EV GFCI protection

[don_resqcapt19]

If this change from the 2026 First Draft Report makes it through the process, it will remove a lot of discussion that has taken place about this issue.

625.54 Ground-Fault Circuit-Interrupter Protection for Personnel.
All receptacles and outlets installed for the connection of electric vehicle charging shall have ground-fault circuit-interrupter protection for personnel.

 

[jaggedben]

This is terrible. It will amount to a ban on hard-wired EVSEs. This is incredibly wrong headed.

It also perpetuates the wrong-headed language of outlets 'having' GFCI which gets us into the how-many-angels-can-dance-on-the-head-of-a-pin discussion of exactly where the outlet is. So, if you think it will end that discussion, you're wrong. It's all the worse if someone actually produces an EVSE with onboard GFCI.

Thanks for the heads up. I'll be making a comment.

 

[wwhitney]

If this change from the 2026 First Draft Report makes it through the process, it will remove a lot of discussion that has taken place about this issue.

Speaking of the process, if a change is in the First Draft, and there are no PCs on it for the Second Draft, is it still subject to a final individual vote? Or is the default outcome that it gets adopted?

Thanks,
Wayne

 

[wwhitney]

This is terrible. It will amount to a ban on hard-wired EVSEs.

How's that? Seems like it just requires GFCI breakers for all EVSEs, whether cord and plug connected or hardwired.

At least until manufacturers of EVSEs start producing models that implement CCID5 instead of CCID20, and further list them as GFCI equipment. At which point we can argue that for a hardwired EVSE, the vehicle connector is the "outlet" and so the GFCI implemented by the EVSE will suffice.

Of course, all this raises the question of whether the EV itself and its onboard charger is always designed and manufactured to have a leakage current of below 5ma. I assume concern that this may not be the case was the reason that EVSEs implement CCID20 instead of CCID5.

Cheers, Wayne

 

[don_resqcapt19]

Speaking of the process, if a change is in the First Draft, and there are no PCs on it for the Second Draft, is it still subject to a final individual vote? Or is the default outcome that it gets adopted?

Thanks,
Wayne

No individual vote if there is no action is taken in the second draft. It might be possible to submit a Notice of Intent to Make a Motion and if that would be accepted there would be debate and an individual vote on that motion at the June 2025 NFPA meeting.

 

[don_resqcapt19]

There was only one negative vote among the panel members and his comment in support of his negative vote said.

Adding language to include hard-wired connections in 625.54 does not add value. Hardwired connections do not present the same end-user shock hazard as plug-to-receptacle connections

There was no one on panel 12 that had comparability concerns with requiring all EV chargers to have GFCI protection.
However since there are large chargers that are supplied by branch circuits with a voltage to ground that exceeds 150 volts, there will have to be a change to address that.

 

[jaggedben]

How's that? Seems like it just requires GFCI breakers for all EVSEs, whether cord and plug connected or hardwired.

Perhaps I'm wrong but I was under the impression that most if not all hardwire EVSE manuals say not to use a GFCI protected circuit.

At which point we can argue that for a hardwired EVSE, the vehicle connector is the "outlet" and so the GFCI implemented by the EVSE will suffice.

I'm very much not looking forward to having that discussion with an AHJ. Between too stupid and too stubborn I think 90% will not listen to that. My second paragraph in my first reply addressed that issue.

 

[don_resqcapt19]

Perhaps I'm wrong but I was under the impression that most if not all hardwire EVSE manuals say not to use a GFCI protected circuit.



I'm very much not looking forward to having that discussion with an AHJ. Between too stupid and too stubborn I think 90% will not listen to that. My second paragraph in my first reply addressed that issue.

That is addressed in a First Revision to 110.3(B).

(B) Installation and Use.
Equipment that is listed, labeled, or both, or identified for a use shall be installed and used in accordance with any instructions included in the listing, labeling, or identification. The instructions included shall not conflict with the minimum safety requirements in this code.

 

[jaggedben]

That is addressed in a First Revision to 110.3(B).

Right! Terrible!

 

[don_resqcapt19]

Right! Terrible!

A very needed change as manufacturers continue to provide instructions that conflict with the code rules. However a lot of those are not really instructions, they just are recommendations.

 

[wwhitney]

Perhaps I'm wrong but I was under the impression that most if not all hardwire EVSE manuals say not to use a GFCI protected circuit.

That may have been true in the past, but my understanding is that now the EVSE listing standard has a ground leakage test, so an EVSE should be compatible with a GFCI protected circuit. The issue, I believe, is the EVSE's "ground assurance test" required to verify an EGC is present, but I would think that could be done with a current of 1 or 2 ma.

Cheers, Wayne

 

[jaggedben]

There was only one negative vote among the panel members and his comment in support of his negative vote said.

His comment was apt. But what's most fascinating to me is that apparently the rest of the panel members voted in favor of a PI that was apparently utterly devoid of a genuine, intelligible statement of a problem or reasoning as to why the change should be made.

 

[jaggedben]

That may have been true in the past, but my understanding is that now the EVSE listing standard has a ground leakage test, so an EVSE should be compatible with a GFCI protected circuit. The issue, I believe, is the EVSE's "ground assurance test" required to verify an EGC is present, but I would think that could be done with a current of 1 or 2 ma.

Cheers, Wayne

Okay, hopefully it's not as bad as I thought, other than the probably very unnecessary cost of GFCI breakers for consumers. And also the likely redundancy of the safety features in the breaker and the EVSE.

 

[jaggedben]

A very needed change as manufacturers continue to provide instructions that conflict with the code rules. However a lot of those are not really instructions, they just are recommendations.

To be clear, I'm not objecting to the change to 110.3(B). I'm objecting to the change to 625 that is being apparently being done without regard to what EVSE manufacturers typically say in their documentation.

 

[ramsy]

However a lot of those are not really instructions, they just are recommendations.

Are instructions in product manuals, not showing the NRTL anywhere, or similar installation guides, considered 110.3(B) listed instructions?

 

[ActionDave]

His comment was apt. But what's most fascinating to me is that apparently the rest of the panel members voted in favor of a PI that was apparently utterly devoid of a genuine, intelligible statement of a problem or reasoning as to why the change should be made.

View attachment 2572516

I think his PI was meant to have the opposite result from what he got.

 

[don_resqcapt19]

I think his PI was meant to have the opposite result from what he got.

Yes, given his posts here and on other sites, I would agree. However, he did put the words "or outlet" in the text that he submitted.

 

[jaggedben]

When I looked again I see there were actually five PIs submitted for this section, and the other four all advocated a deletion or exception or clarification with the intent of continuing to allow hardwired EVSEs to be installed without GFCI. But the CMP chose to adopt the one PI that was submitted with a one-sentence substantiation that seemed to argue against its own revision. <headslap>

The statement of problem in the PI from chargepoint is particularly interesting, especially the last few sentences.