Sunny Island 6048 wiring question

[magic8192]

I have a pair of 120V SI 6048-US inverters and The plan is to wire from the Sunny Island to a 60 amp disconnect to the secondary loads panel with a 60 amp breaker.

Each inverter is wired L1, neutral for inverter 1 and L2, neutral for inverter 2. I have 2 neutral leads and I am not sure of the best way to handle this. I have one connector in the disconnect for neutral. I could put a 2 lug to 1 lug combiner in the disconnect box and wire the L1 neutral and the L2 neutral into that combiner lug?

 

[Hedges]

2x 6048 AC1 loads output --> 2-pole disconnect --> secondary loads panel?

You could join the two wires one way or another. Whatever lug you're considering, maybe a ferrule that takes 2 wires, or replace lug with dual input lug. I've taken a length of wire, stripped insulation off 1/2" in the middle, fed it through a lug and had two ends to use. Putting two wires in one lug isn't approved by code (but seems to work.) Use a split bolt to join two wires, with one sticking out to fit lug, and tape the split bolt.

Feed neutral past disconnect, just land two neutral wires on sub-panel busbar. Daisy-chain neutral from one Sunny Island to the other, make splice there, and run one wire to sub-panel.

My system was already wired with Sunny Boy on a sub-panel with 2 awg L1, L2, Neutral and smaller ground coming from main panel through disconnect. When I installed Sunny Island, I left the Neutral and ground in place, wired output of disconnect to input of Sunny Island, output of Sunny Island to sub-panel. I pulled a neutral and ground back from sub-panel to Sunny Island. I only hooked up Neutral and Ground to AC1 of Sunny Island, not to AC2; they are connected internally. Due to conduit getting full, I pulled a single neutral wire from sub-panel and spliced to two wires in a junction box just before Sunny Island. Used on two phases, Neutral carries difference between L1 and L2, so single wire sufficient.

Why have a disconnect there at all? The breaker disconnects loads panel from Sunny Island. We typically use a disconnect between grid and UL-1741 inverters when required by utility.

I wonder if 60A is too small. Manual says, "A circuit breaker with a rated current of a maximum of 56A must be installed", but SI can deliver 50A continuous, and pass-through from grid or generator can be 56A. The thermal-magnetic breakers typically used in U.S. are supposed to be sized 25% over, so only carrying 80% of rating continuously. 56A x 1.25 = 70A. I used 70A QO270 initially, then switched to 63A DIN mount breaker before the box. I think the magnetic-hydraulic breakers often used in Europe (and similar sold by Midnight in the U.S.) don't need that derating because they don't work by heat.

Section 6.1 of the manual, "Grounding", says "Use an overcurrent protective device for a maximum of 70 A" which I take to mean on input, AC2. I would have thought it would be designed for same on output, but that's not what manual says. I suppose OCP protection needed on output because 56A pass-through from grid/generator plus 50A from inverter would be 106A; unfortunate they didn't design PCB traces for that.

Feeding a sub-panel you way. Do you have Sunny Boys? It is useful to have Sunny Boys wired to output of Sunny Island, but have a load-shed relay disconnecting another sub-panel with loads (in case of low battery.)

 

[magic8192]

2x 6048 AC1 loads output --> 2-pole disconnect --> secondary loads panel?

You could join the two wires one way or another. Whatever lug you're considering, maybe a ferrule that takes 2 wires, or replace lug with dual input lug. I've taken a length of wire, stripped insulation off 1/2" in the middle, fed it through a lug and had two ends to use. Putting two wires in one lug isn't approved by code (but seems to work.) Use a split bolt to join two wires, with one sticking out to fit lug, and tape the split bolt.

Feed neutral past disconnect, just land two neutral wires on sub-panel busbar. Daisy-chain neutral from one Sunny Island to the other, make splice there, and run one wire to sub-panel.

My system was already wired with Sunny Boy on a sub-panel with 2 awg L1, L2, Neutral and smaller ground coming from main panel through disconnect. When I installed Sunny Island, I left the Neutral and ground in place, wired output of disconnect to input of Sunny Island, output of Sunny Island to sub-panel. I pulled a neutral and ground back from sub-panel to Sunny Island. I only hooked up Neutral and Ground to AC1 of Sunny Island, not to AC2; they are connected internally. Due to conduit getting full, I pulled a single neutral wire from sub-panel and spliced to two wires in a junction box just before Sunny Island. Used on two phases, Neutral carries difference between L1 and L2, so single wire sufficient.

I am just going to use something like the pic below, which seems to be accomplishing the same thing you did with splicing 2 wires into one.

Why have a disconnect there at all? The breaker disconnects loads panel from Sunny Island. We typically use a disconnect between grid and UL-1741 inverters when required by utility.

I have to have the disconnect on the outside of the house near the mains coming from the power company. I have the solar disconnect and the battery disconnect - battery disconnect is the input from the inverters to the secondary loads panel.

I wonder if 60A is too small. Manual says, "A circuit breaker with a rated current of a maximum of 56A must be installed", but SI can deliver 50A continuous, and pass-through from grid or generator can be 56A. The thermal-magnetic breakers typically used in U.S. are supposed to be sized 25% over, so only carrying 80% of rating continuously. 56A x 1.25 = 70A. I used 70A QO270 initially, then switched to 63A DIN mount breaker before the box. I think the magnetic-hydraulic breakers often used in Europe (and similar sold by Midnight in the U.S.) don't need that derating because they don't work by heat.

Section 6.1 of the manual, "Grounding", says "Use an overcurrent protective device for a maximum of 70 A" which I take to mean on input, AC2. I would have thought it would be designed for same on output, but that's not what manual says. I suppose OCP protection needed on output because 56A pass-through from grid/generator plus 50A from inverter would be 106A; unfortunate they didn't design PCB traces for that.

Feeding a sub-panel you way. Do you have Sunny Boys? It is useful to have Sunny Boys wired to output of Sunny Island, but have a load-shed relay disconnecting another sub-panel with loads (in case of low battery.)

I used 4 AWG wire because the SI installation manual recommended that. You have me concerned over the 60 amp breaker. I was thinking the 60 amp breaker was correct because the 60 amp disconnect is sized for 4 AWG wire.

I really appreciate you response.

 

[RCinFLA]

I think the input breakers should be same rating as pass through relay's rating in inverter. For Sunny Island that is 60A (actually 56A on inverter's spec's ), dual breaker for two series connected 240/120vac inverters.

 

[Hedges]

I think the input breakers should be same rating as pass through relay's rating in inverter. For Sunny Island that is 60A (actually 56A on inverter's spec's ), dual breaker for two series connected 240/120vac inverters.

Except, SI can pass 56A continuously. It is actually a 3p relay rated 30A. At first I thought 3 poles in parallel, but maybe one used for another purpose and 2x 30A rated contacts in parallel, derated slightly to 56A.

A 60A breaker shouldn't be used for 56A continuous, at least not a thermal breaker. Nor a fuse. Maybe a magnetic-hydraulic breaker.
Sunny Island will limit input current to 56A, so I'm not worried about over-current through relay under normal conditions. If there is a fault, 70A breaker would trip. I happen to have 63A breaker on input because I used DIN not yet another breaker panel to connect 2-pole switch to 2x 2-pole breakers to 4x SI.
SI manual says max 70A breaker on input, so that is kosher.

The SI manual says 56A max on the output. I used 70A initially, now 63A. I used 6 awg, which is good for 70A under moderate temperature conditions. 4 awg is better, can handle 70A in hotter conditions. Not that 70A will flow, but 70A fuse OK with proper NEC derating of wire ampacity.

I am just going to use something like the pic below, which seems to be accomplishing the same thing you did with splicing 2 wires into one.

I've used a 2-hole version on main lugs of a sub-panel. That let me daisy chain to another panel.

I have to have the disconnect on the outside of the house near the mains coming from the power company. I have the solar disconnect and the battery disconnect - battery disconnect is the input from the inverters to the secondary loads panel.

So you have AC2 connected to grid. Isn't that where utility wanted disconnect?
Maybe you have Sunny Boys or other GT PV on AC2, so disconnect there accomplishes the same.
SI is also a UL-1741 inverter and can backfeed the grid under some conditions (DC coupled charging which raises battery voltage above SI's setpoint.) One more reason to have a disconnect between SI and main panel.


You haven't mentioned PV. Or if your SI system is just a battery UPS with no PV.
Especially if only AC coupled PV and no DC coupled PV, I would want to see a load-shed relay. That avoids running the battery down and having to do a black start.

 

[magic8192]

I will try to get a diagram of my setup so it is easier to see what I have.

 

[magic8192]

Here is a basic diagram

 

[Hedges]

"200A transfer switch" feeds 200A secondary panel from grid.
SI AC1 feeds 200A secondary panel.
Do you have an interlock to prevent both from feeding that panel?

SI AC2 also goes to "200A transfer switch", which seems a bit confusing. Is that switch connecting main panel to either SI or secondary panel?

I have a 200A panel with 200A main breaker to grid, and interlocked back-fed breaker to SI AC1 (after load-shed relay). That lets me feed non-critical loads from either grid or inverter, manually switched. Because same panel has breaker feeding SI AC2, I manually shut that off when backfeeding this panel, so inverter doesn't feed its own input.


60A disconnect for PV related stuff. That's fine.

The 60A disconnect between SI AC1 and 200A secondary panel seems unnecessary.

Here is what I might do:

Grid feeds main panel.
Breaker in main panel feeds SI AC2
SI AC1 feeds 60A disconnect switch to PV system.
SI AC1 goes through load-shed relay to 200A secondary panel back-fed interlocked breaker.
Main breaker (interlocked) in secondary loads panel is fed by main panel.

This would connect PV only through Sunny Island, and the loads panel normally through Sunny Island, but option to connect to grid.
If battery gets low, loads panel is disconnected by PV keeps working.

You might want to be able to manually switch PV straight to grid, to keep it working if SI is down.

 

[magic8192]

Transfer switch down feeds secondary loads panel from grid
Transfer switch up feeds generator backup on SI

My battery room is attached to the house and my understanding is that I have to be able to disconnect power to the house in case of fire.
I do not have an interlock.

 

[magic8192]

When the Transfer switch is in the down position, I run the PV grid tied.
The system is currently running grid tied.

The new stuff is the SI/Battery/60 amp safety disconnect/wiring to SI generator backup. (Everything in the top left quadrant of the picture.

 

[Hedges]

Transfer switch down feeds secondary loads panel from grid
Transfer switch up feeds generator backup on SI

...
I do not have an interlock.

If it is possible to close switches so SI AC1 and grid connect together, that can kill the inverter or electrocute the lineman.
If you are lucky it trips a breaker without damaging inverter (but that wouldn't help lineman who touches wires you're feeding, especially high voltage side of a transformer)

A transfer switch to connect SI AC2 to either grid or generator is appropriate, but your switch has connection to 200A panel not generator.
If AC2 can be switched between grid and generator, switch should have a signaling contact to tell SI when generator is connected (so it doesn't permit backfeed.)
Switching between grid and generator must turn AC2 power off for 5 seconds before connecting other source. Because the two sources won't be synchronized, don't want SI or SB driving power out of phase, need 5 seconds to ensure disconnected.
No generator shown.

Or, when you say "Generator" do you mean SI is the "Generator"? But it is AC2, input of SI, not AC1, output of SI. Also AC1 separately goes to the panel through yet another switch.

Unfortunately, SMA system diagrams don't label AC1 and AC2. Only zoomed in picture of AC connections show that, and words distinguish which goes where.

 

[magic8192]

The manual labels AC2 as generator/grid and I was using that loosely, my bad. When I said generator, I mean AC2 of the SI. My idea was that if there was no sun and the battery was low, the SI could swap over to AC2 (which is connected to the grid) and charge the battery. I wasn't aware that I needed a 5 second delay.

I need to rethink the interlock too ugh.

 

[Hedges]

SI AC2 would normally remain connected to the grid. If set up for grid-feed, GT PV (Sunny Boys, etc.) on AC1 produce power, go through SI internal relay, backfeed grid. SI only disconnects from grid if grid is down.

If you have only one AC source (grid, no generator) feeding input AC2 of SI, no need for transfer switch there. Just 70A breaker or fused disconnect. In my system, I have a visible blade disconnect there to satisfy PG&E. The Sunny Boys only feed through Sunny Island and through that disconnect (if Sunny Island goes down, my PV goes down, no separate path to grid.)

The interlock I would install is a "generator input" in the 200A panel. Main breaker goes to grid (main panel). Backfed Generator breaker goes to SI AC1 (through a load-shed relay.) If SI is down, you can throw breakers and loads are on grid. It is about $50 to $60 sheetmetal piece that mounts on the panel.

I don't know if your main panel has any branch circuits, or just a 200A main breaker and cable to the 200A sub-panel.
If no branch breakers, you could use Polaris or similar taps to also feed cables to a fused disconnect (70A fuse max) feeding SI AC2.

Utility wants a disconnect which stops PV backfeed. If your PV can go either to SI or direct to grid, a different disconnect (you show 60A safety disconnect) would do that.

 

[magic8192]

I don't have a generator. The plan was to use the grid as backup to charge the battery if it gets low. Did not know about the 5 sec delay.
My 200 Amp sub panel has a main disconnect that I can interlock with the breaker going to the SI.

Maybe I should just not use the Grid as a backup? I can get an actual generator and hook it up with auto start?

 

[Hedges]

No need for 5 second delay if only one AC source.
Switching AC off/on to AC2 input is OK. Only problem would be switching between grid and gas generator. Both making AC, but out of phase.

It is generally not desirable to run a battery low, recharge from grid, repeat. Cycling wears out a battery. Unless you don't have net-metering, and you're trying to run entirely off PV, charge from grid (or generator) only if PV can't keep up.

No, actual generator costs more to run than grid.
Just leave SI AC2 input permanently wired to grid. It will remain connected to grid when up, disconnect when down. Like a UPS.
If PV inverters do frequency-watts, feed those into AC1 output. They will feed through to grid when grid is up, operate under control of SI when grid is down.

 

[magic8192]

My plan is to run off grid and charge from the PV. Using the AC2 only when PV can't keep up. Lots of cloudy days.
The secondary loads should run off the battery at night.

My understanding is that if the battery is charged and there is not enough load, the PV micro inverters will turn off.

Ok, so I can hook up the grid to the AC2 connection and use that as a back up if that is the only backup source. That is good. Whew!

So all I really need to do is get the interlock setup working. I have a Square D loads panel and my understanding is that you were referring to something like the picture below. I have a main disconnect and breaker locations just like pictured below.

 

[Hedges]

If your micro-inverters do frequency-watts, as Sunny Island ramps up frequency they will ramp down power, and things run smoothly.
Without that feature, as frequency ramps up, microinverters drop offline, wait 5 minutes, reconnects. That may cycle one every 10 minutes or so. I don't think that is good for batteries or relays.

Using grid as generator only, surplus PV is wasted. Do you not want to backfeed it to grid?

Similar to that interlock. The Square-D QO ones I have are sheetmetal rather than cast aluminum.
It seems to work fine. The main breaker is quite stiff, and the interlock has a tab sticking up, to be careful not to slip and cut yourself. Leather work gloves maybe?

Does your planned battery support the usage planned, enough storage and long enough cycle life?
I don't think most batteries are cost effective compared to grid power. Some low-cost ones are. But net metering will be more cost-effective unless credit is reduced vs. charge for power consumed.

 

[magic8192]

If your micro-inverters do frequency-watts, as Sunny Island ramps up frequency they will ramp down power, and things run smoothly.
Without that feature, as frequency ramps up, microinverters drop offline, wait 5 minutes, reconnects. That may cycle one every 10 minutes or so. I don't think that is good for batteries or relays.

My micro inverters drop offline

Using grid as generator only, surplus PV is wasted. Do you not want to backfeed it to grid?

I wasn't aware that you could backfeed PV to the grid using the SI 6048-US? It probably isn't worth it because I get paid producer rates, about 4 cents / kwh

Similar to that interlock. The Square-D QO ones I have are sheetmetal rather than cast aluminum.
It seems to work fine. The main breaker is quite stiff, and the interlock has a tab sticking up, to be careful not to slip and cut yourself. Leather work gloves maybe?

Does your planned battery support the usage planned, enough storage and long enough cycle life?
I don't think most batteries are cost effective compared to grid power. Some low-cost ones are. But net metering will be more cost-effective unless credit is reduced vs. charge for power consumed.

I have 48 280 amp hour LiFePo4 LF280k cells. I expect these to pay for themselves in about 7 - 8 years. I think I will get 10+ years service from them. I have about 43kwh available and my average battery usage per day will be about 15 - 20 kwh.

 

[Hedges]

My micro inverters drop offline


I wasn't aware that you could backfeed PV to the grid using the SI 6048-US? It probably isn't worth it because I get paid producer rates, about 4 cents / kwh

I have 48 280 amp hour LiFePo4 LF280k cells. I expect these to pay for themselves in about 7 - 8 years. I think I will get 10+ years service from them. I have about 43kwh available and my average battery usage per day will be about 15 - 20 kwh.

Would be nice to have GT PV inverters that do adjust power. Even having some that do and some that don't, I don't think makes it well behaved.
Maybe an equal or greater amount of DC coupling would let batteries be at desired charge voltage, with microinverters either 100% on or 100% off.

With GT PV on AC1, if configured for it SI will let them backfeed through its relay to the grid.
If DC coupled PV has an SCC driving voltage higher than SI wants, SI can suck power from batteries and backfeed grid.
SI is UL-1741 but not UL-1741-SA, not allowed for net metering where that is adopted.

If you do cycle that much, it is only 35% to 50% of capacity, so ought to last at least 10 years, hopefully 20.
DIY with individual cells I expect to cost $0.05/kWh over their cycle life. Assuming retail utility rates much higher, this could save money.
PV may cost you $0.05 (DIY) or maybe $0.20/kWh (amortized over 10 years), depending on what you payed to install. Should last 20+ years, bringing cost down to half that.

Ideally you would target keeping batteries around 50% or 65% full, leaving charge for loads and capacity to charge from PV. I don't know a great way to do that, but if SI is informed (or can infer) SoC and does "grid charge" at a somewhat lower % full. Ideally knowing charge/discharge pattern vs. time of day, but that is in other products, not Sunny Island.

 

[magic8192]

Unfortunately I don't have power adjusting inverters, they weren't available when building the original PV system. I will see how it works for a while. I might try something moving forward like using the excess energy for bitcoin mining?

Reconfiguring my PV is not impossible, but will involve a commitment because the PV is about 100 ft from the house and all the wires are buried in a 4 inch pipe.

It is kind of a catch 22 because I want to use my own power to lower the power bill to offset the cost of the system. That is how I sold this to the wife unit.

I still have to configure the battery and BMS and commission the inverters.