Using a three-phase breaker box "L3" as a bus bar for off-grid AIO leg

[hex4def6]

I'm looking to use an AIO battery / PV inverter to power a subset of circuits in my home.

The "easy" way I can think of doing this is adding a breaker to the Main panel (Grid input to AIO), and then having the AIO output feed the sub panel.
This should work, but seems a bit inflexible; I want to replace the main panel anyway, and in the future if I want to move a circuit from the main to sub panel (or vice versa), it will be finicky.

I next thought was using a split-bus panel, but that seems to have gone out of fashion a long time ago.

My final thought was this:
What if I replace my split-phase panel with a 3-phase panel? Now, instead of two hots and a neutral, I'll have 3 hots + neutral. L1 / L2 will be wired to two of the bus bars as usual. The third bus bar will be the "backup" rail, and will be fed from the AIO.

EDIT: Arggg... lost the fancy image I made. Here's the low res cached one....



The benefit of doing this is that I can easily shift loads to the backup rail, simply by moving breaker position.

If I have a separate "critical loads" panel, it's a pain if I want to shift e.g, the garage fridge circuit, since I'd have to move the Romex physically between two panels. In this system, if I want to shift it to being backed up by the AIO, I just move the position so that it's on the "L3" busbar.

My AIO I've got is 5kW, so I'd still have to be careful about which loads I shift, but this makes it a lot more flexible and something you can do pretty easily.

How crazy is this, code-wise?

 

[rodrick]

If you are using 240 volt split how can that work with a 3 phase panel you would only have one leg on backup power

 

[hex4def6]

If you are using 240 volt split how can that work with a 3 phase panel you would only have one leg on backup power

Sure.

But the idea would be breaker in
position 1 = L1 + N
Position 2 = L2 + N
Position 3 = AIO + N

Obviously, this wouldn't work for 240V loads, only 120V.

 

[rodrick]

Do you have 3 power sources

 

[hex4def6]

My AIO is

Do you have 3 power sources

I have a regular split-phase supply. L1 and L2.
I have an AIO which generates 120V. That would go on "L3".

The idea being, you move the breaker position, it changes which phase its connecting to. Connected to "L3" makes it connect to the AIO.

 

[rodrick]

Here’s a picture of my switch from 2 different inverter banks easy diy safety bar

 

[timselectric]

How crazy is this, code-wise?

Not crazy.
You will have to label the panel, identifying that it has more than one power source. How it's configured, voltages, and the location of the disconnects for each source.

 

[timselectric]

The thread title is confusing. Because this is not about a "selector".

 

[timselectric]

Here’s a picture of my switch from 2 different inverter banks easy diy safety bar

It looks like there's only a single bolt on that slide bar. And that the bar can pivot, which wouldn't block someone from turning all breakers on at the same time.

 

[wpns]

I'm looking to use an AIO battery / PV inverter to power a subset of circuits in my home.

The "easy" way I can think of doing this is adding a breaker to the Main panel (Grid input to AIO), and then having the AIO output feed the sub panel.
This should work, but seems a bit inflexible; I want to replace the main panel anyway, and in the future if I want to move a circuit from the main to sub panel (or vice versa), it will be finicky.

I next thought was using a split-bus panel, but that seems to have gone out of fashion a long time ago.

My final thought was this:
What if I replace my split-phase panel with a 3-phase panel? Now, instead of two hots and a neutral, I'll have 3 hots + neutral. L1 / L2 will be wired to two of the bus bars as usual. The third bus bar will be the "backup" rail, and will be fed from the AIO.

EDIT: Arggg... lost the fancy image I made. Here's the low res cached one....
View attachment 272713


The benefit of doing this is that I can easily shift loads to the backup rail, simply by moving breaker position.

How crazy is this, code-wise?

Not crazy function, wise, I see what you’re doing and I get how you got there, but even with labeling you just know that someone in the future is gonna put a 240 V load across L3 and one of the other lines, and then Nobody knows what’s going to happen.

 

[mciholas]

What happens when a 2 pole breaker straddles AIO and either L1/L2 bus bars?

Not good, I imagine.

Your inverter can only be 120 VAC output, which limits it severely.

Does the panel have a 3 phase main breaker? So the inverter is tied to one input on that breaker? This means shutting off that breaker also disconnects the inverter. Your pic just shows main lugs, no main breaker.

If you are okay with manual reconfig and a brief power outage when grid fails, then a transfer switch before the panel allows selecting the grid or the inverter. If grid goes down, you manually drop the branch circuits you don't need and then flip the transfer switch to the inverter. You get to choose which branches you power.

If you want this more automatic, then a backup panel on the inverter output is the way to go, but relatively inflexible for choosing what gets powered. This has the advantage of being automatic and potentially so quick devices don't reboot.

Mike C.

 

[SparkyJJO]

Creative for sure. Electrically would work. But as mentioned the risk is there that a 2 pole breaker be mispositioned and cause trouble. Not sure if the 3rd "phase" rows could be clearly marked somehow when the cover is off to prevent accidents (even by yourself, you know it would be easy to miscount and do that). Maybe the plastic cover behind the 3rd "phase" tabs could be painted a bright color somehow to make it obvious which is which if you wanted to proceed with this idea.

 

[rodrick]

The thread title is confusing. Because this is not about a "selector".

Lost me are we physically moving breakers during a outage to L3 is that the idea

 

[SparkyJJO]

Lost me are we physically moving breakers during a outage to L3 is that the idea

No the L3 is powered by the inverter (which gets its grid power from L1 or L2). The circuits on L3 are the ones that stay powered during an outage.

 

[rodrick]

It looks like there's only a single bolt on that slide bar. And that the bar can pivot, which wouldn't block someone from turning all breakers on at the same time.

When I had double pole breakers it was impossible now with single I’ll have to try if so mods coming thanks

 

[rodrick]

No the L3 is powered by the inverter (which gets its grid power from L1 or L2). The circuits on L3 are the ones that stay powered during an outage.

Ok light comes on it basically a built in critical loads panel that’s normally feed by grid passing through inverter ats that switches during a power outage to inverter power?

 

[SparkyJJO]

Ok light comes on it basically a built in critical loads panel that’s normally feed by grid passing through inverter ats that switches during a power outage to inverter power?

Bingo.

I suspect it would not pass an inspection but if you don't need one then whatever.

 

[wpns]

No the L3 is powered by the inverter (which gets its grid power from L1 or L2). The circuits on L3 are the ones that stay powered during an outage.

Oh wow, I just realized it’s worse than that. When the grid is up and the inverter is in sync with the grid, L1/L3 or L2/L3 might work for 240 V loads, but when the grid goes out, you will lose the sync and then when the grid comes back, you’ll be out of phase, and who knows what the load will do with that?

 

[SparkyJJO]

Oh wow, I just realized it’s worse than that. When the grid is up and the inverter is in sync with the grid, L1/L3 or L2/L3 might work for 240 V loads, but when the grid goes out, you will lose the sync and then when the grid comes back, you’ll be out of phase, and who knows what the load will do with that?

 

[rodrick]

Maybe this needs to be moved to the experimental thread

 

[hex4def6]

Oh wow, I just realized it’s worse than that. When the grid is up and the inverter is in sync with the grid, L1/L3 or L2/L3 might work for 240 V loads, but when the grid goes out, you will lose the sync and then when the grid comes back, you’ll be out of phase, and who knows what the load will do with that?

That is an interesting thought that I hadn't considered. First, this requires a double pole be in an "invalid" position, but that seems like something that could be physically possible, so should be considered.

If the phase are out of sync, you'd never get more than 240V; you'd get some random voltage between 0V (perfectly in sync) and 240V (perfectly out of sync). This probably wouldn't be healthy for something like a compressor / motor which could stall. I can't think of a way that you'd get fireworks though.

I guess I don't know the answer to this question, other than a guess: When an AIO loses grid connection, obviously its now free-running, and over time will drift from perfect lock in both phase and frequency. When grid comes back online, I assume it tries to lock on to the phase / frequency pretty quickly(?). If that's true, the 'fault' condition above would probably only last for a few cycles.

Now that I think about it, I'm pretty sure that's what happens. I once made the mistake of feeding the output of a lead-acid UPS into the input side (snake eating its own tail). It ended up settling at the one extreme of its operating frequency range (47Hz), which makes sense, because it would try to follow its "input", and drift pretty quickly to one extreme (min / max).


Just so we're all following along, this is one of the two invalid positions:

 

[hex4def6]

The thread title is confusing. Because this is not about a "selector".

Yeah, changed the title. Not sure it's improved. Kind of a non-standard setup...

 

[wpns]

That is an interesting thought that I hadn't considered. First, this requires a double pole be in an "invalid" position, but that seems like something that could be physically possible, so should be considered.

Why is L1-L3 or L2-L3 an invalid position in a 3-phase box?

If the phase are out of sync, you'd never get more than 240V; you'd get some random voltage between 0V (perfectly in sync) and 240V (perfectly out of sync). This probably wouldn't be healthy for something like a compressor / motor which could stall. I can't think of a way that you'd get fireworks though.

Might depend on what the beat frequency was, might depend how different loads respond to varying voltages between 0 and 240V at different beat frequencies, might depend on the output impedance of the inverter, I dunno, lots of moving parts, and I'd assert it's going to happen someday, maybe when you are gone.

I guess I don't know the answer to this question, other than a guess: When an AIO loses grid connection, obviously its now free-running, and over time will drift from perfect lock in both phase and frequency. When grid comes back online, I assume it tries to lock on to the phase / frequency pretty quickly(?). If that's true, the 'fault' condition above would probably only last for a few cycles.

Usually they spend 5(?) minutes making sure the grid is stable, and then sync up and reconnect the input, how fast they'd slew to the new frequency/phase is another unknown. I'd hope it wouldn't be too fast (a few cycles), or motors would see a rapid frequency/phase change that might not be good for them. Better or worse than the 0-240V at whatever the beat frequency is"? Beats me, but I don't want to expose any of my loads to it, nor do I want to leave it as a surprise for my kids or whoever I sell the house to.

 

[timselectric]

I suspect it would not pass an inspection but if you don't need one then whatever.

I don't see why it wouldn't, as long as the labeling is correct.

 

[timselectric]

Why

Because it's not fed by 3-phase.
It's fed by two sources. Split-phase from grid and 120v from the inverter.