DC Disconnect

[Sparkbob]

Good morning All, Long time lurker, first time poster. Hope that someday I can learn as much as some of you have forgotten.
My local electrical utility and Fire Dept. are requiring a lockable DC disconnect ahead of the inverter. They want to break both conductors of the PV source circuit in this disconnect. I feel that this is a violation of 2008 NEC 690.13, and does not meet the criteria for the exception.
Thanks in advance for any input.

 

[SolarPro]

I agree with you. There is an additional exception in the 2011 NEC. However, it doesn't appear to justify this either:

"Exception No. 2: A disconnecting switch shall be permitted in a grounded conductor if all of the following conditions are met:

(1) The switch is used only for PV array maintenance.
(2) The switch is accessible only by qualified persons.
(3) The switch is rated for the maximum dc voltage and current that could be present during any operation, including ground-fault conditions."

You could comply with the utility/fire department's requirements by installing a non-isolated (transformerless) inverter, in which case both PV source and PV output conductors are ungrounded.

However, if you break a "marked, grounded conductor" in a grounded PV system and leave that conductor in "an ungrounded and energized state," you are clearly violating 690.13. If the wire is marked white, the NEC expects it to be grounded. This isn't unique to PV systems.

I'd inquire after the source of (or rationale behind) this policy.

 

[ggunn]

Good morning All, Long time lurker, first time poster. Hope that someday I can learn as much as some of you have forgotten.
My local electrical utility and Fire Dept. are requiring a lockable DC disconnect ahead of the inverter. They want to break both conductors of the PV source circuit in this disconnect. I feel that this is a violation of 2008 NEC 690.13, and does not meet the criteria for the exception.
Thanks in advance for any input.

Unless it is an ungrounded array, IMO they are incorrect and it would be a code violation.

 

[Sparkbob]

Thanks for the feedback. I am under the impression that the PV source circuit from the panels is ungrounded until the ground is established in the inverter. Right or wrong?

 

[SolarPro]

To date, most PV systems in the US are grounded systems, meaning that one of the current carrying conductors is intentionally connected to ground at the inverter. While the majority of these systems are negatively-grounded, some are positively-grounded. For example, SunPower modules require a positive ground.

Like any grounded electrical systems, the grounded current carrying conductor must be identified with white insulation or white tape or equal. If the system is grounded, then the Code doesn't want the grounded conductor opened. A conductor that is identified as being grounded should remain grounded. If it is not, it presents a hazard.

The Code also allows for ungrounded PV systems, and the number of and market share of these systems is increasing. In an ungrounded PV system, neither current carrying conductor is connected to ground anywhere in the system. Since neither the PV positive or PV negative is grounded, both ungrounded current carrying conductors have to have OCP and both have to be opened at any dc disconnect.

Just about any question you might have about ungrounded PV systems is covered in this article (or the resources at the end of the article):

http://solarprofessional.com/article/?file=SP5_5_pg26_Fisher

If the system you are describing is an ungrounded PV power system, then the AHJ is correct in requiring that both conductors are opened in the dc disconnect. If the system is a traditional grounded PV system, then this requirement would create a hazardous condition.

 

[jaggedben]

Thanks for the feedback. I am under the impression that the PV source circuit from the panels is ungrounded until the ground is established in the inverter. Right or wrong?

Right. At least in 98% or so of cases. Every grid-tied inverter I've seen that uses a grounded conductor has the GFDI device integrated into the inverter device.

Some systems have GFDI devices separate from the inverter. Mostly these are off-grid capable systems where the GFDI device is in the charge controller or is separate.

Another exception might be if there is a GFDI device in a combiner, but I think this approach is in its infancy, and would only be seen on large systems.

 

[Sparkbob]

To date, most PV systems in the US are grounded systems, meaning that one of the current carrying conductors is intentionally connected to ground at the inverter. While the majority of these systems are negatively-grounded, some are positively-grounded. For example, SunPower modules require a positive ground.

Like any grounded electrical systems, the grounded current carrying conductor must be identified with white insulation or white tape or equal. If the system is grounded, then the Code doesn't want the grounded conductor opened. A conductor that is identified as being grounded should remain grounded. If it is not, it presents a hazard.

The Code also allows for ungrounded PV systems, and the number of and market share of these systems is increasing. In an ungrounded PV system, neither current carrying conductor is connected to ground anywhere in the system. Since neither the PV positive or PV negative is grounded, both ungrounded current carrying conductors have to have OCP and both have to be opened at any dc disconnect.

Just about any question you might have about ungrounded PV systems is covered in this article (or the resources at the end of the article):

http://solarprofessional.com/article/?file=SP5_5_pg26_Fisher

If the system you are describing is an ungrounded PV power system, then the AHJ is correct in requiring that both conductors are opened in the dc disconnect. If the system is a traditional grounded PV system, then this requirement would create a hazardous condition.

Solarpro, Jaggedben, and ggun, thanks for the replies.
This particular system IS a grounded DC system. The rationale for the requirement comes from the fire department, where the reasoning is twofold: 1) The fireman doesn't have to try to operate the integral inverter disconnect if the inverter is on fire, and 2) If all conductors are broken, no current flows, thereby creating a "safer" situation. IMO, still no reason to create a code violation...
Because the grounding of the negative DC conductor occurs in the inverter, opening both conductors ahead of the inverter leaves the grounded conductor "in an ungrounded and energized state", as specifically prohibited by 690.13, correct?

 

[SolarPro]

Correct. The ground connection is through the GFDI circuit, which is located at the inverter. If you open the grounded conductor outside the inverter, the marked grounded conductor is no longer grounded.

 

[ggunn]

Solarpro, Jaggedben, and ggun, thanks for the replies.
This particular system IS a grounded DC system. The rationale for the requirement comes from the fire department, where the reasoning is twofold: 1) The fireman doesn't have to try to operate the integral inverter disconnect if the inverter is on fire, and 2) If all conductors are broken, no current flows, thereby creating a "safer" situation. IMO, still no reason to create a code violation...
Because the grounding of the negative DC conductor occurs in the inverter, opening both conductors ahead of the inverter leaves the grounded conductor "in an ungrounded and energized state", as specifically prohibited by 690.13, correct?

With all due respect to your AHJ, they should be following code rather than trying to figure out what seems safe and what doesn't on their own, especially, as it seems here, when they are relatively unfamiliar with PV systems.

 

[SolarPro]

What is the capacity of this system? If this is a single-phase inverter, I can't think of any circumstance where opening the ungrounded current-carrying conductor doesn't stop all current flow, whether during normal operation or potential fault conditions. If the system is operating normally, then opening the ungrounded conductor opens the PV power circuit and there is no current. If there is a ground-fault, then the 1 A ground fault fuse opens, the inverter shuts down, and the array goes to open-circuit voltage; opening the dc disco doesn't change anything. If there is a direct short at the inverter, then opening the ungrounded conductor opens the circuit and stops the fault current. If there is a direct short upstream in the array, opening a downstream dc disco isn't going to change anything.

Also, if this is a residential application, it is difficult to imagine a dc disconnect that opens a grounded conductor meeting 690.13 Exception 2 in NEC 2011. How do you ensure that the disco is used only by qualified persons and only for array maintenance? If this is a commercial rooftop or an industrial application, then you could locate the dc disco where it is only accessible to qualified persons and switch the grounded conductor as allowed in NEC 2011. In a residential setting, if the fire department can access the switch from the inverter, then so can unqualified persons.

 

[don_resqcapt19]

With all due respect to your AHJ, they should be following code rather than trying to figure out what seems safe and what doesn't on their own, especially, as it seems here, when they are relatively unfamiliar with PV systems.

There are fire service issues that the codes have not yet addressed and some fire departments are addressing them on their own. In some cases this is accomplished by completely prohibiting the use of PV systems. Some have rules on the amount of roof area that can be covered, so that they can safely do ventilation if there is a fire in the building.
There are lots of issues for fire fighters that the codes have not even thought about.

 

[SolarPro]

Apples and oranges, perhaps? Permissible roof coverage isn't an electrical safety issue and is outside the scope of the NEC, so it's understandable why fire departments would want to establish and enforce these policies as they see fit. However, dc disconnect wiring seems to be pretty clearly within the scope of the NEC. It's one thing to have local requirements that exceed the minimum safety requirements found in the NEC. It's another to have requirements that run counter to those found in NEC. Who really wants to take on the liability associated with that?

This gets back to my orignal post. The Code-compliant way for this AHJ to do what they want to do—switch all current carrying conductors—would be to allow only for the installation of ungrounded PV power systems, in which case both current-carrying conductors have to be switched in any disconnect.

 

[G._S._Ohm]

There are lots of issues for fire fighters that the codes have not even thought about.

In that case, who decides?
And in the case of conflicting rules, who evaluates risk vs. benefit vs. cost vs. practicality, for the FD and the others?

 

[don_resqcapt19]

We could go back to what the fire departments really wanted and almost got it passed in one code...a load break disconnect at each and every PV module.

 

[SolarPro]

No need to go back. The excerpt below is from the NEC 2014 draft that was circulated for public comment:

690.12 PV Arrays on Buildings Response to Emergency

Shutdown. For PV systems installed on roofs of buildings, photovoltaic source circuits shall be deenergized from all sources within 10 seconds of when emergency shutdown is initiated or when the PV power source disconnecting means is opened. When the source circuits are deenergized, the maximum voltage at the module and module conductors shall be 80 volts.

 

[ggunn]

No need to go back. The excerpt below is from the NEC 2014 draft that was circulated for public comment:

690.12 PV Arrays on Buildings Response to Emergency

Shutdown. For PV systems installed on roofs of buildings, photovoltaic source circuits shall be deenergized from all sources within 10 seconds of when emergency shutdown is initiated or when the PV power source disconnecting means is opened. When the source circuits are deenergized, the maximum voltage at the module and module conductors shall be 80 volts.

I don't see how that could be done without having remotely operated disconnects for every two modules in every string.

 

[SolarPro]

This would probably involve putting an addressable semiconductor switch in every module j-box. Microinverters and dc-to-dc device can also meet this requirement, but are more complex ways of doing so. A whole range of module-level control solutions will be on the table if this section makes the final draft for the 2014 Code.

 

[jaggedben]

Whether we need module level disconnection/power limitation for fire safety is actually completely irrelevant to the question of whether a grounded conductor should be opened at a disconnect. The AHJ in question simply has it wrong on the latter issue. This would be true whether the former concept were required by the code or not.

 

[Sparkbob]

What is the capacity of this system? If this is a single-phase inverter, I can't think of any circumstance where opening the ungrounded current-carrying conductor doesn't stop all current flow, whether during normal operation or potential fault conditions. If the system is operating normally, then opening the ungrounded conductor opens the PV power circuit and there is no current. If there is a ground-fault, then the 1 A ground fault fuse opens, the inverter shuts down, and the array goes to open-circuit voltage; opening the dc disco doesn't change anything. If there is a direct short at the inverter, then opening the ungrounded conductor opens the circuit and stops the fault current. If there is a direct short upstream in the array, opening a downstream dc disco isn't going to change anything.

Also, if this is a residential application, it is difficult to imagine a dc disconnect that opens a grounded conductor meeting 690.13 Exception 2 in NEC 2011. How do you ensure that the disco is used only by qualified persons and only for array maintenance? If this is a commercial rooftop or an industrial application, then you could locate the dc disco where it is only accessible to qualified persons and switch the grounded conductor as allowed in NEC 2011. In a residential setting, if the fire department can access the switch from the inverter, then so can unqualified persons.

Solarpro, the system in question is indeed a single phase residential system. When I asked the original question, it was intended as a general question that would apply to any residential, grounded DC string system of 2KW - 6KW. Thanks to all for the excellent responses. You have given the answers that I anticipated, I just wanted to make sure that I wasn't missing something in an exception somewhere...