Six-handle rule in 690.14

[Elexit]

It is clear that 690.14 requires a PV system to have no more than six disconnects, and that they are required to be grouped together, just as service disconnects are limited by 230.71. But at what point are we counting? Until recently, we could say that if all inverter outputs are combined in a panelboard, and that panelboard has a main breaker or main switch, we have satisfied this rule, as this AC switch serves to "disconnect all current-carrying conductors of a photovoltaic system from all other conductors in a building or structure."

Then the 2011 NEC specified that this provision applies to the DC disconnecting means. Now we have a problem if we want to install a system with more than six inverters. Or do we?

Look at the definition in 690.2:

Solar Photovoltaic System. The total components and subsystems that, in combination, convert solar energy into electric energy suitable for connection to a utilization load.

If I have eight string inverters, each with its own array, combiner box, and disconnects; and if I open seven of the eight breakers in my aggregation panelboard, the one remaining inverter and array functions as designed. It is a complete system, which could be installed by itself and pass inspection. If this meets the NEC definition of a PV system, as I believe it does, then what we have is a group of eight systems, each of which is allowed to have up to six disconnects.

The problem is that the NEC definition is not as specific as the MPP bus of an inverter. If I have eight buildings on a campus fed from a single medium voltage service, and each building has a PV system with an interconnection in that building, then no one would argue that I don?t have eight systems instead of one. But put all of the arrays and inverters on one building, and the aggregate thereof ?wants? to be called a system.

Presently we are designing a system for two carport structures atop a parking garage. We want to use eight PowerOne Aurora TRIO inverters, and the customer would like them placed in a room at grade level. Our AHJ has been equivocal about what we can do to satisfy the Code requirements, but three inspectors in the department agree that the six-handle rule appears to apply here. Yet I find it hard to believe that the intent of the Code is to restrict design to six or fewer inverters.

If we could install a group of four disconnects on each carport, that would address the intent of this section: to limit the number of operations of the hand, and the amount of time a firefighter spends walking around looking for them. It would not, however, strictly comply with a rule derived from a 1937 precedent designed to apply to energy from a single source (utility transformer).

I've read the 2014 CMP 4 proposals, and I still don't fully understand the requirement. Any ideas?

 

[ggunn]

It is clear that 690.14 requires a PV system to have no more than six disconnects, and that they are required to be grouped together, just as service disconnects are limited by 230.71. But at what point are we counting? Until recently, we could say that if all inverter outputs are combined in a panelboard, and that panelboard has a main breaker or main switch, we have satisfied this rule, as this AC switch serves to "disconnect all current-carrying conductors of a photovoltaic system from all other conductors in a building or structure."

Then the 2011 NEC specified that this provision applies to the DC disconnecting means. Now we have a problem if we want to install a system with more than six inverters. Or do we?

Look at the definition in 690.2:



If I have eight string inverters, each with its own array, combiner box, and disconnects; and if I open seven of the eight breakers in my aggregation panelboard, the one remaining inverter and array functions as designed. It is a complete system, which could be installed by itself and pass inspection. If this meets the NEC definition of a PV system, as I believe it does, then what we have is a group of eight systems, each of which is allowed to have up to six disconnects.

The problem is that the NEC definition is not as specific as the MPP bus of an inverter. If I have eight buildings on a campus fed from a single medium voltage service, and each building has a PV system with an interconnection in that building, then no one would argue that I don?t have eight systems instead of one. But put all of the arrays and inverters on one building, and the aggregate thereof ?wants? to be called a system.

Presently we are designing a system for two carport structures atop a parking garage. We want to use eight PowerOne Aurora TRIO inverters, and the customer would like them placed in a room at grade level. Our AHJ has been equivocal about what we can do to satisfy the Code requirements, but three inspectors in the department agree that the six-handle rule appears to apply here. Yet I find it hard to believe that the intent of the Code is to restrict design to six or fewer inverters.

If we could install a group of four disconnects on each carport, that would address the intent of this section: to limit the number of operations of the hand, and the amount of time a firefighter spends walking around looking for them. It would not, however, strictly comply with a rule derived from a 1937 precedent designed to apply to energy from a single source (utility transformer).

I've read the 2014 CMP 4 proposals, and I still don't fully understand the requirement. Any ideas?

I have designed and my company has installed a system with 42 SMA inverters. We combined their output in an AC combiner panel feeding a single AC disco and no one had a problem with it.

 

[GoldDigger]

I have designed and my company has installed a system with 42 SMA inverters. We combined their output in an AC combiner panel feeding a single AC disco and no one had a problem with it.

But if the inverters were widely spaced (different building areas, sides of building, whatever) then you might get into trouble if you connect each group separately to a backfed breaker on a local subpanel instead of taking them all back to a single (or six) backfed breaker(s) or supply side tap(s). That might require a lot more additional wiring. I think that the argument would indeed revolve around whether you have one 42-inverter system or seven separate six-inverter systems at that point.

 

[ggunn]

But if the inverters were widely spaced (different building areas, sides of building, whatever) then you might get into trouble if you connect each group separately to a backfed breaker on a local subpanel instead of taking them all back to a single (or six) backfed breaker(s) or supply side tap(s). That might require a lot more additional wiring. I think that the argument would indeed revolve around whether you have one 42-inverter system or seven separate six-inverter systems at that point.

That could be a problem, I guess, but all our systems interconnect at a single point on a single handle. It seems to me that if they interconnect in separate subpanels they are separate systems.

 

[jaggedben]

I would totally agree that the six handle rule in 690.14 was intended to apply to DC disconnects, but its hard to make a conclusive argument based on the wording of the code. Maybe looking back through old ROP substantiations would help. Something else which might help is to point out that 705 has its own section on disconnecting means, which does not mention the six handle rule.

Are you going to have a single main disco that would disconnect all 8 inverters, or not? Seems to me it makes a difference.

Beyond that, I can only offer extemporaneous comments about how I think the code should be re-written to avoid such ambiguities.

1) 690 should be completely stripped of all references to AC disconnects except in as much as they:
-refer to 705
or
-refer explicitly to stand alone systems

2) All references to disconnects in 690 should get the acronyms 'ac', 'dc' or 'ac and dc' put in front of them, as applicable. Hopefully there does not actually have to be an instance of 'ac and dc'.

 

[ggunn]

I would totally agree that the six handle rule in 690.14 was intended to apply to DC disconnects, but its hard to make a conclusive argument based on the wording of the code.

We just finished up a 250 kW PV project that has a single AC disco after a combining tap, two AC discos (for two inverters) before the tap, and 10 DC discos (5 per inverter in two enclosures), and it is in the territory of an AHJ well known for its intensive PV system inspections. There was no problem with either the preconstruction interconnection study or the final inspection.

OTOH, the system we put in with 42 Sunny Boys has 42 DC discos (the integrated ones) and no other DC discos at the inverters, but if the AC disco is open, the DC discos at the inverters don't really do anything.

Once you shut off the AC disco there is no current flowing in the DC conductors, and opening DC discos which are next to the inverter(s) only de-energizes the short conductor run(s) between the discos and the inverter(s). Most inverters have an integrated DC disco, anyway, and a DC disco next to an inverter mainly affords compliance to 690.16(B). Disconnecting combiners will de-energize the DC runs to the inverter(s), but they may be scattered all over the site.

The safest installation might have disconnecting combiners out at the array(s) with remotely actuated contactors tied to a single handle next to the inverter(s) or AC disco, but that would be expensive and I don't know personally of any systems designed that way.

 

[jaggedben]

The safest installation might have disconnecting combiners out at the array(s) with remotely actuated contactors tied to a single handle next to the inverter(s) or AC disco, but that would be expensive and I don't know personally of any systems designed that way.

This is something that looks like it may be (more or less) mandated in the 2014 code, BTW.

 

[ggunn]

This is something that looks like it may be (more or less) mandated in the 2014 code, BTW.

Yes, I know.

 

[tallgirl]

This is something that looks like it may be (more or less) mandated in the 2014 code, BTW.

There's a proposal for 2014 which is worse than that -- strings would have to be disconnected to prevent any portion of the string from being above 80 volts. Disconnecting at the combiner won't be enough -- solar panel JBs would have to have some form of disconnect.

 

[jaggedben]

There's a proposal for 2014 which is worse than that -- strings would have to be disconnected to prevent any portion of the string from being above 80 volts. Disconnecting at the combiner won't be enough -- solar panel JBs would have to have some form of disconnect.

I don't think it will be quite that bad. IIRC correctly, the language in the ROC only applies to wiring more than 10ft away from the array. They've not gone quite so far as to require smart JBs on panels.

 

[SolarPro]

Jaggedben is correct. The language was changed from the original proposal. Obviously a j-box level disconnect will meet the new requirement, but so will contactor combiners or similar.