Cable size and how to earth this system?

[HappyHippyDad]

Morning....

I have a friend of a friend who is just looking for some advice regarding his 12V solar set up.

It's a metal shed approx 5m x 3m.

My 2 questions are:

1. Does the cable joining each of the 5 batteries have to be 50mm as well as the cable from battery to 3kW inverter? I'm sure it does but just wanted to be certain.

2. How to earth this system? I have drawn an earth rod coming from the 230V consumer unit but I realise this is no good by itself. I'm just not sure how this gets joined to the neutral (or which neutral) in order for it to become a working earthing system that would allow the RCD's to trip with earth leakage? (See picture below)

I realise I could have an earth free system (as I have in my wooden shed) but with this being a metal shed that didn't seem like a good idea.

The picture below only shows 1 solar panel, please ignore this as there will be more.

 

[davesparks]

When you say referenced to earth could you explain that a little more? 'Referenced' can't mean the same as 'connected' otherwise that would mean just connecting the same earth rod to the 'neutral in' terminal on top of the main switch in the CU.
How do I achieve this 'reference' to earth? How can the inverter already be referenced to earth if it is not in contact with earth in any form?

Referenced does in this case mean connected, you would be installing the N-E link at the point of your choosing. The easiest place is probably as you say, at the DB where you would link the neutral to the earth bar, this is the same as putting the N-E link in a large switchboard as seen in industrial or big commercial installations where they are fed from their own transformer.

 

[Lucien Nunes]

The inverter may already have a neutral-earth connection inside. Or, some other point of its internal circuit may be earthed, to the socket, the case or both. You need to check the manufacturers' data which might or might not say. The connection needs to be solid, and made at the source, then the resulting CPC referenced to earth with the rod. Don't earth via the DC side, treat that as floating although it might not be within the inverter.

As DS says, the problem with having an RCD without an earthed neutral is that it won't react when a single fault or latent leakage path occurs, nor will you get a shock, so you don't know that the fault exists. If it is downstream of the RCD as it is likely to be, i.e. at an appliance or in the wiring rather than in the inverter, then when another fault or leakage occurs to the other pole of the supply you have a fault path that doesn't upset the RCD balance, you do get a shock but the RCD still doesn't trip. What is needed is an earthed neutral so that whenever a fault occurs to the line, the RCD trips regardless of whether another fault or a shock path exists.

E2A - when we say earthed neutral, there are two aspects. Both the MET / CPC and the general mass of earth should be connected to the neutral. If one lived in a plastic bubble with no access to earth, shocks would only be to the CPC / casing of appliances rather than true earth. There would then be no need for the mass of earth to be brought into the equation, only the CPC and neutral would need to be connected in order to provide an equipotential and enable the RCD to detect line-CPC faults. But in the real world, where the mass of earth is accessible to the touch, the CPC should be connected to it via the MET to make the equipotential that of earth rather than some arbitrary voltage. Therefore the result is that the neutral is earthed, as well as being connected to the MET. The RCD will then also protect against a shock path from line to true earth, as well as to the CPC.

 

[HappyHippyDad]

ps... I should have drawn an extra line coming from the inverter output going to the earthing bar in the CU. i.e it would be a normal plug plugged into the inverter which is the supply to the consumer unit. Does this mean the earthing system would be effective if this earth is referenced to earth in the inverter?

 

[HappyHippyDad]

The inverter may already have a neutral-earth connection inside. Or, some other point of its internal circuit may be earthed, to the socket, the case or both. You need to check the manufacturers' data which might or might not say. The connection needs to be solid, and made at the source, then the resulting CPC referenced to earth with the rod. Don't earth via the DC side, treat that as floating although it might not be within the inverter.

As DS says, the problem with having an RCD without an earthed neutral is that it won't react when a single fault or latent leakage path occurs, nor will you get a shock, so you don't know that the fault exists. If it is downstream of the RCD as it is likely to be, i.e. at an appliance or in the wiring rather than in the inverter, then when another fault or leakage occurs to the other pole of the supply you have a fault path that doesn't upset the RCD balance, you do get a shock but the RCD still doesn't trip. What is needed is an earthed neutral so that whenever a fault occurs to the line, the RCD trips regardless of whether another fault or a shock path exists.

E2A - when we say earthed neutral, there are two aspects. Both the MET / CPC and the general mass of earth should be connected to the neutral. If one lived in a plastic bubble with no access to earth, shocks would only be to the CPC / casing of appliances rather than true earth. There would then be no need for the mass of earth to be brought into the equation, only the CPC and neutral would need to be connected in order to provide an equipotential and enable the RCD to detect line-CPC faults. But in the real world, where the mass of earth is accessible to the touch, the CPC should be connected to it via the MET to make the equipotential that of earth rather than some arbitrary voltage. Therefore the result is that the neutral is earthed, as well as being connected to the MET. The RCD will then also protect against a shock path from line to true earth, as well as to the CPC.

Referenced does in this case mean connected, you would be installing the N-E link at the point of your choosing. The easiest place is probably as you say, at the DB where you would link the neutral to the earth bar, this is the same as putting the N-E link in a large switchboard as seen in industrial or big commercial installations where they are fed from their own transformer.

How can you just join the neutral bar in the CU to the earth bar in the CU? I don't mean physically I mean that it would just cause the RCD to trip as you would have continuity between earth and neutral. In a domestic CU if I made a link between the earth bar and neutral bar this would be a N-E fault and immediate RCD trip.
New diagram below with the earth showing going from plug (inverter) to the earth bar in the CU.

 

[HappyHippyDad]

 

[Lucien Nunes]

How can you just join the neutral bar in the CU to the earth bar

You can't. Davesparks actually said:

link the neutral to the earth bar

Meaning the incoming neutral upstream of the RCD, not the neutral bar downstream of it.

Note that the RCD protects against faults on the opposite side of itself to the N-E link because that's how current bypasses it to create an imbalance. Want to protect the installation downstream? Put the N-E link upstream. Linking the neutral and earth bars in the CU downstream of the RCD won't protect the installation, only the inverter upstream of it. Such a link downstream of an RCD will cause an immediate trip on a DNO supply because there is already a link at the substation upstream. Having a link either side of the RCD allows some neutral current to bypass it and create an imbalance.

 

[davesparks]

How can you just join the neutral bar in the CU to the earth bar in the CU? I don't mean physically I mean that it would just cause the RCD to trip as you would have continuity between earth and neutral. In a domestic CU if I made a link between the earth bar and neutral bar this would be a N-E fault and immediate RCD trip.
New diagram below with the earth showing going from plug (inverter) to the earth bar in the CU.

You make the link before the RCD. If you make the link after the RCD it won't work, the RCD certainly wouldn't trip.
If you did it in a domestic installation its completely different, the DNO have already made the N-E link at the substation.

The N-E link being discussed here is effectively the equivalent of the N-E link normally made at the substation by the DNO.

 

[HappyHippyDad]

You can't. Davesparks actually said:

Meaning the incoming neutral upstream of the RCD, not the neutral bar downstream of it.

Note that the RCD protects against faults on the opposite side of itself to the N-E link because that's how current bypasses it to create an imbalance. Want to protect the installation downstream? Put the N-E link upstream. Accidentally putting the N-E link downstream won't cause a trip if there's no link upstream (as there would be on a DNO supply), but now the RCD is protecting against faults in the inverter upstream, not the installation.

Oh I see!! So just like a domestic installation where the neutral bar only has continuity with the earth bar when the main switch is closed as it's connected at some point upstream.
So my connection could be between the neutral in the incoming terminal of the main switch (i.e top) in the CU and the earth bar of the CU?
Could I not just check for continuity between the earth and neutral in the inverter by doing a simple continuity check? I'm guessing not as otherwise you would not have said to check the manufacturers data.

 

[HappyHippyDad]

You make the link before the RCD. If you make the link after the RCD it won't work, the RCD certainly wouldn't trip.
If you did it in a domestic installation its completely different, the DNO have already made the N-E link at the substation.

The N-E link being discussed here is effectively the equivalent of the N-E link normally made at the substation by the DNO.

Same question to you Dave as asked above to Lucien.

I appreciate you both taking the time to explain this to me.

 

[Lucien Nunes]

Could I not just check for continuity between the earth and neutral in the inverter by doing a simple continuity check

You could, but how would you know whether the continuity is via a small interference suppression component on the PCB, or a heavy cable able to withstand the full output. Chances are, if they are connected, it's OK to put an external link anyway, but there is no guarantee on a consumer-grade inverter how things have been done. If you buy a high-end product such as a Victron or Mastervolt, it will be well engineered and documented.

 

[HappyHippyDad]

You could, but how would you know whether the continuity is via a small interference suppression component on the PCB, or a heavy cable able to withstand the full output. Chances are, if they are connected, it's OK to put an external link anyway, but there is no guarantee on a consumer-grade inverter how things have been done. If you buy a high-end product such as a Victron or Mastervolt, it will be well engineered and documented.

I think (I'll use the term loosely) I may be nearing the end of the questions!

So, It would not be a good idea to rely on testing for continuity in the inverter between N and E, therefore i could use a suitable cable to join the neutral and earth as I described in the last post. However, if the inverter does not have a neutral to earth connection in it already is it ok to introduce this connection (i.e N-E link)? Or do all inverters have some form of link between N-E.

 

[Lucien Nunes]

Many cheaper units will have their output floating because they are intended to be used as IT, i.e. with nothing earthed, supplying a single piece of equipment only. This is generally safe as two faults on a single load appliance, from opposite poles of the supply, that do also not connect together within, is very unlikely indeed. Such inverters sometimes state that they must not be used to feed a distribution board or installation. Whether any RFI suppression is defeated by an external link, whether 230V is then imposed on something that doesn't like it, is another matter. It will probably be fine but I'm not prepared to take a wild guess.

 

[HappyHippyDad]

Many thanks to both of you (Lucien and DS), it's been an interesting morning