2008 Art 690 doubts

[PetrosA]

I don't have the exact quote with me, unfortunately. My question concerns backfeeding a subpanel from an inverter and how this affects downsizing of the main breaker. Since code says that the backfeeding breaker (in this case a 2 pole 60A in a subpanel shall be used for calculations for ALL bussbars and conductors, shouldn't the main breaker of a service be downsized to accommodate the backfeed as well as the subpanel's disconnect? I had a solar seller/contractor tell me yesterday that the way I'm doing it is wrong, that only the subpanel disconnect need be downsized and that he's been flagged by inspectors for additionally downsizing the main. He claims to have gotten clarification about this from a number of inspectors as well as the same opinion from classes he's been to. The reasoning is that because there are multiple busses, the heat is dispersed better. That sounds like bunk to me.

Any thoughts?

 

[Smart $]

Only the panel being backfed needs the other supply ocpd's downsized. Off the top of my head, isn't it something like the sum of supply breakers cannot exceed 125% of the panel's rating. The reasoning, I believe, is that both supplies could be feeding the panel (and thus its buses) at the same time and up to their trip ratings. This would not affect the feeding panel's buses.

 

[PetrosA]

Ok, I have the code.

Art. 690.64 (B)(2) "In systems with panelboards connected in series, the rating of the first overcurrent device directly connected to the output of a utility interactive inverter shall be used in the calculations for all busbars and conductors."

This implies that it's not enough to just reduce the OCP for the subpanel as I understand it. Am I misunderstanding this?

The inverter is able to feed all the way back to the utility, so how could that not affect the rating of the main busbar? In theory, you could have a large enough load being drawn from the utility through the main panel and feed additional current from the inverter that would then overload the busbar on the main panel, while the busbar in the subpanel would be fine carrying only the current being fed from the inverter. Why else would they say "all busbars etc."?

 

[bluesmoke...oops]

yes it could feed out, but also consider that your inverter production is going to serve the attached loads of your system before it goes out to the utility. And if it is connected on a subpanel, it's limited in how much amperage it can send back into the system by the overcurrent device. You are allowed to have the two sources c/b's add up to 120% of the bus rating of the panel you are connected to.

 

[elohr46]

Ok, I have the code.

Art. 690.64 (B)(2) "In systems with panelboards connected in series, the rating of the first overcurrent device directly connected to the output of a utility interactive inverter shall be used in the calculations for all busbars and conductors."

This implies that it's not enough to just reduce the OCP for the subpanel as I understand it. Am I misunderstanding this?

The inverter is able to feed all the way back to the utility, so how could that not affect the rating of the main busbar? In theory, you could have a large enough load being drawn from the utility through the main panel and feed additional current from the inverter that would then overload the busbar on the main panel, while the busbar in the subpanel would be fine carrying only the current being fed from the inverter. Why else would they say "all busbars etc."?

I have to agree with you, it has to impact the main buss the same way it does on the sub panel.

 

[bluesmoke...oops]

But unless you are in a fault condition, shouldn't your system only pull as much current as is required from the utility?

Say you need 500 amps from the utility. If you put in an inverter that supplies 60 amps, then you will only be drawing 440 amps from the utility. The system will balance out.

Now if you only need 200 A from the utility, and your inverter can produce say 250 A, then 200A from your inverter will service your loads, and the other 50A will go back out onto your utility grid.

Either way, it's not like your inverter is going to produce 60 A, and your utility is going to push more current in that you can handle into your system. If that was the case, the utility system would do that regardless of the inverter being there.

 

[Smart $]

Ok, I have the code.

Art. 690.64 (B)(2) "In systems with panelboards connected in series, the rating of the first overcurrent device directly connected to the output of a utility interactive inverter shall be used in the calculations for all busbars and conductors."

This implies that it's not enough to just reduce the OCP for the subpanel as I understand it. Am I misunderstanding this?

The inverter is able to feed all the way back to the utility, so how could that not affect the rating of the main busbar? In theory, you could have a large enough load being drawn from the utility through the main panel and feed additional current from the inverter that would then overload the busbar on the main panel, while the busbar in the subpanel would be fine carrying only the current being fed from the inverter. Why else would they say "all busbars etc."?

Let's say for example your subpanel is rated 100A with service-supplied feeder at 60A and inverter-supplied feeder at 60A (120% total on 100A panel). In order for 60A to be fed back to the main panel, it cannot be feeding 60A to the subpanel (current can only travel one direction at a time). The main panel should be rated by calculation to supply 60A to the subpanel, so when it is not and the inverter is feeding 60A back to the main panel, the current on the main panel buses is no more than if it were feeding 60A to the subpanel.

IMO, there is no need for this requirement (at least this stringent). Where properly sized by Chapter 2 requirements, the subpanel's load will prevent the buses from ever seeing more than it is rated for under nominal conditions. It would take several and simultaneous branch circuit faults before the buses would ever see more than they are rated for, and that is what the branch circuit OCPD's are suppose to alleviate.

 

[PetrosA]

So I guess my understanding and interpretation was wrong and the only time you need to downsize the main breaker is when you backfeed directly into the main panel, correct? I don't know that I understand it mathematically, but I also would think that the likelihood of actually overloading the panel to the point where a bus burns out is small...

 

[acrwc10]

So I guess my understanding and interpretation was wrong and the only time you need to downsize the main breaker is when you backfeed directly into the main panel, correct? I don't know that I understand it mathematically, but I also would think that the likelihood of actually overloading the panel to the point where a bus burns out is small...

Wrong, you were correct the first time. The code is clear, it says "ALL busbars and conductors." All means "ALL" not one, or some, or just the first, it means ALL the busbars and conductors. Like it or not it is possible to over feed the main busbar from a distribution panel, fed from an inverter, just as easy as if the inverter were connected to the main. The "if" scenarios are endless, but if the downstream panel is back fed from an inverter and if there is no load on that panel and if it was a really cold day (solar panels can put out 125% of there rating when cold) and the solar panels were at full power production, and if there was a overload on the main panel, then you could be overloading the buss and the main breaker would not shut down.
Lots of "if's" but it is possible, and If it is possible then it could cause a fire. That is why ALL busbars and conductors must be properly rated.

 

[PetrosA]

Wrong, you were correct the first time.

I love it when a wrench gets thrown in Your arguments and the ones in my head make more sense to me logically, but the number of people and inspectors who seem to feel otherwise makes it tough to feel super confident about it.

Your comment about a subpanel with no load exactly describes the situation at the installation I'm working on now. Two 5 kW inverters will feed into two QO230s in a subpanel, and then out through a QO260 to the AC disconnect, then back to the main panel - so while it's a subpanel, it's also kind of an AC combiner as well and will contain no other circuits. In my mind, the breaker backfeeding from this subpanel in the main panel becomes the de facto interface point between the two sources - so the BR260 backfeeding and the 200A main are now potentially feeding 260 amps into a bus rated for 200A. That puts us over the 125% limit before we even get to winter and higher output from the panels.

 

[Smart $]

I love it when a wrench gets thrown in Your arguments and the ones in my head make more sense to me logically, but the number of people and inspectors who seem to feel otherwise makes it tough to feel super confident about it.

While not a total alleviation on your mode of thinking, but FWIW 690.64 in the 2011 NEC points without elaboration to 705.12. Reasoning was to eliminate redundant requirements in 690, 692, 695 and 705; the 705.12(D)(2) requirement does not have the second sentence in either the 2008 or 2011 NEC. Technically, the requirement is still there, just not emphasized... which is all the second sentence of 690.64(B)(2) does.

Your comment about a subpanel with no load exactly describes the situation at the installation I'm working on now. Two 5 kW inverters will feed into two QO230s in a subpanel, and then out through a QO260 to the AC disconnect, then back to the main panel - so while it's a subpanel, it's also kind of an AC combiner as well and will contain no other circuits. In my mind, the breaker backfeeding from this subpanel in the main panel becomes the de facto interface point between the two sources - so the BR260 backfeeding and the 200A main are now potentially feeding 260 amps into a bus rated for 200A. That puts us over the 125% limit before we even get to winter and higher output from the panels.

I'm not going to say you are wrong... and will say you are interpreting the requirement correct, technically. It is just my opinion this requirement is a little too restrictive. As a counterpoint, consider 230.90(A) Exception No. 3, where the sum of two to six service OCPD's is permitted to exceed the ampacity of service conductors without limitation, provided the calculated load is not more than the ampacity of the service conductors. Granted, service conductors are essentially on the outside of buildings... but we still don't want them burning up and causing a fire or safety hazard... which is entirely possible if the system is overloaded. Yet for the type of installation discussed here, we are considering system overload as imminent, which it should never be.

 

[elohr46]

I love it when a wrench gets thrown in Your arguments and the ones in my head make more sense to me logically, but the number of people and inspectors who seem to feel otherwise makes it tough to feel super confident about it.

Your comment about a subpanel with no load exactly describes the situation at the installation I'm working on now. Two 5 kW inverters will feed into two QO230s in a subpanel, and then out through a QO260 to the AC disconnect, then back to the main panel - so while it's a subpanel, it's also kind of an AC combiner as well and will contain no other circuits. In my mind, the breaker backfeeding from this subpanel in the main panel becomes the de facto interface point between the two sources - so the BR260 backfeeding and the 200A main are now potentially feeding 260 amps into a bus rated for 200A. That puts us over the 125% limit before we even get to winter and higher output from the panels.

Again I agree with you, but it begs the question if you are adding a subpanel why not do a line side connection and eliminate the 120% rule of the existing panel entirely?

 

[bluesmoke...oops]

Technically, instead of making the main smaller, you could upsize the bus correct?

 

[Smart $]

Technically, instead of making the main smaller, you could upsize the bus correct?

Yes... but regarding panelboards, that usually means changing out the panelboard entirely. Depending on its calculated load, this may be a necessity if backfeeding the panelboard. After all, PV systems generally have little output at night... so the service supplied OCPD has to be rated for the panel's full load (not to be construed with the panelboards max current rating). Service conductor or feeder tap may be the less costly alternative... especially for existing systems.

 

[bluesmoke...oops]

I'm a design engineer, so for the most part I'm used to just being able to size it as I need to do what I want. Or the existing system has enough capacity in it that adding the amount of PV that they want won't bump up against the 120% rule, no matter what panel you hook it too.

 

[BillK-AZ]

Art 690/2008 vs 2005

Art 690/2008 vs 2005

John Wiles of Southwest Technology Dev. Inst. gives talks on PV systems and the NEC. He has a set of drawings that explain this. (no link available)

Prior to the 2008 version, each panel had to be evaluated per the sum of the breakers feeding it can not exceed 120% of the rating. This meant that if the PV breaker was a 20A in a 100A subpanel (fed by a 100A breaker in the main panel), then the main panel had to be evaluated using the 100A breaker plus the main breaker not exceeding 120% of the main panel rating. Clearly difficult to work with.

The 2008 version wording means that only the 20A breaker needs to be considered when evaluating the main panel (and any other panels between the PV connection and the main). The 2008 version added the requirement that if the sum of the main and backfed breakers exceeds 100% of the panel rating, then the backfed breaker "...shall be positioned at the opposite (load) end from the input feeder location or main circuit location," (Art 690.64(b)(7).

 

[elohr46]

John Wiles of Southwest Technology Dev. Inst. gives talks on PV systems and the NEC. He has a set of drawings that explain this. (no link available)

Prior to the 2008 version, each panel had to be evaluated per the sum of the breakers feeding it can not exceed 120% of the rating. This meant that if the PV breaker was a 20A in a 100A subpanel (fed by a 100A breaker in the main panel), then the main panel had to be evaluated using the 100A breaker plus the main breaker not exceeding 120% of the main panel rating. Clearly difficult to work with.

The 2008 version wording means that only the 20A breaker needs to be considered when evaluating the main panel (and any other panels between the PV connection and the main). The 2008 version added the requirement that if the sum of the main and backfed breakers exceeds 100% of the panel rating, then the backfed breaker "...shall be positioned at the opposite (load) end from the input feeder location or main circuit location," (Art 690.64(b)(7).

So then the 100 amp subpanel feed breaker would have to be positioned opposite end of main panel CB and the 20 amp inverter CB in the subpanel could be positioned anywhere.

 

[BillK-AZ]

Art 690/2008 vs 2005

Art 690/2008 vs 2005

That is correct. But if the PV breaker in this instance was 25A, then it would need to be positioned opposite end of main panel CB.

 

[Smart $]

So then the 100 amp subpanel feed breaker would have to be positioned opposite end of main panel CB and the 20 amp inverter CB in the subpanel could be positioned anywhere.

In his example, there is no mention of the mdu (bus) rating or its mcb rating. As such, we cannot assess the sum or percentage for the mdu.

Regarding the subpanel, the 20A PVI-backfed breaker would have to be positioned opposite the subpanel bus' service-supplied end, i.e. 100A Panel (Bus) Rating < 100A Service-supplied Breaker + 20A PVI-backfed Breaker.

 

[mpd]

the November/December IAEI magazine has a good article called utility interconnections & code requirements by John wiles, there is info on calculating a PV combining load center and a line side connection