Running power to a garden terrace

[martinpalmer8]

I have run some 4mm=B2 twin & earth from a 20A MCB in my consumer unit to a double-insulated switch box by my back door. From the switch box I have run 4mm=B2 SWA cable to a remote garden terrace (30m away).

I now wish to power up a 13A socket, water pump and set of lights on the terrace. The socket, 2-pole pump switch, and lights' FCU will be connected in series at the terrace end of the cable.

I want to RCD-protect the socket, pump and lights, and have driven a

1=2E2m copper earth rod in the corner of the terrace.

I would like to connect this earth rod to the armour of the SWA cable at the terrace end. I would also like to connect the armour at the back door end to the earth terminal of a RCD in the switchbox. This one RCD will protect all the terrace circuits. Is this legitimate?

Or must I run an earth cable from the terrace earth rod back to the switchbox RCD earth?

Or should I have the RCD local to the terrace?

[Owain]

They'll be connected in parallel.

What calculations/tests have you done to demonstrate that this earth electrode is satisfactory? Or are you relying on the Building Control Officer to check this for you under Part P?

RCDs don't have earth terminals.

The usual way of doing this is to earth the cable armour to the house earth meaning the circuit protective conductor is earthed at the supply end. An insulating gland is used at the garden end to isolate the cable armour, and the garden sockets etc are locally earthed.

No.

You can put the RCD at whichever end of the cable you like.

Owain

[martinpalmer8]

You're right - I expressed myself incorrectly. I meant to say that they would be linked on the same radial circuit.

The installation will be done by a NIC contractor. He will provide the necessary certification. If he's doing his job properly, I'm hoping that he will test the electrode.

You're right. My mistake.

OK, but... if a manic gardener cuts digs a pit 0.5m deep next to the terrace and accidentally shorths the swa sheath to live (very hypothetical, I realise), wouldn't it be better for the sheath's earth to be close by. The consumer unit earthing is PME - would that be a shorter ohmic path?

[Chip]

On 1 Jul 2005 13:53:35 -0700,it is alleged that [e-mail] spake thusly in uk.d-i-y:

Not at all hypothetical, it's amazing what a garden fork can do:-)

With a modern CU with type B circuit breakers, a short of that nature should pop the breaker pretty fast, IF the armouring is earthed to the main earth terminal at the feed end.

Assume the scenario that you have the RCD at the load end, therefore the cable is only protected by the MCB, and the armouring earthed to a local rod at the load end, rather than correctly to the main earth terminal in the house as Owain suggested.

It is very unlikely the current would trip the MCB, it would raise the earth of the installation above true ground potential /from the NON rcd protected supply/ and generally be A Bad Thing involving shockage and potential for lethality.

*NB, I tend to take the worst case scenario and run with it. I realize this is highly unlikely, still, if it does, the correct method will avoid the (unlikely) hazard.

[martinpalmer8]

Doesn't that depend on the length of the cable though? A live-to-earth fault at the very end of a long SWA cable could also produce a shock voltage relative to the consumer unit earth. I thought that that was why earth rods were used for outdoor buildings remote from the main building?

My system is TN-C-S by the way.

[John Rumm]

Yup, you could do the sums to be sure (don't have my on-site guide handy or I could look up the figures), but you only need 100A fault current to open a 20A type B breaker instantly (i.e. in its magnetic trip region).

If the screen is going to be connected to the earth rod, then the RCD must be at the souce end of the cable. Otherwise you get the potential problem you highlight where there is not enough phase to earth current to open the breaker.

[bigcat]

how to make the simple complicated.

Length of earth conductor has nothing to do with it. Impedance does, and a local earth rod will be the poorest performer in this respect.

NT

[martinpalmer8]

I'm seriously misunderstanding something here. I thought that the longer the length of the protective conductor from its connection to earth, the greater the earth loop impedance.

The consensus in other threads on this newsgroup is that earth rods should be used for remote locations.

What's going on?

[John Rumm]

This is generally true (at least for earth conductors of the same cross sectional area).

However, you can look at the total earth loop impedance as the sum of two parts: the total resistance of the earth conductor itself, plus the impedance of the earth connection.

In the case of a supplier provided earth you can be fairly sure that the impedance of their earth will be under 1 ohm (possibly well under for a new TN-C-S install).

For an earth stake driven into the ground on a TT setup it is not uncommon for its impedance to be in the tens or even hundreds of ohms range. (the best our earth stake here can manage is about 11 ohms at best)

As a result you are usually required to use an RCD with a TT setup in order to meet the required disconnection times since you will not be able to get enough fault current flowing to earth to open a conventional current operated protective device.

This is true, and is often recommended for a number of reasons. If you export the house earth then you need to also export / create the same equipotential zone at the remote location as you have in the house. This will not always be easy. The long cable lengths can also force loop impedances to creep up to the point where it is not longer to safe to rely on traditional earthing and equipotential bonding alone to give you the required disconnection times in the event of a fault. This coupled with the good liklihood that remote power sockets will also power appliances used outside mandates the use of an RCD for these setups. If you have that anyway then you may as well go TT and save some of the hassle.

[martinpalmer8]

Thank you, John.

Under normal conditions, it seems that the consumer unit earth impedance on my TN-C-S will always be much lower than the earth electrode, so it would be better to use the electricity board's earth in the garage. BUT because there is the slim chance that the electricity board's neutral supply could become disconnected, I can't use their earth in the garage (or anywhere else outside of the house's equipotential zone.)

I am assuming that the garage's earth musn't be connected to the house earth for a PME system, otherwise a "loss of neutral fault" would put a lot of current through the earth electrode.