600Vdc PV Disconnect

[pv_n00b]

What about just manual switches at the array like IMO S132-PEL64R-4. Its good for 32 amps and 2 strings.
View attachment 2572650

Since the switch is considered the RSD initiator having multiple switches is allowed but they have to be grouped together and limited to six.

 

[wwhitney]

It identifies only the circuits installed on or in buildings as requiring RSD. Not the whole PV system, just the circuits on or in a building. If a ground mount system has a circuit that enters a building then only the PV circuits on or in the building are controlled for the purpose of 690.12.

I agree with all the above, but the question is whether "circuit" is a divisible quantity or not. On first reading, I lean towards the idea that it is not, the DC circuit is the entire conductor loop from the inverter to the PV array and back to the inverter. So then 690.12 would mean that a circuit, any part of which is on or in a building, would require rapid shutdown, and 690.12(A), in specifying that "PV system DC circuits" are controlled conductors, is specifying the entire circuit.

Now if "circuit" is divisible, which is also reasonable, then you can read the first sentence of 690.12 as specifying that only the portion of the circuit "on or in" the building is subject to 690.12. Still, even under this interpretation, it would be clearer if 690.12 started off "Portions of PV system circuits installed on or in . . ."

Cheers, Wayne

 

[pv_n00b]

And that's why people have different interpretations of the NEC. To me my reading is perfectly clear. To you yours is. From there it's a numbers game, how many people adopt which interpretation.

 

[ggunn]

And that's why people have different interpretations of the NEC. To me my reading is perfectly clear. To you yours is. From there it's a numbers game, how many people adopt which interpretation.

This. One inspector I had to deal with interpreted the RSD requirements to mean that if there were 10 inverters in a rooftop PV system, he should be able to turn off the breaker to a single inverter and have all the DC to that inverter, and only that inverter, go into shutdown. Anyone who has dealt withPV systems that have multiple SMA Tripower inverters and Tigo RSD boxes knows how problematic that could be, but it is irrelevant. The RSD initiator is the PV AC disconnect, which shuts down all the inverters with their attendant DC circuits; our system was compliant.

But I lost that fight and it was a major PITA to get it working to the inspector's liking.

 

[solarken]

Unless you have an interesting local amendment to 690.13 I'm not seeing how having a manual DC disconnect in the circuit changes what is a controlled conductor. How does this work?

This is how I interpret it.
690.12(A) identifies what are controlled conductors for the purpose of rapid shutdown.
1. PV System DC circuits
2. Inverter output ckts that come from inverters within the array boundary.
For a ground mount, with inverter in basement, there are no conductors meeting 2 above.
For a ground mount with PV DC system disconnect located at the ground mounted array, there are no conductors meeting 1 above that are located on or in buildings. This is because PV system DC circuits are defined as DC conductors in PV source circuits, PV string ckts, and PV DC to DC converter ckts. PV source ckts are defined as between modules, to DC combiners, to converters, up to a DC PV system disconnecting means.
This per the definitions in 2023 NEC.
So putting dc disconnect at ground mounted array removes the conductors from being controlled.
Putting the DC disconnect on the exterior of the house also eliminates rapid shutdown requirement IMO by the Exception in 690.12(A), but I think it could be worded better.

 

[jaggedben]

I see it slightly differently...

RSD requirements only apply to conductors installed on or in buildings. That's in the first sentence of 690.12. From there, controlled conductors are subdivided into two categories: within and without the array boundary. For a groundmount, you have no conductors that are both within the array boundary and on or in a buidling, so we can dispense with worrying about those requirements. However, you've described bringing a PV DC circuit on or in a building to an inverter. So now that portion of the circuit that is on or in the building is controlled. But in my opinion, that only applies to that portion that is on or in the building. By putting a DC disconnect at the point where they come onto the building, or at a readily accessible point before that, you provide an initiation device that allows the contrlled conductors on/in the building to meet the requirements when the disconnect is opened.

One loophole that I don't think was intended is that they don't specify how far away from the building your rapid shutdown switch can be. If it's 1000ft away I think that violates the intent of the requirement, although not the letter. They should have required the initiation device to be at a "readily accessible location outside *and within sight of* the building". i.e. no more than 50ft away. If it were me, I wouldn't push your interpretation too hard with an AHJ if the groundmount were farther away than that.

 

[ggunn]

So in this case the rapid shutdown initiator isn't the service disconnect or the PV AC disconnect? When the power goes out the DC conductors into the building will remain energized; why is this not a problem?

 

[jaggedben]

So in this case the rapid shutdown initiator isn't the service disconnect or the PV AC disconnect? When the power goes out the DC conductors into the building will remain energized; why is this not a problem?

Because 690.12(C)(3) allows the RSD initiation device to be another readily accessible switch.

You have a similar sort of issue when there is battery backup. In that case you also need an initiation device that is separate from the service disconnect or (in some systems) the PV AC disconnect. The code allows this.

 

[solarken]

I see it slightly differently...

RSD requirements only apply to conductors installed on or in buildings. That's in the first sentence of 690.12. From there, controlled conductors are subdivided into two categories: within and without the array boundary. For a groundmount, you have no conductors that are both within the array boundary and on or in a buidling, so we can dispense with worrying about those requirements. However, you've described bringing a PV DC circuit on or in a building to an inverter. So now that portion of the circuit that is on or in the building is controlled. But in my opinion, that only applies to that portion that is on or in the building. By putting a DC disconnect at the point where they come onto the building, or at a readily accessible point before that, you provide an initiation device that allows the contrlled conductors on/in the building to meet the requirements when the disconnect is opened.

One loophole that I don't think was intended is that they don't specify how far away from the building your rapid shutdown switch can be. If it's 1000ft away I think that violates the intent of the requirement, although not the letter. They should have required the initiation device to be at a "readily accessible location outside *and within sight of* the building". i.e. no more than 50ft away. If it were me, I wouldn't push your interpretation too hard with an AHJ if the groundmount were farther away than that.

What about just manual switches at the array like IMO S132-PEL64R-4. Its good for 32 amps and 2 strings.

That is interesting. I have used IMO switches in the past and some manufacturers (like Enphase) use private labeled IMO switches for RSD initiation, but the ones I have encountered before were "DC Isolators" with instructions not to make or break while under load. This one you posted seems to be designed for make and break at rated current and voltage, although it is not clear to me whether it is UL certified for that operation in the US. I will definitely look into these further, as the cost is much less than similar rated safety switches.

 

[solarken]

I see it slightly differently...

However, you've described bringing a PV DC circuit on or in a building to an inverter. So now that portion of the circuit that is on or in the building is controlled.

This is where don't agree. What I am saying is the Disconnect at the ground mounted array means there is no PV DC circuit on the building. The code specifically states that controlled DC conductors are either 1) PV source circuits, 2) PV string circuits, 3) DC-DC circuits between and at the output of DC to DC converter MLPE and , 4) at the output of DC combiners. It states from those circuits TO a dc PV system disconnecting means (the DC disconnect between the array and the inverter). So the DC disconnect is the demarcation point between controlled conductors and non-controlled conductors, the way I read it. Of course it should be written more clearly and explicitly, rather than having to look at definitions. But we are not free to make up our own definition that is different than what is in section 100, so I don't think we can call the conductors after the dc disconnect as PV DC Circuit (PV System DC Circuit).

In other posts in this thread, the idea of circuit divisibility was discussed. NEC does not define "circuit" to my knowledge, but common sense suggests that circuits are indeed separate if there are certain devices between them, i.e. a Disconnect, an OCPD, etc. Just because current serially flows through conductors on either side of a disconnect does not mean they are the same "circuit", IMO. There can be different code requirements on one side or the other. So in general, I think it is proper to think of conductors going between PV modules, DC combiners, DC disconnects, and the inverter as being divisible into different "circuits" with potentially different requirements in the code, even though electrically with switches turned on it may be one big closed DC current loop.

In the end, I think we all agree that ground mount PV arrays with DC disconnect at the array means you do not need to have RSD devices like Tigo or other in that array or interrupting the DC home runs from that array, it just seems we disagree on why that is. The NEC could be improved on this section to clarify why.

 

[jaggedben]

I disagree. Your DC conductors to the inverter are still 'between modules in a PV string circuit and ... electronic power converters", the latter being your inverter. The bit about a "dc PV system disconnecting means" is intended to demarcate the end of the 'PV System' where it directly feeds DC loads via DC feeders and branch circuits (rarely done). Your PV 'system' disconnecting means is an AC disconnect, so that bit doesn't apply to your situation.

Also, unless you are on the 2023 NEC your argument clearly doesn't apply whatsoever. (The NPFA is really so incompetent. I can't believe they got rid of the perfectly good PV Output Circuit definition that had been there for decades, and replaced it with that awful grammatical construction of source circuits and the useless 'string circuit' definition. They keep fixing what isn't broken. They obviously didn't intend to create a loophole like the one you're arguing for here.)

 

[wwhitney]

Also, unless you are on the 2023 NEC your argument clearly doesn't apply whatsoever.

OH is on the 2023 NEC.

Cheers. Wayne

 

[ggunn]

Something I questioned way back in this thread is that this doesn't sound like rapid shutdown at all to me. Rapid shutdown drops the DC voltage in and on buildings when the inverter shuts down or when an initiator marked as such and grouped with the service disconnect is activated; it seems to me that an external DC disconnect doesn't do that. It may make the system NEC compliant but it does not seem to me that it has to do with rapid shutdown.

 

[jaggedben]

.... Rapid shutdown drops the DC voltage in and on buildings when the inverter shuts down or when an initiator marked as such and grouped with the service disconnect is activated; it seems to me that an external DC disconnect doesn't do that. ....

How does it not do that in this case? Also 690.12 does not require that the RSD switch be grouped with the service disconnect.

Please do not make up code rules based merely on your usual way of doing things.

 

[ggunn]

How does it not do that in this case? Also 690.12 does not require that the RSD switch be grouped with the service disconnect.

Please do not make up code rules based merely on your usual way of doing things.

Please do not.... oh, never mind.

 

[wwhitney]

rapid shutdown drops the DC voltage in and on buildings

A disconnect that is between the building and the PV array will do that, the voltage in and on the building will be 0 when the disconnect is open.

an initiator marked as such and grouped with the service disconnect

2023 NEC 690.12(C) has no such grouping requirement. All it says is "For one-and two-family dwellings, an initiation device(s), where required, shall be located at a readily accessible outdoor location."

It may make the system NEC compliant but it does not seem to me that it has to do with rapid shutdown.

What is "rapid shutdown" other than NEC 690.12? If it makes the system compliant with NEC 690.12, it's rapid shutdown.

Cheers, Wayne

 

[solarken]

I disagree. Your DC conductors to the inverter are still 'between modules in a PV string circuit and ... electronic power converters", the latter being your inverter. The bit about a "dc PV system disconnecting means" is intended to demarcate the end of the 'PV System' where it directly feeds DC loads via DC feeders and branch circuits (rarely done). Your PV 'system' disconnecting means is an AC disconnect, so that bit doesn't apply to your situation.

Also, unless you are on the 2023 NEC your argument clearly doesn't apply whatsoever. (The NPFA is really so incompetent. I can't believe they got rid of the perfectly good PV Output Circuit definition that had been there for decades, and replaced it with that awful grammatical construction of source circuits and the useless 'string circuit' definition. They keep fixing what isn't broken. They obviously didn't intend to create a loophole like the one you're arguing for here.)

I agree PV DC source circuits include conductors between modules and electronic power converters. But in my ground mount array example, there is no conductor that connects a module directly to an electronic power converter (inverter). There is a conductor between the module and the DC disconnect. That would be a controlled conductor for the purpose of 690.12 (ignoring the Exception for conductors on buildings for the moment). Then there is a conductor between the DC disconnect and the inverter. That is not a controlled conductor for the purpose of 690.12 because it is not included in the definition of a PV DC source circuit:
"(1) conductors in a string circuit, and
(2) FROM string circuits and dc combiners to dc combiners, electronic power converters, or a dc PV system disconnecting means.

(2) above to me includes:
From string ckt to dc combiner
From dc combiner to dc combiner
from string ckt to electronic power converter
from dc combiner to electronic power converter
from string circuit to dc PV system disconnecting means
from dc combiner to dc PV system disconnecting means.

No where in the above includes the ckt between the dc PV system disconnecting means and the inverter (electronic power converter).

IMO this is in alignment with the intent of rapid shutdown because there is no danger to emergency personnel working on or in a building from those conductors once the disconnect is off.

 

[wwhitney]

Then there is a conductor between the DC disconnect and the inverter. That is not a controlled conductor for the purpose of 690.12 because it is not included in the definition of a PV DC source circuit:
"(1) conductors in a string circuit, and
(2) FROM string circuits and dc combiners to dc combiners, electronic power converters, or a dc PV system disconnecting means.

Jaggedben's point is that your basic DC disconnect between the inverter and the panels is not a DC PV system disconnecting means. The inverter is part of the PV system, and so the DC disconnect does not fulfill the requirements of 690.13. That means conductors between the DC disconnect and the inverter are still controlled conductors.

Seems like the DC disconnect could serve as a rapid shutdown initiation device, it just needs to meet all the requirements in 690.12(C).

Cheers, Wayne

 

[solarken]

Something I questioned way back in this thread is that this doesn't sound like rapid shutdown at all to me. Rapid shutdown drops the DC voltage in and on buildings when the inverter shuts down or when an initiator marked as such and grouped with the service disconnect is activated; it seems to me that an external DC disconnect doesn't do that. It may make the system NEC compliant but it does not seem to me that it has to do with rapid shutdown.

I think the external DC disconnect takes what would have been controlled conductors (if disconnect were not external) for the purpose of rapid shutdown out of the picture for rapid shutdown. In that respect it has to do with rapid shutdown.

 

[solarken]

Jaggedben's point is that your basic DC disconnect between the inverter and the panels is not a DC PV system disconnecting means. The inverter is part of the PV system, and so the DC disconnect does not fulfill the requirements of 690.13. That means conductors between the DC disconnect and the inverter are still controlled conductors.

How is it not? If the disconnect were on the other side of the inverter, it is a PV system disconnect, not a DC PV system disconnect. It has to be on the DC side of the inverter to be a DC PV system disconnect, so it is before the inverter.

The conductors that they are worried about are DC conductors that are always energized, within and leaving the array, and AC conductors leaving the array but only if there were microinverters or string inverters co-located at the array that remain energized when the sun is on the array.