6kW system design

[max2k]

If you use anything other than SMA's rapid shutdown, you lose the Secure Power Supply capability. (Ok, I think there are other array powered options, but all I found was more expensive than SMA). Happy trails... This is why I said the heck with it and went with SolarEdge, which I know and understand, even if I don't love.

What I found peculiar is that single SMA RS controller can control up to 16 RS boxes meaning they solved the problem of controlling RS boxes for completely isolated strings feeding different inverters but despite each RS box having 2 independent 'channels' those channels must be connected to the same inverter making me buy 2 separate RS boxes for my layout.

[sensij]

Another question just came up- county requires now use of Rapid Shutdown. For the system I'm trying to build which includes 2 independent inverters it seems I need to purchase 2 separate Rapid Shutdown Boxes (the ones which are installed next to strings). Each goes for around $350 so this requirement brings up costs by $800 as I'd also need controller + misc wiring hardware. Are there cheaper solutions possible?

If you use anything other than SMA's rapid shutdown, you lose the Secure Power Supply capability. (Ok, I think there are other array powered options, but all I found was more expensive than SMA). Happy trails... This is why I said the heck with it and went with SolarEdge, which I know and understand, even if I don't love.

[max2k]

Another question just came up- county requires now use of Rapid Shutdown. For the system I'm trying to build which includes 2 independent inverters it seems I need to purchase 2 separate Rapid Shutdown Boxes (the ones which are installed next to strings). Each goes for around $350 so this requirement brings up costs by $800 as I'd also need controller + misc wiring hardware. Are there cheaper solutions possible?

[max2k]

negligible IMO

I'd look at just a regular square-D (or eaton or GE or ...) subpanel. This is going to be mounted next to/between the two inverters on a wall, right? If MidNight is cheaper or better in some way - go for it - but I would guess the economy of scale will make square-D cheaper.


For a single string or 2 strings going to independent circuits in the inverter I don't think there is.
The current is already limited by how much the panels can produce (and of course you should use the Isc not Impp. And look at what Isc will be on a really cold day (since cold => higher current)



If it's on the side of the house I think you can do regular AC wiring inside the walls back to the panel. Or from your subpanel back to your main panel.
Of course that may mean some drywall repair - but it's another option for you to think about.
The DC wires probably need to be in conduit (I think they have to be inside metal conduit if it's inside the building)

A water hose can do great at making a hole through dirt or sand to get across underneath sidewalk.
A piece of pipe, a coupling (to make it easier to hit) and a sledgehammer can work too.
The third option is to just remove that section and dig down then pour a new section.
And I've done all 3 of those at different times.

The water definitely winds up being a muddy mess - but if you have time to let it drain away and dry out it can work nicely.

Midnight combiner is anything but cheap . It is nice looking and standing out so I'll probably go with it just for my own sake- I see the point utility is trying to make when they insist on AC disconnect being visible/lockable and next to MSP: this way anyone can quickly turn this off without reading through red labels.

I'll go with wiring through the walls/attic- after running numbers the loss associated with extra length of #8 AC wire will be around 0.2% which can be considered negligible. This way all the EMT can be hidden inside the walls. If I'm not mistaken AC output from inverters can even be run through the walls without EMT- the rest of the house is wired this way.

I used that water technique once while making hole under sidewalk for a drain pipe - worked like a charm. Here it would be a little difficult as I don't have any side of the concrete slab open- it is stretched between house wall and stucco fence, no ground to dig in between. Anyway, it's a moot point now since I'm going to hang inverters on the house wall.

Thank you for the ideas, much appreciated.

[foo1bar]

Cold causes higher *voltage*, but lowers the Isc. Isc on a really hot day would be the worst case current.

Oops - my mistake - that is indeed the case. On one sheet I checked it's 0.03% per degree C.
My understanding is the starting point/characterization point is 25C - so if it gets to 55C (131F) that'd be 30 degrees more.
So 1.0003 ** 30 = 1.0090; so less than 1%
So I think at least for LG's 320W modules you won't see that much difference in Isc even on a pretty hot day.

Voc and max-power makes has a much larger negative % per degree C - so it winds up being more power on a really cold/windy day if you have the same sunlight (same angle, same amount of dust in the air, etc)

[Mike90250]

DC breakers in a combiner box, even for only 2 strings, are handy for any eventual troubleshooting. Makes it easy for a A/B comparison, or if you need to take gear offline.

[sensij]

For a single string or 2 strings going to independent circuits in the inverter I don't think there is.
The current is already limited by how much the panels can produce (and of course you should use the Isc not Impp. And look at what Isc will be on a really cold day (since cold => higher current)

Cold causes higher *voltage*, but lowers the Isc. Isc on a really hot day would be the worst case current.

[foo1bar]

There's some benefit from reliability point of view as the system now consists of 2 independent generators combined only inside AC combiner

negligible IMO

(MidNight AC combiner / Disco)

I'd look at just a regular square-D (or eaton or GE or ...) subpanel. This is going to be mounted next to/between the two inverters on a wall, right? If MidNight is cheaper or better in some way - go for it - but I would guess the economy of scale will make square-D cheaper.

I don't really see a point in using DC fuses/OCPD in this system.

For a single string or 2 strings going to independent circuits in the inverter I don't think there is.
The current is already limited by how much the panels can produce (and of course you should use the Isc not Impp. And look at what Isc will be on a really cold day (since cold => higher current)

After taking long look at the sidewalk (concrete everywhere) I think I'll hang inverters on the side of the house instead to avoid need to run conduits inside/under sidewalk. The only problem with this layout is I need to run AC conduit around door on that side of the house. It's a side door to the built in garage and inverters AC openings are at about middle of its height. I'm trying to avoid creating conduit jungle out of all this.

If it's on the side of the house I think you can do regular AC wiring inside the walls back to the panel. Or from your subpanel back to your main panel.
Of course that may mean some drywall repair - but it's another option for you to think about.
The DC wires probably need to be in conduit (I think they have to be inside metal conduit if it's inside the building)

A water hose can do great at making a hole through dirt or sand to get across underneath sidewalk.
A piece of pipe, a coupling (to make it easier to hit) and a sledgehammer can work too.
The third option is to just remove that section and dig down then pour a new section.
And I've done all 3 of those at different times.

The water definitely winds up being a muddy mess - but if you have time to let it drain away and dry out it can work nicely.

[max2k]

If the wire is sized so that it can handle the maximum current the panels could ever possibly generate then there's no need for a OCPD.



If the GEC/EGC is inside the metal conduit - yes absolutely. (reduces effect of the metal conduit acting as a magnetic choke on current through the grounding wire)
I think you said you're planning to run it outside the conduit, in which case I don't think it's required - but don't see how it'd hurt anything so I would do it.


Are you sure you need a GEC and not just an EGC?


I don't see why it would - Obviously you need to follow proper burial depth rules - which probably means it's actually UNDER the sidewalk by a fair amount (since presumably the conduit will partially be not under cement and therefore need to be at the proper depth for that. (obviously you could be doing a really thick sidewalk and it could be in the sidewalk - but that'd be pretty unusual)


I would connect any metal conduit to the EGC.


I'm not a licensed electrician - so take the above comments and use them to do your own research into what is required by code.

Good luck

Thanks a lot for your responses! Since then I changed design a little and it will be 2 strings x 12 panels each connected to its individual SB 3.8 inverter, 2 inverters / 24 panels in total. While 7.7 model would fit the bill power wise these 2 inverters don't have cooling fans inside so it would be 1 less maintenance item/noise source. There's some benefit from reliability point of view as the system now consists of 2 independent generators combined only inside AC combiner (MidNight AC combiner / Disco). Ideally it will be much more reliable provided failures are statistically independent. Unfortunately I'm buying all equipment at once so this might not be the case after all.

Reading about DC fuses / OCPDs convinced me not to use either as DC rated breaker is completely different creature from AC counterparts due to necessity to fight arcing effect. The DC wires will be sized to sustain continuous Isc with good margins required by NEC anyway so I don't really see a point in using DC fuses/OCPD in this system.

After taking long look at the sidewalk (concrete everywhere) I think I'll hang inverters on the side of the house instead to avoid need to run conduits inside/under sidewalk. The only problem with this layout is I need to run AC conduit around door on that side of the house. It's a side door to the built in garage and inverters AC openings are at about middle of its height. I'm trying to avoid creating conduit jungle out of all this.

Don't worry, I find strange the notion of trying to blame someone from the Internet for any sort of trouble: you're free to say anything and I'm free to use that info or not at my own risk. It's not like I'm forced to act in any certain way, that is purely my own choice.

[foo1bar]

1. Am I correct that DC side does not need OCPD or even fuses as DC wires go from each of 2 strings to the MPPT inputs individually? Inverter has 3 MPPT inputs but I'm going to utilize only 2.

If the wire is sized so that it can handle the maximum current the panels could ever possibly generate then there's no need for a OCPD.

2. Am I required to use grounding bushings to bind GEC to the DC conduit on both ends?

If the GEC/EGC is inside the metal conduit - yes absolutely. (reduces effect of the metal conduit acting as a magnetic choke on current through the grounding wire)
I think you said you're planning to run it outside the conduit, in which case I don't think it's required - but don't see how it'd hurt anything so I would do it.

3. The inverter will be installed 8'-10' from the building on a stucco fence/wall which goes perpendicular to the building. Do I need to run GEC all the way to the inverter or GEC can branch off earlier to connect to the building ground conductor?

Are you sure you need a GEC and not just an EGC?

4. Both DC & AC conduits to the inverter will go inside sidewalk concrete for 4'-5'. Would this present any problem from grounding point of view?

I don't see why it would - Obviously you need to follow proper burial depth rules - which probably means it's actually UNDER the sidewalk by a fair amount (since presumably the conduit will partially be not under cement and therefore need to be at the proper depth for that. (obviously you could be doing a really thick sidewalk and it could be in the sidewalk - but that'd be pretty unusual)

5. The inverter AC conduit will have 4 conductors inside: L1, L2, N & EGC coming from main panel (no GEC). Do I need to bind EGC to the AC & DC conduits on the inverter side?

I would connect any metal conduit to the EGC.


I'm not a licensed electrician - so take the above comments and use them to do your own research into what is required by code.

Good luck

[max2k]

I'm working on 6.4kW grid tied solar install which consists of 20 LG320 panels in 2 strings(10 + 10) and SMA SB 6.0-US string inverter. I'm going to run 6AWG bare copper unsplit GEC wire from the rails next to DC EMT conduit all the way to the building ground electrode. Panels will be grounded to the rails by manufacturer approved grounding bolts. Their DC conductors will go through SolaDeck roof penetration where wires will change from PV to THWN-2 type using terminal block inside SolaDeck and then run down through the attic inside 3/4" EMT conduit (4 DC conductors). Could someone please clarify few points for me:

1. Am I correct that DC side does not need OCPD or even fuses as DC wires go from each of 2 strings to the MPPT inputs individually? Inverter has 3 MPPT inputs but I'm going to utilize only 2.

2. Am I required to use grounding bushings to bind GEC to the DC conduit on both ends?

3. The inverter will be installed 8'-10' from the building on a stucco fence/wall which goes perpendicular to the building. Do I need to run GEC all the way to the inverter or GEC can branch off earlier to connect to the building ground conductor?

4. Both DC & AC conduits to the inverter will go inside sidewalk concrete for 4'-5'. Would this present any problem from grounding point of view?

5. The inverter AC conduit will have 4 conductors inside: L1, L2, N & EGC coming from main panel (no GEC). Do I need to bind EGC to the AC & DC conduits on the inverter side?