micro-inverters or optimizers when shading is major issue ?

[Ampster]

You know anything about this IQ8 technology supposed to be released soon ? I saw some vague claims about support of multi-mode.

Only what their press releases say and what others have said on this forum.

[RShackleford]

Talked to a helpful and knowledgable guy at Tigo, and he confirmed: It is impossible to support 2017 rapid-shutdown requirements and still have SPS functionality. I believe this means that to have SPS in a legal system, the system must be exempt from rapid-shutdown requirements. This means ground-mount (or possibly carport-mount, a grey area).

My problem is that I was planning to run the high-voltage DC underground and then into my house's crawlspace and a utility room where the main panel is, and put the inverter next to it; it's going to be a very close call if I can do that and not have more than 5ft of wiring inside the house (a requirement for the ground-mount exemption from rapid-shutdown).

I guess another option is to put the inverter with the ground-mounted panels. The installation manual (for the Sunny Boy -41 models) states "The product is suitable for indoor and outdoor use". Not sure if that means it can be fully-exposed to weather, or what, but hopefully suitable for the environment underneath the panels. Then run AC to the house and the main panel grid connection. Then it'd be perfectly fine to turn off the system that shuts downs the DC from the TS4's when AC grid fails.

[foo1bar]

I guess another option is to put the inverter with the ground-mounted panels. The installation manual (for the Sunny Boy -41 models) states "The product is suitable for indoor and outdoor use". Not sure if that means it can be fully-exposed to weather, or what,

Yes it can be fully exposed to weather. Inverters are installed outside quite often.
I would use a shaded location whenever possible (ex. northern wall of the house so it's shaded 90% of the time)

If I were you I'd look at losses from having the inverter at the array vs. at the house. You should have less amperage with inverter at the house - so less losses. But how much is that loss? How much would it cost to increase the copper wire size to make the losses equivalent?

[RShackleford]

Yes it can be fully exposed to weather. Inverters are installed outside quite often.
I would use a shaded location whenever possible (ex. northern wall of the house so it's shaded 90% of the time)

Ah, good to know, thanks.

If I were you I'd look at losses from having the inverter at the array vs. at the house. You should have less amperage with inverter at the house - so less losses. But how much is that loss? How much would it cost to increase the copper wire size to make the losses equivalent?

Well, in general, yes I would have greater wiring losses, since the DC voltages are typically higher than 240. But since I have significant shading issues, there may be times that at least one of my strings is down towards the inverter's minimum, more like 100. So better the devil I know, than the ... whatever that saying is. Anyhow, I don't think it'd be much of an issue: it'll be max 50ft from panels to load center in house, and 100ft of 10awg is about 1/10 of an ohm, so drop 2 volts if my system is cooking along at 5kw (not that using 8awg will break the bank at about $0.37/ft, so less than $100).

[RShackleford]

So clearly their 4452 kwh/year estimate includes significant shading losses.

I checked with that installer (nice of them to talk to me, since I pretty much blew 'em off, saying I've decided to DIY) and they confirmed that in fact they did use shading loss of 30% in their PVWatts run.

The fine print says "guarantees annual solar production to be within 10% of proposed production as calculated by PV Watts for a period of five years or the company will reimburse the retail value of the difference in kWh."

So seems like they have pretty high confidence in the PVWatts numbers.

When I run it for 5.25kw peak output, with 30% shading loss and default on the other losses (for a total loss of 38%), slightly sub-optimal azimuth and elevation for site-specific preferences, I get 5517 kwh/yr.

I think I can DIY the install for about $6000 out of my pocket:
$4400 delivered, for 15 of those Solaria 350watt seconds I mentioned awhile back
$750 for TS4-R-O
$1200 for SMA inverter
$1000 for wood etc for the ground-mount structure
$1000 for miscellaneous stuff (wire, conduit, etc)
$200 permits
X 0.7 for the federal tax credit

That's a little over $1 per kwh generated in a year. With net metering and our POCOs time-of-day rates, maybe 1/3 of those kwh are worth 33 cents to me and the other 2/3 about 6 cents. Seems pretty viable to me - payback well under 10 years. Plus, around here at least, the PV installation adds value to your home.

I need to do some Excel-fu to figure out more closely how much those generated kwh are worth. And I'm sure I'm forgetting some things in my cost estimate.

[foo1bar]

I checked with that installer (nice of them to talk to me, since I pretty much blew 'em off, saying I've decided to DIY) and they confirmed that in fact they did use shading loss of 30% in their PVWatts run.

So seems like they have pretty high confidence in the PVWatts numbers.

When I run it for 5.25kw peak output, with 30% shading loss and default on the other losses (for a total loss of 38%), slightly sub-optimal azimuth and elevation for site-specific preferences, I get 5517 kwh/yr.

I think I can DIY the install for about $6000 out of my pocket:
$4400 delivered, for 15 of those Solaria 350watt seconds I mentioned awhile back
$750 for TS4-R-O
$1200 for SMA inverter
$1000 for wood etc for the ground-mount structure
$1000 for miscellaneous stuff (wire, conduit, etc)
$200 permits
X 0.7 for the federal tax credit

$8550 for 5250W = $1.63/W

Pretty good...

I'm guessing the Solaria are through Renvu.

I would look at cheaper panels - you can get 300W 60cell modules for $.65/W. Downside is larger # of panels, so more optimizers, possibly less ideal shading. (18-panels at 300W would be 5.4kw, and $700 cheaper for panels, but maybe 20% more for optimizers and some of the misc. stuff.
Or there are 72-cell modules that are cheaper per watt. I don't think I'd do them for a roof mount DIY. But for ground mount DIY I'd think about it.

[RShackleford]

$8550 for 5250W = $1.63/W

Yeah, and then 30% back from Uncle Sam. Which drops to 26% in 2020, but apparently you can be grandfathered in if you spend some percentage of the total cost by the end of 2019.

I'm guessing the Solaria are through Renvu.

I would look at cheaper panels - you can get 300W 60cell modules for $.65/W. Downside is larger # of panels, so more optimizers, possibly less ideal shading. (18-panels at 300W would be 5.4kw, and $700 cheaper for panels, but maybe 20% more for optimizers and some of the misc. stuff.
Or there are 72-cell modules that are cheaper per watt. I don't think I'd do them for a roof mount DIY. But for ground mount DIY I'd think about it.

I'm kinda space-constrained.

[RShackleford]

Ok, compadres ... a lot of water under the bridge and a lot of decisions made by me. In light of my shading issues, I decided my system makes sense only if I can do it for $1/watt or so; which I'll be able to achieve since I'm getting some $0.36/watt Talesun panels from Renvu, and doing a ground-mount with treated wood.

So, my system will be ground-mount, with 16 Talesun 270watt panels (4.32 kW), and a 3.8kW SMA -41 string inverter (which has 2 MPPTs). I will not require rapid shutdown. I intend to run the DC strings (in buried conduit) to the exterior wall of the room where my main panel lives. I'll mount the inverter and a fused AC disconnect there, and connect the 240vac using line tap connectors (on the cables from the meter to the main breaker). I've confirmed that "secure power" will work without any TS4 modules at all.

I'm assuming that since I'll have ready access to the underside of the panel array, that it'll be a relatively simple matter to add Tigo opitmizers at some point in the future. The TS4-R-O being somewhat obsoleted by the new TS4-A-O (electrically similar but different form factor), I'd hoped the TS4-R-O would drop dramatically in price, but the cheapest I've seen it for is $40. So (with shipping) I could add optimizers to all panels for $674 (another $0.16/watt); no point in adding it to a subset of the panels, as they'll virtually all receive some partial shading from trees during the day. Also confirmed that neither the "access point" nor the "rooftop comm kit" is required for the optimizers to work; but they are required if I wanted module-level monitoring.

[nwdiver]

So, my system will be ground-mount, with 16 Talesun 270watt panels (4.32 kW), and a 3.8kW SMA -41 string inverter (which has 2 MPPTs).

What was the deciding factor for 16? A string of 8 is a bit low. Your DC voltage is only going to be ~250. How long is the run? If you can fit more panels I think that would improve your ROI. Same amount of wire but more energy.

[RShackleford]

What was the deciding factor for 16? A string of 8 is a bit low. Your DC voltage is only going to be ~250. How long is the run?

I thought more than 16 (landscape) would take too much space. Pondered 6x3 portrait and 4x5 landscape and just didn't want something quite that hulking out in the yard.

Not sure what you mean by how long is the run: from the panels to the inverter might end up being 40ft of wire. Probably 12awg THWN, $50 or so, plenty of ampacity for the panel's 9+amp short-circuit output current, 130 mohm or so for the round-trip - I don't see any issue there. Just a few feet from inverter to grid connection.

Yeah looks like 200vdc is about 2% off the sweet spot for inverter efficiency (see curves below), but well above the 100vdc minimum operating point for the MPPT. Maybe this is why, when I messed with SMA's tool for optimizing strings awhile back, I got a result I thought weird at the time, maybe something like 6 and 10 for the string lengths, seemed counterintutiive. I suppose I could test the wiring with one string pretty easily.

[nwdiver]

I thought more than 16 (landscape) would take too much space. Pondered 6x3 portrait and 4x5 landscape and just didn't want something quite that hulking out in the yard.

Not sure what you mean by how long is the run: from the panels to the inverter might end up being 40ft of wire. Probably 12awg THWN, $50 or so, plenty of ampacity for the panel's 9+amp short-circuit output current, 130 mohm or so for the round-trip - I don't see any issue there. Just a few feet from inverter to grid connection.

Yeah looks like 200vdc is about 2% off the sweet spot for inverter efficiency (see curves below), but well above the 100vdc minimum operating point for the MPPT. Maybe this is why, when I messed with SMA's tool for optimizing strings awhile back, I got a result I thought weird at the time, maybe something like 6 and 10 for the string lengths, seemed counterintutiive. I suppose I could test the wiring with one string pretty easily.

Ah... 40' isn't too terrible. Longer strings are generally better for a variety of reasons. For example if you had 1 string of 13 you would only need to run 1 pair of wires, your inverter would operate more efficiently AND line losses as a percentage basis would be slightly lower since the current is the same. 16 is a bit ungainly but it is what it is.

I think 2 string of 8 is ok... it's just that you could run 26 panels with the same amount wire.

[foo1bar]

I thought more than 16 (landscape) would take too much space.

What are you doing for your array?
2 high, 8 wide, landscape orientation?
Because when I first read the above sentence my first thought was "He's doing an array 16 wide and one tall? That's odd..

Pondered 6x3 portrait and 4x5 landscape and just didn't want something quite that hulking out in the yard.

I think bcroe's are 2-high portrait, many long, and that looks good.
Since you have some shading issues you probably want to have more panels in a string - so when panels are shaded or partly shaded there will still be enough voltage for the inverter.
So if it were me, I'd probably look at a 2 high, 10-long portrait orientation setup if 10 is how many can fit in a string. (maybe a 4x5 array - that'd be getting pretty tall, but if you needed it as a shed that could work.)

[RShackleford]

For example if you had 1 string of 13 you would only need to run 1 pair of wires, your inverter would operate more efficiently AND line losses as a percentage basis would be slightly lower since the current is the same. 16 is a bit ungainly but it is what it is.

I think 2 string of 8 is ok... it's just that you could run 26 panels with the same amount wire.

Just one string would save me about $20 on wire; seems a little silly. Ditto for line losses: 80ft of 12awg is about 0.13 ohms, so if the panels are producing short-circuit current I'm losing about ONE volt, out of several hundred.

[RShackleford]

What are you doing for your array?

Doing 4x4, with the panels in landscape mode. So array is about 22 x 13 ft, at 25-degree angle the footprint is 22 x 12 ft. Conceivably I could do 8x2, mighty wide, but a bit less wind uplift to ballast with the footings.

Since you have some shading issues you probably want to have more panels in a string - so when panels are shaded or partly shaded there will still be enough voltage for the inverter

I think it'd be very simple to wire it as two strings (sure, I'd need $40 worth of wire instead of $20) and experiment with configuring it as one string right there where the wires come into the inverter (just tie the two string into one long one by connecting one pair's '+' to the other's '-'). Not sure how legal or advisable it would be, since the panel's open-circuit voltage is 38.5v at STC, so a string of 16 exceeds 600v. This exceeds the inverter's maximum DC input of 600v, and I think 600v is a big breakpoint in the NEC requiring different wiring "methods". OTOH, at NOCT the open-circuit voltage is only 35.5v, so I'd be well below 600v. I wonder which test condition I need to observe ?

[bcroe]

I think bcroe's are 2-high portrait, many long, and that looks good.
Since you have some shading issues you probably want to have more panels in a string - so when
panels are shaded or partly shaded there will still be enough voltage for the inverter.
So if it were me, I'd probably look at a 2 high, 10-long portrait orientation setup if 10 is how many
can fit in a string. (maybe a 4x5 array - that'd be getting pretty tall, but if you needed it as a shed
that could work.)

If there is not much space, a pretty tall array works. But you give up ease of snow removal or
making them tilt. With multiple rows, more, lower arrays do not really take a lot more land, because
the shadow is shorter. 2 high work well here, everything is in easy reach. With both landscape and
portrait mounts here, it is much easier to keep snow off the landscape, about the same height but
my near vertical landscape mounts have an 8 inch space and 40 inch ground clearance.

After this weekends storm, the near vertical panels were completely clear, but the summer angle
panels had 3 inches of snow and ice. Bruce Roe