diy solar

diy solar

Solark 12k with autotransformer only for balancing loads - need sanity check for open neutral scenario

Samsonite801

Solar Wizard
Joined
Oct 15, 2020
Messages
2,979
Would some dear friend here please help me sanity-check a scenario for a minute? I'm not an EE, just some guy here who tries to pay attention sometimes.

So I have a neighbor I regularly do heavy equipment repair work for (on our co-op) who has a Solark 12k (installed in an off-grid situation, solar or generator power only), which is tripping when he gets too much leg imbalance. The cause is understood, we are going to work and try to get the expected loads more balanced-out naturally, by swapping wires out in the breaker box and get close to balanced (with all the expected loads) as we can first.

It has been a popular trend for my other neighbors here (also with Solark 12k's) on my co-op to add in the autotransformers onto the outputs, to provide better line balancing as well.

This neighbor in question here, would also like to add in an autotransformer to help balance loads (for the unexpected loads)... He said he already purchased one for this purpose (same like the other shareholders). I have not seen it yet and do not know the brand/model they are using at this point (it might be the popular SolarEdge model I'm thinking).

In the case of a Solark 12k, it already has a neutral output provided (native split-phase inverter), so the autotransformer is not being added to provide a neutral, only for load balancing.

So in this case, it is being wired in parallel only, using the output of the transformer (L1,L2,N simply wired in parallel), where you have only one 240v dual-pole breaker to parallel it in, or take it out of the circuit.

I am trying to imagine the behavior on what would happen in this topology if a neutral connection was lost between the main panel neutral and the autotransformer (where L1 and L2 would still be paralleled to the autotransformer).

I had thought of just playing it safe and putting in a 3-pole breaker (to have neutral going through it too) between the autotransformer and the breaker box, so the autotransformer is either all in or all out...

But part of me is thinking it does not matter at all.

If the hots trip and neutral is still connected to the autotransformer then nothing bad can happen (also seen in video below), and in the other scenario where the hots stay on but if the neutral was lost to the autotransformer, then what happens there?

Wouldn't the autotransformer simply act like a regular 240v load on the circuit, it shouldn't do anything bad? (since the Solark still provides a neutral to the panel for the appliances, it just wouldn't provide any balancing if the autotransformer neutral wasn't included)...

Just trying to wrap my head around this scenario, and make sure we don't add something in which could fail in a way that can over-volt any appliances. I've also tried read all the threads on this as well, and it seems I found a couple references saying this may not matter here, but it's hard to confirm, since there are a lack of threads discussing using an autotransformer with a split-phase inverter (only for load balancing, not to provide a neutral).

Here is a video showing an example of how these would wire in with a Solark 12k (in parallel topology, to a split-phase inverter):

Thanks, any help in confirming is appreciated...
 
Last edited:
Make sure the firmware is up to date and balance up your loads before adding a transformer.
Yes they do help but also introduce other issues as has been discussed here occasionally !!!!
But yes as long as both hots L1&L2 get tripped at the same time with a dual pole breaker, your 120v loads will be protected.
 
Make sure the firmware is up to date and balance up your loads before adding a transformer.
Yes they do help but also introduce other issues as has been discussed here occasionally !!!!
But yes as long as both hots L1&L2 get tripped at the same time with a dual pole breaker, your 120v loads will be protected.

Yeah thanks for reminding on the firmware, I wasn't thinking anymore about that part since I saw the information about that issue where firmware can help improve imbalance load handling several months ago...

Yeah I was thinking the same thing, that since the input and output of the transformer is the same thing (simple parallel), then as long as both lines are tripping together then it shouldn't do any funky stuff.

And theoretically (in my mind anyways), it seemed if neutral was lost on the autotransformer then it wouldn't matter since the appliances are all still neutral-connected to the Solark (via the main breaker panel), the autotransformer would just act like a simple 240v appliance load and only would not balance loads anymore at that point.
 
Please elaborate.

The intended purpose attempting to provide for, is for line / load balancing only.
If you transfer grossly imbalanced phases to a auto , you remain with similar issues especially neutral levels. You even out the split phase load but you get issues at the output of the auto with floating neutrals.

You actually need a isolation transformer
 
I am trying to imagine the behavior on what would happen in this topology if a neutral connection was lost between the main panel neutral and the autotransformer (where L1 and L2 would still be paralleled to the autotransformer).

If auto-transformer connection is lost, you're back to SolArk supplying, and it will shut down if it can't sustain current & voltage.

If Sol-Ark neutral is lost, auto-transformer will provide derived neutral to loads. Voltage will sag according to winding resistance (as I showed with DMM)


If overloaded, you would like power disconnected. You do not want auto-transformer disconnected!
That can be accomplished by:

Sol-Ark --> breaker --> auto-transformer --> load

If breaker trips, auto-transformer is disconnected but so is load.

The issue you're concerned with is not unique to auto-transformer. It happens with grid if neutral is lost. Need solid connections you can trust.



I had thought of just playing it safe and putting in a 3-pole breaker (to have neutral going through it too) between the autotransformer and the breaker box, so the autotransformer is either all in or all out...

Be aware that auto-transformer Neutral carries current equal to sum of L1 and L2, not difference like for an isolation transformer with split-phase output. L1 and L2 current will always be equal. So 3-pole breaker is one option. If the two windings are completely isolated, just use separate neutral for each, landing on neutral busbar. If winding connections are inaccessible, breaker on L1 and L2 each half what current you want on N would protect it. (The mystery will be what manufacturer actually designed for; would hope recommended breaker protects it. But if neutral wire gauge is only sufficient for current of each individual L1 and L2 pole rather than twice as much, they did it wrong.)

An auto transformer doesn’t really help here

I plan to do this for my stacked 120V Sunny Island inverters.
Each can carry its own rated current, regardless of whether load is single or split phase.
By adding an auto-transformer, I hope to have two inverter which are wired for 120/240V split phase support single phase load greater than what just one could carry.

It will have a relay, disconnected when grid passes through but connected when off-grid.
To be useful, it must have low winding resistance (compared to transformers inside Sunny Island.) It also interacts with the inverter programmed voltage/current characteristics, which are intended to let multiple inverters on one phase share the load.

First test, I plugged a 9kVA auto-transformer in with a long cable. It only carried around 10% of the load. In the future I will place it very close to breaker panel.

If you transfer grossly imbalanced phases to a auto , you remain with similar issues especially neutral levels. You even out the split phase load but you get issues at the output of the auto with floating neutrals.

You actually need a isolation transformer

An isolation transformer will perfectly distribute any single or split-phase load across 240V output of GroWatt. More can be carried than by paralleling an auto-transformer.

Disconnect of neutral from split-phase isolation transformer would be a problem, same as other lost neutral.
 
If you transfer grossly imbalanced phases to a auto , you remain with similar issues especially neutral levels.

Theoretically the loads won't be 'grossly imbalanced' since we are going to try and balance the expected loads up as much as possible, prior to bringing AT into the picture.


You even out the split phase load but you get issues at the output of the auto with floating neutrals.

The output is the same as the input for all three wires on this particular parallel wiring proposed, with the AT and the Solark, so I don't get what you mean by floating neutrals.


You actually need a isolation transformer

Well he already had purchased the AT prior to my discussions with him, I did mention him the possibility of an isolation transformer, and I think he is past the return window for his AT he bought, so I doubt he wants to dump a bunch more money on it at this stage. He just wants to get the Solark to stop tripping so often.
 
Last edited:
If auto-transformer connection is lost, you're back to SolArk supplying, and it will shut down if it can't sustain current & voltage.

Fair enough, that should be fine. We are just trying to make shutdowns not such a regular occurrence as he seems to be indicating now.


If Sol-Ark neutral is lost, auto-transformer will provide derived neutral to loads. Voltage will sag according to winding resistance (as I showed with DMM)


If overloaded, you would like power disconnected. You do not want auto-transformer disconnected!
That can be accomplished by:

Sol-Ark --> breaker --> auto-transformer --> load

If breaker trips, auto-transformer is disconnected but so is load.

The issue you're concerned with is not unique to auto-transformer. It happens with grid if neutral is lost. Need solid connections you can trust.

Should be fairly robust neutral connections with fairly short wire run lengths. I guess there is just as much chance the AT could lose neutral as if the Solark were to lose neutral, either way it seems like neutral is more redundant by having the AT and the Solark both providing a parallel center tap neutral. Can't be worse than one or the other by itself I would imagine.


Be aware that auto-transformer Neutral carries current equal to sum of L1 and L2, not difference like for an isolation transformer with split-phase output. L1 and L2 current will always be equal. So 3-pole breaker is one option. If the two windings are completely isolated, just use separate neutral for each, landing on neutral busbar. If winding connections are inaccessible, breaker on L1 and L2 each half what current you want on N would protect it. (The mystery will be what manufacturer actually designed for; would hope recommended breaker protects it. But if neutral wire gauge is only sufficient for current of each individual L1 and L2 pole rather than twice as much, they did it wrong.)

I am aware of neutral carrying the sum of total current... Good to bring it up as reminder, thanks. I would gauge the wire and 3-pole (if used) breaker for max neutral current carrying.


I plan to do this for my stacked 120V Sunny Island inverters.
Each can carry its own rated current, regardless of whether load is single or split phase.
By adding an auto-transformer, I hope to have two inverter which are wired for 120/240V split phase support single phase load greater than what just one could carry.

It will have a relay, disconnected when grid passes through but connected when off-grid.
To be useful, it must have low winding resistance (compared to transformers inside Sunny Island.) It also interacts with the inverter programmed voltage/current characteristics, which are intended to let multiple inverters on one phase share the load.

First test, I plugged a 9kVA auto-transformer in with a long cable. It only carried around 10% of the load. In the future I will place it very close to breaker panel.


An isolation transformer will perfectly distribute any single or split-phase load across 240V output of GroWatt. More can be carried than by paralleling an auto-transformer.

Disconnect of neutral from split-phase isolation transformer would be a problem, same as other lost neutral.

So really a lost neutral would be a rare occurrence. I just want to make sure I am not bringing in a higher chance of frying appliances into the picture by backing some solution with my supposed rubber stamp.
 
He just wants to get the Solark to stop tripping so often.

A lot of 120V loads on L1, and a lot on L2? Sometimes too much load on one or the other phase?

With a pair of CT around L1 and L2, could determine how much imbalance and shed some loads on the offending phase, or transfer them to the other. Would take a bit of measure and compare logic.


Has anybody measured the current, and determined both VA and W, for their auto- or isolation- transformers?
If my measurements are good (DC read with a cheap clamp ammeter), Sunny Island does a pretty good job of driving a reactive load without dissipating too much power.
But current drawn by no-load transformer sure varies between models.
I'm thinking that a pair of 240V primary windings (of a 240/480 to 120/240V transformer) may be a lot better than some purpose-built auto-transformers.
Transformers designed for cost and minimum materials seem to drive further into saturation, and current shoots up past the sine wave I expect. Running a 240V winding at 120V would minimize that (and power handling would have to be derated according to intended current in those windings; VA rating of transformer was based on full voltage.)
 
A lot of 120V loads on L1, and a lot on L2? Sometimes too much load on one or the other phase?

With a pair of CT around L1 and L2, could determine how much imbalance and shed some loads on the offending phase, or transfer them to the other. Would take a bit of measure and compare logic.

This is exactly what I was planning to do first, I honestly think if I go in there and just balance out the expected loads, that will solve the problem and he may not even need the AT...


Has anybody measured the current, and determined both VA and W, for their auto- or isolation- transformers?

I haven't seen the system yet. That day we were just riding together in the truck as we went out to go fix a broken water pipe out at the batch plant, where he had told me about his issue, which is when the microwave is on and they turn on some other stuff (can't remember whatall devices he said), then it trips it all the time. He told me the other shareholder guy out there (also has a Solark) recommended to put an AT on it like he did and would fix it. I don't even know who did the initial Solark install for him.


If my measurements are good (DC read with a cheap clamp ammeter), Sunny Island does a pretty good job of driving a reactive load without dissipating too much power.
But current drawn by no-load transformer sure varies between models.
I'm thinking that a pair of 240V primary windings (of a 240/480 to 120/240V transformer) may be a lot better than some purpose-built auto-transformers.
Transformers designed for cost and minimum materials seem to drive further into saturation, and current shoots up past the sine wave I expect. Running a 240V winding at 120V would minimize that (and power handling would have to be derated according to intended current in those windings; VA rating of transformer was based on full voltage.)

Thanks, sounds really interesting, I'll keep that in mind in case I ever need to do something like that in the future.
 
Last edited:
Back
Top