SolarEdge 20kW Inverter Grid Connection Riddle

[freddyjones39]

I have just installed a 20kW SolarEdge inverter onto a 480V Wye grid. I am getting the error: Error 59: AC Voltage is too high on line 2 (it is 3 phase, and in Canada).

Here is a bit about the site:
-The grid is 208V stepped up to 480V via a pad mount transformer
-It is a large agricultural operation
-It is mounted on a pump house used for geothermal about 1600 meters from most of the buildings (buried conductor)

Here are some notes about what I have tried:
-I checked the AC Voltage (it was within spec, and we also brought it down well below 480, but the same error persisted)
-Checked the waveform of the grid AC (it was not noticeably dirty)
-I contacted SolarEdge Support (they tried, but to no avail)
-I checked all of the conductor sizes (they were all appropriate)
-I checked all of the connections back to the grid service for tightness (they were tight)
-I replaced the inverter (the same error was given)
-I did a resistance test on the lines (results detailed below)
-I switched around the lines (detailed below)

When I did a resistance test (using 500VDC):
Testing the conductors into the inverter:
-Line 1-G, 2-G: 4.7 MOhm
-Line 3-G: 2.4 MOhm
-Line 1-2: 9.5 MOhm
-Line 1-3: 7.2 MOhm
-Line 2-3: 6.9 MOhm

Testing all conductors inside the building:
-550 MOhm (max of the testing equipment)

Testing the underground conductors:
-Line 1-2: 31 MOhm
-Line 1-3: 20 MOhm
-Line 2-3: 28 MOhm
-Line 1-G: 8.6 MOhm
-Line 2-G: 16.2 MOhm
-Line 3-G: 9.5 MOhm

When I switched the lines into the Inverter around (this is where it gets weird):
(Below, the first set of numbers correspond to the grid leg, and the second set correspond to the inverter leg)
-(1,2,3)->(1,2,3)=Error on Leg 2
-(1,2,3)->(1,3,2)=Error on Leg 1
-(1,2,3)->(2,1,3)=Error on Leg 1
-(1,2,3)->(2,3,1)=Error on Leg 2
-(1,2,3)->(3,1,2)=Error on Leg 2
-(1,2,3)->(3,2,1)=Error on Leg 1

Does anybody have any ideas? It is really throwing me for a loop!

 

[jaggedben]

Don't know if this will help, but a couple observations...

In your last chart, you're getting 'error on leg 2' as long as you're just rotating the legs around the inverter terminals. But you get 'error on leg 1' if you switch the position of 3 with respect to 2 (3 comes before 2 instead of after, if that makes sense.) To put it anothet way, the error you get depends on whether 1,2, and 3 are connected clockwise or counter clockwise, so to speak, but not the positions of each conductor.

The fact that you can rotate either of these configurations around the terminals and get the same error tends to suggest something about the inverter, not the AC source.

2nd observation is that you said this is a wye but you haven't told us anything about the neutral.

 

[coop3339]

I think the problem is the 1600M feed increasing the grid impedance and the inverter driving up the voltage to compensate.

What size is the feeder and transformer? Also is feeder AL or CU?
Does the error happen before or after it connects?

 

[freddyjones39]

Don't know if this will help, but a couple observations...

In your last chart, you're getting 'error on leg 2' as long as you're just rotating the legs around the inverter terminals. But you get 'error on leg 1' if you switch the position of 3 with respect to 2 (3 comes before 2 instead of after, if that makes sense.) To put it anothet way, the error you get depends on whether 1,2, and 3 are connected clockwise or counter clockwise, so to speak, but not the positions of each conductor.

The fact that you can rotate either of these configurations around the terminals and get the same error tends to suggest something about the inverter, not the AC source.

2nd observation is that you said this is a wye but you haven't told us anything about the neutral.

Very observant! I didn't notice the clockwise thing.

Even more importantly: I think you nailed it with the Neutral. I just asked my electrician to check, and sure enough, this is a 480 without a neutral! I don't know why I didn't notice that when I was testing it...

This inverter only supports 3 phase systems with a neutral. So now the next question: Does that mean that I tear down the system or does anybody know of a way to make it work?

 

[freddyjones39]

I think the problem is the 1600M feed increasing the grid impedance and the inverter driving up the voltage to compensate.

What size is the feeder and transformer? Also is feeder AL or CU?
Does the error happen before or after it connects?

My appologies, it should have been 1600 ft (we jump between metric and imperial in Canada).
Feeder is CU ($$!)
The error happens before it connects.
Not sure about the transformer size, but this 20kW system is MUCH smaller than their actual demand (by orders of magnitude).

 

[ggunn]

Very observant! I didn't notice the clockwise thing.

Even more importantly: I think you nailed it with the Neutral. I just asked my electrician to check, and sure enough, this is a 480 without a neutral! I don't know why I didn't notice that when I was testing it...

This inverter only supports 3 phase systems with a neutral. So now the next question: Does that mean that I tear down the system or does anybody know of a way to make it work?

The only way I know of is to install a 1:1 delta to wye isolation transformer with a grounded neutral between the service and the inverter. I'd be interested to hear if there is another way.

 

[pv_n00b]

You need the neutral so you can either swap out the transformer for a 208V delta/480/277V WYE or you can derive a neutral using a grounding transformer.