690.64 (B)(2) - Bus or Conductor Rating

[jfinley]

Take a look at the second part of 690.64(B)(2):

(2nd Part) In systems with panelboards connected in series, the rating of the first OCPD directly connected to the outputof the utility-interactive inverter(s) shall be used in the calculations for all busbars and conductors.

This reads to me that when an inverter is backfed to a panelboard (Either dedicated for the PV system or an existing house panelboard). And then backfed to a main service. That the 120% rule applies to the panelboard 1st and then to the main service.

Is that how you interpret this section?

 

[jaggedben]

I agree. Otherwise, there would be no point in referring at all to "panelboards connected in series."

BTW, in 2011 this is moved to 705.12 (D)(7), which means that it also applies to the positioning of the breaker(s) if the 120% rule is used.

The only thing that's a little vague involves your mention of a panelboard "dedicated for the PV system." In cases where there is such a dedicated solar AC combiner with multiple breakers feeding inverters, do you add up the value of the breakers feeding the inverters? That seems to be the best interpretation but it's not spelled out.

 

[jfinley]

Thanks for the reply.

 

[Smart $]

... That the 120% rule applies to the panelboard 1st and then to the main service.[/SIZE]

Is that how you interpret this section?

Well, yes... but the calculation seldom, if ever, affects anything beyond the second ocpd encountered.

Say for example you're backfeeding an MLO panelboard rated 200A with a 90A breaker and the feeder breaker is 150A (total is 120% of bus). You have to run a 200A rated feeder because of the 120% requirement. Actually this is overkill for nominal operating conditions, but because it is possible a fault could draw full power from both sources before tripping both breakers, it is a must.

Now let's consider how the service panel is affected. Your calculated load on the subpanel is 150A or less because your PV system doesn't provide any power at night. Your service panel has to be rated for those loads also. The maximum backfeed from the PV system is 90A to the service panel (subpanel has a no-operating-load condition and PV system is at max output). When this occurs, the service isn't supplying any of the 150A subpanel load, so the 90A being backfed plus any additional service-supplied current to other loads on the service panel cannot be more than the service panel bus is rated for (which has to be equal or greater than the calculated load of the subpanel, plus any additional loads served by the main panel).

 

[jaggedben]

Believe it or not, I attended a Bill Brooks seminar this morning where we talked about this.

Well, yes... but the calculation seldom, if ever, affects anything beyond the second ocpd encountered.

Right. Actually, Bill said the language was put in the code to clarify that this is so. In the case of your example...

Say for example you're backfeeding an MLO panelboard rated 200A with a 90A breaker and the feeder breaker is 150A (total is 120% of bus).

...the language is there to make clear that when you get back to the main, you can still use 90A and you don't have to use 150A, even though that 150A breaker is 'feeding' the main. Evidently the lack of this language caused people problems with inspectors back in what is now becoming the old days.

You have to run a 200A rated feeder because of the 120% requirement. Actually this is overkill for nominal operating conditions, but because it is possible a fault could draw full power from both sources before tripping both breakers, it is a must.

Well, we discussed this too, and evidently Bill disagrees with John Wiles and thinks you don't have to size the conductor for the sum of the currents from both directions. Personally I think Bill's thinking makes more sense. To continue with your example, the max current that can flow 'forward' is 150A, and the max current that can flow 'back' is 90A. It is not physically possible for both of these currents to flow in both directions in the same part of the same conductor at the same time, even in the case of a fault. No part of the conductor will ever see more than 150A. So if the conductor is properly sized to feed the subpanel through a 150A breaker then you don't have to change it to backfeed. But as far as interpreting what the code requires, I guess that's something of an opinion (and I hope I represented the opinions of the mentioned names accurately).

 

[Smart $]

Believe it or not, I attended a Bill Brooks seminar this morning where we talked about this.



Right. Actually, Bill said the language was put in the code to clarify that this is so. In the case of your example...



...the language is there to make clear that when you get back to the main, you can still use 90A and you don't have to use 150A, even though that 150A breaker is 'feeding' the main. Evidently the lack of this language caused people problems with inspectors back in what is now becoming the old days.



Well, we discussed this too, and evidently Bill disagrees with John Wiles and thinks you don't have to size the conductor for the sum of the currents from both directions. Personally I think Bill's thinking makes more sense. To continue with your example, the max current that can flow 'forward' is 150A, and the max current that can flow 'back' is 90A. It is not physically possible for both of these currents to flow in both directions in the same part of the same conductor at the same time, even in the case of a fault. No part of the conductor will ever see more than 150A. So if the conductor is properly sized to feed the subpanel through a 150A breaker then you don't have to change it to backfeed. But as far as interpreting what the code requires, I guess that's something of an opinion (and I hope I represented the opinions of the mentioned names accurately).

It's still debatable, even with the 2011 changes...

It is true that at any one point in time that current can only travel in one direction on a bus or wire conductor... but that is only correct when current enters that conductor on one end and leaves on the other. If you have a ground fault in the middle of the conductor, this is no longer true, as the conductor acts the same as an individual conductor to each side of the fault. In the case of my example, say you have a non-bolted ground fault somewhere along the feeder. 150A could flow from the service to the fault point and at the same time 90A could flow from the PV system to the fault point... and not trip either breaker. Personally, I do not recommend backfeeding an MLO panel... I just used it as an example.

 

[jaggedben]

It's still debatable, even with the 2011 changes...

Actually, I don't think it's debatable what the code says. The code requires the feeder to be sized for 200 amps like you said. (And I certainly shouldn't have accidentally implied that Bill Brooks disagrees with this, bad wording.)

The point is simply that in actual fact, you won't get 200 amps flowing on the conductor. Maybe there is some reason related to power factor that I don't fully understand that means this isn't exactly so, but as you said, under normal operating conditions it is overkill. The 120% rule is pretty conservative as far as I can see (although with so many middle-fed panelboards out there, probably necessary for panelboards, if not conductors).

say you have a non-bolted ground fault somewhere along the feeder. 150A could flow from the service to the fault point and at the same time 90A could flow from the PV system to the fault point... and not trip either breaker.

Yes, but you still don't have 240A flowing on any part of the conductor. If the fault current isn't enough to trip the 150A breaker without back feed, well then, it's not enough to trip it with back feed to the fault either.

Personally, I do not recommend backfeeding an MLO panel... I just used it as an example.

Well, yes, if it's an existing panel, you have to check the rating of the feeder conductors as well as the panelboard, but I guess I don't see any reason to 'not recommend' it. In any case, it's done all the time.

 

[Smart $]

Actually, I don't think it's debatable what the code says. ...

I didn't mean debating the code, but rather its interpretation and implementation... namely the 120% rule affecting the MDP when a subpanel is backfed.

The rest of your post is otherwise well received by me :thumbsup: