Ground-based Grid-tied Solar system – Self Install

[jacknj]

I am in the planning phase of a ground mount solar system in NJ (where net metering is still provided). I am planning for a 7.3 KW grid-tied solar system (Full South exposure, tilt of 30 degrees) which according to PVWATTs would give me about 9.7 KWh / year. I have included in the PVWATTS model parameters about 4.5% of shading loss – a Sketchup model with Skelion extension provided me with 2.2% overall shading loss. I do have an existing 12 KW generator feeding a subpanel through an ATS switch, and at this time I do not consider battery to be required for the solar system. This sub-panel will be isolated from the solar system in case of outage.
I had been planning for a while to use the SolarEdge inverter/optimizers solution, mostly for the capability to:

• automatically shut down (through the optimizers) the panels to just 1 VDC, and to
• monitor the production of each panel (this appealed to an engineer with data analysis background).

However the reported failure rates of SolarEdge as well as their several weeks of lead time for replacing their inverters (as acknowledged to me by SolarEdge) is making me consider the simpler SMA solution using only one inverter without optimizers, solution that seems to have a better track record in terms of reliability. I understand that SMA offers also the the capability to add optimizers, but this will just add to the complexity of the system without much benefit in terms of production (if you believe their ShadeFix optimization is good enough for my type of shading). For a ground mounted system, I do not need a rapid shutdown as roof mounts do– even though the added safety of such feature might be considered a benefit.
As a homeowner-installer I have reached out in the past weeks to SolarEdge and SMA to get answers to some specific technical questions, and their pre-sale support responsiveness (or lack thereof) left me wondering what to expect in post-sale.
I plan to put the inverter inside the basement on the side of the electrical panels, so there would be no exposure to heat, and may be better reliability.
As I am planning at least 2 strings for the solar system, I have been trying to understand if string monitoring was a possibility for the SMA inverter with Smart Connected. No concrete answer from SMA yet. Is there a known possibility for that? If that is the case, I would configure the 20 panels in 3 strings to get more granular level of monitoring.
My questions:

• Is the SMA system simplicity and reliability a better trade-off in my case?
• Is there a string level monitoring available with an SMA inverter or through post analysis of SMA data?
• Does NEC 2017 require DC disconnect in addition to AC disconnect (required by my power company) for ground based solar system? The code does not seem very clear on that topic. I have also asked the inspector at my local AHJ. Box. if I select the SMA solution I am considering putting a combiner box with a DC disconnect capability at the array.

[solarix]

I recommend SMA string inverters. Wire your array as two strings of ten and you will be able to compare the production of the two and catch any module failures (which are extremely rare). The display on the SunnyBoy inverter will give you voltage and current for each input but this is not available through its online monitoring. These inverters are very reliable and easy to service in case of a failure. Ask yourself this question - Would you rather have a less reliable, individually optimized system where you can easily see the failures, or a very reliable, easily serviced system that has less visibility? PV panels are ultra reliable - don't marry them with the reliability risk of panel optimization.
On the DC disconnect question, the code implies there should be a disconnect at the array if it is "out of sight" of the inverter, but our inspectors around here never require it. If you put in a disconnect at the array, it needs to be rated for high voltage DC as people will tend to turn it off when the current is still flowing. I think you are better off with no disconnect at the array and force people to shut off the inverter first. Shutting of the array disconnect doesn't do much anyway as the array is still hot.

[peakbagger]

I would advocate two string inverters sized for 50/50. When one dies (they will) you still have half capacity. until you replace. Most string inverters have multiple MPPT inputs so with two of them you can have some shading flexibility. Unless pricing has changed, string inverters are generally far less expensive than a optimizers and a SolarEdge inverter.. No need for optimizers on a ground mount array.

IMO no code requirement for local disconnect but a very good idea as it keeps the fire department happy. They really want a positive means of disconnect on PV or they sit back and watch the place burn while they cover the panels. Siemens makes or made some affordable ac/dc disconnect switches. I would also install a Midnight Solar Surge Protection Device at the array so if there is surge it stays out at the array.

[reader2580]

My solar electrician friend recommended staying away from the newer Solaredge inverters without the display. I had purchased a new Solaredge 11.6KW inverter without the display for an expansion and used it for a few months to test it. My friend said it would most likely fail prematurely and to stick with the 3000 watt and 6000 water Solaredge HD Wave inverters with display I already had as they had been in service for several years and would be unlikely to fail prematurely.

I ended up selling the Solaredge 11.6KW inverter on Ebay to recoup some of what I spent and used Enphase microinverters for my new panels.

[bcroe]

My inverters have proven reliability over a decade of hard use.
The last thing I want is something complicated by the internet
and with no display. I managed to pick up a new spare set for
the future, for a bargain price when someone decided they were
obsolete. Bruce Roe

[jacknj]

I recommend SMA string inverters. Wire your array as two strings of ten and you will be able to compare the production of the two and catch any module failures (which are extremely rare). The display on the SunnyBoy inverter will give you voltage and current for each input but this is not available through its online monitoring. These inverters are very reliable and easy to service in case of a failure. Ask yourself this question - Would you rather have a less reliable, individually optimized system where you can easily see the failures, or a very reliable, easily serviced system that has less visibility? PV panels are ultra reliable - don't marry them with the reliability risk of panel optimization.
On the DC disconnect question, the code implies there should be a disconnect at the array if it is "out of sight" of the inverter, but our inspectors around here never require it. If you put in a disconnect at the array, it needs to be rated for high voltage DC as people will tend to turn it off when the current is still flowing. I think you are better off with no disconnect at the array and force people to shut off the inverter first. Shutting of the array disconnect doesn't do much anyway as the array is still hot.

Thanks.
May be the chart in this video makes also your point...

https://twitter.com/i/status/1559535586121236481

Regarding the disconnect, I would consider it might be appropriate to have one if you have a combiner box with fuse. Those boxes with fuses and no disconnect seems dangerous to me. I have looked at boxes with disconnect, fuses, and surge protection but they all seemed to be designed for separate MC4 outputs (which is not my application as I plan for THHN running in one conduit)