Earthing and N-E bonding

[Kumiho]

This subject has been mentioned in other topics but has got submerged in other discussions, so I've begun this new thread.

When operating as an EPS in island mode, typically to supply the whole house in the event of grid failure, there needs to be an effective earth and a Neutral-Earth bond.

The need for an earth that does not rely on the distributor's earth seems to be well accepted in these discussions. One cannot rely on the distributor's earth as, for example, the reason for the grid failure may be that the cable has been cut. However, that is easily sorted by installing a local earth rod connected to the main earth terminal on the incoming distribution board.

The need for N-E bonding is more subtle. When switching to island operation, the supply N-E bond is lost and so a N-E bond should be switched in at the new power source - the inverter. Several documents and YouTube videos show this being done by a separate relay or contactor, but it could equally well be performed by an additional switch section on the changeover switch.

This seems not to be provided on my installation, in island mode there is significant voltage neutral to earth, as others have noted. The inverter clearly does not make this bonding connection automatically

My question is, how have others addressed this issue?

[cosie]

Have you made any progress on this? I'm experiencing the same voltage difference between N and E when in EPS mode.

[Kumiho]

Hi Cosie,
No, I emailed my installer, suggesting that there ought to be a local supplementary earth and an N-E bond. I can install an earth spike myself so that was uncontentious, although my installer said that it is still a bit of a grey area in the world of distributed generation. I think he meant that they don't bother.

He also said that N and E are connected to the same point in the inverter, which is clearly not the case.

I've emailed twice more since, politely, but I have had no further replies. I suspect this is something they just haven't considered or come up with a solution to.

BS7671 and the IET Code of Practice 'Electrical Energy Storage Systems' are quite clear, both a supplementary earth and an N-E bond are required in island mode. The system then operates as TN-S in island mode.

When I get a bit of time to think through the possible fault scenarios I'll decide how bothered I am about the lack of N-E bonding. Generally though, I don't like the L and N system potential floating around relative to earth, it could cause damage to filter components etc. in electronic equipment.

I am an electronics engineer rather than in the electrical trade, but I do understand the concepts.

Incidentally, the IET Code of Practice is an excellent reference for battery + solar system practice, the diagrams are very clear and it is much more digestible than the wiring regs, BS7671. I'd recommend it to all installers and anyone with a technical interest.

I'd be interested to hear on this forum from any installers who have considered this issue

[tony.matthews1]

Came across this diagram in the Fox AIO installation guide, using internal switch mode:

Clearly shows direct bonding of EPS N to grid N (no switch). This avoids EPS floating and generating a signficant N-E potential.

Also shows EPS E not connected. I checked this and EPS E is bonded to grid E inside my H1-6.0-E, so my assumption is that this is so any external fault current is not sent through the inverter EPS connection.

Don't know if this helps?

Tony

[Kumiho]

Thank you Tony, that is interesting.

There are two similar diagrams in the AIO installation guide, one as you post showing permanent N-E bonding in the distribution box and the other showing no N-E link at all.

(The distribution box is probably the EPS box shown in a preceding diagram. This seems to be a Fox ESS accessory, a web search discovered a manual for it.)

The H1/AC1 User Manual on the Fox website also includes the same diagrams, although my paper manual that came with the inverter does not.

Both of these diagrams relate to installations where the EPS is permanently supplying a subset of loads that are required to be kept going on loss of grid, in effect a UPS for some loads. See Ryan Morgan's post 'What is EPS?', this would correspond to what Ryan calls Level 1 or Level 2.

My installation has a manual changeover switch to the whole house supply, so that when the grid dies, the battery and inverter will run the whole house, subject of course to not exceeding the maximum inverter power. This is the same as would be done with a backup petrol generator.
This corresponds to Ryan's Level 3.

The Fox ESS guides and the diagrams direct the installer to follow local (country) installation practice. The diagrams will suit some countries' practices but not others. In the UK, there should be only one N-E bond in the supply.

In a TN-S supply, this is at the DNO supply transformer. In a TN-C-S (PME) supply, there is a combined PEN conductor into the premises, which then splits as separated N and E.

The concern is that when the supply is lost and the PV/battery/EPS system is disconnected from the grid and operating as an island, it becomes a TN-S system. In other words, it takes the place of the DNO supply transformer. The single N-E bond should then be made at the battery/inverter 'generator' while, and only while, the system is operating in island mode.

My installer did get back to me, apologising that my emails had slipped through the net in a premises move. He promised to take this up with Fox and let me know the outcome, when I hear anything useful I'll post here.

[outofmyshed]

I have this same issue. Fox H1 inverter, Whole house EPS, manual changeover switch, TN-C-S system with local earth rod, but floating neutral of ~44V when in EPS mode. So there’s no N-E bonding being done in the inverter.

The question is what to do about it. Making a simple bond between EPS N and common E bar with the changeover switch enclosure wouldn’t allow EPS pass through (Bypass relay setting on) as this connects grid N to EPS N therefore introduces a second permanent N+E bond in the system which is forbidden and would cause RCDs to trip.

Unlike Sunsynk inverters, the H1 doesn’t have a 240V signal available to enact an external contactor. So the only sensible solution seems to be to have an NC single pole contactor (of 25A or so) to be actuated by grid L&N and connecting the earth and neutral bars close to the changeover switch, enacted by the grid deenergising and making a local N-E bond, allowing a safe switch of the changeover to EPS When the grid returns, the contactor would open, breaking the local bond.

Has anyone else solved this in another/better way in the UK with Fox inverters and EPS?

[Dave Foster]

Like you I have a local earth rod but only run the EPS circuits via a manual changeover switch.

I have a relay on the EPS output that automatically bonds the EPS N to Earth but only when the EPS circuit becomes active (grid lost)

[outofmyshed]

Like you I have a local earth rod but only run the EPS circuits via a manual changeover switch.

I have a relay on the EPS output that automatically bonds the EPS N to Earth but only when the EPS circuit becomes active (grid lost)

Thanks Dave, so do you have the inverter Bypass Relay disabled, and also isolate the inverter grid AC before flipping the changeover switch to avoid having two N-E bonds when the grid returns?

Could you sketch out a wiring diagram or describe how the relay is connected and actuated?

[Dave Foster]

Like you I have a local earth rod but only run the EPS circuits via a manual changeover switch.

I have a relay on the EPS output that automatically bonds the EPS N to Earth but only when the EPS circuit becomes active (grid lost)

Thanks Dave, so do you have the inverter Bypass Relay disabled, and also isolate the inverter grid AC before flipping the changeover switch to avoid having two N-E bonds when the grid returns?

Could you sketch out a wiring diagram or describe how the relay is connected and actuated?

Yes I use a manual changeover switch so I have to throw a big switch to enable my EPS driving the critical circuits, that means I can use a NO contactor which is energised ( the coil is connected to EPS L, N ) on the EPS going live which bonds the EPS N & E.

If you are thinking of using an automatic changeover switch or running EPS all the time for critical circuits you should install a NC contactor held open by the grid supply ( the coil is connected to Grid L, N ) - that way it will only engage and bond EPS N & Earth when then grid supply has been lost (at which time you will have lost the supply N+E bond).

[outofmyshed]

News just in on N-E bonding - the G2 inverters sold in the UK as of April 2024 now have an internal N-E bonding relay for EPS mode included. According to my installer (who was in contact with Fox UK on this issue) the first batch of G2s sent out didn't, but newer ones do. So if this is important to anyone installing in the future and requiring EPS, it's worth explicitly checking this.

There is no documentation that I can find that references this change, but I had an H1-5.0-E-G2 installed on Friday (upgrade from an H1-3.7-E G1) and the 170V I was getting between N-E when in EPS mode has gone. Using a socket tester confirmed this.

Interestingly the Bypass Relay setting has also disappeared - it's not clear why, but the EPS circuit appears to be permanently energised even when EPS mode

[Dave Foster]

Thanks for the update, be good to get that confirmed from Foxess and certainly good news... one less thing to worry about on EPS.