DIY grid-tie roof mount in Tucson . . . some first-order questions

[sonoran]

well along the way of my 12.3kW install

Fantastic! and thank you for writing back
Would you mind contacting me in regards to the TEP portion of the equation
Thank you

[rtr688]

Nope! But I'm well along the way of my 12.3kW install.

Did all my own drawings for TEP, then after their approval went to county. They had a few small comments.

Marked where I wanted to install my disconnect and production meter, TEP approved, and away I go.

[sonoran]

Resurrecting old Tucson post.

TucsonMCM I'm going to be tackling a DIY

RTR has MCM reappeared ? ...also a TEP customer

[rtr688]

Resurrecting old Tucson post.

TucsonMCM I'm going to be tackling a DIY and have similar questions to what you've aske here (Pima County DIY/remote inspections/rapid shutdown requirements, TEP interconnection, etc) and I'd love to pick your brain. This winter I'm going to tackle a 8-10kw project on the NE side of town.

Apparently we can't send direct messages through the site. If you're still around let me know. Thanks!

[foo1bar]

Although #4 is possible,

#4 (line side tap) is probably not possible with that panel, unless adding a socket adapter thingy. And that'd probably fall under "Other"

At least I think it's highly likely that the conductor from the socket to the main breaker is a metal bar - not a wire.
And therefore it isn't something you can tap.

IMO #1 (standard backfeed breaker) is likely to be what you'll want to do.
If you want >40A, you could replace the main with a smaller breaker, which would be #4.

[foo1bar]

So approaching this from the other direction, the 120% rule applied to my 200A panel & 200A breaker results in a maximum 40A backfeed breaker. And if I understand correctly that the backfeed breaker must be 125% of the incoming load that means I can have a maximum 32A @ 240V = 7680W system. And I assume this would be a CEC-AC measurement of the system?

It would be the AC wattage of the inverter (really it's the AC amperage of the inverter has to be 32A or less)

This is the reason you'll see "7600W" as one a model size for a lot of manufacturers. (ex. Solaredge has SE7600A and SE7600H,
SMA has "Sunny Boy 7.7-US", Fronius has "Primo 7.6-1", etc.) All of their installation instructions say 32A max output current

If I accept the TEP application systems computations for system CEC-AC (using my enterred components, tilts, azimuth, etc) then 7680 translates to 26 panels (of my tentative make/model), versus the 14 or so panels (= 4.06kw CEC-AC) that I was intending to use. So without any change/derate to my main breaker, I still have ~ +85% future upsize capacity, for example if I purchase an elec vehicle in the future + PV on carport to support it.

Probably more, because usually it usually makes sense to have more DC watts than what the inverter is sized for. Again - installation instructions / data sheet for the inverter will tell you the maximum PV power.

BTW - your POCO (TEP) may have a maximum size they'll allow you to install based on your historical usage. You may be able to get around that by showing them your plans to buy an EV.

Also - there may be a limit at 10kW.
Some places have a 10kW limit for residential installations.
(You might be able to do a 10kW DC array with a 7.6kW inverter - but that might be oversized for your usage.)

[TucsonMCM]

So approaching this from the other direction, the 120% rule applied to my 200A panel & 200A breaker results in a maximum 40A backfeed breaker. And if I understand correctly that the backfeed breaker must be 125% of the incoming load that means I can have a maximum 32A @ 240V = 7680W system. And I assume this would be a CEC-AC measurement of the system?

If I accept the TEP application systems computations for system CEC-AC (using my enterred components, tilts, azimuth, etc) then 7680 translates to 26 panels (of my tentative make/model), versus the 14 or so panels (= 4.06kw CEC-AC) that I was intending to use. So without any change/derate to my main breaker, I still have ~ +85% future upsize capacity, for example if I purchase an elec vehicle in the future + PV on carport to support it.

Seems like I can confidently proceed w/ the standard backfed breaker as the interconnect selection, and it won't meaninfully constrain me for future changes.

[TucsonMCM]

The way my local inspector interprets the code, you can't use "unused load capacity" in your backfeed calcs....

I was wondering about that. I've seen both approaches in online examples, but hadn't made an effort to correlate to NEC revs or other possible explanations. There is some logic to treating the main breaker as the load number, regardless of actual load breakers. So that interpretation implies that one must "commit" to a lower load via a change/derating of the main breaker *IF* you wish to utilize that buss capacity in your allowed PV backfeed.

[solarix]

The way my local inspector interprets the code, you can't use "unused load capacity" in your backfeed calcs....

[TucsonMCM]

Really - a lot of possible options you have for the backfeed - but I'm making a LOT of guesses about what you have and my guesses are probably wrong.

Your guesses in your post were largely correct. See my photos above fyi.

Given my current configuration (200A buss, 200A main breaker, 140A load, and 4kw-ish PV array) it seems that the 120% rule combined with my 60A of unused load capacity translates to a simple backfed breaker as the optimal interconnect, with a huge margin of extra amps for future changes. Unless I'm missing something, I could potentially double the size of my intended PV array to ~8kw (= ~ 50A PV breaker) AND add an additional 40A load circuit to my panel, and still meet the 120% rule. So I'm concluding that there is no need for me to entertain the cost/complexity of a line tap or a change of my main breaker.

[TucsonMCM]

Hey folks, sorry to disappear on you for a month. Had some urgent distractions.

I've looked further into the TEP Application process, and after a couple weeks wait I finally got my login credentials etc for their project management webpages. Their system is heavily centered on professional installers, which requires some inobvious "selections" to get a DIY install moving.

I've confirmed that the export rate $ amount is locked in based on *intitial* application date, and not any subsequent approval dates or other milestones. So all I need to do in the next week or so is get the application submitted. The amount of information I need to submit an application is relatively modest, and changes are permitted (like brand/model of panels) if they don't have a significant impact on scale of system.

One parameter for the application is a selection of interconnect method, and I could use some tutoring/guidance on this selection.

The choices in the application system are:
1) Standard backfed breaker
2) Solar ready panel with standard backfed breaker
3) Main disconnect derate with standard backfed breaker
4) Line side tap - requires TEP performed power kill
5) Other

I perceive that my likely selection is either #1 or #3. Although #4 is possible, I perceive it would add cost/hassle and isn't necessary since my PV output is modest relative to my panel/bus. As noted earlier in the thread, I believe I have a 200A service connection and a 200A bus in my panel, with a 100A interior subpanel and a 40A exterior breaker for a/c, so total load currently at 140A. I've attached some photos of my panel for others to review to confirm or correct my interpretation. The primary supply lines are coming from overheard lines in an alley to a conduit/header at the edge of my roof, which runs thru my eave directly into the panel.

It would be great if folks could clarify the differences between #1 and #3, and also comment on #4 if I'm missing some benefit there. I may want to add another interior subpanel for a future garage/shop circuit, which could increase my load by 20A-40A, so comments also appreciated on the interplay of that and any derating or bus capacity versus PV and load.

Are backfed breakers always housed in their own box/panel as I've seen in some photos online? Or can the by installed in the main panel similar to my A/C and interior subpanel breakers?
ElecPanelLabel4mp.jpg

ElecPanel4mp.jpg

[nwdiver]

My guess is string inverters are likely to be a few percent worse in many partial shading situations.

Under the vast majority of conditions it's likely <1% or possibly even an edge to the string inverter (~no difference). If one cell group is shaded it's the same. If two cell groups are shaded then the benefit could go to the string inverter if 1 of 3 cell groups is unable to maintain the 22v the micro inverter requires. Only when shade falls across all 3 cell groups of the panel does a micro inverter have an edge but at that point it's blood from a stone.

This is what really irks me about Enphase. In some of their promotional material they insinuate that shading part of one panel will reduce the output of the entire array. If each string is on an independent MPPT which should be true for ~all residential string inverters installed after ~2017 then shading part of a panel will have no effect on the unshaded portion of the array because of the bypass diodes.

[foo1bar]

Modern String inverters also handle shade for the most part just as well as micros or optimizers.

My guess is string inverters are likely to be a few percent worse in many partial shading situations. And probably just as good or maybe better in a few specific cases of partial shading.
I think the cost savings (probably) are more than that few percent.
Except when you add in rapid shutdown requirements, I think the cost savings isn't nearly as much.
But I haven't priced things out lately.

For OP's situation I don't know how things would price out between Enphase, Solaredge and SMA w/ rapid shutdown...
The cost of the propietary cable (and my desire for less electronics on the roof) pushed me to solaredge over enphase for my system, but prices change - and my system was bigger.

My distain for micro inverters is mostly due to the fact that they cost more for ~no return. Plus their existence is mostly due to electricians not wanting to put in the effort to learn about series-parallel DC circuits.

If I had wanted to do an SMA system, I believe I would have had less production due to 3+ orientations and shading. For me, enphase was the most expensive option. Solaredge was more expensive than SMA - but gave me per-panel information, and I believe handles multiple orientations and shading better.

Per-panel info is not a huge benefit at the end of the day. IMO it's not nearly as beneficial as the salesman will hype it up. But it is fun. And on a commercial (church) installation it helped me identify that the installers screwed up, not connecting 2 of the >100 panels

[foo1bar]

1) details like this (rapid shutdown) are where I can use some guidance. I can't find a coherent, centralized description of the requirements. UL-1703 for the panels and UL-1741 for the inverters are mentioned. Disconnect requiremenets are described but I don't see "rapid" anywhere. I'm not even clear (yet) on what rev of NEC applies. Is the applicable NEC the determining document for rapid shutdown?

Rapid shutdown is described in 690.12 of 2017 NEC (Tucson I *believe* is using 2017 NEC - based on a very quick web search - do your own confirmation of that)

Each of the inverter manufacturers are going to have a thing about how their system is (or can be) rapid shutdown compliant.
ex:
https://www.sma-america.com/newsroom...lications.htmlhttps://www.sma-america.com/newsroom...lications.html
from searching "site:sma-america.com rapid shutdown"

Probably a supply house doing "kits" will help with supplying you the right equipment to do rapid shutdown.

I went with Solaredge system (partial shading, multiple orientations) and although I didn't need rapid shutdown (earlier NEC version) I know it does rapid shutdown because that was one of it's selling points.

3) I think I'm good to go here. I have a 200A main panel, which was an upgrade circa ~2005 when a new AC unit was installed. Its conveniently placed under a large overhang, near one edge of the intended array placement. There are a total of 140A of load breakers on the 200A service. I had one installer give me a quote before I settled on DIY, and they acted "happy" w/ the main panel and didn't include any upgrades or mods in their quote. If I understand the 120% rule, then an array of my size will clearly fit.

It really really depends on how your main panel is setup. IMO a few diagrams (and potentially some pictures) can communicate a lot more than words.
The 120% rule says that if you have a 200A main breaker and a 200A bus for that breaker you can put a 40A breaker at the far end (200A * 1.2 - 200A = 40A)
IF it's a 225A bus, then that equation is 225A * 1.2 - 200A = 70A.

You may not even need to look at the 120% rule though.
If it's a 200A main breaker and 200A bus - and 140A of load you can possibly add a 60A backfeed breaker... Under older NEC I think it was 705.12(D)(2)(3)(c) – Sum of inverter and load OCPDs.
I don't see that in 2017 NEC though.... Ah - I think it may now be NEC 705.12(B)(2)(3)(c) in 2017 NEC. (with a requirement for a warning label)
nfpa.org has free access to the 2017 NEC (and other versions)


Of course you also potentially could use 120% rule in that situation and downsize the main breaker - so 150A main with 200A bus = 200A * 1.2 - 150A = 90A max backfeed. Swapping out a main breaker is serious matter though.... I'd triple check that I'd torqued that thing correctly.

Really - a lot of possible options you have for the backfeed - but I'm making a LOT of guesses about what you have and my guesses are probably wrong.

[nwdiver]

Not a lot of microinverter fans here, eh?

I've got some modest shade issues in my situation (neighbors trees) that may get marginally worse or better, depending on what they do w/ one off the trees. But for the sake of comparison, lets assume that I would use optimizers if I went w/ a string inverter. Its worth noting that a legit installer (Technicians for Sustainability here in Tucson) included optimizers on both their string inverter proposals.

Is the "microinverters cost more" mantra really true on smaller systems? The optimizers I've seen run $70-ish while an Enphase IQ7 is $125-ish. Assuming my system is 12 panels, that's a $660 delta in those components, I can't buy a 3000W inverter for $660. I'm ignoring lots of details in cabling, monitoring, etc which could add up. But on the primary components, it seems like micros can be competive.

And yes, they are easier to design with. Despite @bcroe's passive-aggressive insult to my intelligence, ease of design and install is a benefit regardless of your capabiities. I can TIG weld a bike frame reasonably well, but does that mean I'm stupid to buy a mass produced one that cost more than raw tubes?

Modern String inverters also handle shade for the most part just as well as micros or optimizers. You can get a 3kW string inverter for ~$900 that also includes wifi monitoring. 3kW of Enphase IQ7s might cost ~$1125-ish but then you need to buy the proprietary trunk cable at ~$25 per panel then if you want monitoring that's another ~$400. So you'll end up paying ~50% more for micro inverters vs a string inverter for ~no tangible benefit aside from the real estate saving on a wall somewhere.

My distain for micro inverters is mostly due to the fact that they cost more for ~no return. Plus their existence is mostly due to electricians not wanting to put in the effort to learn about series-parallel DC circuits. Tiny AC inverters cost more but they were more comfortable with AC so that's what they bought....