DC Disconnect at Readily Accessible Location

Status
Not open for further replies.

shortcircuit2

Senior Member
Location
South of Bawstin
2014 NEC 690.13(A) requires a DC disconnect at a readily accessible location nearest the point of entry of PV system conductors.

Does the exception allow us to blow by the nearest readily accessible location and put the DC disconnect further inside the building?
 

Carultch

Senior Member
Location
Massachusetts
2014 NEC 690.13(A) requires a DC disconnect at a readily accessible location nearest the point of entry of PV system conductors.

Does the exception allow us to blow by the nearest readily accessible location and put the DC disconnect further inside the building?


Are you sure you have the correct reference?

In the 2014 NEC, 690.13 is about arc fault protection of DC circuits. And the contentious 690.12 is about the 10 ft (or even 5 ft) limit on energized conductors leaving the vicinity of the array. It is unclear how this distance needs to be measured in the wording of the code.

690.12 would nullify any requirement of needing an additional DC disconnect where there (A) isn't a combiner, (B) isn't an inverter, and (C) isn't already a means of meeting rapid shutdown.
 

shortcircuit2

Senior Member
Location
South of Bawstin
Are you sure you have the correct reference?

In the 2014 NEC, 690.13 is about arc fault protection of DC circuits. And the contentious 690.12 is about the 10 ft (or even 5 ft) limit on energized conductors leaving the vicinity of the array. It is unclear how this distance needs to be measured in the wording of the code.

690.12 would nullify any requirement of needing an additional DC disconnect where there (A) isn't a combiner, (B) isn't an inverter, and (C) isn't already a means of meeting rapid shutdown.

Yes 690.13(A)...690.11 is arc fault.

690.12 Rapid Shutdown and 690.13(A) DC Disconnect are separate requirements.
 

jaggedben

Senior Member
Location
Northern California
Occupation
Solar and Energy Storage Installer
2014 NEC 690.13(A) requires a DC disconnect at a readily accessible location nearest the point of entry of PV system conductors.

Does the exception allow us to blow by the nearest readily accessible location and put the DC disconnect further inside the building?

Yes.

Edit: Note that the exception has a misprint in the printed version. It's supposed to refer to 690.31(G), not (F). Refer to previous codes for corroboration. I don't know if the NFPA has published an official errata on this, but if not then contact them and ask them to.
 
Last edited:

jaggedben

Senior Member
Location
Northern California
Occupation
Solar and Energy Storage Installer
690.12 would nullify any requirement of needing an additional DC disconnect where there (A) isn't a combiner, (B) isn't an inverter, and (C) isn't already a means of meeting rapid shutdown.

690.12 doesn't nullify any requirements in 690.13. I suppose one could argue that if the Rapid Shutdown control is readily accessible, and results in disconnecting the conductors at a location that complies with 690.13, then both sections' requirements are met. But 690.12 doesn't require the Rapid Shutdown control to be readily accesible, so if it is not, then you still need the 690.13 disco.

FWIW, around here both the practice and the AHJs expectation is that we install everything according to 690.31(G) (misprint in the code where it says 'F' in 690.13) and the integrated DC disconnect counts for meeting 690.13. We're still on the 2011 version, but these sections have only been re-numbered, not otherwise changed.
 
Last edited:

Carultch

Senior Member
Location
Massachusetts
690.12 doesn't nullify any requirements in 690.13. I suppose one could argue that if the Rapid Shutdown control is readily accessible, and results in disconnecting the conductors at a location that complies with 690.13, then both sections' requirements are met. But 690.12 doesn't require the Rapid Shutdown control to be readily accesible, so if it is not, then you still need the 690.13 disco.


That's kind of my point. If you install rapid shutdown with an initiation device that is readily accessible (anywhere), then you meet 690.13 without even doing anything in particular to meet it.

I can't think of a reason why you wouldn't install rapid shutdown at a non-readily accessible location. Other than intentionally ignoring the intent of 690.12, because 690.12 says nothing where it has to be.
 

shortcircuit2

Senior Member
Location
South of Bawstin
Yes.

Edit: Note that the exception has a misprint in the printed version. It's supposed to refer to 690.31(G), not (F). Refer to previous codes for corroboration. I don't know if the NFPA has published an official errata on this, but if not then contact them and ask them to.

NFPA has published Errata on this one...http://www.nfpa.org/Assets/files/AboutTheCodes/70/Errata70-14-4.pdf

Since the language of 690.31(G) has been in the Code (2005 690.31(E) required raceways), I've always interpreted the section to allow a means to run the DC inside a building to a location where the DC Disconnecting Means could be installed at a readily accessible location...and as soon as you got to that location, where ever it may be. I didn't interpret this section to allow installation of DC conductors to extend past that 1st readily accessible location. I thought it gave an installer flexibility where penetrations from a roof where made, and the DC conductors where routed to the 1st readily accessible location for the DC Disconnect. The 2011 NEC added MC for even more flexibility.

Why even have the language of 690.13(A) if 690.31(G) allows the DC to run extensively through a building?

IMO, 690.31(G) doesn't allow you to blow by that 1st readily accessible location.
 

jaggedben

Senior Member
Location
Northern California
Occupation
Solar and Energy Storage Installer
NFPA has published Errata on this one...http://www.nfpa.org/Assets/files/AboutTheCodes/70/Errata70-14-4.pdf

Since the language of 690.31(G) has been in the Code (2005 690.31(E) required raceways), I've always interpreted the section to allow a means to run the DC inside a building to a location where the DC Disconnecting Means could be installed at a readily accessible location...and as soon as you got to that location, where ever it may be. I didn't interpret this section to allow installation of DC conductors to extend past that 1st readily accessible location. I thought it gave an installer flexibility where penetrations from a roof where made, and the DC conductors where routed to the 1st readily accessible location for the DC Disconnect. The 2011 NEC added MC for even more flexibility.

I don't see how your getting that from the language of the code. It's just not written that way. To my mind, the idea is that it's not safe to run PV circuits inside of buildings unless a) they can be de-energized by a disco at a readily accessible location, or b) they are run with a restricted set of wiring methods that offer more protection. If you do (a) or (b), the other is not required.

Why even have the language of 690.13(A) if 690.31(G) allows the DC to run extensively through a building?

Because: wiring methods. You can run PV circuits in Romex inside a building if that wiring can be disconnected from the PV array at a readily accessible location. I would never do it, but that's what the code allows.

IMO, 690.31(G) doesn't allow you to blow by that 1st readily accessible location.

Sure it does. In fact, if you look at the following portion it's really clear that that was the intent.

...from the point of penetration of the surface of the building or structure to the first readily accessible disconnecting means. The disconnecting means shall comply with 690.14(A), (B). and (D). ...

2011 section references (sorry, my 2014 is at home). Note that 690.14(C), the requirement to be outside or nearest the entry, is not included.
 

shortcircuit2

Senior Member
Location
South of Bawstin
So then in your interpretation...we can run PV DC conductors inside a building in a raceway from the roof down through the building passing by many readily accessible locations for the DC Disconnect and exit the building again locating the Inverter with its DC Disconnect on the outside of the building? Doesn't seem like a safe practice even in a raceway.

Under your 2011 Code these DC conductors are live with the DC Disconnect off at the inverter and they run all through the building.

Under 2014 these DC circuits can be >1000volts DC.
 

iwire

Moderator
Staff member
Location
Massachusetts
So then in your interpretation...we can run PV DC conductors inside a building in a raceway from the roof down through the building passing by many readily accessible locations for the DC Disconnect and exit the building again locating the Inverter with its DC Disconnect on the outside of the building?

I don't know about going back outside with it although I do not see why not.

Of the few PV jobs I have run I ran the DC inside the building in EMT about 250' to the inverter located inside the building. No disconnect until it reached the inverter.

This passed the NEC inspection and the engineers inspection for the MA rebate program. However this was under the 2008 NEC so it's possible things changed.
 

shortcircuit2

Senior Member
Location
South of Bawstin
2011 section references (sorry, my 2014 is at home). Note that 690.14(C), the requirement to be outside or nearest the entry, is not included.

I see they didn't include 690.14(C) in the 2011 NEC 690.31(E).

Then we don't need to follow 690.14(C)(2) Marking, 690.14(C)(3) Suitable for Use, 690.14(C)(4)Maximum Number of Disconnects or 690.14(C)(5) Grouping... either?

2011 NEC references above...
 

jaggedben

Senior Member
Location
Northern California
Occupation
Solar and Energy Storage Installer
So then in your interpretation...we can run PV DC conductors inside a building in a raceway from the roof down through the building passing by many readily accessible locations for the DC Disconnect and exit the building again locating the Inverter with its DC Disconnect on the outside of the building? Doesn't seem like a safe practice even in a raceway.

Under your 2011 Code these DC conductors are live with the DC Disconnect off at the inverter and they run all through the building.

Under 2014 these DC circuits can be >1000volts DC.

Do the requirements in 690.31(G) not make it safe enough? By all means if you think that the disconnect makes things so much safer that you just feel you should install it no matter what, then install it. But I think the code says, in a nutshell, that 690.31(G) makes it safe.

I see they didn't include 690.14(C) in the 2011 NEC 690.31(E).

Then we don't need to follow 690.14(C)(2) Marking, 690.14(C)(3) Suitable for Use, 690.14(C)(4)Maximum Number of Disconnects or 690.14(C)(5) Grouping... either?

2011 NEC references above...

Away from the book still... will look again later.
 

jaggedben

Senior Member
Location
Northern California
Occupation
Solar and Energy Storage Installer
I see they didn't include 690.14(C) in the 2011 NEC 690.31(E).

Then we don't need to follow 690.14(C)(2) Marking, 690.14(C)(3) Suitable for Use, 690.14(C)(4)Maximum Number of Disconnects or 690.14(C)(5) Grouping... either?

2011 NEC references above...

Back with the books...

So yes, that's the way they wrote it in 2011. And clearly they screwed up, because if you look at 2014 they fixed the language with respect to marking and suitability. But they confirmed leaving out 690.13(A) [2014 reference]. So the first readily accessible disconnect does not need to meet those location requirements if you install the wiring according to 690.31(G). That was clearly the intent of 2011, to leave out that particular requirement.

My take is that the requirements for no more than six disconnects apply generally for disabling the system, regardless of the location, in both 2011 and 2014; i.e. somewhere you have to have a means to disconnect the system, grouped in one location, consisting of no more than six handles. And I'd say that in 2011 the marking requirement applies generally; i.e. you have to have a disconnect somewhere that is marked. Absolving the remotely located disconnect from the suitable for use requirement in 2011 was clearly a mistake, although one could fall back to 110.3 to make sure that it is suitable.
 

Carultch

Senior Member
Location
Massachusetts
Back with the books...

So yes, that's the way they wrote it in 2011. And clearly they screwed up, because if you look at 2014 they fixed the language with respect to marking and suitability. But they confirmed leaving out 690.13(A) [2014 reference]. So the first readily accessible disconnect does not need to meet those location requirements if you install the wiring according to 690.31(G). That was clearly the intent of 2011, to leave out that particular requirement.

My take is that the requirements for no more than six disconnects apply generally for disabling the system, regardless of the location, in both 2011 and 2014; i.e. somewhere you have to have a means to disconnect the system, grouped in one location, consisting of no more than six handles. And I'd say that in 2011 the marking requirement applies generally; i.e. you have to have a disconnect somewhere that is marked. Absolving the remotely located disconnect from the suitable for use requirement in 2011 was clearly a mistake, although one could fall back to 110.3 to make sure that it is suitable.

I thought the "six disconnect rule" only applies to service disconnects. And that it is unrelated to just about anything on the DC side.

I'm thinking large commercial applications, probably with 6 or more inverters and dozens of DC combiners. In a pre-AFCI/rapid shutdown world, how would you ever disconnect the system with just 6 throws of the hand?

Throw 1: the master AC switch for the PV system
Throw 2 - 6: the DC switch on inverters 1 thru 5
And we need throw 7 to shut off the DC side of inverter 6

That still doesn't fully solve the problem, because even though current-wise it is off, voltage-wise it is still energized. You'd need to run all over the roof to shut off each DC combiner. And that just gets the DC feeders de-energized. The source circuit wiring is still at substantial voltage, until the sun goes down.
 
Last edited:

ggunn

PE (Electrical), NABCEP certified
Location
Austin, TX, USA
Occupation
Electrical Engineer - Photovoltaic Systems
I thought the "six disconnect rule" only applies to service disconnects. And that it is unrelated to just about anything on the DC side.

I'm thinking large commercial applications, probably with 6 or more inverters and dozens of DC combiners. In a pre-AFCI/rapid shutdown world, how would you ever disconnect the system with just 6 throws of the hand?

Throw 1: the master AC switch for the PV system
Throw 2 - 6: the DC switch on inverters 1 thru 5
And we need throw 7 to shut off the DC side of inverter 6

That still doesn't fully solve the problem, because even though current-wise it is off, voltage-wise it is still energized. You'd need to run all over the roof to shut off each DC combiner. And that just gets the DC feeders de-energized. The source circuit wiring is still at substantial voltage, until the sun goes down.
I have always interpreted the rule as maximum 6 handles for DC for each inverter and 6 handles for AC. AC is usually not a problem because no matter how many inverters you have there typically is a single disco that connects their combined output to the service - one handle to rule them all. How this will shake out for DC now that the rapid shutdown rules are coming into force is still a bit murky.
 

Carultch

Senior Member
Location
Massachusetts
AC is usually not a problem because no matter how many inverters you have there typically is a single disco that connects their combined output to the service - one handle to rule them all

Except where you are put in a situation where you have to physically diversify the interconnected power to multiple customers or multiple applications.
 

ggunn

PE (Electrical), NABCEP certified
Location
Austin, TX, USA
Occupation
Electrical Engineer - Photovoltaic Systems
Except where you are put in a situation where you have to physically diversify the interconnected power to multiple customers or multiple applications.
If you have to split your inverters between interconnection points they are separate systems, and a single inverter cannot connect to a service at more than one point or to more than one service.
 
Status
Not open for further replies.
Top