Where's the "Outlet" under the 2020 NEC -- for outdoor hardwire EVSEs

[brycenesbitt]

Can't say every mfg is incorporating this or not but everyone that I've installed has an incorporated an interlock that will de-energize the output cable if it is not connected to the vehicle. This it seems would then make a moot point for GFCI for the hardwired EVSE...

That interlock is required by UL 2231 and IEC 62196. So they're all going to have it. The power is OFF on an EVSE unless the car and EVSE have a complicated handshake via the pilot signal. The issue of GFCI is indeed moot: I hope the code writers realize this.

 

[retirede]

That interlock is required by UL 2231 and IEC 62196. So they're all going to have it. The power is OFF on an EVSE unless the car and EVSE have a complicated handshake via the pilot signal. The issue of GFCI is indeed moot: I hope the code writers realize this.

I would t call it a complicated handshake. The EVSE simply detects whether or not there is a connection to the control pilot (CP) pin. If the detection is successful, it’s plugged in and allowed to energize.

The GFCI protection could be useful in the case of a damaged charging cable.

 

[brycenesbitt]

The GFCI protection could be useful in the case of a damaged charging cable.

Ok, it's not super complicated. But there are more steps: including a cable test. And if that fails, in the case of a damaged charging cable the EVSE under UL 2231 and IEC 62196 has GFCI protection. The second GFCI is not just redundant, but harmful, in my view.
Note: Toyota and others are switching from J1772 to the NACS connector, which will bring a more sophisticated handshake.

 

[jaggedben]

What does the connector have to do with the handshake? I thought there were connector adapters so clearly there is some cross-compatability there.

 

[wwhitney]

Ok, it's not super complicated. But there are more steps: including a cable test. And if that fails, in the case of a damaged charging cable the EVSE under UL 2231 and IEC 62196 has GFCI protection. The second GFCI is not just redundant, but harmful, in my view.

Again, in the context of an NEC forum, the NEC definition of GFCI should apply. Most (almost all?) EVSEs do not implement GFCI under that definition. They implement CCID20, which has a 20ma trip threshold.

Note: Toyota and others are switching from J1772 to the NACS connector, which will bring a more sophisticated handshake.

I thought NACS as a connector is protocol agnostic; surely they will implement the usual J1772 protocol over NACS. They could implement a superset backwards-compatible protocol as well, but doing that would be independent of the physical connector.

Cheers, Wayne

 

[brycenesbitt]

Again, in the context of an NEC forum, the NEC definition of GFCI should apply. Most (almost all?) EVSEs do not implement GFCI under that definition. They implement CCID20, which has a 20ma trip threshold.

That's indeed the problem. The NEC is poorly meshing with UL 2231 and IEC 62196. That should be fixed in the NEC, lest it affect the reliability and reputation of EV charging among the public.

 

[wwhitney]

That's indeed the problem. The NEC is poorly meshing with UL 2231 and IEC 62196. That should be fixed in the NEC, lest it affect the reliability and reputation of EV charging among the public.

Maybe. The CMP in charge of Article 625 specifically added 625.54 requiring GFCI protection for receptacles supplying EVSEs because of their concern over the use of portable EVSEs. The user would be interacting with the receptacle to plug-in and unplug their portable EVSE. In which case GF detection within the EVSE itself isn't going to protect the user at all. So I doubt the CMP will agree with you.

You could try submitting a PI to add an exception to 625.54 for single receptacles that supply an EVSE that is permanently fastened in place. But you'll have to wait 2.5 years to do that, until submissions for the 2029 NEC open up.

Cheers, Wayne

 

[don_resqcapt19]

That's indeed the problem. The NEC is poorly meshing with UL 2231 and IEC 62196. That should be fixed in the NEC, lest it affect the reliability and reputation of EV charging among the public.

There are lots of ongoing issues between the product standards and the NEC, mostly because the standards writers think that they are more important than the NEC, however the AHJ can only enforce the code as adopted, nothing in a product standard can change that.
As far as fixing it in the NEC, there was nothing submitted to make changes in this area for the 2026 code. There were a number of PIs associated with more clearly defining an outlet. The next chance to make a change in the NEC will be for the 2029 code, however the process to make a change in a product standard takes about the same amount of time as it does to make a change in the NEC.

 

[don_resqcapt19]

Maybe. The CMP in charge of Article 625 specifically added 625.54 requiring GFCI protection for receptacles supplying EVSEs because of their concern over the use of portable EVSEs. The user would be interacting with the receptacle to plug-in and unplug their portable EVSE. In which case GF detection within the EVSE itself isn't going to protect the user at all. So I doubt the CMP will agree with you.

You could try submitting a PI to add an exception to 625.54 for single receptacles that supply an EVSE that is permanently fastened in place. But you'll have to wait 2.5 years to do that, until submissions for the 2029 NEC open up.

Cheers, Wayne

And there are multiple members that represent standards writing organizations on that code making panel and as I recall they did not raise the GFCI requirement as an issue.

 

[jaggedben]

Maybe. The CMP in charge of Article 625 specifically added 625.54 requiring GFCI protection for receptacles supplying EVSEs because of their concern over the use of portable EVSEs. The user would be interacting with the receptacle to plug-in and unplug their portable EVSE. In which case GF detection within the EVSE itself isn't going to protect the user at all. So I doubt the CMP will agree with you.

You could try submitting a PI to add an exception to 625.54 for single receptacles that supply an EVSE that is permanently fastened in place. But you'll have to wait 2.5 years to do that, until submissions for the 2029 NEC open up.

Cheers, Wayne

The discussion upthread was about hard wired EVSEs with no receptacle involved. It makes sense to require a GFCI receptacle where the installer cannot control what a user plugs into it. It makes far less sense to require a GFCI breaker if a hardwired EVSE meets an acceptable standard. The code should leave open the latter option, one way or another.

 

[wwhitney]

The discussion upthread was about hard wired EVSEs with no receptacle involved. It makes sense to require a GFCI receptacle where the installer cannot control what a user plugs into it. It makes far less sense to require a GFCI breaker if a hardwired EVSE meets an acceptable standard. The code should leave open the latter option, one way or another.

OK, so now we're back to 210.8(F). The CMP that added it decided the acceptable standard is a Class A GFCI. There are basically zero EVSEs that comply with that standard, so the result is that a GFCI breaker is required, regardless of the answer to the titular question of this thread.

You or the OP could try a PI to add an exception for EVSEs, saying that CCID20 would be acceptable. Then if the outlet location is also clarified to be the vehicle connector, it would be fine to use a regular breaker.

Cheers, Wayne

 

[jaggedben]

OK, so now we're back to 210.8(F). The CMP that added it decided the acceptable standard is a Class A GFCI. There are basically zero EVSEs that comply with that standard, so the result is that a GFCI breaker is required, regardless of the answer to the titular question of this thread.

You or the OP could try a PI to add an exception for EVSEs, saying that CCID20 would be acceptable. Then if the outlet location is also clarified to be the vehicle connector, it would be fine to use a regular breaker.

Cheers, Wayne

Yeah something like that.

But I will go back to what I stated at the end of post 52:

I also think the trend of requiring "outlets" to have GFCI is a bad trend that is going to lead to further confusion and controversy of the kind we're witnessing here. It was started with 210.8(F) and now unfortunately has been expanded to change how 422.5 handles things, in a bad way that reduces options for manufacturers and installers. It doesn't appear to be being done in coordination with manufacturers or UL and I can ascertain for no good reason for it. It used to be clear that equipment "at the outlet" could provide the required protection (without having to ask whether it's part of the premises wiring or where exactly the outlet is). Now that's getting muddied.

If I have the time, I will make a comment on every 2026 first revision that furthers this nasty trend. Code requirements should simply state exactly which equipment can contain the required protection, as they typically did before the last couple cycles. This whole devolution into philosophizing about the location of the outlet does a disservice to providing clear requirements.

 

[tortuga]

I'll weigh in that an EVSE is not utilization equipment. I think of them like a control panel, motor controller, VFD or transformer. I think its reasonable to use art 400 or a chapter 3 wiring method to go from an EVSE to the outlet.

Since UL 2331 used the term CCID (charge current interrupting device), that would be a "UL 2331 compliant default off CCID".

Thanks for the pointer to UL 2331. UL 2331-2 (Table 6) does specify the trip time allowed for a CCID5, and it matches the trip time allowed for a Class A GFCI under UL 943, (20/I)^1.43 seconds, with I in mA. So now the discussion is a whole lot less academic:

Under the 2020 NEC, there is no requirement for GFCIs to be listed. Per the definition in Article 100 (which just references trip time), a UL 943 Class A GFCI and a UL 2331-2 CCID5 are both "GFCI". So if we have an EVSE with CCID5, now the location of the outlet is determinative as to whether an outdoor hardwired installation at a dwelling unit complies with 210.8(F).

The 2023 NEC added the requirement to 210.8 that the GFCI be a "listed Class A GFCI". So for the outlet location to matter under the 2023 NEC, we'd need a CCID5 EVSE that also is listed under UL 943 to provide Class A GFCI.

Yes its required to be listed but not UL listed and nothing in the NEC definition requires it to be 'UL 943' listed. If a competing IEC standard or another standard were to emerge that had an equivalent result then I think it would be up to an AHJ at that point. I think this is why the NEC push to have mor and more common things listed like a zip tie is a problem.

 

[retirede]

Ok, it's not super complicated. But there are more steps: including a cable test. And if that fails, in the case of a damaged charging cable the EVSE under UL 2231 and IEC 62196 has GFCI protection. The second GFCI is not just redundant, but harmful, in my view.
Note: Toyota and others are switching from J1772 to the NACS connector, which will bring a more sophisticated handshake.

Cable damage that merely exposes the conductors without severing them will pose a hazard that won’t be detected by the EVSE’s cable test.

 

[brycenesbitt]

Cable damage that merely exposes the conductors without severing them will pose a hazard that won’t be detected by the EVSE’s cable test.

Right, but we just extensively went over that the EVSE has an earth leakage/imbalance device

• Fault sensitivity of Min 15ma - Max 20ma trip for protection against electric shock of personnel. (NEC 625.22) (SAE J1772) (UL 2231)
• Retry charging up to 4 times after a 15 minute delay per event. (UL 2231)

It's the EVSE's job to monitor it's own freaking output cable. Putting an upstream GFCI on that is not helping the situation.
-
The NEC has muddied the waters. There was no need to mess with the global standards for this type of equipment. I'll do what I can to get the NEC to leave well enough alone for the next code cycle -- it may be an impossible tilting at windmills but I'll try. It's too important that these things work and not cut off in random ways that undermine people's transportation.
-
Imagine if the NEC required a gas vapor detection device on any gas fuel car with an electrical system, that then drained the gas tank if it semelled something. That's the kind of impact that turning off a charger has.
-
Note that UL 2231 retry requirement: that's sensible given the potential for water in the cable or ice or whatever. Give it a few cautious tries, and alert the owner on their cell phone if the vehicle can't charge. The EVSE is in a far better position to communicate the problem to the operator, compared to some pre-Internet dinosaur of a "listed" GFCI device.
-
That upstream GFCI trips and the EVSE can't even call home to complain. That's not the way to run a modern computer controlled smart device. The NEC committees don't seem to have seen this from a very modern lens yet.

 

[shortcircuit2]

"Utilization Equipment: Equipment that utilizes electric energy for electronic ...[etc.] or similar purposes."
A GFCI receptacle plainly utilizes electric energy for electronic purposes, and therefore fits under those definitions.
As retirede pointed out in post #32, "The NEC does not limit what comprises utilization equipment based on the power it consumes."

Interesting.
This comment gives credence to my opinion that an Inverter is an outlet. And in many old arguments on this forum that a PV Disconnect was not Service Equipment, because it did not feed power to premises wiring and an inverter was not an outlet. My position was that an Inverter did consume power for electronic purposes.

So an Inverter is an outlet and may require GFCI under 210.8(F)

A little off the OP in #1, but may I add this to the discussion?

 

[retirede]

My position was that an Inverter did consume power for electronic purposes.

If you take that position to its end, every control device with electronics becomes utilization equipment. Think a simple dimmer switch or occupancy sensor.

 

[jaggedben]

Interesting.
This comment gives credence to my opinion that an Inverter is an outlet. And in many old arguments on this forum that a PV Disconnect was not Service Equipment, because it did not feed power to premises wiring and an inverter was not an outlet. My position was that an Inverter did consume power for electronic purposes.

So an Inverter is an outlet and may require GFCI under 210.8(F)

A little off the OP in #1, but may I add this to the discussion?

Once we spend the entire semester of our second year college philosophy course on the conceptual analysis of 'supply' in the definition of an outlet, then it may still depend on the particular inverter. Some use AC power for startup, some don't.

More seriously, I think it's just another piece of evidence that the linguistic construction in 210.8(F) ('outlet shall have GFCI') is just way too broad and sweeping an overreach, and wasn't coordinated with the industry. 90.4 paragraph 3 if we have to.

 

[don_resqcapt19]

Once we spend the entire semester of our second year college philosophy course on the conceptual analysis of 'supply' in the definition of an outlet, then it may still depend on the particular inverter. Some use AC power for startup, some don't.

More seriously, I think it's just another piece of evidence that the linguistic construction in 210.8(F) ('outlet shall have GFCI') is just way too broad and sweeping an overreach, and wasn't coordinated with the industry. 90.4 paragraph 3 if we have to.

Have you ever got a written special permission based on 90.4(C)? I believe they are very rare.

 

[jaggedben]

Have you ever got a written special permission based on 90.4(C)? I believe they are very rare.

90.4(C)?
I believe the answer to your question is no, but what I was talking about doesn't require written permission per the code.
I should have said 90.4 paragraphs 2 or 3.