Grid tie system with small critical loads battery backup

mciholas said:

The single neutral line to the inverter is carrying the neutral currents from the GRID input/output (varies) to the inverter *AND* the LOAD output.

Click to expand...

Either/ or
Not and
The inverter can't be a source and load, at the same time.

mciholas said:

The single neutral between main and backup panels is carrying the neutral currents from both the inverter GRID input/output (varies) *AND* the backup panel if it is in grid mode.

Click to expand...

Correct, but it's still one circuit from one source.

mciholas said:

NEC 200.4(A):

Neutral conductors shall not be used for more than one branch circuit, for more than one multiwire branch circuit, or for more than one set of ungrounded feeder conductors unless specifically permitted elsewhere in this Code.

Click to expand...

It's not two circuits. It's two optional sources for one circuit.
The interlocks keep them from being used simultaneously.

mciholas said:

No, my inverter has a neutral connection for the GRID and the LOAD that are separate.

Click to expand...

Most of the newer hybrid AIO's have a single neutral bar.
Your should be internally common. (Please verify)
You have two options.
1. Only use the input neutral terminal. (Assuming that it's the larger one)
2. Run both to the terminal block in the trough.
(#1 would be my choice)

mciholas said:

Again, simply tell me how you wire those four points together and not combine neutral current from two circuits and without creating a neutral loop? I can't see anyway to do this.

Click to expand...

See above

mciholas said:

The simplest would be to send every neutral to the backup panel neutral bar.

Click to expand...

I wouldn't do that. Because you can't isolate the backup panel for maintenance.
Without shutting down the inverter. And there may be times when that would not be desirable.

mciholas said:

The separate inverter neutrals are basically tied together at both ends, so that's electrically one wire so they will share current and not be segregated.

Click to expand...

Again, this would be parallel conductors. (Not recommend)
And requires keeping the backup panel neutral bar active, when maintenance is done. (Unless you also shut down the inverter)
A warning label would be needed.

mciholas said:

I am going to be wiring this up in a few days, so I am keen to figure this out.

Click to expand...

Hopefully this cleared up some questions.
I would definitely use a trough. It makes everything simpler and looks much cleaner.

The electrons in L1, L2, N, and (in the event of a fault) G all want to be in the same box. They know if the others aren't with them.
(Trust me in this - I See Electromagnetic Fields!)

timselectric said:

Either/ or
Not and
The inverter can't be a source and load, at the same time.

Correct, but it's still one circuit from one source.

It's not two circuits. It's two optional sources for one circuit.
The interlocks keep them from being used simultaneously.

Click to expand...

With interlock of sub-panel set to "grid", it can put 60A in neutral.

The inverter has 2 load ports, second one could be set to supply a load, could also be single-phase and put another 60A in neutral, 120A total.

(Of course, inverter could charge its batteries and draw 60A, that ought to be in L1/L2 only, zero in neutral. But it is a Black Box, don't think we're allowed to assume zero neutral current.)

So I think either size neutral wire for 120A up to where it branches (possibly at neutral bar in sub-panel), or run separate neutral from Main to Sub and from Main to Inverter.

Mine is separate because inverter is connected through separate conduit branch. I'll run inverter grid and load connections through a single conduit, Red, Black, White, Green, Black Red.

Hedges said:

With interlock of sub-panel set to "grid", it can put 60A in neutral.

Click to expand...

This would also require the interlock in the main panel be set to bypass.

Hedges said:

The inverter has 2 load ports, second one could be set to supply a load, could also be single-phase and put another 60A in neutral, 120A total.

Click to expand...

The breaker feeding the circuit will also limit the neutral current.
You can't have more current on the neutral, than what is on an individual leg.

You two guys are so freaking smart about all this stuff it makes me realize just how far I have to go.

Years of experince I guess.

I did a bunch of different things in the Navy 30 years ago and do things around the house, but for the last 30 I have been riding a desk working at a computer otherwise.

timselectric said:

This would also require the interlock in the main panel be set to bypass.

The breaker feeding the circuit will also limit the neutral current.
You can't have more current on the neutral, than what is on an individual leg.

Click to expand...

I don't think so. Main panel doesn't have an interlock.
It has 60A breaker feeding sub-panel, and 60A breaker feeding inverter.

If inverter gets load attached to Load2 port as well, that could carry 60A bypass on L1, and sub-panel could carry 60A on L1.
That's 120A on L1, 120A on N.

Hedges said:

Main panel doesn't have an interlock.

Click to expand...

It should, in this situation.

Hedges said:

If inverter gets load attached to Load2 port as well,

Click to expand...

There has been no mention of using the load 2 port, up to this point.

But, that would change everything.
You could upsize the neutral for the first section. But I'm not sure if it would get passed an inspection.
Electrically it would be fine. And wouldn't be confusing to anyone else working on it.
I usually try to follow industry standard practices. Because you never know who might be working on it next. Even if that someone is me, I don't want to have to remember something odd. I get enough of that, following others. lol

There should not be any reason to not have an interlock in the main panel.
The only reason for switching to bypass is for maintenance on the inverter.
I don't see any reason to ever want both on.

main panel interlock? please explain that @timselectric , how does my installation require one of those? normal operations for me are one breaker pair does the grid tie, and the other feeds the critical subpanel.
EDIT: Your post above cleared that up, thank you.
What I keep seeing is, this critical load panel is not an automatic fail over the way people keep drawing it. It has an interlock breaker set that requires human input.
I know this is a nice discussion but it has gone in the weeds for me. It is no longer on topic. My topic.
I don't want too much.
I want to be out of town and not have the food spoil, and if I am home I want 3 circuits of light switches to work.
I want to provide power to a 1/3 HP hydronic pump.

timselectric said:

The inverter can't be a source and load, at the same time.

Click to expand...

The inverter can be a load to the grid and a source to the backup panel at the same time.

This is such a basic thing that it is freaking me out you aren't seeing that.

Maybe you are doing some magic semantic handwave that a device with TWO 240 VAC connections is somehow ONE circuit. I just don't see any justification for that interpretation.

There are times the grid port will load more power than the load port puts out (say grid charging battery). There will be times the grid port is loaded less than the load port (say some PV or battery but not enough to power all the backup, deficit from grid). And hopefully, commonly, the grid port will source power while the load port also sources power (excess PV input). The grid and load connections can go basically all over the place power wise.

The inverter can and *must* be able to operate as a source and a load at the same time. This is fundamental to its operation.

timselectric said:

Correct, but it's still one circuit from one source.

Click to expand...

No, it is two, grid and load.

timselectric said:

It's not two circuits. It's two optional sources for one circuit.
The interlocks keep them from being used simultaneously.

Click to expand...

There is no interlock on grid and load to the inverter. They both operate at the same time.

There is no interlock that prevents the inverter from using the grid connection while the backup panel uses the grid.

Thus the shared neutral in those connections crosses two circuits.

The only interlock is the backup panel can be powered from the main panel or from the inverter load output. That is the only interlock in the system.

timselectric said:

Most of the newer hybrid AIO's have a single neutral bar.

Your should be internally common. (Please verify)

Click to expand...

It probably is, would be odd to be otherwise.

timselectric said:

You have two options.
1. Only use the input neutral terminal. (Assuming that it's the larger one)
2. Run both to the terminal block in the trough.
(#1 would be my choice)

Click to expand...

#1 violates the wiring diagrams in the inverter manual. They always show the load neutral connection being used. Not following the manual violates the certifications.

#2 creates a neutral which has a loop. The loop is closed at the inverter with the grid neutral and load neutral are assumed to be internally connected.

timselectric said:

Hopefully this cleared up some questions.

Click to expand...

Not really. I'm far more confused now than when we started because I don't actually understand what you think is the right connections and/or what you are telling me violates code, the manual, or both.

Mike C.

Only reason for interlock on main panel is if you want to feed one of its loads from the inverter, during a grid failure.
e.g. operate the garage door, finish the load of washing that is all soapy, before hanging clothes out to dry.

Your interlock on sub-panel can be a bypass, direct to grid, for inverter repair. Normally fed by inverter so you have UPS for critical loads.
If you never connect anything to Load2, neutral won't be overloaded. Inverter will charge with L1/L2, zero current on N. I used two separate N wires, so they won't be overloaded regardless.

My concern with relying on inverter is there are failure modes that can leave it shut down. In some ways, grid is a more reliable supply for freezer.
I would like a deadman switch that falls back to grid if inverter output is off.

timselectric said:

This would also require the interlock in the main panel be set to bypass.

Click to expand...

There is no interlock in the main panel.

If your backup panel has all loads on one phase, then the neutral will have 60 amps in it.

timselectric said:

The breaker feeding the circuit will also limit the neutral current.

Click to expand...

There are two breakers feeding these combined neutrals.

The neutral from main to backup can have both the backup panel loads (in grid mode) and the inverter acting either as load (battery charging) or source (PV generation). In unbalanced ways, this can also create 60 amps.

You could get 120 amps in the neutral with the common tie point shared neutral concept.

timselectric said:

You can't have more current on the neutral, than what is on an individual leg.

Click to expand...

You can because the neutral is shared among TWO circuits now. The SUM of those breakers is in play.

Mike C.

Hedges said:

Only reason for interlock on main panel is if you want to feed one of its loads from the inverter, during a grid failure.

Click to expand...

Correct.

An illegal way to do this is to:

1. Turn off main panel main breaker. this disconnects you from the utlity.
2. Turn off every breaker in the main panel you do not want to power.
3. Turn off the main panel inverter breaker.
4. Turn on the backup panel grid breaker.
5. Defeat the backup panel interlock and turn on both the backup panel main breaker and the inverter load breaker.

A possible legal way to do this is to make another link between backup and main panel. This requires a source breaker in the backup panel, a back feed breaker in the main panel, and an interlock between main panel main breaker and the backup panel back feed breaker. This back feed breaker would have to be in the top right breaker slot due to the interlock requirements.

Hedges said:

Your interlock on sub-panel can be a bypass, direct to grid, for inverter repair. Normally fed by inverter so you have UPS for critical loads.

Click to expand...

That is what it is, a way to put the backup panel on grid in case of inverter failure or inverter service. If you didn't have this bypass, then the circuit becomes easier but you have lost redundancy.

Hedges said:

If you never connect anything to Load2, neutral won't be overloaded. Inverter will charge with L1/L2, zero current on N. I used two separate N wires, so they won't be overloaded regardless.

Click to expand...

I'm not sure we can assume balance on the inverter demands on the grid input. Seems logical, but it is still an assumption.

Hedges said:

My concern with relying on inverter is there are failure modes that can leave it shut down. In some ways, grid is a more reliable supply for freezer.
I would like a deadman switch that falls back to grid if inverter output is off.

Click to expand...

I thought about some sort of relay that drops out if the inverter load output falters. The relay would transfer backup panel to grid if/when the inverter load output fails.

A downside is that the relay will take power to energize nearly all the time. It would remove the interlocks on the backup panel and would make fail over automatic.

It has the same neutral question as the interlocks, though.

Mike C.

Hedges said:

My concern with relying on inverter is there are failure modes that can leave it shut down. In some ways, grid is a more reliable supply for freezer.
I would like a deadman switch that falls back to grid if inverter output is off.

Click to expand...

In my scenario I want grid power that fails over in a grid outage. I don't want to run on inverter and batteries as the normal mode of operation. I am thinking this is best handled with a relay to the subpanel with a delay for re-making to grid.
I do not care about UPS-like fail over.
If the power goes out, I am happy if 5 seconds later the fridge and freezer get some clean power.
When the power comes back on, often there is a little drama so I would want to delay that return to service.
If I have some electronics that want to be protected by a UPS I will have them on an individual UPS.

Can I get this device in a UL version?



Put this before the backup panel and have it automatically switch from grid or inverter. You can switch it to manual to force the choice, or it can be set to automatically switch on source failure. Rates 100 amps. Uses a servo motor, so not fast

Basic operation of a similar device:


The Europeans have so much nicer stuff sometimes.

Mike C.

skyking1 said:

What I keep seeing is, this critical load panel is not an automatic fail over the way people keep drawing it.

Click to expand...

It's automatic through the AIO.
The interlock is for a bypass for maintenance. (Which is always recommended)

timselectric said:

It's automatic through the AIO.

Click to expand...

Not if the AIO fails.

The ability to pass grid through the AIO depends on the AIO functioning to at least some small degree.

Mike C.

OK. An AIO string inverter seems to be cost effective when compared to microinverters and a piece of backup hardware.
The load paths from grid to critical load panel is linear and goes through the AIO.
If the grid fails the AIO starts up and supplies only the critical load panel.
If I build the panels up on a gutter like this,
Image from reddit.
my main power is coming up from basement slab , on right tagged with some red.
AIO is green,
Subpanel is blue on left.
In the gutter I run the single neutral path and the ground path through there and tee off of them where I tagged it with Tan, using a bus bar for ground and a Polaris for neutral.
This ensures no neutral loop condition.



In the inside premise wiring, I propose using a different color of wire jacket for the critical loads circuits. Say white for regular and yellow for the Critical Loads (CR)
Where the CR lighting circuit occupies the same duplex switch plate as another circuit, use a common plate but a divided box so there is no possibility of looping there.
I say this because I just want those main navigation lights to work on backup, not every light circuit in the house. This helps with the " load to be served" rules.
If I try to serve them all off CR, the thing gets bloated for no need.

@timselectric does this look like a good plan to pass inspection without a lot of grief?
My big concern is the regulatory certification for the AIO. I don't want to put in a functional piece that does not pass all the requirements of the 2020 NEC.
I am in Washington state and they are at NEC 2020 until sometime in 2025 TBD.
I will do everything possible to get this done in 2024

mciholas said:

The inverter can be a load to the grid and a source to the backup panel at the same time.

Click to expand...

Not the neutral.
Any loads in the backup panel will be fed by either the grid or the inverter. (Not both)

mciholas said:

No, it is two, grid and load.

Click to expand...

Not when referring to the neutral.

mciholas said:

There is no interlock on grid and load to the inverter. They both operate at the same time.

Click to expand...

Correct

mciholas said:

There is no interlock that prevents the inverter from using the grid connection while the backup panel uses the grid.

Click to expand...

Correct, through the inverter.
But there should be an interlock at the main panel. So that inverter and bypass are not both on.

mciholas said:

The only interlock is the backup panel can be powered from the main panel or from the inverter load output. That is the only interlock in the system.

Click to expand...

I thought that you mentioned an interlock in the main panel.
But I see that you just detailed the operation as the same.

>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)<

mciholas said:

The inverter can and *must* be able to operate as a source and a load at the same time. This is fundamental to its operation.

Click to expand...

Again, not in regards to the neutral.
It's only a return path for whichever the source of the load is, at the time.

mciholas said:

#1 violates the wiring diagrams in the inverter manual. They always show the load neutral connection being used. Not following the manual violates the certifications.

Click to expand...

I disagree

mciholas said:

#2 creates a neutral which has a loop. The loop is closed at the inverter with the grid neutral and load neutral are assumed to be internally connected.

Click to expand...

I agree
But you seemed bound to running parallel paths. I gave you one that wouldn't hurt anyone.

mciholas said:

Not really. I'm far more confused now than when we started because I don't actually understand what you think is the right connections and/or what you are telling me violates code, the manual, or both.

Click to expand...

The manuals wiring diagrams were fine. Until they tried to add the bypass switch. This is where they introduced the multiple neutral paths.

I'm sorry that I can't explain it in a way that you would understand.
I'm running out of ways to say it.
I feel like we are just repeating ourselves, back and forth.
You really have to understand how current flows. Especially regarding the neutral.
I guess, that you should just do what you feel comfortable with.
And hopefully the parallel neutral path will never be a problem.

Some AIO may have NC DPST relay to grid and would pass it through if inverter is down.
My Sunny Island has NO DPST relay. It has to operate for 5 minutes monitoring grid before connecting.

A DPDT relay would have the functionality desired, keep loads fed through AIO/UPS so long as it is operating, fall back to grid if AIO fails. However, I don't like DT relay to switch between two sources. I've had relays draw an arc and short the two inputs. I want two separate interlocked relays, one opening for a few seconds before allowing other to close.

Or, a big DPDT knife type transfer switch, spring loaded to return to grid when solenoid or motor is unpowered.

Relays do draw current while holding. But pull-in current requirement is much higher than what is needed to hold.
Relays (and drivers) are available with economizer function for DC coil, reduce to a couple percent as much power for hold
AC coils inherently have economizer functionality (but I haven't measured). When pulled in, core contacts to complete magnetic circuit, and inductance becomes much higher so current draw lower. Need to put a scope current probe on it to test.

Loops? I don't care about no stinkin' loops!
Current will take the path of lowest impedance, Z = sqrt(R^2 + X^2) = |<complex impedance R + jX>|
So long as each path has sufficient ampacity, I ain't worried.

skyking1 said:

OK. An AIO string inverter seems to be cost effective when compared to microinverters and a piece of backup hardware.
The load paths from grid to critical load panel is linear and goes through the AIO.
If the grid fails the AIO starts up and supplies only the critical load panel.
If I build the panels up on a gutter like this,
Image from reddit.
my main power is coming up from basement slab , on right tagged with some red.
AIO is green,
Subpanel is blue on left.
In the gutter I run the single neutral path and the ground path through there and tee off of them where I tagged it with Tan, using a bus bar for ground and a Polaris for neutral.
This ensures no neutral loop condition.

View attachment 215801

In the inside premise wiring, I propose using a different color of wire jacket for the critical loads circuits. Say white for regular and yellow for the Critical Loads (CR)
Where the CR lighting circuit occupies the same duplex switch plate as another circuit, use a common plate but a divided box so there is no possibility of looping there.
I say this because I just want those main navigation lights to work on backup, not every light circuit in the house. This helps with the " load to be served" rules.
If I try to serve them all off CR, the thing gets bloated for no need.

@timselectric does this look like a good plan to pass inspection without a lot of grief?
My big concern is the regulatory certification for the AIO. I don't want to put in a functional piece that does not pass all the requirements of the 2020 NEC.
I am in Washington state and they are at NEC 2020 until sometime in 2025 TBD.
I will do everything possible to get this done in 2024

Click to expand...

If I understand correctly. You won't be installing any bypass.
Your system will just be grid (main panel) to AIO to Critical loads panel.
If this is correct, you don't need to do anything special.
The wiring is linear and simple.
I'm sorry if the discussion drifted away from your needs.

No worries. I will install that bypass breaker in the critical loads, and serve it with a breaker from grid, so I can yank the AIO out of there with minimal drama. None of that makes loops if done right.
Clicky clicky and then de-energize the AIO and send if off or even change it out for the latest and greatest, etc.
Thank you @Hedges for this imporant point!

Some AIO may have NC DPST relay to grid and would pass it through if inverter is down.

Click to expand...

That is a must-have criteria. It needs to just send grid. Does anybody have any recommendations for that?

That is exactly why I was proposing a simple relay.

I think this was part of the discussion and may have gotten lost.
There will be a neutral wire from the main panel to the subpanel's neutral bus bar.
There will be a ground wire from the main panel to the subpanel's ground bus bar.
The neutral and ground bus bars in the subpanel are separate bus bars and are not bonded or connected.

How is the neutral from the AOI which is bonded to the neutral bus bar in the subpanel handled when the AOI is pulling power from the grid or from the batteries/array.

Is there a possibility of the subpanel having two neutral paths back to the main, the first hardwired and the second through the AOI

Hedges said:

Some AIO may have NC DPST relay to grid and would pass it through if inverter is down.

Click to expand...

This only works if the "inverter is down" detection logic works. If the invert fails such that it still energizes that relay, your backup panel is dark and the relay won't drop to grid.

Ideally this means this relay is powered by the inverter AC output directly so there is no firmware or logic in the decision, but this has negative consequences in that it may chatter if the inverter output is unstable (blips on and off). Thus you usually have some firmware that observes the voltages a while before switching, and thus this fail over relay can fail in firmware.

Hedges said:

A DPDT relay would have the functionality desired, keep loads fed through AIO/UPS so long as it is operating, fall back to grid if AIO fails. However, I don't like DT relay to switch between two sources. I've had relays draw an arc and short the two inputs. I want two separate interlocked relays, one opening for a few seconds before allowing other to close.

Or, a big DPDT knife type transfer switch, spring loaded to return to grid when solenoid or motor is unpowered.

Click to expand...

That is what is nice about the DIN ATS I pictured above. It has a motor to drive the switch contacts to either input. Thus no power usage during operation (big relays are power hogs). And this also means it can have beefy contacts since the motor can drive them easily. It also provides manual control, too.

I really wish I could find something like that in a UL certified way. That would make things way better since it would add automatic redundancy, remove the breaker interlock on the backup panel, and free up 2 breaker slots in the backup panel.

Hedges said:

Loops? I don't care about no stinkin' loops!

Click to expand...

Me, neither, this is lot of ado about nothing that has no meaningful safety or functional impact.

Mike C.

timselectric said:

Not the neutral.
Any loads in the backup panel will be fed by either the grid or the inverter. (Not both)

Click to expand...

This is getting ridiculous.

Let's assume what I think you are advocating, a single point neutral bar with 3 connections: main panel, backup panel, inverter (combined grid and load neutrals into one wire).

The neutral wire to the backup panel is only the loads on the backup panel, No problem there.

The neutral wire from neutral bar to the main panel is potentially the combination of the inverter grid neutral current *AND* the backup panel neutral current (if the backup panel is in grid mode). TWO CIRCUITS, ONE NEUTRAL.

The neutral wire from neutral bar to the inverter is potentially the combination of the inverter grid neutral *AND* the inverter load neutral currents (if the backup panel is in inverter mode). TWO CIRCUITS, ONE NEUTRAL.

How is it that you don't see that?

#1 violates the wiring diagrams in the inverter manual. They always show the load neutral connection being used. Not following the manual violates the certifications.

Click to expand...

I disagree

Click to expand...

Not connecting to the inverter load neutral is an obvious and clear violation of the manual wiring diagrams. The UL certification is based on installing the inverter per the manual.

And hopefully the parallel neutral path will never be a problem.

Click to expand...

There is no mechanism for it to be a problem.

I see all sorts of possible problems with a shared neutral for two circuits, however. Or not connecting the neutral on the inverter load.

Mike C.

mciholas said:

Me, neither, this is lot of ado about nothing that has no meaningful safety or functional impact.

Click to expand...

I gave you the code reference.
And safety is the reason.
I'm sorry that you can't grasp this.