Convincing a seasoned EE

[jgatzka]

I'm trying to convince an EE hired by the property owner that there will be very little to no current flow on the Neutral so as not to have to take an 0.8 derate in a 277/480 combiner panel wiring (no loads, just my 3 SMA Core 1 inverters). He's telling me 277V loads inside the building will cause current to flow on the Neutral of my combiner panel wiring from the inverters to the load. Am I crazy in thinking that this happens at the utility transformer and my inverters are not capable of adjusting generation based on equip or lighting loads?

 

[ramsy]

I'm trying to convince an EE hired by the property owner that there will be very little to no current flow on the Neutral so as not to have to take an 0.8 derate..

Unless you can prove otherwise, inverter harmonics, computers, and lighting loads propagate harmonics back to the source, regardless that source being POCO transformers or solar inverters.

NEC 310.15(B)(5)(c) counts nonlinear neutrals as current carrying conductors.

Am I crazy in thinking that this happens at the utility transformer and my inverters are not capable of adjusting generation based on equip or lighting loads?

Did you ever check combiners for harmonics?

 

[electrofelon]

He's telling me 277V loads inside the building will cause current to flow on the Neutral of my combiner panel wiring from the inverters to the load.

Perhaps he doesnt understand how these inverters work and that the neutral is just for voltage sensing purposes and does not carry current.

This has come in in regards to neutral conductor sizing. Some manufacturers specifically state that the neutral is only for voltage and phase sensing and some do not. IIRC SMA is one that does not spell this out.

 

[jaggedben]

I'm trying to convince an EE hired by the property owner that there will be very little to no current flow on the Neutral so as not to have to take an 0.8 derate in a 277/480 combiner panel wiring (no loads, just my 3 SMA Core 1 inverters). He's telling me 277V loads inside the building will cause current to flow on the Neutral of my combiner panel wiring from the inverters to the load. Am I crazy in thinking that this happens at the utility transformer and my inverters are not capable of adjusting generation based on equip or lighting loads?

No you are not crazy. It's not that hard to draw up a diagram which includes the inverter and utility transformer and show by Kickoff's law that the transformer will supply the neutral current. Give it a shot. The key assumption is that the current on each leg of the inverter output is always equal. the inverter dictates that.

 

[jaggedben]

Unless you can prove otherwise, inverter harmonics, computers, and lighting loads propagate harmonics back to the source, regardless that source being POCO transformers or solar inverters.

NEC 310.15(B)(5)(c) counts nonlinear neutrals as current carrying conductors.

That section refers to 'the major portion of the load'. There is no load on an interactive inverter output circuit. I would say that section doesn't apply (whether or not the manufacturer would supply info pertinent to 705.95(B) which would override if applicable).

Did you ever check combiners for harmonics?

Have you? Just curious.

 

[ramsy]

Have you? Just curious.

The industry standard for PV inverters is less than 5% total harmonic distortion (THD) at rated operation, IEEE 1547, UL 1741 and FCC Part 15B.

https://www.solectria.com/site/assets/files/1482/solectria_harmonics_noise_pv_inverters_white_paper.pdf

 

[tortuga]

I'm trying to convince an EE hired by the property owner that there will be very little to no current flow on the Neutral so as not to have to take an 0.8 derate in a 277/480 combiner panel wiring (no loads, just my 3 SMA Core 1 inverters). He's telling me 277V loads inside the building will cause current to flow on the Neutral of my combiner panel wiring from the inverters to the load. Am I crazy in thinking that this happens at the utility transformer and my inverters are not capable of adjusting generation based on equip or lighting loads?

The minimum sized neutral conductor code will allow from a AC combiner panel to a panelboard I'd say is based on the NEC 215.2(A)(2) size of the required ECG.
I would addendum the owner with a smaller neutral option.

 

[jaggedben]

The minimum sized neutral conductor code will allow from a AC combiner panel to a panelboard I'd say is based on the NEC 215.2(A)(2) size of the required ECG.
I would addendum the owner with a smaller neutral option.

But if it is a question of number of current carrying conductors then the line conductors would need to be larger. I think that is the issue.

 

[Carultch]

The key assumption is that the current on each leg of the inverter output is always equal. the inverter dictates that.

In general, that is the case with most 3-phase inverters. It is common that they only use the neutral for the instrumentation purposes of measuring voltage and phase. NEC705.28(B)(2), formerly 705.95(B), allows an EGC-sized neutral if the manufacturer confirms that this is the case for that particular inverter. We're talking milliamps of current. A clamp-on ammeter calibrated to the current you expect on the ungrounded wires, is not even going to be sensitive enough to pick up the neutral current, when only used for instrumentation purposes.

The one exception I know about is the Fronius IG Plus units, that you can no longer get. This working principle could still apply to other inverters, I haven't ruled out that possibility completely. It also would apply if you are combining a triplexed group of single phase inverters operating phase-to-neutral, because their arrays won't necessarily perform uniformly. Fronius had a 11.4kW 208V version that applied its outputs across two phases each in a delta-manner, and a 12 kW 277V version that applied its outputs phase-to-neutral. The inverters were built with 3 internal power stages, each producing the phase-specific AC ouptut. There was a mix-mode where it would cycle through operating only 1 or 2 of the power stages at the same time to be consistent with the DC power available. So if you had 8 kW dc sourcing a 12 kW inverter, it would only operate 2 out of 3 of the power stages in order to improve efficiency. There was an alternative to mix-mode called balance-mode, which would uniformly operate all 3 power stages in the event that the utility requires balancing.

Because of this, the 12kW wye inverters in this family would require a full size neutral. In mix-mode, the inverter could potentially put the full load amperes on the neutral conductor. It would never put the full load amperes on all 4 conductors simultaneously, so it would be unrealistically conservative to treat the neutral as a CCC for derating purposes. Once all 3 phase wires are loaded, the outputs would be equal, and the neutral current would be zero. 12kW at 480V is small enough that you'd put it on a 20A circuit, so the individual inverter would require all 5 conductors to be uniformly sized at min #12Cu anyway. But where it would be significant, is for the aggregate output, after you combine 10 of these inverters in a 200A panel. It would make imply that the aggregate neutral would need to be #3/0 Cu instead of #6 Cu.

 

[electrofelon]

In general, that is the case with most 3-phase inverters. It is common that they only use the neutral for the instrumentation purposes of measuring voltage and phase. NEC705.28(B)(2), formerly 705.95(B), allows an EGC-sized neutral if the manufacturer confirms that this is the case for that particular inverter. We're talking milliamps of current. A clamp-on ammeter calibrated to the current you expect on the ungrounded wires, is not even going to be sensitive enough to pick up the neutral current, when only used for instrumentation purposes.

The one exception I know about is the Fronius IG Plus units, that you can no longer get. This working principle could still apply to other inverters, I haven't ruled out that possibility completely. It also would apply if you are combining a triplexed group of single phase inverters operating phase-to-neutral, because their arrays won't necessarily perform uniformly. Fronius had a 11.4kW 208V version that applied its outputs across two phases each in a delta-manner, and a 12 kW 277V version that applied its outputs phase-to-neutral. The inverters were built with 3 internal power stages, each producing the phase-specific AC ouptut. There was a mix-mode where it would cycle through operating only 1 or 2 of the power stages at the same time to be consistent with the DC power available. So if you had 8 kW dc sourcing a 12 kW inverter, it would only operate 2 out of 3 of the power stages in order to improve efficiency. There was an alternative to mix-mode called balance-mode, which would uniformly operate all 3 power stages in the event that the utility requires balancing.

Because of this, the 12kW wye inverters in this family would require a full size neutral. In mix-mode, the inverter could potentially put the full load amperes on the neutral conductor. It would never put the full load amperes on all 4 conductors simultaneously, so it would be unrealistically conservative to treat the neutral as a CCC for derating purposes. Once all 3 phase wires are loaded, the outputs would be equal, and the neutral current would be zero. 12kW at 480V is small enough that you'd put it on a 20A circuit, so the individual inverter would require all 5 conductors to be uniformly sized at min #12Cu anyway. But where it would be significant, is for the aggregate output, after you combine 10 of these inverters in a 200A panel. It would make imply that the aggregate neutral would need to be #3/0 Cu instead of #6 Cu.

I started a thread a while back and stated which inverters said what. This is two years old now so might be out of date:

inverter neutral size update

Near the top of my list of frustrations with the PV industry is the Inverter neutral conductor (note for the purpose of this I am referring to mid to large size PV systems using 480V string inverters). I do have a bit of good news I would share in case any of y'all were not aware. Please help...

forums.mikeholt.com

 

[Carultch]

This has come in in regards to neutral conductor sizing. Some manufacturers specifically state that the neutral is only for voltage and phase sensing and some do not. IIRC SMA is one that does not spell this out.

For the CORE-1 inverters, SMA does spell this out. It also applied to the Tripower inverters that preceded them. They just don't spell it out in their manual as I'd prefer, and you have to ask on of their applications engineers for documentation.

 

[electrofelon]

For the CORE-1 inverters, SMA does spell this out. It also applied to the Tripower inverters that preceded them. They just don't spell it out in their manual as I'd prefer, and you have to ask on of their applications engineers for documentation.

Maybe this has changed a little, but in the past it seemed most manufacturers that didn't state this in their manuals, would never commit to anything or provide documentation. IIRC, SMA was particularly dodgy about this. Maybe things have changed in the last couple years.

 

[tallgirl]

Maybe this has changed a little, but in the past it seemed most manufacturers that didn't state this in their manuals, would never commit to anything or provide documentation. IIRC, SMA was particularly dodgy about this. Maybe things have changed in the last couple years.

There's a real resistance to documenting too much because then people do things they really shouldn't. Many inverters need the neutral so they "know" where the middle of the wave is. It might not even be needed to carry any current. But there may also be failure modes in which half of the inverter output fails and suddenly you've got a neutral that's carrying current for some length of time before the inverter shuts down..

The documentation should tell you what you need. Do what it says.

 

[tallgirl]

No you are not crazy. It's not that hard to draw up a diagram which includes the inverter and utility transformer and show by Kickoff's law that the transformer will supply the neutral current. Give it a shot. The key assumption is that the current on each leg of the inverter output is always equal. the inverter dictates that.

No, there's no absolute requirement that the current be identical. Indeed, if the neutral is there to be used, it may well not be equal. Recall that the inverter is trying to raise the voltage of its inputs as a way of moving energy out. Thus, the inverter is monitoring voltage, not current.

 

[jaggedben]

No, there's no absolute requirement that the current be identical. Indeed, if the neutral is there to be used, it may well not be equal. Recall that the inverter is trying to raise the voltage of its inputs as a way of moving energy out. Thus, the inverter is monitoring voltage, not current.

Yeah I get that. Didn't mean to present the assumption as a foregone conclusion for any inverter. However, with regard to the OP's question, his main issue seems to be convincing the EE involved that the utility transformer can supply the unbalanced neutral current for the loads on the premise. The assumption that inverter output is always balanced would be valid for proving that conceptually. Once that is out of the way, it comes down to the details of manufacturer documentation, but I'd say the code is pretty non-explicit about whether small imbalances or failure modes need to be considered with regard to current carrying conductor de-ratings.

 

[jaggedben]

...
The documentation should tell you what you need. Do what it says.

Unfortunately I think one would be very lucky if a manufacturer's standard documentation gave explicit guidance on whether the neutral meets NEC definitions of a current carrying conductor. One tends to see boilerplate instructions to follow local codes for conductor sizing.

 

[tallgirl]

Unfortunately I think one would be very lucky if a manufacturer's standard documentation gave explicit guidance on whether the neutral meets NEC definitions of a current carrying conductor. One tends to see boilerplate instructions to follow local codes for conductor sizing.

Without violating any NDAs, the topology of the output section of the inverters I'm most familiar with for residential use can wind up producing a fair bit of the rated output in a somewhat unbalanced manner. I wouldn't bother undersizing the neutral unless the installation instructions explicitly say you can do so. The microinverters I worked on had all 4 conductors -- L1, L2, N, G -- sized the same and we were paying for the wiring in the cable, which were UL-listed, etc.

 

[tallgirl]

Yeah I get that. Didn't mean to present the assumption as a foregone conclusion for any inverter. However, with regard to the OP's question, his main issue seems to be convincing the EE involved that the utility transformer can supply the unbalanced neutral current for the loads on the premise. The assumption that inverter output is always balanced would be valid for proving that conceptually. Once that is out of the way, it comes down to the details of manufacturer documentation, but I'd say the code is pretty non-explicit about whether small imbalances or failure modes need to be considered with regard to current carrying conductor de-ratings.

Is the concern that somehow, in some magical manner, the neutral current could exceed the rating / ampacity without L1 or L2 doing so?

As a side note, I loved to watch the power change direction and the power factor go from 100% to -100%, and sometimes hang out between 20% and -20%. PV equipment is rather ... weird.

 

[jaggedben]

Is the concern that somehow, in some magical manner, the neutral current could exceed the rating / ampacity without L1 or L2 doing so?

Well, to review, the OP has a three phase system and what is at issue is interpreting whether to apply the language in NEC 310.15(B)(5)(c), which affects the size of all the conductors in the feeder, not just the neutral. (So it's rather a bit different than dealing with microinverters.)

I would be interested if you, ramsy, or anyone else can put some math on how harmonics would manifest in this situation when the inverter is configured (essentially) as an even current source on all three phases.

 

[electrofelon]

Without violating any NDAs, the topology of the output section of the inverters I'm most familiar with for residential use can wind up producing a fair bit of the rated output in a somewhat unbalanced manner. I wouldn't bother undersizing the neutral unless the installation instructions explicitly say you can do so. The microinverters I worked on had all 4 conductors -- L1, L2, N, G -- sized the same and we were paying for the wiring in the cable, which were UL-listed, etc.

And I dont really care much about the neutral on resi systems, its probably not even worth my time to try and figure out if I can reduce. My big issue (as I say in that thread from two years ago post #10) is commercial systems say 300k and up. I think all of the 300kw-2.5MW I have worked on had a giant full sized neutral everywhere, 400' to 400 amp combiner panelboards, to inverters, everything. My stance is that these larger three phase string inverters should be designed to not even have a neutral at all. IF you were not aware or didnt read through that link, there are a few manufactures that state in the manual you can skip the neutral and install a jumper between the EGC and neutral terminal on the inverter. So it is clearly very feasible, and should be universal.