This question has been coming up more frequently lately. I answered this many times at my previous company SMA America back in the early 2010s. Rapid shutdown wasn’t even a twinkle in the NEC patriarch’s eyes. Commercial rooftop solar started to take off and some people were leery about using microinverters and optimizers on such large systems. The largest central inverter at the time was the SMA 250kW Sunny Central. Huge for its time, revolutionary even, but by today’s standards it is a tick pulled off a gargantuan 2 MW central inverter, but still not the right inverter for some people.
Installers familiar with string inverters wanted to start stacking them up in these larger applications and the Code watchdogs and utility wienies started to ponder the implications. The National Electric Code and local utilities like PG&E in California had a lot to say about it, and it was good. Since most string inverters back then were single phase (sometimes referred to as split phase, meaning they had 2 hots, a neutral and ground), and most commercial buildings are three-phase (3 hots, a neutral and ground), people started asking questions. Oh, I should have started with a disclaimer, this post is going to get technical and very Codey.
Is it allowable for a single phase inverter (with L1 and L2 output) 2 pole breaker to connect to a 3-phase panel that has L1, L2, and L3 busbars?
This is a valid question considering commercial PV designs had 10 to 20 single phase inverters speced in. The obvious and easiest solution would be to install PV inverters in sets of three so that all phases would be accounted for, meaning no phase on the three phase panel would not be connected to at least one PV inverter output on any leg. Why the big fuss? Phase imbalances.
I am a big fan of citing my sources and did so in my book, The Battery Powered Home. The last section is devoted to my sources so that you know I am not pulling something out of my cloaca or parroting misinformation. However, that level of detail tends to make reading a little longer than normal.
The Code had guidelines for these imbalances to keep three phase motors from burning up if there was too large of a discrepancy between any of the three phases. Motors have manufacturing tolerances and if one leg of a three phase panel had more loads on it than the others, then the voltage would be lower and that phase of the motor would cause premature degradation. PG&E had a 6000W imbalance for this application and so the inspectors and AHJ adopted it for solar applications- as they were used to doing. Some people referenced phantom articles in the Code that stated 2% imbalance between any two lines but in another reference, it said 3%.
The imbalance stipulation did make it into the Code in Solar PV Systems section 690.63, then it was moved to 705.100, and now sits comfortably in 705.45 Unbalanced Interconnections. Before we get into the Code snoozefest, I want to explain the tech behind PV inverters and their grid interconnection.
PV inverters, regardless of how many are sending current to their respective breaker, do not know that there are two, three, or 20 other inverters on that panel. If you are confused about my use of the word “panel”, I am referring to the main service entrance of the installation, the breaker box, the Can, the AC distribution panel, the panel board, et. al. I am not referring to this:
That is a PV module, not a panel. And that’s all I am going to say about that in this post. If you call it a panel, you are wrong.
“But, people call them….”
No, you are wrong.
Multiple inverters feeding the same panel will not have a negative effect on the voltage unless:
1. There is an existing “significant” voltage imbalance.
2. The PV inverters are not installed in groups of three.
Notice the quotes. That leaves a lot of room for interpretation- which can be good or bad. Good for the installers who like these applications, but bad for the installers depending on the utility or AHJ interpretation of “significant”.
NEC 2020 705.45 Unbalanced Interconnections
(A) Single Phase. Single-phase power sources in interactive systems shall be connected to 3-phase power systems in order to limit unbalanced voltages at the point of interconnection to not more than 3 percent.
Informational Note: For interactive power sources, unbalanced voltages can be minimized by the same methods that are used for single-phase loads on a 3-phase power system. See ANSI/C84.1-2016, Electric Power Systems and Equipment — Voltage Ratings (60 Hertz).
(B) Three Phase. Three-phase power sources in interactive systems shall have all phases automatically de-energized upon loss of, or unbalanced, voltage in one or more phases unless the interconnected system is designed so that significant unbalanced voltages will not result.
So, if you are installing single phase inverters on a three phase panel, it would be wise to see how much imbalance there is before you even get started slapping on these two-pole breakers. Let’s see how much imbalance there is from the utility lines with no loads on the panel. This is an easy exercise, but perhaps not so practical depending on the building. You will cut power to the panel and then measure the utility side voltage for all phases. Let’s use some not-so-realistic numbers for our calculations. FYI: Lower phase voltages mean more loads on that phase.
Phase A – B = 480V
Phase B – C = 475V
Phase A – C = 478V
I apologize in advance for the math. It is hard on the eyes, but an important part of this exercise.
Voltage Imbalance = 100 x Max Voltage Deviation from Average / Average Voltage
Average Voltage = 480 + 475 + 478 / 3 = 477.6
Max Voltage Deviation from Average = 477.6 – 475 = 2.6
Voltage Imbalance = (100 x 2.6) / 477.6 = .5%
This means the voltage from the utility is within 3% so now we need to perform this same exercise with the loads on. Turn on the three phase breaker and bust out that multimeter. We get different results with that main breaker on (voltages exaggerated for dramatic effect).
Phase A – B = 478V
Phase B – C = 475V
Phase A – C = 450V
Let’s plug in these new numbers:
Average Voltage = 478 + 475 + 450 / 3 = 467.6
Max Deviation from Average: 467.6 – 450 = 17.6
Voltage Imbalance = (100 x 17.6) / 467.6 = 3.7%
This panel exceeds the 3% rule, and if you think the utility will excuse that 0.7% I have a bridge in New York I’d like to sell you. But what are the implications of this voltage drop for our application?
First, if there are any three phase motors in this building they will suffer more wear and tear and will likely need to be serviced sooner than if the voltages were within the #5 tolerance.
Second, if you add single phase PV inverter breakers to this panel that are not divisible by three, it could make the imbalances worse. For example, installing three 40A two-pole breakers on this panel would not significantly contribute to the voltage imbalance. Probably. However, installing 10 or 11, probably would. Maybe. But here is the cool part:
The single phase inverters could actually help reduce the voltage imbalance on Phases C – A! So, in this case, I would put the two pole breakers for inverters #10 and #11 on Phase C – A and ensure the other inverter breakers were divided evenly amongst the other three phases. Done and done!
Yes, Virginia, you can install single phase inverters on a three phase panel as long as you do not cause any significant imbalance and keep it under 3%.