705.12 (D) question

[sketchy]

If I am backfeeding a subpanel and follow the rules for that interconnection, do all connections made upstream of the subpanel need to follow the same rules? I.E. does the subpanel feed breaker in the main service panel need to follow 705.12 too? I don't think it does but my boss disagrees. Any references to back up your point or directions to existing posts would b e appreciated.

 

[ggunn]

If I am backfeeding a subpanel and follow the rules for that interconnection, do all connections made upstream of the subpanel need to follow the same rules? I.E. does the subpanel feed breaker in the main service panel need to follow 705.12 too? I don't think it does but my boss disagrees. Any references to back up your point or directions to existing posts would b e appreciated.

Your boss is correct. 125% of the inverter nameplate current (Imax) contributes to the 120% rule in every panel between the PV system and the service, and every breaker that has the potential for backfeeding a panel needs to be at the opposite end of the busbar from the feed from the utility if 1.25 X Imax plus the rating of the OCPD protecting the bus is between 100% and 120% of the busbar rating.

 

[sketchy]

Can you direct me to the code section that states this?

 

[ggunn]

Can you direct me to the code section that states this?

You are looking at it. It talks about inverter current coming into a panelboard. It doesn't matter if the inverter is connected directly to the panelboard or if the current from it is coming in though a subpanel. I have been designing PV systems for 9 years in many jurisdictions and every AHJ I have dealt with interprets the code this same way.

Think about it. If you have a 200A main panel with a 200A main breaker, the code says you cannot bring in more than 32A of PV current into it through a backfed breaker. If you connected 50A of PV to a 200A subpanel with a 150A main breaker, which is OK, and connected the subpanel to the MDP through a 200A breaker, would that look OK to you? You are still supplying 62.5A of potential PV current plus 200A from the main breaker to a 200A bus. You cannot get around the code that way.

Your boss is correct.

 

[pv_n00b]

Can you direct me to the code section that states this?

It states in 705.12(D) that it applies to all distribution equipment fed by a primary source and one or more inverters and that covers all equipment from the PV connection in a subpanel back to the utility.

 

[jaggedben]

There used to be language that stated that if panelboards were wired in series, the rating of the breaker closer to the inverter was to be used in the calculations. That language implied that the rules applied to all panelboards in series. Then the code changed to using 125% of inverter output instead of breaker ratings, and that language was no longer necessary and was removed.

I do tend to think that the code should state somewhere that when equipment is wired in series that the conditions at each location apply to a connection at that location. But as you can see from my last sentence, it's kind of hard to put the idea in language that is crystal clear. As ggunn says, I've been doing this for 9 years and that is how everyone in the industry understands it.

 

[pv_n00b]

I agree with jaggedben. Basically what I see on the ground is when interconnecting at a subpanel following that back to the service entrance often shows that upgrades need to be done to existing equipment to meet code. No one wants to do that since the whole idea of interconnecting at a subpanel is usually about using the closest connection to the existing electrical system and being done.

At this point, I've probably heard just about all the justifications for not having to change out existing equipment. The best one I have heard is that all the output from the PV system will be used locally and never export out of the subpanel. I love that one.

 

[shortcircuit2]

I agree with the posts here and would add that if for some reason the ampere ratings of all overcurrent devices in the upstream panelboard don't exceed the rating of the busbar...the solar breaker can input anywhere in the upstream panel. 2017 NEC 705.12(B)(2)(3)(c)

 

[ErikS]

I am a bit late to this conversation but was hoping add on to the initial question. The responses make a lot of sense but I see the following situations where I wonder if another argument can be made.
Assume a 400 Amp Main Panel Board (400A main breaker and 400A Bus Bar Rating).
3 breakers are located in this Main Panel Board that feed Sub-Panel Boards:
Panel Board A - 225A
Panel Board B - 100A
Panel Board C - 150A

Let's say I interconnected 100A of PV (80A x 125%) into subpanel C (which has a 250A bus bar rating).

Let's further assume that there aren't any active loads on Panel Board C and all of the theorectical 100Amp of Solar generation feeds into the Main Panel Board. The maximum loads at the Main Panel Board would be the loads from Panel Board A and B = 325A. The bus is rated for 400Amp however.
It seems to me that the 100A of solar interconnected through Panel C could not cause the maximum current on the Bus of the Main Panel Board to exeed 400A (or even 325 Amp).

Wouldn't this meet or exceed the intend of the code?
Any input would be appreciated. I wonder NEC 2014 705.12 (D) (2) (d) would allow for an engineering analysis that comes to the conclusion that this is acceptable.

 

[jaggedben]

Interesting question, but I think it's hard to make a statement about meeting the intent of the code here. If the CMP wanted to allow what you're proposing, they might have written the rule to say "neither the sum of the loads nor the sum of the sources shall exceed the busbar." Instead, they wrote "the sum of the loads and the sources shall not exceed the busbar." (Paraphrasing.) There are potentially a lot of scenarios which might look good on paper but haven't been tested with real thermal loading in real panelboards. I think the CMP is going more conservative for that reason.

 

[ggunn]

I wonder NEC 2014 705.12 (D) (2) (d) would allow for an engineering analysis that comes to the conclusion that this is acceptable.

FWIW, I have invoked 705.12(D)(2)(3)(d). The AHJ basically said, "OK, but it's your license on the line if something goes wrong." I did my homework and I was confident in the design, but still it made me a little nervous. You'll likely need to get a PE to bless it and take the risk.

 

[ErikS]

FWIW, I have invoked 705.12(D)(2)(3)(d). The AHJ basically said, "OK, but it's your license on the line if something goes wrong." I did my homework and I was confident in the design, but still it made me a little nervous. You'll likely need to get a PE to bless it and take the risk.

Thanks for the feedback ggunn and jaggedben. I am actually having this situation right now and am dealing with the AHJ. I proposed the engineering analysis and will keep you posted on the response I get.

 

[pv_n00b]

I am a bit late to this conversation but was hoping add on to the initial question. The responses make a lot of sense but I see the following situations where I wonder if another argument can be made.
Assume a 400 Amp Main Panel Board (400A main breaker and 400A Bus Bar Rating).
3 breakers are located in this Main Panel Board that feed Sub-Panel Boards:
Panel Board A - 225A
Panel Board B - 100A
Panel Board C - 150A

Let's say I interconnected 100A of PV (80A x 125%) into subpanel C (which has a 250A bus bar rating).

Let's further assume that there aren't any active loads on Panel Board C and all of the theorectical 100Amp of Solar generation feeds into the Main Panel Board. The maximum loads at the Main Panel Board would be the loads from Panel Board A and B = 325A. The bus is rated for 400Amp however.
It seems to me that the 100A of solar interconnected through Panel C could not cause the maximum current on the Bus of the Main Panel Board to exeed 400A (or even 325 Amp).

Wouldn't this meet or exceed the intend of the code?
Any input would be appreciated. I wonder NEC 2014 705.12 (D) (2) (d) would allow for an engineering analysis that comes to the conclusion that this is acceptable.

Consider this, let's say we have a panel with a 400A bus, a 400A main breaker at one end, and a 400A PV backfeed breaker at the other, 320A of PV output. There is no way to have at any point on the bus more than 400A even if we somehow have 720A in loads attached. So people will ask, why can't we do this? The other reason is heating of the enclosure due to I^2R loss. If you have 400A from the main that adds a certain amount of heat to the enclosure, that heating has been taken into account in the UL Listing. Add more heat from the PV backfeed current and now we overheat the enclosure, which since molded case breakers are thermally operated could cause tripping at less than the rated current and long-term overheating will damage components. So it's not just the current at a point in the bus that needs to be considered.

 

[ggunn]

FWIW, I have invoked 705.12(D)(2)(3)(d). The AHJ basically said, "OK, but it's your license on the line if something goes wrong." I did my homework and I was confident in the design, but still it made me a little nervous. You'll likely need to get a PE to bless it and take the risk.

Sorry to followup myself but be advised that the language in 705.12(D)(2)(3)(d) is pretty specific concerning what you can apply it to, i.e. multiple-ampacity busbars and center fed panelboards. You can't use it to override the NEC on what would otherwise be a violation.

 

[wwhitney]

Here is a crazy possible solution to ErikS's example: install the breakers in the panel in the following electrical order:

400A Main Breaker
225A Breaker for Panel A
150A Breaker for Panel C with 100A of PV backfeed
100A Breaker for Panel B

Now for analysis purposes divide the busbar in the panel into two sections, above the 150A Breaker and below the 150A Breaker, and apply 705.12(D)(2) to each section separately. The panelboard ex the 100A Breaker satisfies 705.12(D)(2)(3)(c). Call the busbar below the 150A Breaker a feeder, and then it qualifies under 705.12(D)(2)(1)(b).

I admit this is a stretch. However, you could just set a 100A enclosed breaker next to the current panel to feed Panel B, and replace the 100A breaker in the main panel with some feed through lugs to feed the 100A enclosed breaker. That arrangement would allow the main panel to qualify under 705.12(D)(2)(3)(c). It would be a physical manifestation of the division for analysis purposes I proposed, and the electrical connectivity graph would be identical. Thus if that arrangement with an enclosed breaker satisfies 705.12(D)(2), so should the original arrangement without the separate enclosure for the 100A breaker.

[Perhaps I should have posted this idea yesterday. I believe it is logical, but it may not be practical in terms of AHJ acceptance.]

Cheers, Wayne

 

[jaggedben]

While in principle I agree that a panelboard busbar can be part of a feeder, I do not believe you can exempt a panelboard busbar from any of the 705.12(D)(2)(3) rules by calling it that. It's not either/or.

As both pv_noob and I have said, lacking actual testing of thermal loading in panelboard busbars it is perhaps not safe to simply follow Kirchoff's law logic. The code, so far, has been deliberately written more conservatively than that, because (to my understanding) such testing hasn't been done. Otherwise the opposite-end 120% rule could be the opposite-end 200% rule.

 

[wwhitney]

Yeah, my proposed analysis doesn't fit the actual wording of 705.12(D)(2)(3). Yet moving one breaker into its own enclosure without changing the connectivity graph would fit the actual wording.

As a mathematician, it seems to me that 705.12(D)(2)(3) should be rewritten along the following lines:

Every connection to the busbar of a panel board is categorized as a source only (starting at the panelboard, no loads downstream of that connection), a sink only (no sources downstream of that connection), or a combined source/sink. Sinks are assigned the value of the first downstream breaker, and sources are assigned the lesser value of the breaker protecting the connection or 125% of the source output current. Then 705.12(D)(2)(3)(a) and (c) can be combined as:

(a) All possible combinations of treating each combined source/sink as either a source or a sink shall be considered, and for each combination either (i) the sum of all sources or (ii) the sum of all sinks shall be less than or equal to the busbar rating. [Or in negative language, there is no division of the combined source/sinks into sources and sinks such that the sum of all the sources and the sum of all the sinks both exceed the rating of the busbar.]

Then (b) can rephrased as:

(b) All possible sources are connected at the ends of the busbars, the sum of the sources at one end does not exceed the busbar rating, and the sum of all the sources does not exceed 120% of the busbar rating.

I'm not sure if (a) can be expressed simply enough for inclusion into the NEC, I don't think I did a great job at that myself.

Cheers, Wayne

 

[ErikS]

Impressive
After a bit back on forth with the plan checker I decided to perform a Line Side Tap. More coordination efforts, more cost, but acceptable to the AHJ.

Thanks again for all the input.

 

[ggunn]

Yeah, my proposed analysis doesn't fit the actual wording of 705.12(D)(2)(3). Yet moving one breaker into its own enclosure without changing the connectivity graph would fit the actual wording.

As a mathematician, it seems to me that 705.12(D)(2)(3) should be rewritten along the following lines:

Every connection to the busbar of a panel board is categorized as a source only (starting at the panelboard, no loads downstream of that connection), a sink only (no sources downstream of that connection), or a combined source/sink. Sinks are assigned the value of the first downstream breaker, and sources are assigned the lesser value of the breaker protecting the connection or 125% of the source output current. Then 705.12(D)(2)(3)(a) and (c) can be combined as:

(a) All possible combinations of treating each combined source/sink as either a source or a sink shall be considered, and for each combination either (i) the sum of all sources or (ii) the sum of all sinks shall be less than or equal to the busbar rating. [Or in negative language, there is no division of the combined source/sinks into sources and sinks such that the sum of all the sources and the sum of all the sinks both exceed the rating of the busbar.]

Then (b) can rephrased as:

(b) All possible sources are connected at the ends of the busbars, the sum of the sources at one end does not exceed the busbar rating, and the sum of all the sources does not exceed 120% of the busbar rating.

I'm not sure if (a) can be expressed simply enough for inclusion into the NEC, I don't think I did a great job at that myself.

Cheers, Wayne

As an engineer, I must design systems which are compliant with the NEC as it is written. To that end, I strive to avoid end runs and, um, creative interpretations of code language, especially where they would be sleight of hand attempts to circumvent the spirit of the code. This is, IMO, one of those cases.

If you are successful at rewriting the code, I will comply with it.

 

[wwhitney]

To that end, I strive to avoid end runs and, um, creative interpretations of code language, especially where they would be sleight of hand attempts to circumvent the spirit of the code. This is, IMO, one of those cases.

Thank you for your response.

That is the crux of the matter, does ErikS's original configuration comply with the spirit of 705.12(D)(2)(3)? If so, then creative interpretations would not be circumventing the spirit of the code language, only the letter of it.

I believe that in ErikS's original configuration, no more than 400A (nominal, assuming breakers trip at their rating) can flow into the busbar in any normal operating situation. Is there more to the spirit of 705.12(D)(2)(3) than that?

If you are successful at rewriting the code, I will comply with it.

So do you think the procedure I described meets the intent and spirit of 705.12(D)(2)(3)? And do you see how to express it more simply and clearly?

Cheers,
Wayne