Generator wiring for split phase with Autotransformer

[BLUE STEEL]

Hello TF members,

I'm almost finished a split phase electrical upgrade on my trawler and I'm looking to confirm the correct 240v generator wiring for my system.

12kva Northern Lights genset> 240v 60Hz Multiplus> 100A Autotransformer > Main AC Distribution Panel.

The generator has 2 wiring options that output "240vac 4 wire single phase".

I am currently wired in the second option, split phase, and have disconnected the white neutral. Ground is still connected to terminal V1, with. a jumper to U2.

Is this correct for my application? Or should I be feeding the Multiplus with option 1?

 

[rslifkin]

Either should be acceptable. You can either wire for just 240V single phase and use that, or wire for 120/240 split phase (which is what you have now) and just use L1 and L2, but not neutral (that config will give 120V from L1 or L2 to N, but 240V from L1 to L2).

If you're using the Victron Autotransformer, be aware that the 100A version can still only handle 32A on the neutral for 30 min, 28A continuous. So you're limited in how much unbalance between the 120V legs you can handle on the output (max of approx 30A difference between legs). So make sure your loads are distributed such that you'll never end up with a 40A load on one leg and 0 on the other, for example.

 

[BLUE STEEL]

Thanks for this.

I was concerned that the ground wire which is still connected to V1 (and jumped to U2) would be considered a bonding location. I've pulled the G-N jumper from the Iso transformer and want to be certain that the generator is also wired correctly since my bond is now after the Autotransformer.

Noted about the max 28A continuous imbalance....I'm slowly replacing high draw items with 240V appliances. I believe the 32A version has a max of 32A even in pass-through - so limited to 7680watts?

 

[rslifkin]

Yes, the 32A version can only pass 32A. If you needed more power capacity and more imbalance capacity you could run 2 units in parallel though.

As far as neutral/ground bond, it should always be bonded at a source, so it should be bonded at the generator (even though nothing after the autotransformer will see that bond).

 

[twistedtree]

You have hit on a surprisingly complex problem, with a number of contradictory requirements.

The first question is whether you have any loads, either 120V or 240V before the input to the inverter, or are 100% of the loads on the inverter output?

The requirements for your isolation transformer and your generator are that their neutrals be bonded to ground, and there are two ways to do it.

First options is to individually bond the iso transformer's and generator's neutrals, and switch their neutrals such that only one connects to the rest of the boat at a time, just like the L1 & L2 switching. Your drawing currently only shows L1 & L2 being switched, so you would need to add the neutral to those switches.

The second option is to tie together the neutrals from the iso transformer, generator, and all other boat neutrals to a central neutral bus, and bond that bus to ground.

In either case, your power sources will always operate with a ground reference which is required.

Where things get messy is when you introduce the autotransformer. If you have taken approach #1 above with the boat's bonding, then you need a switched bonding connection for the AT such that it's bonded when on inverter power, and unbonded when the inverter is in passthrough. But that creates a problem in passthrough because the AT's neutral isn't really bonded. Its inductively held more or less at ground reference by way of the generator or iso transformer, each of which has a bonded neutral. But there is no fault current path from the AT neutral to ground as you would expect in a bonded system. The actual AT neutral voltage will depend on the L1 and L2 windings ratios in the AT relative to the iso transformer or generator, whichever is providing power. I have seen cases where the AT neutral can be elevated by 5V or more relative to ground.

If you take approach #2 above for bonding, you will have an electrically bonded AT neutral which is good, but the system is then subject to circulating currents, again based on the relative windings ratios of the equipment involved. I have seen these currents run as high as 5-10A.

Getting to work acceptably comes down to having well matched windings ratios in all the equipment, something that nobody ever pays any attention to. In theory, the windings in a split phase device like an iso transformer or a generator will be exactly the same between L1-N and L2-N, such that the L1-N voltage is exactly the same as the L2-N voltage, i.e. the N is an exact center tap in the L1-L2 winding. But sometimes they are off a bit, and in some autotransformers they are intentionally off. Unfortunately, the Victron AT that you show is one where the windings are intentionally asymmetric, and as a result it is actually unsuitable for use in this application. It's fine if you want to create 240V from 120V, or 120V from 240V, but it doesn't work well when wired as a split phase device in parallel with another split phase device like an isolation transformer or generator.

 

[BLUE STEEL]

Thanks TT, really appreciate your considerable knowledge and time to respond.

The first question is whether you have any loads, either 120V or 240V before the input to the inverter, or are 100% of the loads on the inverter output?

There are no loads connected before the inverter, which has a 50A transfer switch for when shore/generator power is available. I have to manually manage loads while inverting to keep below 2400W continuous.

Your drawing currently only shows L1 & L2 being switched, so you would need to add the neutral to those switches.

The neutral wires are still present from both the Iso transformer, and generator, as are the 3 pole switches. I could easily re-connect them if option 1 is a more advantageous wiring approach given the hardware I have.

The inverter is currently wired to internal ground relay in the AT, using the 24V control wires.

I suppose my goal was to use option 2, but you got me wondering if that is ideal with the hardware I have?

 

[twistedtree]

Thanks TT, really appreciate your considerable knowledge and time to respond.



There are no loads connected before the inverter, which has a 50A transfer switch for when shore/generator power is available. I have to manually manage loads while inverting to keep below 2400W continuous.



The neutral wires are still present from both the Iso transformer, and generator, as are the 3 pole switches. I could easily re-connect them if option 1 is a more advantageous wiring approach given the hardware I have.

The inverter is currently wired to internal ground relay in the AT, using the 24V control wires.

I suppose my goal was to use option 2, but you got me wondering if that is ideal with the hardware I have?

I think either option 1 or 2 is fine. I personally prefer Option #2 because I think a common neutral and single point of bonding is easier to keep track of, and less prone to getting messed up down the line. But that's a personal choice.

With the AT, Option 1 will give you an elevated neutral on the AT in pass through mode, and option 2 will give you circulating currents in pass through mode. In both cases, the trick is to minimize them. There is active work going on in ABYC to address this, but it's been slow going. I expect the guidance for maximum neutral voltage lift or circulating currents will be in the 1-2% of the circuit capacity. In my experience, the Victron AT will never come close to meeting that because it's windings are intentionally asymmetric. The only way you can get the numbers down to an acceptable level will be with a symmetric winding AT.

There also is a whole other can or worms around over current protection for these devices. The capacity of the inverter, especially in pass through mode, and even more again if the inverter can boost pass through current, least to a large breaker on the inverter output leading to the loads. But the AT capacity is typically much lower because it's only rated for the max imbalance between L1 & L2, and never needs to carry the full output load of the inverter system. That calls for a breaker that can be 1/2 or even as low as 1/4 the rating of the main output. The catch is that these to overcurrent devices need to somehow be interlocked such that a trip of the AT breaker will in turn trip the main breaker, Otherwise if only the AT breaker trips, you end up with a split phase system with a disconnected neutral, and that is very likely to lead to equipment damage, and very well could cause a fire. In a nutshell, and 120V devices that are connected will see anywhere between 0V and 240V, and that can be disastrous.

Victron's packaging makes this easy because it includes the "main"breaker", and then there is electronic sensing of the AT itself and it remotely trips the breaker if there is a problem. So it's a nice package, except it's asymmetric.

I have solved this on one boat by using an aux contact on the AT breaker to trigger a shut trip on the main breaker. It works well, but is a bit of a tricky and expensive pile of ABB parts to make it work.

Honestly, the best solution is a native split phase inverter, but those are few and far between, I'm afraid.

 

[BLUE STEEL]

I'd like to complete this with option 2, central bonding. I failed to show the neutral bus on my drawing.

The boat has single pole 120v breakers, and all of the neutral conductors run to a neutral bus bar(s). The ground wires from those circuits run to a ground bus bar. (pic attached)

From another thread, you mention for central bonding: "In this case you would not bond at the gen, isos, or inverter output, would have a common neutral bus throughout the boat (what you currently show), and you would bond at a single point, typically at the main electric panel"

Which brings me back to the V1-U2 grounding jumper in the generator: Should I remove that, and leave the green wire connected to V1, which runs to the ground bus?

Then finally, create the bond in the panel and disable the Inverter/AT grounding relay?

 

[twistedtree]

I'd like to complete this with option 2, central bonding. I failed to show the neutral bus on my drawing.

The boat has single pole 120v breakers, and all of the neutral conductors run to a neutral bus bar(s). The ground wires from those circuits run to a ground bus bar. (pic attached)

From another thread, you mention for central bonding: "In this case you would not bond at the gen, isos, or inverter output, would have a common neutral bus throughout the boat (what you currently show), and you would bond at a single point, typically at the main electric panel"

Which brings me back to the V1-U2 grounding jumper in the generator: Should I remove that, and leave the green wire connected to V1, which runs to the ground bus?

V1-U2 remain jumpered together, but there should be no green wire grounds on those terminals. BUT, the generator chassis/frame should have a green wire AC ground. You just don't want any connection from neutral (V1-U2) to ground inside the generator.

Then finally, create the bond in the panel and disable the Inverter/AT grounding relay?

Correct. And don't use the AT or Inverter ground relay.

You will also need to make sure there is no ground bonding in the isolation transformer on teh secondary side. The chassi should be grounded to the boat's AC ground system, but there should not be an connection from the transformers neutral to ground.

When you think yo have it all set up correctly, you should be able to remove your central bonding wire and check for continuity between neutral and ground, and there should not be any. You should only have continuity if you reinstall the ground bonding wire.

But beware that with the Victron AT, you will have a bunch of circulating current anytime you are on shore power or generator. When I tested a 100A Victron AT, I measured 7A. That's a lot.

 

[BLUE STEEL]

Thank you!

V1-U2 remain jumpered together, but there should be no green wire grounds on those terminals. BUT, the generator chassis/frame should have a green wire AC ground.

This makes total sense to me now, and corresponds with the 2 wiring options for 240vac generator output from my OP.

The newer wiring diagrams (for the newer model) from Northern Lights confirm this a little more clearly.

 

[BLUE STEEL]

with the Victron AT, you will have a bunch of circulating current anytime you are on shore power or generator. When I tested a 100A Victron AT, I measured 7A

7A circulating current, at what voltage? Where will this value be measured from?

I believe the 32A and 100A Victron Autotransformers have the same windings, with a different main breaker rating. So conceivably this effect will be replicated equally on both models.

Hopefully this thread will be of use to others with similar hardware and configurations.

 

[twistedtree]

7A circulating current, at what voltage? Where will this value be measured from?

I believe the 32A and 100A Victron Autotransformers have the same windings, with a different main breaker rating. So conceivably this effect will be replicated equally on both models.

Hopefully this thread will be of use to others with similar hardware and configurations.

The test is to be on shore power, inverter in passthru, and no loads on the AT output. Out a clamp meter on the AT neutral and I predict you will measure about 7A. That is current circulating between the iso and AT transformers because of the difference in windings rations. The ISO wants to see 120V and 120V on the two legs, but the AT is trying to create something like 115v and 125v. The current is driven by the devices trying to “correct” each other.

With a symmetric AT, you will see maybe 1A of circulating current in a 50A circuit.