705.12 Load-side connections & a dedicated OCPD

[kamomeno68]

I have a question regarding the requirement that "Each source interconnection of one or more power sources installed in one system shall be made at a dedicated circuit breaker or fusible disconnecting means" (NEC2020 705.12(A) or NEC2017 705.12(B)(1)). My question is generally about the meaning of the word "dedicated". An inspector provided me a photocopy of a page from a John Wiles' article which states, about this code section: "Multiple PV systems may be connected to an existing utility-supplied electrical system, but each connection must be through a dedicated OCPD and a disconnect. 'Dedicated' means that no loads are to be placed on this circuit between the breaker and the PV system ac output." (emphasis mine). I've also seen this perspective expressed by Bill Brooks during a Mike Holt code panel discussion. My specific question is about the meaning of "no loads", and might be best expressed in an example. We design and install commercial pv systems, often with multiple string inverters. The inverter output breakers are usually combined in a single PV AC Aggregation Panelboard, which then connects to the utility-supplied electrical system via a dedicated breaker or fused-disconnect. Most of these projects are designed with a DAS system as well, and that DAS system is powered from a load breaker in the same PV AC Aggregation Panelboard. Depending on the project, there is often another load breaker installed in the PV AC Aggregation Panelboard to feed a service outlet. Here is an example diagram:



Specific questions: (1) Does the presence of the 15A/3P load breaker violate the "dedicated OCPD" requirement? (2) Is the fused AC disconnect the dedicated OCPD, or are the individual inv breakers in the panelboard the dedicated OCPDs? (3) Would the answer change if the panelboard had a MB?

 

[SceneryDriver]

My reading is that it wouldn't, though I tend to be mighty practical about these things. I presume that the DAS (distributed antenna system?? or does DAS stand for something else?) is directly related to operation of the solar array and its inverters; I don't think any reasonable AHJ would say it has to be fed from another panel. Don't some of those systems HAVE to be connected relatively close to the inverters they communicate with over the power conductors, anyway?

Likewise, if the intent of the receptacle is to provide the code-mandated receptacle for service and maintenance activities at the panel, it too should be allowed.

What you can't do is power a water heater, pump, electric range, general receptacles, lighting, etc... from the panel.



SceneryDriver

 

[Carultch]

My reading is that it wouldn't, though I tend to be mighty practical about these things. I presume that the DAS (distributed antenna system?? or does DAS stand for something else?) is directly related to operation of the solar array and its inverters; I don't think any reasonable AHJ would say it has to be fed from another panel. Don't some of those systems HAVE to be connected relatively close to the inverters they communicate with over the power conductors, anyway?

Likewise, if the intent of the receptacle is to provide the code-mandated receptacle for service and maintenance activities at the panel, it too should be allowed.

What you can't do is power a water heater, pump, electric range, general receptacles, lighting, etc... from the panel.

DAS usually stands for Data Acquisition System in this context. It usually means an auxiliary load directly associated with the PV system, for purposes of collecting performance data & metering, and uploading it to the internet.

 

[Carultch]

Specific questions: (1) Does the presence of the 15A/3P load breaker violate the "dedicated OCPD" requirement? (2) Is the fused AC disconnect the dedicated OCPD, or are the individual inv breakers in the panelboard the dedicated OCPDs? (3) Would the answer change if the panelboard had a MB?

(1) No, it does not violate the dedicated OCPD requirement. This requirement means that each individual inverter circuit must be connected to a dedicated OCPD/disconnecting means. You cannot gang up multiple inverters that would individually require a 40A breaker, on the same 80A breaker. If you were to do this, you would need a subpanel between that 80A breaker and the two inverters in question, and that subpanel would need two 40A breakers dedicated to each of the inverters.

The exception to this is microinverters, that are allowed to be combined with specific paralleling trunk cables and equipment, listed for the application, prior to being connected to a breaker that is assigned to an entire branch circuit of numerous microinverters.

(2) The individual breakers are the "dedicated OCPD" for each inverter.

(3) No, the answer wouldn't change if a main breaker were added to the panelboard.

 

[jaggedben]

If you look through NPFA code revision docs it is clear that the CMP has never intended to require that all sources on a site connect through a single point which has no loads connected. But they've struggled to convey their intent clearly. It's the reason they changed 'The' back to 'Each' in one cycle. I don't know why they just don't add 'no loads' language to make it clear. Brooks proposed it at one point and they agreed in principle but still didn't insert the language.

 

[kamomeno68]

Thank you all for your input. Practically speaking, I think the system as designed is perfectly safe, but it is frustrating how many nuances can be pulled out of code language sometimes. To address two of the comments above:
(1) Yes, as Carultch pointed out, DAS means Data Aquisition System, and is just the monitoring system for the pv system.
(2) Regarding the last statement above from SceneryDriver "What you can't do is power a water heater, pump, electric range, general receptacles, lighting, etc... from the panel", I'm not sure that I agree. I would think that either loads are safe (in the context of the design shown) or they aren't. Generally speaking, I wouldn't think the type of load should matter. I can understand how it might complicate production metering if that metering is done on the line-side of the AC disconnect shown, and a utility entity might have an issue, and I can see how putting a non-pv-related load there might be impractical (if the AC disconnect was opened to lock-out the pv system, the load would lose power also) but I don't see a problem from an electrical safety point-of-view if that load was a pump, for example, and not just a load related to the pv system.

For another perspective, I did find the following comment about this code reference in the recent "PV and the NEC, based on the 2020 NEC" by Bill Brooks and Sean White: "Discussion: In the past, some have interpreted the NEC to say that each inverter needed a dedicated OCPD, which was especially inconvenient for microinverter enthusiasts. Multiple inverters can have a single OCPD as long as those inverters have been listed and tested to work safely with a single OCPD. This would be reflected in the installation instructions. We cannot, however, have a combination of a load and an interactive inverter on a branch circuit breaker. This means that you cannot have a PV system that you plug into an outlet. There have been some good ideas that have been taken off the books because of this rule. This is also for good reason. If a homeowner started plugging in a new PV system to a power strip every week, before they reached a MW, something would catch fire."

Although the use of an outlet is an extreme example, I interpret this to support Carultch's comments above. As long as the inverters in a system and the loads in a system are behind their own dedicated OCPD's, then it feels to me like the intent of this code requirement is being met. Thank you all again for your input.

 

[jaggedben]

I agree with your understanding.

One thing I might add, apropo of your comment about metering, is that a utility or other entity might want all sources to be connected through a single disconnect or meter, and this would preclude most normal loads from being connected on the load side, and maybe monitoring loads as well depending on their opinion. But none of that is about safety and none of it is in the NEC.