PV Disconnect N/G bonded

[Fred B]

It's kind of implicit in the installation instructions. The 'line' side should be understood to be the side which remains energized when the disconnect is opened. Many people new to solar either don't understand that interactive inverters de-energize their outputs when disconnected, or they get confused by the typical direction of power flow, which they should ignore in this case.

It seems for this, power is in some ways in both directions as on one hand it is sending power to the utility company, but on the other it is sending power back thru to the inverter that shuts down in the event utility power is lost. Thus the question because is it line or load at any given moment. And it would seem any breaker used must be rated for back feeding. And is there a code reference that requires the back feed breaker be secured in place, like I've seen in other applications?

 

[ggunn]

I am not aware of any code rule that requires the utility connections to be on the top. Can you cite the rule?

Well, if a fused disco to a supply side interconnection were wired with the utility on the load side of the switch, would you want to change the fuses?

 

[ggunn]

It seems for this, power is in some ways in both directions as on one hand it is sending power to the utility company, but on the other it is sending power back thru to the inverter that shuts down in the event utility power is lost. Thus the question because is it line or load at any given moment. And it would seem any breaker used must be rated for back feeding. And is there a code reference that requires the back feed breaker be secured in place, like I've seen in other applications?

In the 2020 NEC see 705.12 (D) and (E). The language is in earlier NEC versions as well but in a different place.

 

[romex jockey]

In the 2020 NEC see 705.12 (D) and (E). The language is in earlier NEC versions as well but in a different place.

more fat for the PV fire>>>>

705.12(B)(4) Suitable for Backfeed. Circuit breakers, if backfed, shall
be suitable for such operation.

Informational Note: Fused disconnects, unless otherwise
marked, are suitable for backfeeding.

(5) Fastening. Listed plug-in-type circuit breakers backfed
from electric power sources that are listed and identified as
interactive shall be permitted to omit the additional fastener
normally required by 408.36(D) for such applications.

~RJ~

 

[jaggedben]

.... Thus the question because is it line or load at any given moment. ....

To repeat what I said above, just forget about all that and ask yourself: "Which terminals are going to remain energized when the disconnect is opened?".

...

As something of an aside, after years of trying to educate everyone that solar inverter disconnects and breakers are really basically the same as any other circuit, I'm now increasingly installing with energy storage systems with disconnects that truly are energized from both sides. Which makes me increasingly annoyed with all the warning labels that I've been asked to put on solar disconnects in the past, which are essentially crying wolf.

 

[Fred B]

more fat for the PV fire>>>>

705.12(B)(4) Suitable for Backfeed. Circuit breakers, if backfed, shall
be suitable for such operation.

Informational Note: Fused disconnects, unless otherwise
marked, are suitable for backfeeding.

(5) Fastening. Listed plug-in-type circuit breakers backfed
from electric power sources that are listed and identified as
interactive shall be permitted to omit the additional fastener
normally required by 408.36(D) for such applications.

~RJ~

That just drew my attention to 705.12(B)(2)(c). This is the first time I recall instruction to tallying the sum of the ampere ratings of all overcurrent devices on panelboard both load and supply devices and shall not exceed the rating of the busbar (just 10- 20Amp breakers will fill the capacity of the panel). Normally it is viewed the individual ratings are not counted but actual loads. I don't know of many panels that are not overloaded from that perspective of 705 and then would not allow adding into that panel a parallel power source.

 

[jaggedben]

Assuming a the disconnect is a breaker, there are very few marked line and load, and no breaker that is permitted to be used on a solar system could be marked line and load, as such markings mean that the breaker is not suitable for back-feeding.

If the disconnect, is a knife switch, then the top would be the line per code rules.

If it's a bolt on breaker that's correct, but if it's a plug on breaker then it raises the issues of hold down provisions mentioned in other posts. For interactive inverters the utility should always be considered the line side. That's the best rule of thumb.

 

[jaggedben]

That just drew my attention to 705.12(B)(2)(c). This is the first time I recall instruction to tallying the sum of the ampere ratings of all overcurrent devices on panelboard both load and supply devices and shall not exceed the rating of the busbar (just 10- 20Amp breakers will fill the capacity of the panel). Normally it is viewed the individual ratings are not counted but actual loads. I don't know of many panels that are not overloaded from that perspective of 705 and then would not allow adding into that panel a parallel power source.

Note that that is only one of three options for qualifying a panelboard for backfeed. You can pick one, you don't have to meet all three.

 

[Fred B]

Note that that is only one of three options for qualifying a panelboard for backfeed. You can pick one, you don't have to meet all three.

So let's add a wrinkle. If you are using a meter/main as the tapping point for the parallel service, if trying to us option (c) do go upstream to the primary distribution panel an count all those loads as well, or as it seems just the load associated with the individual busbars in the meter/main?

 

[jaggedben]

So let's add a wrinkle. If you are using a meter/main as the tapping point for the parallel service, if trying to us option (c) do go upstream to the primary distribution panel an count all those loads as well, or as it seems just the load associated with the individual busbars in the meter/main?

Your wording is a bit confusing. A meter main usually cannot be 'tapped' without violating the listing, unless you're using that word in a less technical fashion. Also what distribution would be upstream of a meter main?

Anyway, each panelboard that is in series between the interactive source (inverter) and the primary source (utility, usually) needs to meet one of those three options. They do not all need to follow the same one, you can pick and choose. Hope that answers your question.

 

[romex jockey]

So let's add a wrinkle. If you are using a meter/main as the tapping point for the parallel service, if trying to us option (c) do go upstream to the primary distribution panel an count all those loads as well, or as it seems just the load associated with the individual busbars in the meter/main?

I so it all the time via>>

Buy Electrical Supplies, Components, & Parts Online - Westway Electric Supply

We offer electrical equipment and supplies to any consumer at wholesale prices. Huge selection. Shop our inventory online or call 888-349-1196.

www.westwayelectricsupply.com

~RJ~

 

[ggunn]

So let's add a wrinkle. If you are using a meter/main as the tapping point for the parallel service, if trying to us option (c) do go upstream to the primary distribution panel an count all those loads as well, or as it seems just the load associated with the individual busbars in the meter/main?

No. The provision in the code only applies to load and supply breakers directly attached to the panel busbars. One interesting wrinkle, though, comes in when you are connecting two phase loads and supplies to a three phase panel. To qualify the bus you need only total up the loads and supplies connected to each busbar rather than just totaling all the breaker ratings and comparing the sum to the bus rating.

 

[ggunn]

If it's a bolt on breaker that's correct, but if it's a plug on breaker then it raises the issues of hold down provisions mentioned in other posts. For interactive inverters the utility should always be considered the line side. That's the best rule of thumb.

The issue is if the plug on breaker pulls loose from the bus whether it is still hot or not. For a breaker from a grid tied utility interactive inverter, compliance with UL1741 ensures that the inverter shuts down when it stops seeing the grid, so a hold down is not required.

 

[don_resqcapt19]

If it's a bolt on breaker that's correct, but if it's a plug on breaker then it raises the issues of hold down provisions mentioned in other posts. For interactive inverters the utility should always be considered the line side. That's the best rule of thumb.

Should does not equal shall and rules of thumb are not enforceable code.

 

[Carultch]

The issue is if the plug on breaker pulls loose from the bus whether it is still hot or not. For a breaker from a grid tied utility interactive inverter, compliance with UL1741 ensures that the inverter shuts down when it stops seeing the grid, so a hold down is not required.

There is an exception for listed utility interactive inverters to not require this separate fastener on the plug-on breaker. In 2014, it was 705.12(D)(5). I didn't see it get removed from the later versions of the NEC, and it probably moved to somewhere in 705.12(B). The inverter is subordinate to another source forming the voltage waveform of the grid, and behaves just like any other load in terms of whether it is on or off, and what scope of the AC circuit is energized. The only difference is that the current waveform is designed to be 180 degrees out of phase with the voltage waveform, in order for it to put power onto the grid instead of draw power from the grid.

Where this rule regarding a separate fastener for the breaker would need to be applied, is if you are using a plug-on breaker as the main breaker of an AC combining load center, even though power ordinarily flows through it in this application the way it was originally intended. It is the fact that the plug-on breaker has its wired terminals remaining energized after unplugging the breaker, that makes it require a separate fastener.

 

[jaggedben]

Should does not equal shall and rules of thumb are not enforceable code.

The hold-down rule in 408 is certainly enforceable code. The rule of thumb I suggested is simply guidance on how to apply the code.

 

[ggunn]

The inverter is subordinate to another source forming the voltage waveform of the grid, and behaves just like any other load in terms of whether it is on or off, and what scope of the AC circuit is energized. The only difference is that the current waveform is designed to be 180 degrees out of phase with the voltage waveform, in order for it to put power onto the grid instead of draw power from the grid.

I have no idea what you are trying to say with that, but yes, backfed breakers connecting PV utility interactive inverters to the grid by code do not require hold downs.

 

[hhsting]

For all practical purposes, a disconnect for a line side solar tap is a service disconnect and you bond in all service disconnects.
The 2023 code will take a big step in clearing this up as the Correlating Committee has directed such disconnects will be called service disconnects in the 2023 code. They are connected to the utility supply as have all of the hazards as any other disconnect connected to the utility supply.

I am not sure that the 2017 and earlier codes directly addressed this issue, but the 2020 code does in 250.25.

I wasn’t aware 2020 code address this issue?
I am also not aware that 2023 code will address this issue or not? Where do you see this?

Is their anywhere in NEC 2020 it address to group the building AC main service disco with main AC solar disco tapped ahead of building main service disco?


I am dealing with same situation and now our AHJ is now trying to modify and put in local amendment that AC solar main disconnect is service disconnect. I am not sure if this is good idea or not.

 

[jaggedben]

I wasn’t aware 2020 code address this issue?
I am also not aware that 2023 code will address this issue or not? Where do you see this?

Is their anywhere in NEC 2020 it address to group the building AC main service disco with main AC solar disco tapped ahead of building main service disco?


I am dealing with same situation and now our AHJ is now trying to modify and put in local amendment that AC solar main disconnect is service disconnect. I am not sure if this is good idea or not.

It's not addressed in the 2020 NEC. There were proposals to address it that didn't make it into the final version.

I think treating main disconnects for supply side solar connections as service disconnects is a fine idea. I've always thought so. The fault currently and grounding issues from a physics point of view are for all practical purposes the same. (We are only talking about supply side connections here. Treating load side solar connections as service disconnects would be supremely stupid. Just to be clear.)

 

[don_resqcapt19]

I wasn’t aware 2020 code address this issue?
I am also not aware that 2023 code will address this issue or not? Where do you see this?

Is their anywhere in NEC 2020 it address to group the building AC main service disco with main AC solar disco tapped ahead of building main service disco?


I am dealing with same situation and now our AHJ is now trying to modify and put in local amendment that AC solar main disconnect is service disconnect. I am not sure if this is good idea or not.

My comment about the 2020 addressing this issue in 250.25 was in reply to the bonding and grounding part of the question. My comment says that you treat the line side tap for the solar disconnect like a service disconnect for the purposes of bonding and grounding under the 2020 code.

As far as the 2023, I have sat in on a number of Task Group meetings and Code Panel meetings. There was a specific Task Group set up to submit comments on the issue of the line side solar disconnect being a service disconnect. The TG has members from CMPs 4, 5 and 10. However they receied direction from the Correlating Committee that it WILL be a service disconnect in the 2023 code and the TG is working out the details and language changes needed to accomplish that. It appears that CMP 10 will have purview over the disconnect and the the purview of CMP 4 will start at the load terminals of the disconnect.

However, the issue is not clear in the 2017 code. You work for the AHJ, so you or your boss makes the call.