Use Main Panel as Critical Loads Panel?

[ricardocello]

Ok cool. I was confused because the OP’s inverter is 240 only so would have a lost neutral in that case.

So if the choice is risk of lost neutral vs extra loading of AT if it runs in parallel with grid, I’d rather take the second one

Honestly, the most straightforward path for the OP is to use the Growatt ATS-US Autotransformers, which integrate in the transfer switch.
No need to worry about Neutrals, and it looks like the AT will never be in parallel with the grid.
Growatt has thought this through. Unless 24A is not sufficient power for the OP's loads.
Growatt ATS-US

 

[zanydroid]

No need to worry about Neutrals, and it looks like the AT will never be in parallel with the grid.

It will be when on grid. If grid is on the ATS relays inside will prefer it and connect it to AT. And the AT neutral has to bond to utility neutral due to code and to get the N-G bond.

But Growatt designed it so you can blame them instead of yourself for a home grown approach if it doesn’t work amazing (it has to be safe enough for them to be selling at this scale).

Neutral is not switched, just on a tied breaker with the AT bypass (which hopefully is good enough to prevent lost neutral since unless the breaker fails a specific way the other legs will disconnect on tripping neutral breaker)

 

[ricardocello]

It will be when on grid. If grid is on the ATS relays inside will prefer it and connect it to AT. And the AT neutral has to bond to utility neutral due to code and to get the N-G bond.

I stand corrected. I really wish there were detailed wiring diagrams with these things. That’s one area where Victron really does well, putting out nice colored graphics with different system configurations and every wire shown.

 

[zanydroid]

I stand corrected. I really wish there were detailed wiring diagrams with these things. That’s one area where Victron really does well, putting out nice colored graphics with different system configurations and every wire shown.

I had to find two different versions of the manual, random PowerPoint, and scrutinize some YouTube videos that had the wiring box open to figure out what the schematic probably was

On top of that there are two very similarly named SKUs for autotransformer from growatt, and it’s not clear either growatt or their resellers keep it straight. One has the ATS and breakers in it the other has only terminal blocks. Seems critical to keep separate, but I’m not the one running a business here so what do I know…

 

[lapsmith]

If you've moved down to the 5k, maybe that makes sense as Victron AT gets you some more amperage.
Also, the ATS-US is UL1741. Victron only has CE/IEC certifications (as usual).

One reason for moving down to a 5kw autotransformer is that if my loads are anything bigger, it will kill the 10kw battery in no time.

The Growatt ATS-US says it can handle a unbalanced load of 3000w. Victron says maximum continuous neutral current is 28A (32A for 30 minutes). If watts = V x A, then the Victron is slightly better with 3360w. But it also has a thermal shutdown which I like. Also, I would not be using the ATS function of the Growatt so there is no reason to select it other than price.

 

[lapsmith]

Does your inverter put out 25A at 240V? If so, it could support 50A at 120V. But the autotransformer is only good for 28A continuous, 32A intermittent, on neutral.

Yes. But I'm okay with the lower current on 120V. Battery is mainly for 220V well pump*. I measured a 60 amp surge current with a fluke peak clamp on ammeter so I hope the inverter and transformer can handle it. They don't list any specs for that. If it can't then a soft start device may be in my future.

Another reason for the battery is so I can get output from the panels when the grid is down and the sun is shining. I don't think it is possible without a battery.

Auto-transformer in parallel with the grid is a separate issue. I looked at the schematic and thought I might modify it to do dynamic bonding of center tap to neutral if I used it.

I was thinking I'd have to switch L1 and L2 as well to keep the autotransformer from being in parallel with the grid, but do I only have to switch N? If so, The Victron has a N-G bonding relay that I can probably repurpose for that
EDIT: Never mind. My inverter doesn't have a neutral output. So I'll need a relay in either L1 or L2 (both for safety?), but it will have to be a bigger amp rating one than the N-G they supply, I imagine.

 

[lapsmith]

I was guessing. I can believe 3V or 6V in isolation, but I can’t believe they can sell a UL listed product if it stays even 1V if connected to utility transformer like their wiring diagram allows

What I saw on the Victron forum was in that range. And that's with no load. You would think they could match windings better than that. The odd part is that some saw a much smaller difference.

 

[lapsmith]

On top of that there are two very similarly named SKUs for autotransformer from growatt, and it’s not clear either growatt or their resellers keep it straight. One has the ATS and breakers in it the other has only terminal blocks. Seems critical to keep separate, but I’m not the one running a business here so what do I know…

According to Signature Solar, the version they sell only has the Din rail connector block. I didn't ask them to open one up to verify, but considering the price is about $200 less than it used to be, I'm sure that's the case. The Growatt is $300 from SS and I found the Victron for $450 on Amazon. So considering everything, I just ordered the Victron 32a since it was the last one they had (everyone else is selling it for around $600). It can be returned, but hopefully I won't have to.

 

[Hedges]

Yes. But I'm okay with the lower current on 120V. Battery is mainly for 220V well pump*. I measured a 60 amp surge current with a fluke peak clamp on ammeter so I hope the inverter and transformer can handle it. They don't list any specs for that. If it can't then a soft start device may be in my future.

Transformer isn't in the circuit for 220V surge. If it was, voltage would sag that much more (multiple of "regulation" x 2, e.g. 3% regulation is from +3% to -3%, so 6% voltage drop at 1x rated current, 30% at 5x.

Inverter ought to have surge spec, and you care about 1 second not 30 milliseocnds.

Another reason for the battery is so I can get output from the panels when the grid is down and the sun is shining. I don't think it is possible without a battery.

Some work batteryless, but then no surge beyond PV wattage.

I was thinking I'd have to switch L1 and L2 as well to keep the autotransformer from being in parallel with the grid, but do I only have to switch N? If so, The Victron has a N-G bonding relay that I can probably repurpose for that
EDIT: Never mind. My inverter doesn't have a neutral output. So I'll need a relay in either L1 or L2 (both for safety?), but it will have to be a bigger amp rating one than the N-G they supply, I imagine.

With grid passed through relays, neutral of grid is the neutral.
I've thought the N-G bonding relay could be repurposed to bond load neutral to AT neutral when off-grid.
Maybe there would be an over-voltage glitch while neutral switches. But I figured it would avoid paralleling AT with grid (which may or may not be a problem for you, could measure current, and rely on thermal protection if anything changes.)

 

[Hedges]

One reason for moving down to a 5kw autotransformer is that if my loads are anything bigger, it will kill the 10kw battery in no time.

If watts = V x A, then the Victron is slightly better with 3360w.

Not sure it even has a larger transformer, but it has a fan.

I'd like to know what the no-load current is of each. Inverter has to deal with that as reactive load. The Victron began life as 220V isolation transformer, so with only 120V across each winding there should be plenty of core, low no-load current. Growatt similar size transformer, however, so may also be low. (Transformers designed for grid use are optimized for price, skimp on materials and higher no-load current.)

 

[zanydroid]

One reason for moving down to a 5kw autotransformer is that if my loads are anything bigger, it will kill the 10kw battery in no time.

What is the maximum surge current for starting the appliances on the critical loads panel? Average current might be OK for he 10kWh battery even with a high surge.

 

[lapsmith]

What is the maximum surge current for starting the appliances on the critical loads panel? Average current might be OK for he 10kWh battery even with a high surge.

60 amps at 240v for the well pump is likely the max. But that would be for milliseconds. Only other surge load would be refrigerator, which I'd wager is much less.

 

[lapsmith]

Inverter ought to have surge spec, and you care about 1 second not 30 milliseocnds.

You would think, but this is Growatt. Nothing on their datasheet or elsewhere that I could find.

 

[lapsmith]

With grid passed through relays, neutral of grid is the neutral.
I've thought the N-G bonding relay could be repurposed to bond load neutral to AT neutral when off-grid.
Maybe there would be an over-voltage glitch while neutral switches. But I figured it would avoid paralleling AT with grid (which may or may not be a problem for you, could measure current, and rely on thermal protection if anything changes.)

What is the issue with paralleling the AT with the grid?

If neutral was switched off and only L1and L2 of the AT were connected to the grid (no loads on the AT) does that help?

What would happen if the the neutral that is created by the autotransformer were accidentally connected to the grid neutral while the grid is on? Would I have to check with a scope to see if they were in phase? I ask because I planned on using a 3 pole manual transfer switch to disconnect the inverter's backup output to the generator breaker and what if I forget to switch the transfer switch to off when turning on the main breaker? (L1 and L2 would not be connected to grid becuse of the generator interlock, but N would be).

 

[Hedges]

If spec sheet and manual don't give surge (I looked, didn't find) assume rated VA is all you get. 11,400 VA.
Without an easy-start, surge will be 5x rated current. 11,400 / 5 = 2280 VA, assume that's the highest rated motor you can start. 9.5A at 240V, which is less than my small central A/C.

If centertap is disconnect and AT just connects to L1 and L2, near zero current flows, not a problem.

If centertap is connected to N but the other two transformer connections are disconnected from L1 and L2, zero current flows, no problem.

If grid feeds through inverter and AT is on output of inverter, with center tap connected to N of loads and N of grid, that's where it might draw current trying to rebalance.

 

[zanydroid]

What would happen if the the neutral that is created by the autotransformer were accidentally connected to the grid neutral while the grid is on? Would I have to check with a scope to see if they were in phase? I ask because I planned on using a 3 pole manual transfer switch to disconnect the inverter's backup output to the generator breaker and what if I forget to switch the transfer switch to off when turning on the main breaker? (L1 and L2 would not be connected to grid becuse of the generator interlock, but N would be).

If the AT is in the circuit somehow, it would have 60Hz in sync with the grid. You could put a clamp meter on the neutral to see how much leaks through after the balancing. As well, on L1 and L2.

The most dangerous situation would be L1/L2 visible to loads but floating neutral (not connected to anything) or wrong neutral (connected to the wrong isolated 120/240 system. And this is basically the same as the floating neutral case except more difficult to reason about because there are several combinations to reason about. This is why I think less switching / possibility for independent neutrals is better, and that might be part of why NEC only lets approves certain combinations. If I was setting up a system like this I might consider allowing for the possibility of AT connected in parallel with grid, in exchange for simplifying the switching).

 

[lapsmith]

If spec sheet and manual don't give surge (I looked, didn't find) assume rated VA is all you get. 11,400 VA.
Without an easy-start, surge will be 5x rated current. 11,400 / 5 = 2280 VA, assume that's the highest rated motor you can start. 9.5A at 240V, which is less than my small central A/C.

The well pump draws about 9 amps continuous so it will be close. Fingers crossed. Unfortunately, my grid voltage is usually around 250, so using the 5 times rule of thumb, I might only get 9.1 amps.

If grid feeds through inverter and AT is on output of inverter, with center tap connected to N of loads and N of grid, that's where it might draw current trying to rebalance.

Just to confirm the scenario, the AT is being fed by the grid through the output of the inverter, i.e. L1 and L2 with no neutral coming from the inverter. The only connection on the output side of the inverter is from its neutral to neutral of the grid/loads. So my follow up question is, is there anyway to guestimate the current due to the attempt to rebalance? I'm trying to get a feel for how bad it might be. Is it a function of load imbalance between L1 and L2?

 

[lapsmith]

If the AT is in the circuit somehow, it would have 60Hz in sync with the grid. You could put a clamp meter on the neutral to see how much leaks through after the balancing. As well, on L1 and L2.

I can surely do that once everything is installed. It may be better than trying to estimate a value.

The most dangerous situation would be L1/L2 visible to loads but floating neutral (not connected to anything) or wrong neutral (connected to the wrong isolated 120/240 system. And this is basically the same as the floating neutral case except more difficult to reason about because there are several combinations to reason about. This is why I think less switching / possibility for independent neutrals is better, and that might be part of why NEC only lets approves certain combinations. If I was setting up a system like this I might consider allowing for the possibility of AT connected in parallel with grid, in exchange for simplifying the switching).

Got it. In my case, I don't think I'd ever have a floating neutral...if I (or someone else) forgets a switch, the worst I think I could have is the AT connected in parallel. If that isn't a deal breaker, then I can live with it. In fact, I could use a two pole manual transfer switch to insure the neutral from the AT is always connect to the neutral of the grid/loads.

 

[ricardocello]

So my follow up question is, is there anyway to guestimate the current due to the attempt to rebalance? I'm trying to get a feel for how bad it might be. Is it a function of load imbalance between L1 and L2?

Q1: Do you have your own grid transformer on a pole or in a box not shared with any other properties? (i assume yes)

Q2: Do you have other 120V loads electrically connected to the grid but not served by the backup load panel?

If no, then your AT in parallel shouldn’t be too bad, but it will depend on the unknown impedances of the grid transformer, and the long cable run to the house.

If yes, then those 120V loads will be balanced out too, not necessarily in a predictable way.

 

[lapsmith]

Q1: Do you have your own grid transformer on a pole or in a box not shared with any other properties? (i assume yes)

Yes

Q2: Do you have other 120V loads electrically connected to the grid but not served by the backup load panel?

If you mean while connected to the grid, yes. There are other subpanels on the property. If you mean while the grid is down, no. Even though I am using my main panel as critical loads panel, no other panels or loads will be fed by the AT while on backup. Note that I won't be using the AT to power any loads while the grid is up.

The cable run to the transformer is 350MCM. I previously checked the voltage drop from the grid xfrmr to the house with a 5kw load and it was a few tenths of a volt.