Public input for 705.12(B)(2)(b)

[amluto]

I'm thinking of proposing a change to 705.12(B)(2)(b), and I'm hoping for some input.
It seems to me that the current rule makes complete sense for basic panelboards
that have a main breaker and a bunch of identical branch breaker provisions on the
load side of the main breaker. It doesn't seem to anticipate panelboards that have
provisions for a sub-feed breaker and where the secondary power source is
connected using a branch breaker instead of a sub-feed breaker.

The current language is:

Where two sources, one a primary power source and the other another
power source, are located at opposite ends of a busbar that contains
loads, the sum of 125 percent of the power source(s) output circuit
current and the rating of the overcurrent device protecting the busbar
shall not exceed 120% the ampacity of the busbar. The busbar shall be
sized for the loads connected in accordance with Article 220. A
permanent warning label shall be applied to the distribution equipment
adjacent to the backfed breaker from the power source that displays the
following or equivalent wording:

WARNING:
POWER SOURCE OUTPUT CONNECTION,
DO NOT RELOCATE THIS OVERCURRENT DEVICE

The warning sign(s) or label(s) shall comply with 110.21(B).

Exception: Equipment with multiple ampacity busbars or center fed
panelboards are not addressed by this provision.
​

I propose to change this to:

Where two sources, one a primary power source and the other another
power source, are located on a busbar that contains loads and the
primary power source is at one end of the busbar, the sum of 125 percent
of the power source(s) output circuit current and the rating of the
overcurrent device protecting the busbar shall not exceed 120% the
ampacity of the busbar. The busbar shall be sized for the loads
connected in accordance with Article 220.

If the backfed breaker from the power source is at the opposite end of
the busbar from the primary power source and no devices or taps can be
installed beyond it, then a permanent warning label
shall be applied to the distribution equipment adjacent to the backfed
breaker from the power source that displays the following or equivalent
wording:

WARNING:
POWER SOURCE OUTPUT CONNECTION,
DO NOT RELOCATE THIS OVERCURRENT DEVICE

If the distribution equipment allows additional devices or taps to be
installed between the backfed breaker from the power source and the end
of the busbar opposite the primary power source, then all of the
following are required:

(i) No taps may be made between the backfed breaker from the power
source and the end of the busbar opposite the primary power source.

(ii) The sum of the ratings of any overcurrent devices between the
backfed breaker from the power source and the end of the busbar opposite
the primary power source must not exceed the rating of the busbar.

(iii) A permanent warning label shall be applied to the distribution
equipment adjacent to the backfed breaker from the power source that
displays the following or equivalent wording:

WARNING:
POWER SOURCE OUTPUT CONNECTION,
DO NOT RELOCATE THIS OVERCURRENT DEVICE.
NO TAPS MAY BE MADE ON THE BUS BEYOND THIS OVERCURRENT DEVICE,
AND THE SUM OF THE RATINGS OF ALL OVERCURRENT DEVICES
BEYOND THIS DEVICE MUST NOT EXCEED THE RATING OF THE BUS.

The warning sign(s) or label(s) shall comply with 110.21(B).

Exception: Equipment with multiple ampacity busbars or center fed
panelboards are not addressed by this provision.
​

Substantiation: I think the current language is somewhat dangerous and
unnecessarily restrictive.

Imagine a 200A top-fed main breaker panel that has an unused provision
for a sub-feed breaker at the bottom. Under the current language, it is
probably acceptable to install a 40A breaker (for a 32A power source) at
the bottom of the panel just above the sub-feed provision as long as
it's labeled with a warning sign saying "WARNING: ... DO NOT RELOCATE
THIS OVERCURRENT DEVICE". The installation would be safe as is.
However, a future electrician could see that sign, and, without thinking
about it hard enough, install a 200A subfeed breker at the bottom. The
result would still be safe (no part of the busbar could be overloaded),
but it would no longer comply. If sub-feed lugs were to be installed,
the result would be unsafe: the bus between the lugs and the 40A breaker
could be overloaded.

On the other hand, there are simple installations that would be safe but
are not permitted. Imagine the same 200A panelboard with a 40A breaker
for the secondary power supply at the lowest normal breaker position
and a 200A sub-feed breaker at the very botton. This would be entirely safe:
the bus between the 40A breaker and the sub-feed breaker would be protected by
the sub-feed breaker and the bus above the 40A breaker would be protected
as it would be if the breakers were in the other order.

 

[amluto]

The language I wrote was a bit messy. This version may be better.

Where two sources, one a primary power source and the other another
power source, are located on a busbar that contains loads and the
primary power source is at one end of the busbar, the sum of 125 percent
of the power source(s) output circuit current and the rating of the
overcurrent device protecting the busbar shall not exceed 120% the
ampacity of the busbar. The busbar shall be sized for the loads
connected in accordance with Article 220.

If the backfed breaker from the power source is at the opposite end of
the busbar from the primary power source and no devices or conductors
can be installed on the busbar beyond it, then a permanent warning
label shall be applied to the distribution equipment adjacent to the
backfed breaker from the power source that displays the following or
equivalent wording:

WARNING:
POWER SOURCE OUTPUT CONNECTION,
DO NOT RELOCATE THIS OVERCURRENT DEVICE

If the distribution equipment allows additional loads to be connected
between the backfed breaker from the power source and the end of the
busbar opposite the primary power source, then the following conditions
must be met:

(1) The backfed breaker must be installed as far as practicable from
the primary power source.

(2) The sum of the ratings of the overcurrent devices protecting loads
connected between the backfed breaker from the power source and the
end of the busbar opposite the primary power source must not exceed
the rating of the busbar.

(3) A permanent warning label shall be applied to the distribution
equipment adjacent to the backfed breaker from the power source that
displays the following or equivalent wording:

WARNING: POWER SOURCE OUTPUT CONNECTION, DO NOT RELOCATE THIS
OVERCURRENT DEVICE. THE SUM OF THE RATINGS OF OVERCURRENT DEVICES
CONNECTED BEYOND THIS DEVICE MUST NOT EXCEED THE RATING OF THE BUS.

The warning sign(s) or label(s) shall comply with 110.21(B).

Exception: Equipment with multiple ampacity busbars or center fed
panelboards are not addressed by this provision.
​

 

[Smart $]

The 2017 NEC does not have that Exception according to the online version.

While I understand your concern, it mostly involves an installation to being taken out of compliance after original installation. I say a slight change to the sign wording would be sufficient...

WARNING: POWER SOURCE OUTPUT CONNECTION.
2020 NEC §705.12(b)(2)(b) COMPLIANT. DO NOT RELOCATE
OR ADD LOADS WITHOUT VERIFYING NEC COMPLIANCE.

 

[amluto]

The 2017 NEC does not have that Exception according to the online version.

Whoops, I need to fix that.

While I understand your concern, it mostly involves an installation to being taken out of compliance after original installation. I say a slight change to the sign wording would be sufficient...

WARNING: POWER SOURCE OUTPUT CONNECTION.
2020 NEC §705.12(b)(2)(b) COMPLIANT. DO NOT RELOCATE
OR ADD LOADS WITHOUT VERIFYING NEC COMPLIANCE.

That would address the safety concern but not the usefulness concern.

Panels with sub-feed breakers are unusual in residential applications but aren't particularly rare, and allowing such a panel to have a utility-interactive inverter attached at the farthest normal breaker position next to the sub-feed breaker seems both safe and useful. My main intent would be to explicitly allow such an installation under the 120% rule.

This could be done with a narrower change along the lines of "Exception: If the backfed breaker is connected to the busbar as possible from the primary source, then a single sub-feed breaker may be installed beyond it so long as the sub-feed breaker's rating does not exceed the busbar's rating." It seems worthwhile to make the exception more general, though.

 

[Smart $]

Whoops, I need to fix that.



That would address the safety concern but not the usefulness concern.

Panels with sub-feed breakers are unusual in residential applications but aren't particularly rare, and allowing such a panel to have a utility-interactive inverter attached at the farthest normal breaker position next to the sub-feed breaker seems both safe and useful. My main intent would be to explicitly allow such an installation under the 120% rule.

This could be done with a narrower change along the lines of "Exception: If the backfed breaker is connected to the busbar as possible from the primary source, then a single sub-feed breaker may be installed beyond it so long as the sub-feed breaker's rating does not exceed the busbar's rating." It seems worthwhile to make the exception more general, though.

Doesn't an end-of-bus sub-feed breaker (or feedthrough) contradict the opposite end condition. The backfed breaker is technically NOT at the opposite end as there would be loads being fed both above and below.

 

[GoldDigger]

Doesn't an end-of-bus sub-feed breaker (or feedthrough) contradict the opposite end condition. The backfed breaker is technically NOT at the opposite end as there would be loads being fed both above and below.

That is why the subfeed breaker must be rated at or below the bus rating.
The problem in general is that if you allow multiple breakers beyond the backfed breaker they can add up to more than the bus rating.

Sent from my XT1585 using Tapatalk

 

[Smart $]

That is why the subfeed breaker must be rated at or below the bus rating.
The problem in general is that if you allow multiple breakers beyond the backfed breaker they can add up to more than the bus rating.

Sent from my XT1585 using Tapatalk

And what I'm saying is if you feedthrough directly or through a sub-feed breaker, you are no longer backfeeding at the opposite end of the bus in a normal breaker position.. and thus subject to 705.12(B)(2)(3)(a) requirements* which has it covered.... and all of this is subject to change in the future and thus the outlet (or prevention) is provided in the sign verbiage.

*I'm hoping I'll get a deal on a NEC 2017 pdf & softbound like I did last cycle. Hasn't happened yet so not able to conveniently quote the section.

 

[amluto]

And what I'm saying is if you feedthrough directly or through a sub-feed breaker, you are no longer backfeeding at the opposite end of the bus in a normal breaker position.. and thus subject to 705.12(B)(2)(3)(a) requirements* which has it covered.... and all of this is subject to change in the future and thus the outlet (or prevention) is provided in the sign verbiage.

*I'm hoping I'll get a deal on a NEC 2017 pdf & softbound like I did last cycle. Hasn't happened yet so not able to conveniently quote the section.

Assuming I transcribed it right from the free online version, it says:

The sum of 125 percent of the power source(s) output circuit current and the rating of the overcurrent device protecting the busbar shall not exceed the ampacity of the busbar.

My reason for proposing the change is that I think that this text disallows a useful and safe installation. Imagine a panel with a 400A busbar, a 400A main breaker on top, a bunch of breaker provisions, and an empty sub-feed provision at the bottom. Let's say that the power source output current is 80A and the power source is connected to a 80A breaker at the bottom-most non-sub-feed position in the panel.

Under 705.12(B)(2)(3)(b), I expect most AHJs to accept this installation as long as the sub-feed provision is unused. It would certainly be okay if the sub-feed provision were not present because 400A + (1.25 * 64A) = 480A, which is 120% of the busbar ampacity. Let's think about why this is safe. The middle of the busbar is protected because current flowing downward is limited by the 400A main device and current flowing upward is limited by the (continuous) 64A output capacity of the power source. Assuming that there aren't terrible harmonics or excessive reactive current going from one load to another, there won't simultaneously be current flowing up the bus and down the bus, so the middle of the bus is protected. The top of the bus is protected by the 400A main device (the panel would presumably never pass UL testing if the contacts to the 400A main device could be overloaded), and the sideways bus stubs at the bottom where the 80A device is attached are protected because the panel's instructions will say something like "the sum of the ratings of devices on each pole may not exceed 100A". This is the point of 705.12(B)(2)(3)(b).

Now suppose that someone installed a 400A sub-feed breaker. (This would be a bit silly, but the panel could plausibly allow this.) This would not be allowed under the 2017 NEC as far as I can tell. Nonetheless, I would argue that it is still safe: everything from the 80A backfed breaker up is protected for the same reason that it would be without the sub-feed breaker. The contacts to the sub-ffeed breaker are protected by the sub-feed breaker. The bus between the 80A backfed breaker and the sub-feed breaker is also protected by the sub-feed breaker. (Even if there were two sub-feed breakers, it would still be safe as long as the sum of the ratings of the sub-feed breakers did not exceed 400A.) I would like for this case to be permitted.

The installation that should not be allowed is for feed-through lugs to be installed in the sub-feed position. With the power supply running at full output, the feed-through lugs could draw 384A continuously or 464A for a few hours without tripping any of the overcurrent devices, either of which exceeds the rating of the busbar. (If the whole installation were 100% rated, then the 464A short-term draw would still exceed the rating.)

 

[GoldDigger]

Unfortunately the whole 120% rule is a compromise between physics and CYA.

Sent from my XT1585 using Tapatalk

 

[amluto]

Unfortunately the whole 120% rule is a compromise between physics and CYA.

Isn't the whole NEC more or less this compromise? That doesn't necessarily mean that a more lenient rule here couldn't still achieve the CYA goal.

 

[Smart $]

Isn't the whole NEC more or less this compromise? That doesn't necessarily mean that a more lenient rule here couldn't still achieve the CYA goal.

I'd say less a compromise than it is more... and for that part that is more, the compromise is usually on the conservative side. The point here is that the compromise has already been made when you compare to 705.12(B)(2)(3)(a) to (b)...i.e. 120% vs. 200%. You will be asking for more leniency than 120% provides. I'm willing to bet that your amended version, even if it presents an excellent theoretical substantiation, the CMP will demand that you provide substantiation through tests made in cooperation with one or more NRTL.

I would like the stipulation to allow 200% as long as sum of load OCPD's does not exceed the bus rating. Just look at any service with two to six disconnecting means and understand that the sum of the service OCPDs can be greater than the service conductor ampacity as long as the service conductor ampacity is not less than required for the load.

 

[amluto]

I'd say less a compromise than it is more... and for that part that is more, the compromise is usually on the conservative side. The point here is that the compromise has already been made when you compare to 705.12(B)(2)(3)(a) to (b)...i.e. 120% vs. 200%. You will be asking for more leniency than 120% provides. I'm willing to bet that your amended version, even if it presents an excellent theoretical substantiation, the CMP will demand that you provide substantiation through tests made in cooperation with one or more NRTL.

I neither have the capability to make that happen nor the funding to pay an NRTL to do it for me. (Nor do I know exactly what would be tested. Would they set up a 400A panel as I described, max it out, point a thermal camera at it, and make sure that everything is safe until breakers start tripping?) I could submit to the code making process anyway and see what happens, though.

I could, just maybe, arrange for a test like that at a very much nationally-non-recognized lab, and it might not even cost that much. Ebay has the right kind of panelboards for ~$200 and used breakers aren't all that expensive. The tricky part would be finding or borrowing a power supply that can generate a circulate a few hundred amps of 60Hz AC.

I would like the stipulation to allow 200% as long as sum of load OCPD's does not exceed the bus rating. Just look at any service with two to six disconnecting means and understand that the sum of the service OCPDs can be greater than the service conductor ampacity as long as the service conductor ampacity is not less than required for the load.

I don't know the history, but I assumed that the 120% maximum was about limiting the potential nasty effects of harmonics. Suppose your solar inverter generates an output current that is far from being a sine wave. Then I could imagine a load (especially something like a conventional dimmed lighting load) that draws a higher-than-expected fraction of its current from the inverter. If you plug the top of the panel into the service, put the inverter at the bottom, have a lot of dimmed lighting loads connected near the top and have a lot of linear loads connected near the bottom, I could imagine harmonic currents in the middle of the panel that are a decent fraction of the sum of the two supply currents. With the 120% rule, you'd have to have a very abusive load indeed to cause a problem, but with 200%, there could be less margin for error.

For what it's worth, I bet you can damage a normal panel with just one supply with abusive enough loads. Just plug a few big capacitor banks in at one end of the busbar some inductors with roughly equal (but opposite) reactance at the other end of the panel. You'll get reactive power flowing through the busbar and there's no practical limit that I can see on how much reactive current could flow relative to the size of the main OCPD. Of course, you could probably only do this intentionally, and the NEC probably has something to say about capacitor banks, too.

As for services, I don't know what the NESC says about it, so I don't really know what keeps people from exceeding calculated load an overheating a service conductor. My house has an overhead service drop that I'm quite confident is not adequately protected by the single OCPD at the service entrance. I think it is, however, indirectly protected by the distribution transformer on the pole -- if I believe the stenciled numbers on that transformer, it can't supply enough current to damage the insulation on my service drop. I hope that either the NESC or the power company's common sense requires that something protect at least the utility-owned conductors.

 

[don_resqcapt19]

...
That would address the safety concern but not the usefulness concern. ...

The code scope is limited to the former. The latter is outside the scope of the NEC and is a design issue, not a code issue.

 

[GoldDigger]

As a general rule POCO does not try to protect secondary (service) conductors. If they burn up, it is outside.

Sent from my XT1585 using Tapatalk

 

[Smart $]

As a general rule POCO does not try to protect secondary (service) conductors. If they burn up, it is outside.

True, but service entrance conductors, though also outside or considered such for the most part, fall under NEC jurisdiction. I'm fairly certain there are many multi-disconnect services out there that the sum of the OCPDs is greater than 120% of the service entrance conductor's ampacity.

 

[pv_n00b]

Amluto, I like your proposal in principal. I would simplify it by just adding the allowance of a feed through breaker in a panel as an exception.

 

[GoldDigger]

True, but service entrance conductors, though also outside or considered such for the most part, fall under NEC jurisdiction. I'm fairly certain there are many multi-disconnect services out there that the sum of the OCPDs is greater than 120% of the service entrance conductor's ampacity.

And, because most feeder or branch circuits are not loaded to breaker capacity, even in commercial use, I do not see any inherent problem in the sum being greater than the conductor ampacity. NEC does not allow that when there is a single breaker, and a POCO may well throw in a 100% or 120% requirement even though the service disconnect is after the service point. But not necessarily with the expectation that it will somehow protect their conductors or transformer secondaries.