Syonyk's Solar Status: Rough-in completed

[Syonyk]

Updates on my solar project, because my blog is going to end this weekend until Blogger fixes their crap or I get around to static hosting:

Rough in inspection is done. There are a few labeling issues I need to resolve (in the realm of “I hadn’t installed the labels because I wanted to verify them”), which isn’t a big deal Trench was fine, didn’t even get a tape measure stuffed in it, despite all my work to make sure it was “more than 18 inches” in all the exposed spots.

Adding outlets for backup power is perfectly fine, so I’ll get hardware for those and run them this week. Fun with liquid tight flex conduit, I think. I can use plastic hardware cloth to protect the wiring on the panels, and the PV wires need to be in conduit up to the first panel, which is fine. I think I’ve got some 3/4" Sch80 laying around I can use, or I’ll mess with the liquid tight flexible stuff. Either way.

So, making more progress. The rails will be mounted today or tomorrow (I’ll probably have them ready to tighten down today, but having someone at the end eyeballing them is quite helpful to make sure they’re straight), I can smoosh the dirt back in this weekend, and generally get onto the actual panel mounting soon.

[Syonyk]

Rails are firmly mounted and aligned (the L feet have slots to allow you to adjust for minor roof variations, and we made use of that to adapt to some minor A-frame variations).

Ground is mostly run, though, as is the story of this project, I ran short of wire (6AWG bare copper in this case). There’s “functionally grounded,” and “solidly grounded,” and then what my system is, which is “obscenely grounded.” I’ve got two ground rods per frame, with a “loop” to all the rails.

For no reason beyond “Sure, I’ll play with something new!”, I’m using the liquid tight flexible conduit for protecting the PV wires, with an outdoor metal box to transition to the PV wire gland. Should work, I hope!

Trenches are mostly filled. I spent the morning with the blade on the tractor chasing around and pushing dirt. There are still some areas I need to fill by hand as the tractor won’t fit in there, but most of the work is done. And the kids got a ride on the tractor, after much protesting and asking “Daddy done?” every time I shut the tractor off for some reason…

Rails aren’t up in this picture, but I do like the overview.

[milfox]

You should let the kids get some paint on the storage 40’ container, and maybe throw some panels on top of it.

[Syonyk]

Paint and a mural, maybe. Panels on top of it, no. I don’t think you fully appreciate the regulatory hell that this project has been. It took a year to work through various rejections of the plans (leading to me not having backup batteries because the advice I got was essentially, “nothing battery based will be approved through JS, he doesn’t even approve Tesla systems because he doesn’t like them and will just nitpick them to death”).

In any case, if I want to randomly throw up some extra panels, I get to go through the entire process again.

Then, if I change the system after it’s implemented, I lose my net metering grandfathering (assuming I’ve got this all done before Dec). I could add another system through another meter, but… I probably won’t.

Point is, I can’t just randomly change stuff. Once this is done and approved, it’s set in concrete and will stay that way indefinitely beyond repairs.

==============

Lessons from yesterday: PV wire diameters matter.

12AWG/600V PV wire is 5mm outside diameter.
12AWG/1000V PV wire is 6mm outside diamater.
10AWG/1000V PV wire is about 7.5mm outside diameter.

You can’t fit a pair of 10AWG/1000V PV wires through a 1/2" conduit gland, or through 1/2" conduit fittings. The 5mm stuff is on order…

Given that the rest of the wiring is 12AWG, I don’t mind having a bit more 12AWG in the system over 10AWG.

[milfox]

I didn’t say tie them into the system. Though that’d make sense. Containers are gravity ground anchors after all.

Paint the box? Definitely.

Use as many admin loopholes as you can? (like a battery ‘generator’ via a ‘generator’ input into the system) … hell yeah.

[Syonyk]

Once I hit a plug, (voice of Cartman)I can do whatever!

Current status: Wiring is done, out to the MC4 connectors at the end of the various runs.

The original plan, as of about yesterday, was a work day tomorrow to mount panels. Of course, yesterday evening, I discover that via a few family connections, I’m connected via a plausible series of events to someone who has Covid. So, work day cancelled, will be pushed back a few weeks, and at that point, I may as well just hang panels with my wife. We can probably do a few an evening, kids will love being out in the dirt. I’ll work on it tomorrow with family labor and see what we can do. Moving 54 panels solo sucks, but… I’ve moved 'em before. Plus there are still 8 to get out to my new office array. And a trench to clean out. And… ugh.

Anyway. Contact with Idaho Power gave me the proper parameters to use for the inverters - they want UL 1741 SA, or California Rule 21, take my pick. And 5 minutes between seeing valid AC and inverting.

I’m going to aim to have the panels mounted this week, but it depends an awful lot on weather.

[Syonyk]

So… it turns out, panel mounting doesn’t suck as much as I’d realized. We were able to get it done this weekend with just the people around the property who are all hanging out on the hill for potential covid exposure reasons. Two people can do it, four people makes it halfway fun!

Ends still need to be trimmed, and the main A frames aren’t actually wired yet (I figured I’d hang the panels, then go back and connect them), but the south frame is wired up and I’ve been able to power up the inverter for some firmware updates and configuration. Also, playing with the backup outlet (I’ll test grid sync once everything is wired up). I can charge the car directly from solar, if I want!

The south array, with 6 panels, will run around 1200W currently (on the backup outlet) but that cuts out with the slightest cloud. The main arrays, at 24 panels each, ought to handle 2000W pretty easily in almost all reasonably sunny conditions (and will be able to export far more to the grid - this is just the backup outlets).

[Syonyk]

Well, pencils down. I’ve got an inspection scheduled for tomorrow for the electrical side.

[Syonyk]

… well, that was easy.

System is signed off for the code bits, now I just have to wait for the power company to come out, swap the meter, validate the 5 minute lag between AC applied and feeding, and turn it on.

[Vertiginous]

Congrats on passing code. Are you going to be running fairly detailed datalogs once it’s operational?

[Syonyk]

They log a bunch themselves, and the Sunny Portal Cloud Based Bullshit logs as well, but I intend to do some local logging. Just have to figure out how…

[Syonyk]

Looks better with the ends trimmed. You can see some of my green mesh peeking through at the ends - that’s to keep fingers out of the wiring if someone decides to go under the array. It is rather excessively grounded (two ground rods per array, one at each end, looped to all the rails), so I’m not terribly concerned about kids under it, though I’ll discourage that as I’d rather them not bump into panels by mistake.

[Syonyk]

Well, one moment of sheer panic later, solar is online, approved, net meter swapped in, and purring away at over 7kW on a clear day. I’ll have good graphs from future days this week for the different arrays.

The inspector from the power company showed up and looked over stuff, then informed me that he couldn’t approve it, because my disconnects weren’t the sort you could visually verify as open, and circuit breakers weren’t a permitted AC disconnect. I’m using Midnite Solar disconnects as combiners, which have an external lever to toggle circuit breakers internally.

Redoing this with disconnects would require an awful lot of rework on the system, likely new permits, inspections, etc, so… I more or less requested, “Prove it… because I don’t recall reading anything like this, though the website is hard to navigate at times.”

Turns out, he was thinking of older guides. The current guide, which he had a copy of and I found on the website, simply requires “The switch must be manually operable with a visible ON and OFF indication, and capable of being locked in the off position. Draw-out or other types of disconnects are not acceptable.” Based on this, a circuit breaker based disconnect is perfectly acceptable, so we got past that and moved on.

The main check is simply that the inverters take >5 minutes from seeing AC power to kicking on. Mine, despite a 5 minute setting, take… enough longer that I was beginning to worry about them, but kicked on at 5.5 minutes or so. I guess they were just lazy about looking for grid power or something.

In any case, after deciding that my disconnects do meet the guidelines, I got my meter swapped (for the same physical type, just different programming), and things are online and purring away!

Woo!

[Vertiginous]

Solid accomplishment - not just the build but the saga of navigating a serious bureaucracy. Pour yourself a well-deserved cold one. The tax breaks and net metering returns are icing on the cake at this point.

[Canem]

Such a comfort, knowing the Idaho Power people work so hard to stay up to date with their own standards…

Glad it worked out though, and with almost two months to spare before the deadline!

[Syonyk]

Woo, 54kWh generated today, and I’ve pushed about 50kWh to the grid in the past 2 days. I’ll build up what I can before winter hits!

[Syonyk]

I have most of a full day’s data, and it demonstrates the difference between south facing panels and my A-frames.

This is the south facing panel, and represents a “typical” solar facing. We’re in the wrong timezone, so solar noon is about 1:30 PM. This is 6 panels, 1770W nameplate, peak production of around 1450W or so. Today, 10.37 kWh, or 1.72kWh per panel.

My east A-frame inverter demonstrates the perks of the A-frames - a far wider production curve. Data logging glitched briefly in the morning, but you can see how it rises to a good power rating far faster (these are twice the scale of the north frame, 6kW vs 3kW inverters), peaks around 3kW, but is above 2.5kW for most of the solar day. You can see it’s slightly “chopped” at the end of the day, and that’s shading. This frame produced 22.7kWh today out of 24 panels, so 0.94kWh per panel. The sun is rising/setting south of east/west, so there’s not as strong an advantage to the east/west panels right now, but in the summer, when the sun is rising well north of east/west, these will be online far earlier when south panels are still literally backlit. Also, on a cloudy day, panel area is king, so I can still pull energy out of the sky on darker days.

The west array is a bit lower, physically, than the east array, so has shading a bit longer in the morning. It went from 200W to 2200W over 15 minutes, give or take, as the shading disappeared. The curve is a bit smoother in the evening, as it’s not being shaded then. 21.68kWh today (because of morning shading - the two run basically identically when fully illuminated), so 0.90kWh per panel.

Again, the south panels are more efficient per panel, but the overall curve of power delivery is far flatter with the east-west panels, and they’ll really come into their own on both cloudy days and in the summer when the sun is rising to the north and is up at 35 degrees by the time it hits due east.

Total production looks like this…

Again, fairly flat and wide compared to a south facing curve. There’s actually a bit of a mid-day peak that was intended mostly for vehicle charging. Peak is about 7.5kW right now, but it holds 6kW from 10:30 AM to 4PM.

Total production on a clear day was 54.4kWh, which is rather substantially in excess of our daily energy use (we didn’t drive anywhere today). Trying to build up and bank hours for winter. I expect I’ll be pulling net power from the grid this winter, just because I won’t have enough time to build up the kWh bank, but I expect after next year, I shouldn’t have to pay for power for 25 years.

At this point, I basically just let it run and try to ignore it… instead of watching my power delivery curves. I may get a battery powered leaf blower to blow some dust off the panels on an occasional basis.

[Symbioquine]

From a monitoring perspective, it’s awesome that you have a “load” that will take all the power you can throw at it! My main challenge monitoring off-grid systems is the data I have is more a function of both battery charge deficit and panel-angle/cloud-cover/etc.

[Syonyk]

Yeah, the grid is kind of nice that way!

Looks like a lead acid charge curve, though I’m not sure if you drop to float or not there. I think that little drop from the absorb curve might be float, but if it is, looks like you’re dropping a bit early. Of course, if it’s sealed, that may be the best for it - I’m more used to flooded lead acid, where “beating the crap out of them” is the way to long life.

[Symbioquine]

Yeah, it’s a 100AH SLA battery and you called it on the drop to float;

That’s with the default settings a Epever 3210AN thinks are good for a 100AH SLA battery… I’m still learning this stuff

Hope I’m not co-opting this thread - just though the commentary on monitoring advantages of the “infinite load” of a grid-tied system vs off-grid were worth highlighting.