2023 NEC 705.12(B)(6) (Engineering Supervision)

[wwhitney]

For the 2023 NEC and the 2026 NEC I submitted various PIs to expand the allowances in 705.12(B) on load side interconnections in ways that will not overload the panelboard bus, but which do not fit into any of the existing categories. All of them have been rejected; I think the primary reason has been, to paraphrase, "705.12(B) is complicated enough, let's not add further language for corner cases."

However, this time around the CMP basically said "we see what you are trying to do, and it makes sense, but if you want to do that, just use 705.12(B)(6)." So those of you who are engineers [or who can get an engineer to sign off on the design under 2023 NEC 705.12(B)(6), aka 2020 NEC 705.12(B)(5)], can do any of the following:

1) Use a panelboard with at most 3 busbar connections, each of which is protected at the busbar ampacity. [Of course, you can do the same thing with a feeder interconnection with 705.12(A), so it's silly that to use a busbar instead of a wire-type feeder you need engineering supervision.]

2) Ignore one breaker of the smallest size present when applying 705.12(B)(3). So you can add one breaker for monitoring equipment or whatever to your combiner panels, as long as that breaker is not larger than any other breaker present. This is due to a counting argument--the extra breaker can be either a load or a source, but not both, so if you do the accounting "current in = current out" there's no way to get both sides to exceed the busbar rating due to one extra breaker not larger than any other breaker.

3) When you know what is connected to each breaker, label each breaker "source only," "load only," or "source or load." Then as long as the sum of all the sources, or the sum of all the loads, is no more than the busbar ampacity, you know the busbar can't be overloaded.

4) Take advantage of a 100% rated OCPD. E.g. if you have a PV-only service entrance with a 200A 100% rated service disconnect/OCPD, then you can connect that to a 200A MLO panel (all such are rated for 200A continuously) connected to (5) 50A breakers, each connected to an inverter with a maximum inverter output current of 40A.

Cheers, Wayne

 

[wwhitney]

One more possibility, which is a bit more aggressive. The one ingredient whose validity I'm not certain of is the assumption that each breaker and busbar connection has the same resistance R as far as the I²*R heating it causes. So that might require more thought or further confirmation before using this method:

Do a busbar connection heating accounting for your proposed configuration. The heating in the panelboard is a result of the I²R losses in both the busbar itself and in the connections and within the breakers. The UL standard tests a worst case of full rated current coming in, on (all of?) the busbar, and exiting via a breaker. For the simple model where each breaker and its connection has the same resistance R, this gives a total source connection heating of 100% * I_rated² * R, which I will just call 100% heating (the worst case load connection heating is the same, and the worst case busbar heating is 100% rated current on all of the busbar).

705.12(B)(2) aka the 120% rule allows a worst case source connection heating of 102.56% (100% busbar current primary source plus 16% another load; the worst case load connection heating is likewise 102.56%, and the busbar heating worst case is effectively unchanged). Here are some other configurations that would have worst case source connection heating of at most 102.56% under this model:

90% primary source plus 46% secondary source gives 0.9² + 0.46² = 102.16%. So a 100A panelboard bus, a 90A main breaker, and an interconnected source at the opposite end of the busbar with at most a 46A output current.

88.9% primary source plus 48% secondary source gives 102.07%. So a 225A panelboard bus, a 200A main breaker, and an interconnected source at the opposite end of the busbar with at most a 108A output current.

87.5% primary source plus 50% secondary source gives 101.56%. So a 200A panelboard bus, a 175A main breaker, and an interconnected source at the opposite end of the busbar with at most a 100A output current.

Cheers, Wayne

 

[tortuga]

I think it would be better it have your PI's in the code, very few EE's want to deal with residential which is what 75% of this applies to.

 

[petersonra]

I think the design allowances in the code have gotten out of hand. They need to be simplified so it is obvious what is being allowed without some kind of engineering help.

 

[wwhitney]

I think the design allowances in the code have gotten out of hand. They need to be simplified so it is obvious what is being allowed without some kind of engineering help.

Well, 705.12(B) is a fairly tricky topic. So if you have a suggestion as to how to replace the several different rules there with a single overall approach that is simple enough for electricians to understand in the field, and can be explained using less text than we currently have in 705.12(B), that would be very welcome.

Cheers, Wayne

 

[wwhitney]

I think it would be better it have your PI's in the code

Given the CMP's responses, that's apparently not going to happen, and they even explicitly directed me to 705.12(B)(6). Hence this thread is the best I can do.

Cheers, Wayne

 

[petersonra]

Well, 705.12(B) is a fairly tricky topic. So if you have a suggestion as to how to replace the several different rules there with a single overall approach that is simple enough for electricians to understand in the field, and can be explained using less text than we currently have in 705.12(B), that would be very welcome.

Cheers, Wayne

Its a general comment, not just for this type of thing. For instance, table 250.122 could have a note after it indicating the exceptions to the size rules for rods, plates, and CEEs.

In my opinion, there is no reason for the PV panelboard rules to be so complex. It appears to my uninformed eyeballs that the rules are based around an arbitrary 125% figure that seems to have been made up out of thin air.

 

[jaggedben]

...

3) When you know what is connected to each breaker, label each breaker "source only," "load only," or "source or load." Then as long as the sum of all the sources, or the sum of all the loads, is no more than the busbar ampacity, you know the busbar can't be overloaded.

...

A bit unclear here if the 'source or load' breakers need to be added to the sum.

It seems to me that, at least in the case where there are only two sources, one of which is 'source or load', you can ignore the 'source or load' breaker and check if the sum of the 'load only' breakers is no more than the ampacity of the busbar. (This also works if the 'source or load' connection is on feed through lugs.)

 

[wwhitney]

It seems to me that, at least in the case where there are only two sources, one of which is 'source or load', you can ignore the 'source or load' breaker and check if the sum of the 'load only' breakers is no more than the ampacity of the busbar. (This also works if the 'source or load' connection is on feed through lugs.)

Maybe. I agree after we observe that a further requirement is that the sum of the "load only" breakers (which we might as well treat as a unit), as the sum of the "source only" breakers (likewise) each have to individually be less than the busbar ampacity.

Then if the single "source or load" breaker is a source, the sum of the loads is less than the busbar ampacity, while if the "source or load" is a load, the sum of the sources is less than the busbar ampacity.

I guess that is what I meant to say originally, that you could consider the cases of source or load separately for each such breaker. E.g. 200A busbar, 150A grid connection (source or load), 150A battery inverter/charger (source or load), 50A source only, 200A load only. The possibilities for the first two breakers are S/S (load is at most 200A), S/L (sources are at most 200A), L/S (sources are at most 200A) and L/L (sources are at most 50A).

Cheers, Wayne

 

[jaggedben]

Maybe. I agree after we observe that a further requirement is that the sum of the "load only" breakers (which we might as well treat as a unit), as the sum of the "source only" breakers (likewise) each have to individually be less than the busbar ampacity.

Typo in there somewhere? Doesn't quite make grammatical sense to me. Did you mean the sum of the 'load only' *or* the sum of the 'source only'. In the case where there are only 'source only' and 'load only' breakers, then whichever is less of the sum of the loads or the sum of the sources limits the current.

...
I guess that is what I meant to say originally, that you could consider the cases of source or load separately for each such breaker. ...

Okay, right. So something like: For all permutations where a 'source or load' breaker is considered either as a source or a load, either the sum of the sources or the sum of the loads shall not be more than the ampacity of the busbar.

FWIW this is the most common situation I'm encountering in residential projects where I wish I had a new or modified rule. Common real world example: 200A busbar protected by 200A main (utility side) breaker, a connection to the 'source or load' microgrid with backed up loads, PV and battery, and no more than 200A other 'load only' breakers in the main (i.e. a few non backed up loads). Actually it should still work if I just have a grid-tie PV (or PV+battery) system, just less likely that the load breakers won't add up to more than 200A.

 

[pv_n00b]

I think the design allowances in the code have gotten out of hand. They need to be simplified so it is obvious what is being allowed without some kind of engineering help.

I think professional engineers should have more latitude to provide alternate solutions to what the code requires. I have found that working with larger more complex projects shows that the NEC could never be written to cover all possible situations and still be usable.

 

[petersonra]

I think professional engineers should have more latitude to provide alternate solutions to what the code requires. I have found that working with larger more complex projects shows that the NEC could never be written to cover all possible situations and still be usable.

I agree.

 

[ggunn]

I think professional engineers should have more latitude to provide alternate solutions to what the code requires. I have found that working with larger more complex projects shows that the NEC could never be written to cover all possible situations and still be usable.

But any PE who invokes language in the code for such alternatives needs to be aware that doing so thereby shifts the responsibility from the NEC to his or her license should something go wrong. It is something I would never do lightly.

 

[pv_n00b]

I agree with Tortuga, this is mostly an issue with residential and small commercial that professional engineers will not be available for. In larger projects the service entrance is not a panelboard and where I am we do SWBD interconnections under engineering supervision as the rule, not the exception.
For existing panelboard service entrances I have always looked at it like this:

• The main service disconnect OCPD is less than or equal to the bus rating of the panel board
• The sum of the load OCPDs can be greater than the panelboard and main OCPD rating
• The calculated load must be less than or equal to the panelboard and main OCPD rating
• The calculated load may have diversity that injects uncertainty in the loading. It may be possible for the actual load to exceed the calculated load but this would trip the main OCPD.
• In an existing system if the owners have not had a problem with the main OCPD tripping then the actual load is not exceeding the panelboard rating

If this is all true then a PV back feed up to 80% of the main OCPD rating would not cause any problems or be dangerous, as long as nothing changes. And the PV OCPD could be anywhere in the panelboard. But, and here's the problem the NEC spends so much time and complexity solving, if later someone comes in and adds a load that pushes the panelboard loading above the main OCPD rating because it can draw from both the main source and the PV source, then that can overheat the panelboard. This is what drives all the complexity. If you can figure out how to solve that simply then all the problems go away.

 

[pv_n00b]

But any PE who invokes language in the code for such alternatives needs to be aware that doing so thereby shifts the responsibility from the NEC to his or her license should something go wrong. It is something I would never do lightly.

Where I am we are doing NEC 691 PV and the PEs are responsible for most design decisions, very little is prescriptive from the NEC. It's kind of why we are PEs. So taking on engineering supervision work from the NEC is just part of the job. I'm as happy as anyone to use the NEC as a shield, but when work has to be done the PE does the work.

 

[wwhitney]

Typo in there somewhere? Doesn't quite make grammatical sense to me. Did you mean the sum of the 'load only' *or* the sum of the 'source only'.

No I meant "and" but mistyped "as". 200A busbar, 200A source or load, 200A source only, 200A load only. Can't increase the size of any of those connections. [Well, the 200A source or load could be bumped up and be OK for this analysis, but that's the grid connection in any grid-tied system, and 408.36 is going to limit that to the busbar rating. (*)]

Okay, right. So something like: For all permutations where a 'source or load' breaker is considered either as a source or a load, either the sum of the sources or the sum of the loads shall not be more than the ampacity of the busbar.

Right.

Cheers, Wayne

(*) Unless we say that as 705.12(B)(6) is in Chapter 7 it is allowed to amend Chapters 1-4, and so it overrides 408.36? Or is that a bridge too far?

 

[jaggedben]

No I meant "and" but mistyped "as". 200A busbar, 200A source or load, 200A source only, 200A load only. Can't increase the size of any of those connections. ...

I think I disagree. The source can be larger than the busbar if the load isn't, and vice versa. We are used to the latter, but the former works just as well.

...
(*) Unless we say that as 705.12(B)(6) is in Chapter 7 it is allowed to amend Chapters 1-4, and so it overrides 408.36? Or is that a bridge too far?

I think we already do this, effectively. In 705.12(B)(2), (3) and (4).

 

[petersonra]

But any PE who invokes language in the code for such alternatives needs to be aware that doing so thereby shifts the responsibility from the NEC to his or her license should something go wrong. It is something I would never do lightly.

Isn't that what being a PE is all about?

 

[wwhitney]

I think I disagree. The source can be larger than the busbar if the load isn't, and vice versa. We are used to the latter, but the former works just as well.

My example was "200A busbar, 200A source or load, 200A source only, 200A load only."

If we try to bump this up to "200A busbar, 200A source or load, >200A source only, 200A load only," then this can overload the busbar when the "source or load" is a load Or if we try to bump this up to "200A busbar, 200A source or load, 200A source only, >200A load only" then this can overload the busbar when the "source or load" is a source.

So both the "source only" and the "load only" need to be capped at the busbar ampacity. Well, unless the "source or load" plus one of the others are less than the busbar ampacity; then the 3rd one can be unlimited.

Cheers, Wayne

 

[ggunn]

Isn't that what being a PE is all about?

Of course, but we need to be aware of the risk we take on when we color outside the lines.