2020 NEC 705 question

[ggunn]

In a meeting this morning, an electrician working for a PV company for whom I am doing some consulting work challenged me on my interpretation of 705, specifically about AC inverter combiner panels. They have a 600A combiner panel which has (2) 200A and (3) 80A inverter breakers along with a couple of 15A load breakers for system monitoring, and it is connected on the line side of the service disco via a 800A PV AC disco with 700A fuses. My contention is that this is a violation of 705.12(B)(3) because the combiner busbar is not qualified under any of its provisions; the applicable code is 705.12(B)(3(3) and the sum of all the breaker ratings must be less than that of the busbar. Their busbar is 600A and the total of the breaker ratings is 670A.

His contention is that since the combiner panel is interconnected on the line side of the main service disconnect, 705.12 does not apply and that it should be under 705.11, which has no provision for qualifying the busbar, so he should be able to load it up all he wants.

It went round and round after that, but I will omit the "he said, I said" and leave it at that. I am confident in my interpretation, but as I always say, I know a lot but I don't know everything. What do you think?

 

[shortcircuit2]

The combiner panel is Load-Side of the 800-amp PV AC Disco (Service Disconnect)...and how can you fuse a 600-amp panel with 700-amp fuses?

 

[ggunn]

The combiner panel is Load-Side of the 800-amp PV AC Disco (Service Disconnect)...

That was part of the "he said , I said". I said it.

 

[jaggedben]

My opinion is that you are right and he is completely wrong.

That said, even if we were to agree by some roundabout out-of-date argument that the panel isn't on the load side of the service disconnect, he still couldn't load it up 'all he wants'. The 640A of PV breakers would have to be the result of next-size-up rounding and 125% of inverter continuous output would still have to be 600A or less.

 

[ggunn]

My opinion is that you are right and he is completely wrong.

That said, even if we were to agree by some roundabout out-of-date argument that the panel isn't on the load side of the service disconnect, he still couldn't load it up 'all he wants'. The 640A of PV breakers would have to be the result of next-size-up rounding and 125% of inverter continuous output would still have to be 600A or less.

Numbers aside, what it comes down to is if 705.12(B)(3) needs to be satisfied when inverter outputs are combined and fed through a disconnect to the service conductors. I see no electrical difference, from the perspective of the panel, between bus protection by the disconnect on the service conductors and that by a breaker in the MDP.

 

[jaggedben]

Yes definitely no electrical difference, which really should be all that matters.

 

[ggunn]

Yes definitely no electrical difference, which really should be all that matters.

Thanks. I appreciate your input, but I wish others would weigh in.

If anyone knows any reason, either documented in the NEC or from an electrical perspective, why an inverter combiner panel connected through a fused PV AC disconnect to the service conductors should NOT be required to have a bus qualified under one of the provisions of 705.12(B)(3), please tell me.

 

[wwhitney]

If anyone knows any reason, either documented in the NEC or from an electrical perspective, why an inverter combiner panel connected through a fused PV AC disconnect to the service conductors should NOT be required to have a bus qualified under one of the provisions of 705.12(B)(3), please tell me.

So, from an electrical perspective, obviously not. A "PV disconnect" is the same as a service disconnect.

However, using the text of the 2020 NEC, and playing devil's advocate, one could argue that the PV disconnect required by 705.11(C) is not a service disconnect. For example, 705.11(C) provides allowances for unfused conductors inside the building that are in excess of what is allowed for service conductors. In that case, 705.12 would not apply, as then there is no service disconnecting means between the PV inverters and the utility.

I don't agree with that argument, but it's not 100% clear under the 2020 NEC. I believe the 2023 NEC clarified this by eliminating the extra allowances in 705.11(C) and basically saying the PV disconnect is a service disconnect.

Cheers, Wayne

 

[jaggedben]

The argument that the PV disconnect is not a service disconnect lost its biggest supporting premise in the 2017 cycle, when the definition of a service was changed to be agnostic about the direction of energy flow. (A service now delivers energy 'between' utility and premises rather than 'from' and 'to' them respectively.) So the argument that the PV disconnect is not a service disconnect based on direction of energy flow can no longer be made.

That leaves arguments that for example phrasing such as 'on the supply side of the service disconnecting means' implies that service disconnecting means only exist on other circuits. This is weak implication IMO and not fully logical since the supply-side connection is still on the supply-side of all service disconnects if the PV disconnect is a service disconnect. It is at best reading between the lines.

Add to this that in 2023 the CMP expanded the phrasing in 705.11 and 705.11(C) to allude more directly to article to 230, and the argument has less going for it than ever.

 

[ggunn]

So, from an electrical perspective, obviously not. A "PV disconnect" is the same as a service disconnect.

However, using the text of the 2020 NEC, and playing devil's advocate, one could argue that the PV disconnect required by 705.11(C) is not a service disconnect. For example, 705.11(C) provides allowances for unfused conductors inside the building that are in excess of what is allowed for service conductors. In that case, 705.12 would not apply, as then there is no service disconnecting means between the PV inverters and the utility.

I don't agree with that argument, but it's not 100% clear under the 2020 NEC. I believe the 2023 NEC clarified this by eliminating the extra allowances in 705.11(C) and basically saying the PV disconnect is a service disconnect.

Cheers, Wayne

Up front, this is not a rebuttal to your post.

IMO, interpretation of the NEC should be enhanced by the application of critical thinking. I said this earlier, but from the perspective of the panel busbar I see no difference between the protection provided it from current feeding a fault from the utility by a breaker in an MDP or by a fused disconnect on the service conductors.

The PV AC disconnect itself is under 705.11, but the individual inverters on the busbar are all connected on the load side of the disco OCP, so 705.12 applies in the panel to which they are connected. An individual inverter connected to the service conductors is under 705.11 but not an array of them in a combiner panel protected by a fused disconnect.

The electrician, again IMO, is attempting to exploit what he sees as a loophole in the NEC, which I agree can be seen as an ambiguity, but which I do not agree should be taken advantage of. Unless someone can convince me that his interpretation is the correct one, I cannot seal the design.

 

[wwhitney]

Up front, this is not a rebuttal to your post.

Right, I don't buy the argument I posted. You were just looking for contrary arguments, so I gave it a shot. Here's a bit more along those lines:

The crux of the matter is whether the "overcurrent devices" required by 705.11(C) for the "power source output circuit conductors" connected to the supply side of the service disconnecting means in accordance with 705.11 and 230.82(6) is a "service disconnecting means" for the purpose of applying 705.12.

230.70, which requires service disconnecting means, says "Means shall be provided to disconnect all ungrounded conductors in a building or other structure from the service conductors." So if the "power source output circuit conductors" are ever "in" the building, they need a 230.70 service disconnecting means, and 705.12 would apply on the downstream (non-utility) side of that service disconnecting means. The 705.11(C) "overcurrent devices" would typically be that required 230.70 service disconnecting means.

But if the "power source output circuit conductors" are all outside the building, seems like 230.70 does not require a service disconnecting means, and it is plausible to say that the 705.11(C) "overcurrent devices" are not a service disconnecting means. This is supported somewhat by the language in 230.82 "Equipment Connected to the Supply Side of Service Disconnect." Contrast the subsections:

(5) Conductors used to supply load management devices, circuits for standby power systems, fire pump equipment, and fire and sprinkler alarms, if provided with service equipment and installed in accordance with requirements for service entrance conductors.
(6) Solar photovoltaic systems, fuel cell systems, wind electric systems, energy storage systems, or interconnected electric power production sources, if provided with a disconnecting means listed as suitable for use as service equipment, and overcurrent protection as specified in Part VII of Article 230.

First note that the language at the end of (5) "provided with service equipment and installed in accordance with requirements for service entrance conductors" sounds a lot more like "install it like a service" than the language at the end of (6). Second note that (6) requires a disconnect "listed as suitable for use as service equipment," rather than requiring a disconnect "as specified in Part VI of Article 230," which is the Part that includes 230.70 requiring service disconnecting means.

The upshot is that when the power source output circuit conductors are never "in" the building, it is plausible to say that the 230.82(6) disconnecting means, and the 705.11(C) "overcurrent devices" are not "service disconnecting means" for the purposes of applying 705.12.

Cheers, Wayne


Cheers, Wayne

 

[ggunn]

Right, I don't buy the argument I posted. You were just looking for contrary arguments, so I gave it a shot. Here's a bit more along those lines:

The crux of the matter is whether the "overcurrent devices" required by 705.11(C) for the "power source output circuit conductors" connected to the supply side of the service disconnecting means in accordance with 705.11 and 230.82(6) is a "service disconnecting means" for the purpose of applying 705.12.

230.70, which requires service disconnecting means, says "Means shall be provided to disconnect all ungrounded conductors in a building or other structure from the service conductors." So if the "power source output circuit conductors" are ever "in" the building, they need a 230.70 service disconnecting means, and 705.12 would apply on the downstream (non-utility) side of that service disconnecting means. The 705.11(C) "overcurrent devices" would typically be that required 230.70 service disconnecting means.

But if the "power source output circuit conductors" are all outside the building, seems like 230.70 does not require a service disconnecting means, and it is plausible to say that the 705.11(C) "overcurrent devices" are not a service disconnecting means. This is supported somewhat by the language in 230.82 "Equipment Connected to the Supply Side of Service Disconnect." Contrast the subsections:

(5) Conductors used to supply load management devices, circuits for standby power systems, fire pump equipment, and fire and sprinkler alarms, if provided with service equipment and installed in accordance with requirements for service entrance conductors.
(6) Solar photovoltaic systems, fuel cell systems, wind electric systems, energy storage systems, or interconnected electric power production sources, if provided with a disconnecting means listed as suitable for use as service equipment, and overcurrent protection as specified in Part VII of Article 230.

First note that the language at the end of (5) "provided with service equipment and installed in accordance with requirements for service entrance conductors" sounds a lot more like "install it like a service" than the language at the end of (6). Second note that (6) requires a disconnect "listed as suitable for use as service equipment," rather than requiring a disconnect "as specified in Part VI of Article 230," which is the Part that includes 230.70 requiring service disconnecting means.

The upshot is that when the power source output circuit conductors are never "in" the building, it is plausible to say that the 230.82(6) disconnecting means, and the 705.11(C) "overcurrent devices" are not "service disconnecting means" for the purposes of applying 705.12.

Cheers, Wayne


Cheers, Wayne

Wouldn't that logic also say that any fused disconnect on the outside of a building that feeds a panel which is also outside the building would not be considered a service disconnecting means?

 

[wwhitney]

Wouldn't that logic also say that any fused disconnect on the outside of a building that feeds a panel which is also outside the building would not be considered a service disconnecting means?

Yes, under that logic, the use of the word "in" in 230.70 means that if you had a service that is connected only to equipment that is not "in" any building or other structure, 230.70 would not mandate any service disconnecting means.

But now I'm going to demolish that argument. 240.21 requires that OCPD be located at the point of supply of conductors. The 705.11 connection could possibly comply with this by either intercepting the service conductors with an OCPD with double line side lugs (for incoming and outgoing), or via an MLO service panel, or a 2020 NEC compliant version thereof with multiple compartments, one for each OCPD. Otherwise, the only way to comply with 240.21 is to rely on the allowance of 240.21(D) for Service Conductors. You can't use 240.21(B) as the service conductors are not Feeders.

So if you have any conductors between the (usual) service conductors and the 705.11(C) OCPD, they must be service conductors to comply with 240.21. That means the 705.11(C) OCPD has to be a service disconnect per the definition of "Service Conductors." Therefore 705.12 applies downstream (away from the utility) from any such 705.11(C) OCPD / service disconnect.

This line of thought makes me wonder how the "line side tap conductors aren't service conductors" idea ever held any water, even prior to the 2020 or 2023 NEC.

Cheers, Wayne

 

[wwhitney]

To put the argument more simply: let's call any premises wiring system conductors that are connected to the utility and which have no SC/GF protection within the premises wiring system "unfused". The only way in which "unfused" conductors can comply with 240.21 is under 240.21(D), which requires them to be Service Conductors. And the definition of Service Conductors requires them to terminate at a Service Disconnecting Means. So every OCPD supplied by "unfused" conductors is a Service Disconnecting Means.

Cheers, Wayne

 

[ggunn]

To put the argument more simply: let's call any premises wiring system conductors that are connected to the utility and which have no SC/GF protection within the premises wiring system "unfused". The only way in which "unfused" conductors can comply with 240.21 is under 240.21(D), which requires them to be Service Conductors. And the definition of Service Conductors requires them to terminate at a Service Disconnecting Means. So every OCPD supplied by "unfused" conductors is a Service Disconnecting Means.

Cheers, Wayne

Logically put. In that view, then, any AC inverter combiner panel that is connected to such service conductors via an OCPD would have to comply with one of the provisions of 705.12(B)(3) because the individual inverters connected to it are load side interconnected. That makes sense to me.

 

[electrofelon]

I agree that the combiner panel follows load side rules. I think it is absurd to say that just because there is a line side connection, that that means everything downstream from that you can just throw away all the load side busbar rules. Also, as was mentioned he has a pretty blatant 408.36 violation.

 

[shortcircuit2]

A "PV disconnect" is the same as a service disconnect.
I believe the 2023 NEC clarified this by eliminating the extra allowances in 705.11(C) and basically saying the PV disconnect is a service disconnect.

Cheers, Wayne

AGREED!
Service Disconnect is a Service Disconnect is a Service Disconnect

So a "PV Disconnect" is a "Service Disconnect"...and needs to comply with rules in the Code that apply to both.

Are the conductors between the PV Disconnect/Service Disconnect "Feeder Conductors" or "Power Souce Output Conductors" or both when they termintate in that AC Combiner Panel?

 

[shortcircuit2]

To put the argument more simply: let's call any premises wiring system conductors that are connected to the utility and which have no SC/GF protection within the premises wiring system "unfused". The only way in which "unfused" conductors can comply with 240.21 is under 240.21(D), which requires them to be Service Conductors. And the definition of Service Conductors requires them to terminate at a Service Disconnecting Means. So every OCPD supplied by "unfused" conductors is a Service Disconnecting Means.

Cheers, Wayne

Agree.
The OP is referring to a commercial installation. 2023 section 230.70(A)(1) applies.

Consider the arragement a one-family dwelling installation. 2023 section 230.70(A)(1) applies.

Also, 2023 Code removed the 2020 10ft allowance in 705.11(C)(1) in dwelling units. It added new section 705.11(D) which requires the "Service Disconnect" to comply with Parts VI through VII of Article 230.

Section 230.85 is in Part VI of Article 230 which means supply-side source connections ahead of the normal service disconnect for a dwelling in accordance with 230.82(6) need to terminate in a 230.85 Emergency Disconnect in an "outdoor location" per 230.85(A)(1). No more taps inside the dwelling.

 

[Tulsa Electrician]

Does the PV combiner panel have a 600 amp main in it?
Or is it main lug.
No mention of a main. Just want to make sure as I read along.
Thanks

 

[jaggedben]

He said it was connected to the service through a disconnect with 700A fuses.

The use of 700A fuses with a 600A panel is a tell that someone just didn't keep track of what they were doing.