating the need for this conductor to be bonded.

The local inspector would not get past Art. 100, the definition of "service".

To his great credit, the state building official did get past that point. However, he considers the points between the service conductors and the inverter output terminals as the "inverter output circuit". Therefore, none of the provisions in Art 230 apply. Clearly, this is not the intent of the code, but by the letter of the code, I admit things do get murky. There are multiple references that imply that the portion of the circuit between the tap and the line side of the disconnect are indeed "service entrance conductors", and this is backed up by common sense.

However, as I sketch out the logic chain, unless I can get the State of NJ to jump from 705 to 230, we will be prohibited from bonding the neutral at what the majority of the industry considers a disconnect that, while not explicitly a "service" disconnect, should be treated as such. My sincere hope is that we would not be creating a safety hazard by complying with the state's interpretation.

Any links to references or publications that deal with the safety hazards resulting from failing to bond the grounded conductor at each enclosure would be greatly appreciated.

As always, thanks for the peer review. This sort of conversation is critical to the uniform application of the NEC.

 

[texie]

Well, you know how I feel from my post in the other thread. I'll say it again, 250.24(C) says it all. It also makes it clear "to each service disconnecting means". I know that many would say this is not a service, but in my view it interfaces with the service and 250.24(C) clearly applies.
Let's look at it another way. What if you just had a disconnect there for future use, maybe for a future inverter supply side tap, maybe for a future 240 load to a pump that does not need a neutral, what ever it maybe. 250.24(C) applies as the neutral must come to any disconnect and be bonded that isolates the service conductors from anything on the premises must comply with 250.24(C).

 

[GoldDigger]

Well, you know how I feel from my post in the other thread. I'll say it again, 250.24(C) says it all. It also makes it clear "to each service disconnecting means". I know that many would say this is not a service, but in my view it interfaces with the service and 250.24(C) clearly applies.
Let's look at it another way. What if you just had a disconnect there for future use, maybe for a future inverter supply side tap, maybe for a future 240 load to a pump that does not need a neutral, what ever it maybe. 250.24(C) applies as the neutral must come to any disconnect and be bonded that isolates the service conductors from anything on the premises must comply with 250.24(C).

On the other hand, it is equally important that there not be redundant grounded to ground bonds. So if there is a ground to neutral bond at the GTI's AC disconnect, the neutral wire going to the GTI must not interconnect with any neutral wire which was bonded elsewhere.
Along the same lines, since the PV does not constitute a separately derived system, the PV disconnect will not be interrupting the neutral, and if that service neutral wire is effectively bonded (and not switched) at some other service disconnect, there is no strong reason to bond it at the PV disconnect too.

I see the merit in both arguments, but unless you are looking forward to a time when the PV tap wires, including the neutral, are physically disconnected somewhere, I am not convinced that there is a practical difference between the two situations.

As a separate issue, the PV supply-side tap neutral will have to be sized to carry the full fault current from either the GTI side or the POCO side, but not both simultaneously, so I do not see a need to oversize that neutral nor an opportunity to reduce it. With no bond at the disconnect to the ground electrode system, the neutral may be called upon to carry the full fault current from any possible line to neutral or line to ground fault.

 

[texie]

Here is a piece by John Wiles on this subject. It coincides with my position. http://www.iaei.org/blogpost/927663/157476/Supply-side-PV-Utility-Connections

 

[c_picard]

It's important to note that we commonly see "redundant" grounded to ground bonds wherever we encounter multiple enclosures in a service. Think of an apartment complex, for example. The Grounded conductor is required to be bonded at EACH disconnect, not just one.