Array frame earthing

[paulsamuel1984]

Hi all,

Quick question on earthing.

I am installing solar panel on 2No farms.

1st farm is metal shed, it it is a PME supply. I have bonded the array frame to the building structure (didn't see the need in separate spike).

2nd farm is a wooden shed, metal roof. Again a PME supply. I have bonded the array frame to a separate earth spike.

I am aware that PME supplies are not recommended on agricutlural premises.

Is this an acceptable solution for the earthing arrangements at both sites?

PS

 

[BruceB]

It looks sensible, but it is not trivial and to an extent depending on how you earth/bond decides which sort of fault or mistake you are protecting against and you can rarely protect everyone against everything.

On the first, the array framework needs to be in good contact with the metal of the shed (to ensure no voltage difference) and the metal of the shed needs to be bonded to MET if PME. I would say it is all within EPZ.

On the second, again the array framework needs to be in good contact with the metal of the shed roof (to ensure no voltage difference). If you can touch the metal roof from within the EPZ then arguably that makes the roof and array within the EPZ and you could/should bond to MET rather than separately to a stake.

There is always an element of judgement for the person on site in these situations.

Presumably you have 300mA rcd also.

Regards
Bruce

 

[paulsamuel1984]

Thanks BruceB,

I agree, this is far from trivial.

Yes, we have a 300mA RCD installed. Also meeting the 100mA per inverter recommendation by the inverter manufacturer.

But I am having some issues with the RCD tripping out, the installers have however installed a regular RCD, wheras I would like to see a type A as recommended. Could this be the reason for the tripping?

Any suggestions to avoid the nuicanse tripping and what the cause could be?

PS

 

[BruceB]

Do not misread the inverter manufacturers instructions. They say if you have to have an rcd then it should be at least 100mA, but they do not require one. A single 300mA rcd is likely to be fine. Is it that one tripping?

 

[paulsamuel1984]

Yes, I am aware that it is not a requirement to have an RCD from SMA.

Yes it is the one that is tripping. As I said, currently we have not yet got a type A installed. We have, in fact installed over 3No sites. 2 on metal sheds (bonded to shed), the other on the wooden shed (rod into ground).

We have had the RCD trip out on the 2No sites on the metal sheds. Only once about 2 hours after initial power up. The first thing that I am thinking is quality of earth we are obtaining from bonding to the metal shed??

Any thoughts?

 

[BruceB]

Nor direct experience to offer I regret, but there is a good chance you are on the right lines. I believe there have been reports on here of SMA TL inverters being sensitive to quality of earths and quality of bonding. So that would be a good place to start.

One or two of the suppliers on here seem knowledgable on type A/B/AC rcds so perhaps they will offer something. Nothing leaps to the front of my mind though.

Regards
Bruce

 

[SibertSolar]

@samuel (Paul?) - if the RCD is tripping then it's likely that your focus should be on the earthing as, like Bruce, I believe you are on the right lines there. The RCD tripping suggests that there is a higher than expected leakage current present, or there's an AC leakage current fault - the same will happen to a Type-A RCD anyway I would suspect. It's unlikely to be an excessive DC leakage current that's causing the problem as this would actually saturate the trip coil of a Type-AC RCD and cause it to never trip under actual fault conditions at all.

I think the first thing to do is check for any potential difference between the shed metalwork bond and the MET

Is there anything else on the circuit that the RCD is protecting?

 

[paulsamuel1984]

Thanks for the response Andy / Bruce,

Nothing else on the circuit that the RCD is protecting.

Moving the subject of the original post slightly, I have been talking to SMA regarding types of RCD to use and I have queried the following.

BS 7671 712.411.3.2.1.2 states: where the converter is, by construction not able to feed d.c. fault currents into the electrical installation, an RCD of type B according to IEC 62423 is not required.

SMA declaration states that the inverters are not capable of feeding-in DC failure currents to the grid due to the circuit design.

the SMA declaration then goes on to talk about type A RCD's. What I am questioning, and no one from SMA can answer, if the inverter is not capable of feeding-in DC failure currents then why the need for a type A???

Surely a type AC RCD sould suffice, in accordance with 7671 and the new guidlines (soon to be released).

PS

​