Newbie here doing a 20.8kW DYI install (80 sunMAX 260W panels with microinverters)

[DanS26]

When I was designing my PV ground mount system, I was very much attracted to the micro-inverter setup. Mainly because of the panel level monitoring. During my research the Enphase documentation constantly warned of the voltage drop (rise) problem if the inverters were located far from the main service entrance. Because I had a TED system monitoring my electrical usage, I also knew that I had high average voltage from the POCO.......same issue now faced by the OP.

My solution was to design a dual string inverter system, which although was a compromise on individual panel monitoring, gave me the flexibility to deal with the high voltage problem. By locating the string inverters close to service entrance I was able to mitigate the voltage rise problem. By having dual string inverters, I know immediately if a panel goes bad.......I just don't know which one unless I investigate further. Here's how.......if one inverter starts under producing, I now know which set of strings is involved.....then I compare those strings and narrow the problem down to 12 of the 72 panels......then I find the problem panel. Simple and effective.

[pclausen]

In my case, my POCO is a small rural co-op with 35k members. They have 4,700 miles of line with an average of 8 meters per mile and about 30 substations total. They told me they like to set the voltage to 125V for each leg to neutral at the substations, to ensure that customers at the end of each line receive at least 120V. I'm 3.7 miles (straight line) from my substation, so that is probably more like 5 miles actual.

240V + 5% = 252V, so being at 247V on average means I'm within code by most standards. I have 200' of MCM500 cable between the shop and the house meter base, so that keeps the voltage drop there to a minimum. However, I have a 220' run of 350MCM from my meter base to the transformer on the pole that is the property of the POCO. I have a potential for generating 20kW during optimal conditions. So that's a 1.3V drop from the pole to the meter base, in addition to whatever drop I have between the meter base at the house and the microinverters at the shop. I can't do anything about the 220' run between the meter base and the pole, and the total path from the microinverters back to the meter base is optimized as much as it can be I think:

Microinverter String
|
45' #10Cu
|
Shop Sub-Panel
|
4' 3/0Cu
|
Shop Meter Base
|
200' MCM500Al
|
Splice
|
5' MCM300Cu
|
400A disconnect
|
5' MCM300Cu
|
Meter Base

I'll be receiving my solar Sense (sense.com) power monitor tomorrow, so I'll be able to monitor everything going forward, tracking utility line voltage, consumption and generation. So I'll be able to document any issues and present them to the POCO. The POCO is very friendly and go out of their way to assist their 130 and counting solar customers. I was at a meeting hosted by the POCO a few weeks back, and about 100 of the 130 solar customers were present.

[bcroe]

OK, the other part of the puzzle is what the Indiana Administrative Code says the nominal voltage set point may be. Seems like it shouldn't
exceed 120V or 240V. Then 6 % tolerance, but setting it at 125V will surely cause peaks above that level at times.

As SunEagle says, the PV grid tie inverters have pretty strict requirements to stay on line. If the PoCo is going to operate this way, it appears
the requirements need some relaxation or adjustment to suit the real world. The concept that the max inverter output voltage must be tightly
restrained is certainly debatable; for residential PV the inverter CANNOT set the line voltage, only react to it. SunEagle solved his problem
by reprogramming the inverter. At one point it appeared high line combined with V rise in my wiring would shut me down; fortunately here the
inverters had 277V operational capability. I must admit, the high line is now letting me run more voltage and less current, lowering my AC
losses in a 600' loop. But reprogramming dozens of micro inverters is a big job; is it even possible? Bruce Roe

[Mike90250]

maybe it's time for a pair of 480V transformers for your local run? Does Iron cost less than copper ? But then you have all the internal losses in the transformers....
Are you served with a nearby pole with a 6KV transformer ? How far have they run the 240V wires? Out here, remote houses get their own local transformer

[bcroe]

maybe it's time for a pair of 480V transformers for your local run? Does Iron cost less than
copper? But then you have all the internal losses in the transformers....

Yes that works. But now you have the losses of 2 transformers, instead of the single stage auto transformer. And you are transforming
100% of the energy instead of an adequate small percentage. Maybe 20 times the losses. Not to mention space and cost, perhaps
10 times as powerful, and at both ends instead of just one. Bruce Roe

[pclausen]

Had 6 inverters drop out today and they will stay down until they see the line voltage drop below something like 245V I think, so that likely means they won't be back online until tomorrow morning.



I'm going to open another dialogue with sunMAX/Ubiquiti about modifying the thresholds. They indicated previously that it would not be difficult, but that they were not allowed to due to UL requirements for North America. They sell these inverters worldwide and they are programmed differently based on the region they are deployed in. I think the parameters are all stored in the solar gateway and pushed out to the inverters from there. That's how I upgraded the firmware on them all when I first got them online and realized they were running pre-release BLE firmware.

[FNG AZ]

https://help.ubnt.com/hc/en-us/artic...rror-Codes-#ac frequency discontinuity

Maybe you had a power bump?

[pclausen]

Thanks for that link! I'm surprised I didn't see that before when I went looking. I did reach out to them and they are willing to reprogram my inverters to a higher upper limit, but they want a letter from my POCO first stating they are ok with it. This way Ubiquiti are not liable if something goes wrong. They have run into that before. I have a note into the POCO explaining the situation and requesting a letter. We'll see how they respond.

Once I get my sense power monitor online, I should be able to better track what is happening on the line side and see if there any any major power fluctuations.

[foo1bar]

Had 6 inverters drop out today and they will stay down until they see the line voltage drop below something like 245V I think, so that likely means they won't be back online until tomorrow morning.

Turn on a few giant resistive heaters to warm up the shop and see if that reduces the voltage enough that they can come back online.
(or if you don't have heaters, but have a number of halogen lights, a bunch of those would have similar effect (and add some heat to the shop too).

[pclausen]

I got 12 250W metal halide lights in the ceiling as well as a pair of 500W halogens for when I need light "right now" before the MH units get up to speed. That will at least give me a 4000w load. 4 of the 6 inverters have come back online again already on their own. I'm not at home today, so I'll try that trick over the weekend it if happens again.

[SunEagle]

Turn on a few giant resistive heaters to warm up the shop and see if that reduces the voltage enough that they can come back online.
(or if you don't have heaters, but have a number of halogen lights, a bunch of those would have similar effect (and add some heat to the shop too).

I think FNG AZ has a point. The screen does list an AC Over Frequency fault. Not sure if that is a reason for the inverter disconnecting or an after affect of the disconnect.

The one issue I always had was that those inverter monitors would provide the information but were way too slow to show exactly when it happened, while the software would be sensing the grid power parameters and acting down in the "10 cycle" range which is pretty fast.

[pclausen]

I do have a 7.5hp compressor in the shop that will cut on once or twice a day to maintain pressure. Could that be enough to trigger an AC frequency event?

The sense power monitor that I'm getting does more than 1 million samples a second (main reason is to detect the unique power signature of my individual loads throughout the house and track them all individually) but my hope is that I can also study that data to look for abnormalities on the line side, not just the load side.

[Mike90250]

I do have a 7.5hp compressor in the shop that will cut on once or twice a day to maintain pressure. Could that be enough to trigger an AC frequency event?.....

Easiest is to trigger it manually and see if the inverters drop offline. But a Frequency glitch, they should recover and be back online in 5 minutes.

[pclausen]

I'll try that over the weekend or maybe tomorrow.

One of the ubnt engineers told me this about the inverters:

--- cut ---
The inverters will connect after monitoring the AC line for 5 minutes, IF everything is within the right ranges (59.32 to 60.48 Hz, 220.5 to 253.5V). There is actually a tighter requirement on reconnection vs tripping so that there's some hysteresis if there are true grid problems.
--- cut ---

So if the compressor did trigger the inverters going offline, the likely reason they did not come back on within 5 minutes (it took hours), was that the voltage was above the tighter requirement for re-connection after an "AC Event".

I'll try to map the mac addresses of the inverters that went offline, to the string(s) those panels were on. I should have taken a snapshot of the cloud management console that showed the assigned panel locations that were out. Might just wait for the next event and do that to make less work on myself.

[bcroe]

You are never going to cause a frequency shift. Hope "they" reprogram the inverters; its a lot better to do it yourself. Having your
big new investment shut down is not good. Maybe a consideration of any inverter purchase should be if you can program it.
Bruce Roe