diy solar

diy solar

Grid tie system with small critical loads battery backup

skyking1

Solar Enthusiast
Joined
Mar 28, 2024
Messages
113
Location
Washington State
I have been reading other threads with a 'critical loads" search.
My initial deployment is 5~8Kw of roof mounted panels, depending on the roof real estate I have.
My goals are to have a minimal critical loads subpanel that can be backed up with an inverter and batteries.
I can keep that load down to 240V@50A, a 6 space box with (4) 15A 120V breakers and a pair of 20's.
That is the breakered load, the actual loads during a power outage would be much less. (Turn off the lights when not in the room, etc. )
Originally I was thinking of microinverters for the built in RSD feature, but the backup aspect may be better served with a string inverter/charger.
I may be wrong on that and would welcome any input.
I will get a net metering agreement and it's a 1:1 with banking from March to March, so there is no advantage for running off the battery bank. I just want the lights to work and the food to not spoil, and my hydronic pumps to work in an outage.
If we are away and all the lights are off, the refrigerator and freezer loads won't be much draw and it could go for days if I prep to keep the loads low.
 
Many ways to skin this cat.

Easiest: get a grid interactive battery backup inverter, eg eg4 18kpv $4900, solark 15k $7000. Or wait for the new growatt sph 10000tl for about $2600
 
An 18kpv or Sol-Ark 15k would be massive overkill for a minimal critical loads subpanel. Would recommend performing at least a minimal power demand study of the proposed "critical" circuits to quantify your needs but based on the description something like a Schneider 4048 would be more than adequate. If net metering is your goal then a UL1741SB compliant inverter is required.

Alternatives: SMA, Outback, or Enphase IQ8 microinverter system with Enphase batteries.
 
An 18kpv or Sol-Ark 15k would be massive overkill for a minimal critical loads subpanel. Would recommend performing at least a minimal power demand study of the proposed "critical" circuits to quantify your needs but based on the description something like a Schneider 4048 would be more than adequate. If net metering is your goal then a UL1741SB compliant inverter is required.

Alternatives: SMA, Outback, or Enphase IQ8 microinverter system with Enphase batteries.
He said 50a at 240v
 
I originally did something similar. Have 9k of PV on the roof and a SolarEdge grid tie system with 1:1 net metering. Put up a Growatt SPF6000DVM because I needed 240 for the well pump, and 12 Kwh of batteries. Basically acted as a UPS with charging from the grid. Your more complicated and potentially expensive item is a sub panel to break out your critical loads.
 
Easiest: get a grid interactive battery backup inverter, eg eg4 18kpv $4900, solark 15k $7000. Or wait for the new growatt sph 10000tl for about $2600
Or Amensolar N3H-X10-US available now for $2200.

There are other grid interactive, 10 KW class, sub $3K, string inverters out there.

Add panels and battery to complete the system.

If you want fast backup (like a big UPS), then you have to find units with fast switchover times. If the inverter can pick up the critical loads in under, say, 15 ms, then your devices won't reboot or reset.

Mike C.
 
An 18kpv or Sol-Ark 15k would be massive overkill for a minimal critical loads subpanel. Would recommend performing at least a minimal power demand study of the proposed "critical" circuits to quantify your needs but based on the description something like a Schneider 4048 would be more than adequate. If net metering is your goal then a UL1741SB compliant inverter is required.

Alternatives: SMA, Outback, or Enphase IQ8 microinverter system with Enphase batteries.
Thanks. I do think the loads can go down on the critical loads panel. I need to get the minimal hydronic pumps sized and go forward from there.
 
I originally did something similar. Have 9k of PV on the roof and a SolarEdge grid tie system with 1:1 net metering. Put up a Growatt SPF6000DVM because I needed 240 for the well pump, and 12 Kwh of batteries. Basically acted as a UPS with charging from the grid. Your more complicated and potentially expensive item is a sub panel to break out your critical loads.
Fortunately this is new construction, so It will be wired from the start for that subpanel.
One thing about the All In Ones that bothered me was taking that critical loads grid power through the AIO in the EG4. I have read some about using a relay so if the AIO needs service, it is not sitting in the path for the grid power directly.
 
Or Amensolar N3H-X10-US available now for $2200.

There are other grid interactive, 10 KW class, sub $3K, string inverters out there.

Add panels and battery to complete the system.

If you want fast backup (like a big UPS), then you have to find units with fast switchover times. If the inverter can pick up the critical loads in under, say, 15 ms, then your devices won't reboot or reset.

Mike C.
There are several available in the 10kw class. I didn't mention them because I wouldn't consider them
 
One thing about the All In Ones that bothered me was taking that critical loads grid power through the AIO in the EG4. I have read some about using a relay so if the AIO needs service, it is not sitting in the path for the grid power directly.
You should have a bypass option so you can run the critical loads panel from grid directly while isolating the inverter.

A relatively simple way to do this is an interlock breaker in the critical panel. The grid feeds the main breaker which is usually off, the inverter feeds the inverter load breaker in the the top right column, usually on, and there is a metal slider that prevents both from being on at the same time.

For example:


1715781833519.png

If the inverter needs service or has failed, turn off the inverter load breaker, slide the interlock, turn on the main grid breaker. This puts the critical load panel directly onto the grid and now the inverter can be serviced while those loads are satisfied.

To be clear, the critical panel main breaker is not the "main" breaker, it is fed by another breaker in the main panel.

An alternative is to put in a two position transfer switch. These are fairly ugly and big, though.

Mike C.
 
Last edited:
Thanks that is a fairly elegant solution. Yes I know the critical loads gets a breaker out, and in your example the inverter into the main panel gets a separate breaker going in.
In the pass through scenario the same breaker in the main panel serves both purposes. That was the configuration I wanted to avoid.
 
I always prefer an interlock over a transfer switch. But it can be an issue if you also feed the grid through the inverter. It creates Parallel neutral paths.
Which means that you will have neutral current running through the inverter, even when bypassed. This possess a hazzard when working on the inverter.
 
Thanks Tim. These are the things I want to avoid. It seems like a seperate unit like the Schnieder has merit in that situation.
 
Thanks that is a fairly elegant solution. Yes I know the critical loads gets a breaker out, and in your example the inverter into the main panel gets a separate breaker going in.
In the pass through scenario the same breaker in the main panel serves both purposes. That was the configuration I wanted to avoid.
I don't understand what you're trying to avoid here.
 
In the pass through scenario the same breaker in the main panel serves both purposes. That was the configuration I wanted to avoid.
No, there are two breakers in the main panel.

One breaker feeds the critical panel "main" breaker. It is the grid only path.

Another breaker, usually at the end of the bus bar, serves the inverter grid input. It is the inverter path breaker.

My diagram:

1715783580338.png

In "inverter" mode:

Main panel critical panel breaker: ON or OFF (doesn't really matter)
Main panel inverter grid breaker: ON
Critical panel "main" breaker: OFF (interlocked)
Critical panel inverter load breaker: ON (interlocked)

In "grid only" mode:

Main panel critical panel breaker: ON
Main panel inverter grid breaker: OFF (to fully isolate)
Critical panel "main" breaker: ON (interlocked)
Critical panel inverter load breaker: OFF (interlocked)

All neutrals are tied together, all grounds are tied together. Only the main panel has ground neutral bond.

In "grid only" mode, the inverter is isolated from all AC connections and can be serviced safely.

Mike C.
 
I always prefer an interlock over a transfer switch. But it can be an issue if you also feed the grid through the inverter. It creates Parallel neutral paths.
Which means that you will have neutral current running through the inverter, even when bypassed. This possess a hazzard when working on the inverter.
How so?

If there is a second neutral path, like my diagram above, then disconnecting neutral at the inverter doesn't result in any meaningful voltage on the neutral conductors.

The inverter diagrams show neutral loops when wired up per their manuals. How else would you wire it?

I must not be understanding the issue, please elaborate.

Mike C.
 
How so?

If there is a second neutral path, like my diagram above, then disconnecting neutral at the inverter doesn't result in any meaningful voltage on the neutral conductors.

The inverter diagrams show neutral loops when wired up per their manuals. How else would you wire it?

I must not be understanding the issue, please elaborate.

Mike C.
There are two neutral paths in parallel.
1. From the main panel, through the inverter, to the sub panel.
2. From the main panel, directly to the sub panel.
These two paths will always be sharing the neutral current.
The correct way to do this is with a single neutral to each point, from a central connection.
Example:
A wire trough under the 3 pieces of equipment. (Main panel, sub panel, inverter)
A neutral terminal block in the trough. And a single neutral from the terminal block to each piece of equipment.
 
There are two neutral paths in parallel.
Each of which is capable of carrying the full load.

These two paths will always be sharing the neutral current.
Unless one becomes disconnected. Then the other carries all the load and is rated for it.

The correct way to do this is with a single neutral to each point, from a central connection.
I'm still unaware of the hazardous condition if this isn't the case.

The wiring diagrams in the manuals don't seem to require the wiring layout you propose.

For example, EG4 18KPV:

1715792395569.png

Neutral goes through the inverter and through the 2 pole transfer switch (thus neutral is not switched). Why is that wrong?

Mike C.
 
Each of which is capable of carrying the full load.


Unless one becomes disconnected. Then the other carries all the load and is rated for it.


I'm still unaware of the hazardous condition if this isn't the case.

The wiring diagrams in the manuals don't seem to require the wiring layout you propose.

For example, EG4 18KPV:

View attachment 215559

Neutral goes through the inverter and through the 2 pole transfer switch (thus neutral is not switched). Why is that wrong?

Mike C.
You have to consider the source of the information, before you take it as qualifying.
EG4 is still learning and growing.
And China is not a good source.
They don't know what they don't know.
Whenever you introduce a second source, there are always special considerations.
and parallel paths should always be avoided.
i explained why, above.
I can only suggest the safest way of doing it. But in the end, it's up to you to decide how to proceed.
 
Sorry, I see no explanation of why neutral loops are dangerous or violate code. Can you be a bit more explicit about this?

Mike C.
For code.
All current carrying conductors of a circuit must travel together.
The parallel paths travel in different directions.

For the hazzard.
If you lift one of the neutrals, while loaded. There is a possibility of a spark, while the current shifts to the remaining conductor. And if you become the completion of the circuit, you will be the current carrying conductor.

When you break a circuit in the middle. One side should be the source end , and the other should be the load end.
With a parallel conductor , both ends are potentially both.
 
He said 50a at 240v

He also said, "small", and "minimal"

:ROFLMAO:


I always prefer an interlock over a transfer switch. But it can be an issue if you also feed the grid through the inverter. It creates Parallel neutral paths.
Which means that you will have neutral current running through the inverter, even when bypassed. This possess a hazzard when working on the inverter.

For one, I have no concern whatsoever about parallel paths, if each is big enough for 100% of the current.

Inductance will favor closest path.

For code.
All current carrying conductors of a circuit must travel together.
The parallel paths travel in different directions.

With both paths present, it will have a path co-routed with other current carrying conductors.

For the hazzard.
If you lift one of the neutrals, while loaded. There is a possibility of a spark, while the current shifts to the remaining conductor. And if you become the completion of the circuit, you will be the current carrying conductor.

The spark would only be generated by the inductance of the wire. Due to any inductance, current will have already selected the neutral closest to L1 and L2.


Or you can route it the way I did; neutral doesn't go through the inverter.
My conduit from main panel to protected loads panel has a Tee off to the inverter. AC1 (inverter output) L1 & L2 go through the Tee, and AC2 (inverter input) go through the same Tee. A single N goes through the Tee.

If wires branch, or go to end of conduit, land on breaker panel, return through same conduit to exit other way through Tee, it works fine. (similar to running 2-wire Romex off to a switch; current goes out and back adjacent.)
 
For one, I have no concern whatsoever about parallel paths, if each is big enough for 100% of the current.

Inductance will favor closest path.



With both paths present, it will have a path co-routed with other current carrying conductors.



The spark would only be generated by the inductance of the wire. Due to any inductance, current will have already selected the neutral closest to L1 and L2.
A percentage of the current will be carried by both conductors. The percentage will be affected by resistance, inductance, and voltage drop.

Or you can route it the way I did; neutral doesn't go through the inverter.
My conduit from main panel to protected loads panel has a Tee off to the inverter. AC1 (inverter output) L1 & L2 go through the Tee, and AC2 (inverter input) go through the same Tee. A single N goes through the Tee.
This sounds similar to what I described?
 
This sounds similar to what I described?


The correct way to do this is with a single neutral to each point, from a central connection.
Example:
A wire trough under the 3 pieces of equipment. (Main panel, sub panel, inverter)
A neutral terminal block in the trough. And a single neutral from the terminal block to each piece of equipment.

Yup. Tee connection formed in wiring trough.
(I hadn't read all the words and formed a picture in my mind. When I provide pictures to clarify my writing, it is usually an image stolen from the web.)

An upside of trough vs. conduit Tee like I used is much greater cross section for additional runs. And ability to tap off more conduit sections.

See, I'm learning from you guys on this forum!
I now have a 24" trough under my house where four electrical panels connect, and will put a 36" one between my Sunny Islands and the garage loads panel.
I also used a 3R trough above PV combiner panel at my old place.
 
For code.
Can you give the NEC code reference number for this requirement?

All current carrying conductors of a circuit must travel together.
Then a common neutral point violates this requirement. At the common neutral point, neutral currents will combine but the line1 and line2 wires won't. So the current doesn't travel together any more. The neutral wire from the common point to the main panel is a combination of multiple different circuits now. To say it another way, if you lift the neutral from the main panel to the common neutral bar, you are interrupting multiple circuits.

Tell me how to modify my drawing to achieve what you are asking for because I see no possible way to do it.

1715803825798.png

If you lift one of the neutrals, while loaded. There is a possibility of a spark
The voltage differential is millivolts due to the low resistance of the alternate neutral path, thus no spark, no hazard.

And if you become the completion of the circuit, you will be the current carrying conductor.
The other neutral path is vastly less resistance than you will ever be, thus no hazard.

When you break a circuit in the middle. One side should be the source end , and the other should be the load end.
Grid interactive inverters can do both at the same time. They can sink and source power at any given moment so the current can flow in either direction.

Mike C.
 
Can you give the NEC code reference number for this requirement?
300.3 (b)Screenshot_20240515_162907_Chrome.jpg


Then a common neutral point violates this requirement. At the common neutral point, neutral currents will combine but the line1 and line2 wires won't. So the current doesn't travel together any more. The neutral wire from the common point to the main panel is a combination of multiple different circuits now. To say it another way, if you lift the neutral from the main panel to the common neutral bar, you are interrupting multiple circuits.
Lifting any neutral only interrupts the circuit (feeder) it's running with.
Tell me how to modify my drawing to achieve what you are asking for because I see no possible way to do it.
Draw it as I described above.
The voltage differential is millivolts due to the low resistance of the alternate neutral path, thus no spark, no hazard.
And if the other neutral has already lost connection (no way of knowing), you are going to be in for a surprise.
The other neutral path is vastly less resistance than you will ever be, thus no hazard.
That's assuming that it's in play.
Grid interactive inverters can do both at the same time. They can sink and source power at any given moment so the current can flow in either direction.
Current still only flows in one direction. (One or the other at any time)
And when the connection is lost, grid-tied inverters are designed (required) to disconnect.
 
Draw it as I described above.
You can't.

You can vaguely say how to do it, but when you actually try, you find out it can't be done like you think it can be done.

Any place you put this common neutral bar necessarily means some circuit is sharing a neutral wire with another circuit.

The only way I can see it would actually work is to put in a 3 pole transfer switch and switch the 2 lines *and* the neutral to the backup panel so that neutral either follows the main panel breaker line or the inverter load line. This of course means the interlock breaker concept can't comply with your interpretation of the code since the power comes from two different places but your aren't switching the neutral to align with that.

Not trying to argumentative here, but I've yet to see any way you can make the neutral behave the way you think it should and use a breaker interlock for transfer to a backup panel.

Mike C.
 
You can't.

You can vaguely say how to do it, but when you actually try, you find out it can't be done like you think it can be done.
I have installed many systems in real life, this way.

Adjust your drawing so that the wiring is ran across the bottom of everything.
Like a ladder diagram (if that's familiar).
The neutral will be drawn from the left most piece of equipment to the right most piece of equipment. And a single tap to the center piece of equipment.
The Line conductors will travel as you would expect. (Without taps)
 
Adjust your drawing so that the wiring is ran across the bottom of everything.
The neutral will be drawn from the left most piece of equipment to the right most piece of equipment. And a single tap to the center piece of equipment.
The Line conductors will travel as you would expect. (Without taps)
Like this?

1715808575100.png

The main to backup, the main to inverter, and the inverter to backup lines all travel in different conduits. When the inverter supplies power to the backup panel, the neutral currents now go along the grid feed, to the main panel, and then to the backup panel. This violates your rule.

I await a more precise wiring explanation of how neutrals should be wired. There are four neutral points to be wired: main panel, backup panel, inverter grid, and inverter load. I see no way to wire them such that neutral ONLY carries the current of its circuit with no loops. A backup panel that can be powered from two sources necessarily introduces a neutral loop if you don't switch the neutral. I don't see a way around that.

Like I said, the only solution that meets your objective is a 3 pole transfer switch which negates a breaker interlock system.

Mike C.
 
That wire you have running between "Main" 60A breaker and "Inverter" Grid terminal?
Just run it through same conduit as "Inverter" Load1 terminal, through Sub-panel, back to main panel. Then down inside main panel to "Main" 60A breaker.

Two red, two black, one white, one green. White and green can land on busbars in each panel.

(or if not in conduit, in a trough as Tim says).
 
Like I said, the only solution that meets your objective is a 3 pole transfer switch which negates a breaker interlock system.
Don't break the neutral, that would turn it into separately derived systems.
Which just adds more complications.
 
Two red, two black, one white, one green. White and green can land on busbars in each panel.
I wasn't aware you could share neutral between two circuits.

When the inverter is running, the grid input and the load output are using the same neutral wire in the conduit. Is that okay?

Mike C.
 
That is something to consider.
As drawn, Sub panel either is powered by main panel, otherwise by inverter. Doesn't seem overloaded.

If Load2 was also connected, then both sub panel and whatever is on load 2 would pull current through it, so size accordingly.

I think you could run a second neutral through same conduit, simply flying by sub panel without connecting.
That would still work when inverter supplies sub panel; path would be back to main, then turn around and come through conduit a second time to reach sub. That's about what I've done.
 
Yes
Input and output are not using it at the same time.
Well, they seem to be. To illustrate, I've drawn what I think you guys said:

1715811103631.png

One conduit goes from man to backup, carries one neutral, one ground, and two sets of hots (backup panel or inverter grid). Technically, both could be used at the same time (backup on grid power only, inverter charging from grid or pushing back PV).

Another conduit goes from backup to inverter, carries one neutral, one ground, and two sets of hots (inverter load, inverter grid). These will generally be in use at the same time for normal operation (backup on inverter load, inverter grid either sucking power during dark or pushing PV back to grid during day).

This doesn't feel right to me, but let me know if it is, or if I failed to capture the directive.

Mike C.
 
Better size the neutral for 120A, between main and sub panel.
I don't know if NEC would have any issues, but that works electrically.

I would draw the panel boxes bigger and route the lines representing wires through them, to make it clear they never separate from N and G.
 
One conduit goes from man to backup, carries one neutral, one ground, and two sets of hots (backup panel or inverter grid). Technically, both could be used at the same time (backup on grid power only, inverter charging from grid or pushing back PV).
Backup on grid and charging, yes.
Not pushing back and charging. (This would be either/ or)
Another conduit goes from backup to inverter, carries one neutral, one ground, and two sets of hots (inverter load, inverter grid). These will generally be in use at the same time for normal operation (backup on inverter load, inverter grid either sucking power during dark or pushing PV back to grid during day).
This is also either /or.
(Referring to neutral currents only)

At any given time, current only flows in one direction on a conductor.
(From the higher voltage side to the lower voltage side)
 
Back
Top