Anti-islanding a battery backup system without solar

[ProgPrep]

Hi everyone... Maybe my Google search skills are poor, so please feel free to point me to another thread...

I want to have DIY battery backup in the basement for the fridge and lights, which are on a different floor. The electricity is sent up to the floor via a subpanel with a breaker that connects to the main service and a breaker that then sends 220 two phase up to the floor.

I'm thinking to get a nice LiFePO4 battery or two and add a breaker to this subpanel that is connected to a 110v inlet box attached to this new breaker. (Since this is only for fridge and some lights, I will only need a single phase inverter and can use the leg that powers the fridge on the upper floor.)

As it is, the system is pretty DIY, and if power goes out, I would have to go downstairs and turn off the breaker that goes to the main (so I dont backfeed the grid) and then plug the inverter into the inlet.

As of now I'm just considering an inverter charger that can be left in charge mode most of the time and can also charge batteries via genset.

Does anyone know how I could make this process automatic? So with the desired setup above: 1) automatically switch from charge to inverter mode and 2) automatically anti-island?

Thanks!!!

P.S. - I might also want to keep the hallway lights running which are on a different subpanel in the basement. I assume I can put a second inlet near the hallway lights subpanel and run the same setup just with a power strip from the inverter feeding the two inlets? Just checking if there could possibly be any ground-fault problems with this...?

 

[svetz]

How about a small automatic transfer switch (ATS) for an RV and connect it to your own distribution panel (or circuit transfer switch)? That is you're making an automatic "circuit transfer switch" as a critical loads panel.

The ATS has two inputs: power from the grid and power from the inverter. The ATS's function is to switch between inputs based on when power is sensed on one of the inputs (e.g., output is connected to the grid if the grid is up, otherwise shifts to the inverter).

The output from the ATS would be connected to a distribution panel and be the circuits you want to be powered when the grid is off. This way you can't backfeed the grid, but can power anything you like.

You can also find ATSes for DIN boxes and do it all in a single box. One word of caution, if you go that route be sure the parts are certified by some trustworthy agency (e.g., UL, ETL, CE, ESA, CSA), a lot of non-certified components are sold cheap but won't actually run at the rated power.

Just checking if there could possibly be any ground-fault problems with this...?

GFCIs trip when the round-trip current isn't equal, that is current is running through some external circuit (doesn't have to be the ground wire).

A good analogy for this is a water hose. Think of a pump pushing water through a hose (the hot wire), it travels down the hose, does some work turning a mill, and then the water returns via a second pipe (the neutral wire). If the gallons-per-minute of water (current) out isn't equal to the gallons-per-minute back, then there's a leak and the GFCI trips.

So, there shouldn't be a problem if everything is wired correctly.

 

[ProgPrep]

Hi svetz, thanks much for your input... This is **almost** perfect, excepttttt because of my layout, my transfer switch will be located on a different floor and I can't run a second romex line to the upper floor to have a critical loads panel... Also because of this same layout and backup needs, I've decided to run 120 volts up the two legs of 240 volts to power the upper floor, so even if I got an ATS capable of sending 240 up, I already bought the inverter which is single phase... Which leaves me with a very expensive paper weight drying machine and split system AC.

It would have been perfect if I splurged on an split phase inverter that could have powered my drying machine and A/C on bypass mode... Then I could have done this.... with a warning to the people not involved in the backup system if they turn on the dryer one of the breakers in the line will probably trip and/or the batteries will discharge in 10 minutes if they didnt know the mains was down.

Don't worry though, I'll reply to this thread if anything else comes up though!

 

[macdad22]

Prog,

Your best bet is to use an offgrid hybrid solar/charger/inverter that is similar or perhaps one that GroWatt makes. The key to this is to have the AC input coming from the grid, the subpanel fed ( this would be the items you want on battery backup ) from the AC output on the inverter. When utility AC is up the inverter will be supplying the power as a pass through plus charging you batteries... when the GRID AC power fails or you shut it off it should switch in 10 milliseconds or less to the battery backup.

I have this installed in my house with Zero solar today and it works great. You do need to understand what your kwh consumption to make sure that you have enough hours of backup based upon your local utility reliability.

Good luck,

Chris

 

[boondox]

Whatever you decide to do, do NOT set up a manual system where you turn off one breaker before turning on another. This is dangerous and illegal. If you back feed the grid during an outage your power will go back to the line transformer and get stepped up to 12kV or whatever your local system uses. Either do a proper transfer switch OR a separate system for the battery loads.

 

[ProgPrep]

Just wondering if anyone can comment on a manual transfer switch that seems to be the most common solution to preventing backfeed to the grid versus something like this: rotary switch...

Looks like the rotary switch serves the same purpose as the transfer switch, minus the built in breakers. But I dont see much info about a rotary switch like this. My biggest concern would be poor construction and shorts in the device itself.

EDIT: Actually, there is no output on this... Wondering how people set this up, maybe ignoring ground?? i guess if you wanted one output you would have to jump 2(4) of the terminals to the common out.

 

[boondox]

Not knowing how those switches are made, I couldn't say. Real transfer switches are made with "break before make" contacts to make sure that one is cleanly disconnected before the next connects. No way I would go cheap on a transfer switch, get a real UL approved, up to code switch for the job.

 

[svetz]

... Which leaves me with a very expensive paper weight drying machine and split system AC....

We see that time and again on the forums, it's why you generally see recommendations to plan stuff out before buying regardless of how great the deal is. Great deals generally cycle back around, but being stuck with a paperweight isn't much of a bargain.

...I've decided to run 120 volts up the two legs of 240 volts to power the upper floor, so even if I got an ATS capable of sending 240 up, I already bought the inverter which is single phase...

I think you can still make that work, just put the inverter's hot to both L1 and L2 on the ATS. Just be careful and if not 100% sure what you're doing ask until you are or hire an electrician (the expense is shockingly low compared to fire damage, hospital bills, or worse).

Whatever you decide to do, do NOT set up a manual system where you turn off one breaker before turning on another. This is dangerous and illegal.

I just want to echo my support for this, it's too easy for mistakes to happen (which is why the NEC rules exist for it, mistakes did happen).

Just wondering if anyone can comment on a manual transfer switch that seems to be the most common solution to preventing backfeed to the grid versus something like this: rotary switch...

Rotary switches will work fine, you just need enough poles to switch what you want. But again, beware of anything not listed by one of the labs (e.g., UL) less you get one that won't carry the rated power or fuse shut.

The grounds shouldn't be switched, they should be combined. Since you're using GFCIs you might need to switch the neutrals.

So, for example, the switch you listed isn't UL, but has enough poles, so the diagram is saying in position 0 everything is off. In position 1, terminals 1/2, 5/6, 9/10, and 13/14 are shorted. In position 2, 3/4, 7/8. 11/12. and 15/16 are shorted.

So you could wire it as:

Terminal	P1 shorts to	P2 Shorts to	Connection
1	2		L1 from Grid
2	1		L1 to Distribution
3		4	Hot from Inverter
4		3	Jumper to Terminal #2
5	6		L2 from Grid
6	5		L2 to Distribution
7		8	Jumper to #3
8		7	Jumper to #6 (note that 240V devices on the distribution won't work)
9	10		Neutral from Grid
10	9		Neutral to Distribution
11		12	Neutral from Inverter
12		11	Jumper to #10

So, in position 1, the grid is active and L1/2/N go from terminals 1/5/9 to terminals 2/6/10.
In position 2, the inverter is active and hot goes to 2 and 6 powering both legs and neutral to 10.
The jumpers are safe. For example in position 1, the jump from 12 to 10 doesn't do anything as terminal 12 is open.

If that doesn't make sense just bear in mind an electrician can do it. Double-check everything with an ohm meter before powering, you wouldn't be the first person to come across a rotary switch with the wrong label glued to it.

 

[ProgPrep]

So, in position 1, the grid is active and L1/2/N go from terminals 1/5/9 to terminals 2/6/10.
In position 2, the inverter is active and hot goes to 2 and 6 powering both legs and neutral to 10.
The jumpers are safe. For example in position 1, the jump from 12 to 10 doesn't do anything as terminal 12 is open.

If that doesn't make sense just bear in mind an electrician can do it. Double-check everything with an ohm meter before powering, you wouldn't be the first person to come across a rotary switch with the wrong label glued to it.

Wow, I had to draw out the diagram to visually get it. Thanks for the help!!

I am a little worried that it is not UL because it's not like I'm powering a few lights, I intend to power the entire floor with selective breakers on and telling the residents to not use their hot water tea kettle, hair dryer, etc... while on backup power... So to minimize potential for human error, it would be nice to know that the rotary switch of indeterminable quality won't melt the zinc screws holding it to the wall...

I also decided to get something I just saw in a youtube video, because on a different note, I'm not sure where the top floor is grounded because there is no ground line incoming to the subpanel, where N/G is bonded on the main panel. So I got myself a three light tester to get that situation straight.

Lastly, about measuring the resistance... I assume that means turning off all power, and putting the ohm meter in various scenarios, such as making sure that there is no resistance between one incoming legs and also complete resistance on the other incoming legs with the switch in its respective positions.

Thanks!!!

 

[svetz]

...I am a little worried that it is not UL...

That's healthy, don't put non-certified gear in your house. It's just not worth it in the grand scheme of things to save a few bucks and worry about it ever after.

...I'm not sure where the top floor is grounded...

Some houses are a wiring nightmare, if you start finding oddities like that think about hiring a professional to come in and tell you what's going on and what it'll take to make the house safe.

Lastly, about measuring the resistance... I assume that means turning off all power, and putting the ohm meter in various scenarios, such as making sure that there is no resistance between one incoming legs and also complete resistance on the other incoming legs with the switch in its respective positions.

Yes. I typically test them while they're still sitting in the box. Test every terminal to every other at all switch positions and make sure it matches the specs.