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

[pclausen]

I think that was caused by the thin cloud layer. This afternoon it was pretty clear and here's what the voltages looked like around 2pm:



So pretty close. You can tell from the above shot which panels are on the roof at a 18.4 degree angle, and which ones are on the ground at a 31.5 degree angle. Here's a shot from the "cloud" showing the panels from about the same time. (note that the browser cuts off the last few panels as there are 24 in the "D" array). So Arrays A, B and C are the roof mounted ones, and D is the ground mount.



So the roof ones are all between 147 and 152W. The ground mount ones are between 182 and 187, except D13 and D19, but those are both on the far left edge and I think they caught a little but of shade.

Cleaned up the cat5e run into the shop:



And on the array side:



I drilled a hole in a 1/2" cab just large enough for the cat5 cable to slide through. I'll silicone the cap (and the entrance through the shop wall).

Made 63 kWh today with a somewhat cloudy morning. Not too bad for December I think!

[gbynum]

Your panels look like a mix ... they are either about 23V or about 30V. Is there a connection choic or "tap" type thing?

[pclausen]

I'll check on the next clear day around mid-day which might be tomorrow. Today it has been raining all day, so the panels will be nice and clean.

So they are setting the new upper trip limit to 272.3V and the reconnect limit to 262V. I have a feeling that will solve my issue for good.

[bcroe]

The hardware specs state the Nominal Voltage/Range is 240V (211-264V). So I suspect that they plan to raise the upper limit from the current UL imposed 253.5V, to something a lot closer to the hardware limit of 264V. That's a 10.5V window. Upping it to say 260V is probably more than sufficient, but if they feel going all the way to 264V in my custom profile is safe, that would give me even more headroom. I'll ask them what upper voltage they are defining in my profile.

By 3pm yesterday, all 80 panels were back online again.

I do wonder why there is such a variance in the power produced between the panels as none have any shading that time
of day. Maybe the very thin cloud layer that was present can produce that effect?

Inverters here are 277V capable, yours might be too giving a lot of headroom.

Cloud edges can cause some intensity effects, but should average out over time. Its hard to make any realistic evaluations
under clouds, they are so variable. I suggest waiting for peak time on a clear day to do comparisons. If you don't trust the
readouts, go out with a clamp on ammeter and compare panel outputs. A DC clamp on will show direct panel performance;
an AC on micro outputs (if the individual wires are split) will give overall. I'd look for 5% match over a short time, 10% is marginal.
Bruce Roe

[pclausen]

The hardware specs state the Nominal Voltage/Range is 240V (211-264V). So I suspect that they plan to raise the upper limit from the current UL imposed 253.5V, to something a lot closer to the hardware limit of 264V. That's a 10.5V window. Upping it to say 260V is probably more than sufficient, but if they feel going all the way to 264V in my custom profile is safe, that would give me even more headroom. I'll ask them what upper voltage they are defining in my profile.

By 3pm yesterday, all 80 panels were back online again.



I do wonder why there is such a variance in the power produced between the panels as none have any shading that time of day. Maybe the very thin cloud layer that was present can produce that effect? The above screen cap is from the solar gateway on the roof, so the RSSI to the ground mount inverters is very weak. Fortunately there's another solar gateway attached to the ground based array, which shows the opposite. So both gateways shows all inverters on the "cloud whitelist". I wish there was a way to have each gateway only show the arrays that I assign to it. This would make pushing firmware updates much easier. As it is, I have to prep a list of MACs for each gateway, and then push out updates to only those. Fortunately this is not something I have to do very often.

[DanS26]

That's a great idea Bruce! So you would use a 2x4 or something to hold the panel against the upper stops until you got the fasteners in place below each panel? But then when you slide the next row of panels up against the top row, won't the fasteners be too tight to allow the lower row panels to slide in all the way? At least I think that would be an issue with the Northridge fasteners.

Ubiquiti is going add add a new country code to their system this afternoon. Once done, I can simply go my my site in their mobile app, select the new country code "UL-CVEC" and presto, all my inverters will pick up the config file for this new 'Country" and allows for a higher voltage threshold. I had 19 panels drop out today around noon time. Btw, the issue was NOT related to my compressor.

I wish you the best......if the voltage rise is outside of the hardware parameters, software adjustments will not work.

My wife calls it pessimism.....but all of us DIY'ers call it realism.

[DanS26]

I'm going to suggest a procedure, of sliding the top row up from the bottom first, then do lower rows the same way.
Bruce (who does almost everything alone) Roe

"Better, faster, simpler, safer".....experience is a very good teacher.

[bcroe]

That's a great idea Bruce! So you would use a 2x4 or something to hold the panel against the upper stops until you got the fasteners in place below each panel? But then when you slide the next row of panels up against the top row, won't the fasteners be too tight to allow the lower row panels to slide in all the way? At least I think that would be an issue with the Northridge fasteners.

Ubiquiti is going add add a new country code to their system this afternoon. Once done, I can simply go my my site in their mobile app, select the new country code "UL-CVEC" and presto, all my inverters will pick up the config file for this new 'Country" and allows for a higher voltage threshold. I had 19 panels drop out today around noon time. Btw, the issue was NOT related to my compressor.

Very glad the voltage problem is nearly solved; otherwise it sort of defeats the whole project. Here there were 2 V monitors
to be concerned with, line-to-line voltage, and line to neutral voltage. I don't have your rail system, but sliding panels up from
the bottom ought to be simpler than handing them over the top from a mobile platform. Probably slower, but a lot less man
hours. Rows of panels here aren't adjacent, there is a big gap for snow to drop through. Just got half a foot of heavy stuff;
will report on that after some pictures are processed. Bruce Roe

[pclausen]

That's a great idea Bruce! So you would use a 2x4 or something to hold the panel against the upper stops until you got the fasteners in place below each panel? But then when you slide the next row of panels up against the top row, won't the fasteners be too tight to allow the lower row panels to slide in all the way? At least I think that would be an issue with the Northridge fasteners.

Ubiquiti is going add add a new country code to their system this afternoon. Once done, I can simply go my my site in their mobile app, select the new country code "UL-CVEC" and presto, all my inverters will pick up the config file for this new 'Country" and allows for a higher voltage threshold. I had 19 panels drop out today around noon time. Btw, the issue was NOT related to my compressor.

[bcroe]

I need your tractor! Took us 5 people to put up row 5, two sitting on the top of 8' ladders to lift it over the rail and
3 to hand up the panel to us. You are making great progress.

I'm going to suggest a procedure, of sliding the top row up from the bottom first, then do lower rows the same way.
Bruce (who does almost everything alone) Roe

[pclausen]

The PVC conduit dips down in the center to about 20", but you're right, right, where it comes off the concrete footer at the shop and array, it is only a few inches. I plan to add some 6x6 treated posts on the array side to make the path between the shop and array level by adding fill dirt, and then a few inches of #57 gravel on top.

[foo1bar]

The "23 foot" conduit - it looks like it's buried under only a few inches of dirt. And it looks like it's PVC.

PVC conduit should be 18" or 24" deep. (or under 4" of concrete or 12" down when under 2" thick concrete.)

See table 300.5 of NEC.

[tyab]

I need your tractor! Took us 5 people to put up row 5, two sitting on the top of 8' ladders to lift it over the rail and 3 to hand up the panel to us. You are making great progress.

[pclausen]

Well, after the wife and kids went to bed, and having enjoyed a "few" single malts, I decided to complete connecting the ground mount array, despite the rain that started a few hours ago.

So here are the End Runs connected and me cheating and using a temporary cable for the gateway:



And a shot of the shop sub-panel. Like I said before, plenty of slack in the wiring in case the inspector does not want me to run though the meter base j-box.

[pclausen]

Looks like a great install. do you do contract work & travel ? And it helps to have the machines to do the work.

Not at this time Mike. I have a full time boring IT management job, but if I could make it work financially, I would seriously consider it!