Design Review - Growatt SPF 5000 ES Grid Backup & Neutral Bonding

[automatikdonn]

How does the ground relay play into the SPF 5000 ES neutral protection discussion, if at all? The relay appears to be designed to work with their units only. I’m curious to hear how you'd integrate the Victron AT into your system…

From the manual:
“2.4. Ground Relay
In most energy systems for residential or mobile use a TT configuration is used, which means the Neutral of the AC system is grounded. This is to ensure the GFCI breaker will trip in case there is an earth failure (The connection N/PE will ensure a current will flow in the secondary circuitry which is detected by the GFCI).


As an autotransformer is also able in some systems to “create” a neutral different from the incoming neutral, a ground relay is included in the unit. This relay can be controlled by a Multi/Quattro (by disabling the internal ground relay in the inverter/charger through VE Config or Victron Connect and interconnecting the “ground relay” terminal block in Multi/Quattro and Earth terminal block in Autotransformer by Positive and Negative conductors. In some of the diagram examples in Section 2.4.1 through 2.4.3 it is mentioned when it is possible to use the ground relay.”

Of course the built in temp protection is an advantage as well…

In my particular setup, I’m not using any grid input, because it causes problems I just don’t have time to solve at the moment. The ground bond is still done in my main service panel. I use a protran/2 panel to connect from solar to circuits.

I know a lot of people aren’t a fan of manual switches, but I think a properly sized and architected system should have very little interventions needed.

As for handling things with the victron at, I would have to look at the relay function of it. I believe it gets the proper signaling from a victron inverter so the grounding is handled appropriately. Victron makes some really nice stuff. I have a 6kw setup for my RV that is almost entirely victron.

If that unit is just looking for a NO/NC contact, you can easily get the growatts to support that function.

 

[Hedges]

Having thought through the whole Auto-Transformer thing A LOT, I've come up with what I think is the best, and KISS, solution:

Don't use an auto-transformer. Use an isolation transformer. Take in 240V from L1/L2 of the inverter and convert to isolated 120/240V split-phase. Bond the center-tap (neutral) to ground, same ground as main panel. Transformer VA rating must be sufficient for full inverter output and full pass-through from grid. It just works.

I haven't read through all pages and all 261 previous posts of this thread, just some pages. But I did all of a couple other threads on the topic, and found multiple safety and equipment hazards.

I may yet add an auto-transformer to my system, but for other purposes. Not to establish a neutral; the inverters are stacked 120V each. Auto-transformer would be for load balancing, and to create missing phase if all inverters except one master area allowed to sleep for power savings. It would be disconnected when on-grid.

 

[ChrisFullPower]

Having thought through the whole Auto-Transformer thing A LOT, I've come up with what I think is the best, and KISS, solution:

Don't use an auto-transformer. Use an isolation transformer. Take in 240V from L1/L2 of the inverter and convert to isolated 120/240V split-phase. Bond the center-tap (neutral) to ground, same ground as main panel. Transformer VA rating must be sufficient for full inverter output and full pass-through from grid. It just works.

I agree with that being the best solution but the whole reason this problem even exists is because people are trying to save a buck and get a cheap inverter. I don’t think spending twice as much for a proper transformer is something they’re really interested in doing.

 

[neuropa3]

How about a schamatic without grid tie. My system will be totally OFF Grid.

 

[rick777]

Having thought through the whole Auto-Transformer thing A LOT, I've come up with what I think is the best, and KISS, solution:

Don't use an auto-transformer. Use an isolation transformer. Take in 240V from L1/L2 of the inverter and convert to isolated 120/240V split-phase. Bond the center-tap (neutral) to ground, same ground as main panel. Transformer VA rating must be sufficient for full inverter output and full pass-through from grid. It just works.

I haven't read through all pages and all 261 previous posts of this thread, just some pages. But I did all of a couple other threads on the topic, and found multiple safety and equipment hazards.

I may yet add an auto-transformer to my system, but for other purposes. Not to establish a neutral; the inverters are stacked 120V each. Auto-transformer would be for load balancing, and to create missing phase if all inverters except one master area allowed to sleep for power savings. It would be disconnected when on-grid.

Sounds like an interesting solution. I'm a total neophyte at electrical work and while I think I understand the design by automatikdonn and others on this forum, it's too much for me to try to tackle at my current level of knowledge. If you have any interest in trying to flesh out your isolation transformer solution, I would greatly appreciate it. I suspect I'm not alone on that score. Any ballpark idea of cost if I were to use three 5Kes' with 15KW capacity?

Although I don't think I can use the auto-transformer design, I really want to thank the guys who put that together. It was a great learning exercise for me!

 

[Hedges]

Autotransformer has the benefit of higher efficiency - only power converted between 240V/120V goes through it and encounters loss. Whatever passes straight from inverter to load does not. Also only has to be half as big, because only half the power goes through it. That saves weight and $$.

You can find lots of transformers at various vendors. look for 240/480V primary, 120/240V secondary. At least 15kVA for your situation. Copper is better, but aluminum is cheaper. Many vendors offer used as well, can cost less.

Something like this, for instance:

SQUARE D SORGEL 15KVA 1PH TRANSFORMER 15S3HIS HV 240X480 LV 120/240 | eBay

Find many great new & used options and get the best deals for SQUARE D SORGEL 15KVA 1PH TRANSFORMER 15S3HIS HV 240X480 LV 120/240 at the best online prices at eBay! Free shipping for many products!

www.ebay.com

A toroid transformer would be more efficient. I saw a 10kVA one with list price $1000. But it wasn't packaged, ready to go.

 

[danbussep3]

Sure it would. You have a couple methods to do it, easiest method being to get a 4 pole contactor. Just make sure the breakers and wiring is sized appropriately for your workload.

This one is a bit speny but is UL. The next version down would probably be just fine too. If your workload runs these inverters at 100% load 24/7, I would be looking at an entirely different setup. ( of which i have a very large power system in the works)

McMaster-Carr

McMaster-Carr is the complete source for your plant with over 595,000 products. 98% of products ordered ship from stock and deliver same or next day.

www.mcmaster.com

What about this contactor?

McMaster-Carr

McMaster-Carr is the complete source for your plant with over 595,000 products. 98% of products ordered ship from stock and deliver same or next day.

www.mcmaster.com

 

[rick777]

Autotransformer has the benefit of higher efficiency - only power converted between 240V/120V goes through it and encounters loss. Whatever passes straight from inverter to load does not. Also only has to be half as big, because only half the power goes through it. That saves weight and $$.

You can find lots of transformers at various vendors. look for 240/480V primary, 120/240V secondary. At least 15kVA for your situation. Copper is better, but aluminum is cheaper. Many vendors offer used as well, can cost less.

Something like this, for instance:

SQUARE D SORGEL 15KVA 1PH TRANSFORMER 15S3HIS HV 240X480 LV 120/240 | eBay

Find many great new & used options and get the best deals for SQUARE D SORGEL 15KVA 1PH TRANSFORMER 15S3HIS HV 240X480 LV 120/240 at the best online prices at eBay! Free shipping for many products!

www.ebay.com

A toroid transformer would be more efficient. I saw a 10kVA one with list price $1000. But it wasn't packaged, ready to go.

Thank you.

 

[automatikdonn]

What about this contactor?

McMaster-Carr

McMaster-Carr is the complete source for your plant with over 595,000 products. 98% of products ordered ship from stock and deliver same or next day.

www.mcmaster.com

That will surely work.

 

[Taulli1]

Well I have a youtube video I put together as a demo. I only have the neutral leg disconnect circuit together at the moment. The video is still processing, but here is the link.

This is a prototype and not even close to being ready for production. Please heed this warning.

I think this circuit covers two events. Loss of neutral and loss of one leg on the 240v circuit. The voltage protection relay is monitoring the loss of a neutral leg, while the contactor needs 240v to operate. The loss of proper service in any case is handled in this circuit design. At least from parts that I know of that exist in the real world. There is also purpose in having two auto transformers, mainly being yet another mechanism to mitigate a disaster. I highly doubt I will use even the one to full potential, the second transformer is there mainly for the sake of redundancy. They are so cheap I would feel dumb for not having two.

Would love to get some feedback on what everyone thinks.

Next up I will be giving the auto bonding a go round.

Can you confirm or correct what I have found after researching this topic?

Main (grid) panel, has G/N bond. Each sub panel must not be bonded.

Separate power source panels, a generator load distribution panel, for instance, should have G/N bond. And a solar inverter is the same, so a separate ground rod, with a G/N bond at the primary distribution panel.

This is what I found from NEC discussions on the topic, I am not stating as fact, just observation. What do you think?

I love your circuit design btw, planning to copy it.

 

[cdevidal]

Having thought through the whole Auto-Transformer thing A LOT, I've come up with what I think is the best, and KISS, solution:

Don't use an auto-transformer. Use an isolation transformer. Take in 240V from L1/L2 of the inverter and convert to isolated 120/240V split-phase. Bond the center-tap (neutral) to ground, same ground as main panel. Transformer VA rating must be sufficient for full inverter output and full pass-through from grid. It just works.

I haven't read through all pages and all 261 previous posts of this thread, just some pages. But I did all of a couple other threads on the topic, and found multiple safety and equipment hazards.

I may yet add an auto-transformer to my system, but for other purposes. Not to establish a neutral; the inverters are stacked 120V each. Auto-transformer would be for load balancing, and to create missing phase if all inverters except one master area allowed to sleep for power savings. It would be disconnected when on-grid.

What are some isolation transformers that would be recommended for a 3000W 120V load?

 

[Hedges]

Here's the first suitable match I found on eBay.
2x 240V primary windings, 2x 120V secondary windings, 3000VA.
However, I'm not sure secondary windings in series for 120/240V split-phase could deliver 3000VA at 120V. Certainly at least 1500VA.
If secondary windings are put in parallel it will deliver 3000VA at 120V only, no isolated 240V.
Maybe look for 6000VA if you want to load with 3000VA 120V one moment, 3000VA 240V the next.

Daykin Transformer 3000va 240/480-120/240v TF3000EP | eBay

Find many great new & used options and get the best deals for Daykin Transformer 3000va 240/480-120/240v TF3000EP at the best online prices at eBay! Free shipping for many products!

www.ebay.com

 

[fmeili1]

I just found this (new?) off-grid EG4-3000EHV-48 inverter. This offers also an integrated bypass mode for grid used as backup with the following feature (see manual page 7).

"When the inverter is working in PV mode, battery mode, or standby mode, the output neutral is connected to the ground of AC input (neutral/ground bonded). When the inverter is working in AC mode, neutral of output is disconnected to grounding of AC output and connected to neural of AC input."

I think, this is exactly what we are looking for. It's a 120V inverter which could be stacked up to 12 units also in split phase mode. So no auto transformers are required and the neutral/ground bonding problem does not longer exists. For a large (complete house) setup e.g. with about 25kW inverter power it would cost just $800 bugs more (instead of 5x SPF5000ES for $900 each + 2x auto transformers for $350 each). It would cost 8x $750 to get nearly the same power (24kW instead of 25kW for the SPF5000US). No hassle with the additional wiring for the auto transformers and the complicated problems with the ground/neutral situation with SPG5000US (specially since option 24 is not longer available to be able to find out if the SPF5000US is running in bypass or not to control an external neutral/ground relay). From my understanding this will be a much easier and safer setup.

What do you think?

 

[Solar Guy Richard]

I just found this (new?) off-grid EG4-3000EHV-48 inverter. This offers also an integrated bypass mode for grid used as backup with the following feature (see manual page 7).

"When the inverter is working in PV mode, battery mode, or standby mode, the output neutral is connected to the ground of AC input (neutral/ground bonded). When the inverter is working in AC mode, neutral of output is disconnected to grounding of AC output and connected to neural of AC input."

I think, this is exactly what we are looking for. It's a 120V inverter which could be stacked up to 12 units also in split phase mode. So no auto transformers are required and the neutral/ground bonding problem does not longer exists. For a large (complete house) setup e.g. with about 25kW inverter power it would cost just $800 bugs more (instead of 5x SPF5000ES for $900 each + 2x auto transformers for $350 each). It would cost 8x $750 to get nearly the same power (24kW instead of 25kW for the SPF5000US). No hassle with the additional wiring for the auto transformers and the complicated problems with the ground/neutral situation with SPG5000US (specially since option 24 is not longer available to be able to find out if the SPF5000US is running in bypass or not to control an external neutral/ground relay). From my understanding this will be a much easier and safer setup.

What do you think?

Personally I love it! ....not biased at all

 

[Averageaviator]

After all this mess with my 5000ES's and grounding, I wish Signature solar would let us return them......I'm exhausted from trying to make these work safely and without extensive/expensive solutions like the ones in this thread. Maybe they would let us swap them for the new EG4 inverters ??

 

[Struc]

After all this mess with my 5000ES's and grounding, I wish Signature solar would let us return them......I'm exhausted from trying to make these work safely and without extensive/expensive solutions like the ones in this thread. Maybe they would let us swap them for the new EG4 inverters ??

Why do you think those are any better? They also bond neutral & ground while on battery and release it when on grid power. Pretty much every inverter on the market today has some varying amount of trouble involved with the neutral & ground bond.

 

[fmeili1]

Why do you think those are any better? They also bond neutral & ground while on battery and release it when on grid power. Pretty much every inverter on the market today has some varying amount of trouble involved with the neutral & ground bond.

In my understanding, the important difference is that the SPF5000US has a permanent connection between "AC in GND" and "AC out GND" (in all modes). The EG4 disconnects the GND in and GND out this if it runs in bypass mode. While in inverter mode the "GND out" could be connected with the grid AC in (hot1 and hot2) because the hot's are not connected so the "GND in" which is in fact a neutral&GND could be connected with the auto transformers neutral without any issue. The problems arise only if the inverter switch to bypass and if the two neutrals (one from the grid and one from the autotransformer) will stay connected, the both transformers (the pole transformer and the auto transformer) will work in parallel, which is a problem (like already discussed).

In case the SPF5000US would provide a "clean" external signal like a dry contact which only activates if inverter runs in bypass mode would allow to implement this disconnect externally with a relay. But unfortunately the existing dry contact of the SPF5000US does not provide this signal (only batt low signal).

 

[Hedges]

Why do you think those are any better? They also bond neutral & ground while on battery and release it when on grid power. Pretty much every inverter on the market today has some varying amount of trouble involved with the neutral & ground bond.

It seems that inverters with neutral connection, either European models with 230V line + neutral or US models with dual 120V inverters for L1 + L2 + neutral are fine. Only problem seems to be using a 230V floating L1 + L2 inverter for on & off grid split-phase operation. That can be made to work with isolation transformers.

I understand some of the (MPP? GroWatt?) inverters have two 120V inverters inside which should be fine. Other big names like Outback and SMA have individual 120V inverters which are stackable for split-phase or 3-phase. Victron appears to have a good solution for connecting floating 230V inverter to split-phase.

So the only problem is some low-cost imports repurposed for US split-phase grid.
Mobile systems of course have a particular UL standard to comply with which addresses bonding of neutral to ground internally only when shore power isn't present.
Good off-the-shelf solutions exist. Sow's ears can be turned into silk purses with varying difficulty.

 

[Struc]

In my understanding, the important difference is that the SPF5000US has a permanent connection between "AC in GND" and "AC out GND" (in all modes).

You're confusing the ground and the neutral. The SPF5000US does not have a connection between the AC in Ground and AC out Ground.

The problem with the design of this system (and a lot of other systems) is the neutral / ground bond. There should only be 1 in any electrical system. That's a problem in the design, depending on if you are running in bypass mode off the grid, or running in inverter mode off the batteries. This situation happens with many inverters on the market today, including the one you mentioned. If you are running 100% off grid, then the problem can be minimized.

 

[fmeili1]

You're confusing the ground and the neutral. The SPF5000US does not have a connection between the AC in Ground and AC out Ground.

The problem with the design of this system (and a lot of other systems) is the neutral / ground bond. There should only be 1 in any electrical system. That's a problem in the design, depending on if you are running in bypass mode off the grid, or running in inverter mode off the batteries. This situation happens with many inverters on the market today, including the one you mentioned. If you are running 100% off grid, then the problem can be minimized.

Hi Struc,
mhh, that's interesting. Please can you clarify and/or verify some details which I thought, are right. Maybe you can point me to the points, where I'm wrong. Here are my assumptions and my understanding so far:

1. From the pole, there is a wire "Hot1", a wire "Hot2" and one "neutral&GND" wire.
2. Because, the neutral from the pole is already grounded at the pole, it's a neutral&GND wire which comes into the house (and not only a neutral).
3. The Hot1 (or Hot2) from the pole is connected (through the meter) to the "AC in" clamp with the label "L" of the GW inverter.
4. The Hot2 (or Hot1) from the pole is connected (through the meter) to the "AC in" clamp with the label "N" of the GW inverter.
5. The neutral/GND from the pole is connected to the "AC in" clamp with the label "GND" of the GW inverter.
6. In the US model of the GW SPF5000US there is no connection between the "AC in" clamp labeled "N" and the "AC in" clamp labeled "GND" (which is required and important here in the US).
7. In the GW inverter, there is a permanent connection (independent of the current running mode of the GW inverter) between the "AC in" clamp labeled "GND" label and the "AC out" clamp labeled "GND".
8. The GW inverter clamp "AC out" labeled "L" is connected to one side of the auto transformer.
9. The GW inverter clamp "AC out" labeled "N" is connected to the other side of the auto transformer.
10. The middle connect of the auto transformer will be the neutral, which is created by the auto transformer.
11. The GW inverter clamp "AC out" labeled "GND" is connected to middle connect of the auto transformer.
12. The 3 wires from the auto transformer (Hot1, Hot2, neutral) runs to the existing main breaker panel of the house
13. The existing main breaker panel of the house creates (split) the neutral to a separate neutral and GND which runs to all consumers in the house (this should be the only point in the house where the neutral/GND should be split in two separate clamps AND grounded to the earth with a stake).
14. This works great and with no issues, if the inverter is running in inverter mode (because HOT1 and HOT2 from the pole are not connected at all with the house)
15. But when GW inverter is in bypass mode, the HOT1 and HOT2 form the "AC in" clamps are switched directly (via an internal contactor of the GW) to the "AC out" terminal labeled "L" and "N".
    1. In this mode, the pole transformer is parallel connected to the auto transformer which is bad.
    2. To solve this problem a neutral/GND disconnect relay/contactor would be required to disconnect the auto transformer middle clamp (self generated neutral) from the rest of the system - but only if running in bypass mode!
    3. In my understanding, the GW inverter does not disconnect the "AC in" clamp labeled "GND" from the "AC out" clamp labeled "GND" when it runs in bypass mode to achieve this.
    4. The GW inverter also does not provide an external dry contact mechanism which signals "I'm running in bypass mode now" to be able to implement this neutral disconnect feature with an external circuit. This is in my understanding the core of the problem.

Please tell me in which of these points I'm wrong.

Thanks a lot for clarification.

Frank