Hi John

Thanks for taking the time, once again

John Rumm coughed up some electrons that declared:

Tim S wrote:

snip general agreement

At (or for that matter in) the main CU seems sensible to me.


Splendid. Until I raised the subject on the IET forum, I was going to leave
the MET back at the supply origin (because that's where it always is). As
BT used to say, it's good to talk! I'm completely happy I understand why it
should be at the CU instead now.

However, that leaves the reverse problem at the other end, eg big fault,
pipes go up to 100+V, exposed conductive parts like the switch-fuse case
don't. Ditto pipes outside and in shower see the 100+V potential
difference with respect to the floor slab.

Maybe I am missing something, but I am not sure I follow what the
problem is...

Your switch fuse and CU will be linked by something like 52mm^2 of
earthing conductor[1] so any voltage rise on the CU's earthing terminal
is going to be reflected on the casework of the switchfuse pretty
closely. (I guestimate somewhere just north of 5V for a fault current of
1kA)


[1] (assuming XLPE 3 core with armour area alone of 62mm^2 giving a
copper equivalent of that over 2.255 or 27.5mm^2 before you add the
25mm^2 of the core)

You're using a k1/k2 conversion for Cu equiv CSA? I was taking my data from
Prysmian's BS5467 datasheet which says that (for 25mm2 3-core)
conductor@20C is 0.727 milliOhms per meter and armour@20C is 2.5
milliohms/m, but the end result is in the same ballpark.

Think I had a brain spasm. The R2 of my distribution circuit is about
9milliOhms, so I agree (now that I see) that the max PD end to end is 7.5V
at my max fault current (790A due to Ze mostly).

Someone else confused me by looking at R1/R2 like a potential divider with
230V across it, hence the 100 odd volts.

*** So should I bond gas and water to the means of earth back at the
meter as well, directly to EDF's earth terminal? BTW - this is a TN-S
setup, Ze of 0.19 Ohms

Can't see the need. There are quite a number of mitigating factors here
as I see it. Firstly the potential difference (in both senses) between
CU and EDF's terminal is small. The chances of being in contact with the
metalwork of the switchfuse while a fault is present is small. The
chances of any fault taking even a second to clear on a 17th edition
type install that is festooned with RCDs is pretty slim anyway, let
alone 5 secs.

Said like that, I agree. The only fault that could take 5s to clear is a
nail though the distribution cable - but that's an unlikely fault.

I came up with another way round: put a Type S 100mA DP RCCD in place of
the main switch in the switch fuse enclosure, and thus reduce the
disconnection time of the distribution circuit to 0.15s which avoids all
these issues. Unconventional with a TN-S system, but harmless at worst...

*** Overkill? Or not?

Probably is overkill in this circumstance.

Random thoughts gratefully received...

Random, and worth what you paid for em ;-)


Excellent - this and BigWallop's post is exactly the sort of clarity I was
after.

Amazing how easy it is to get all complicated and confused...

Many thanks indeed - I shall now order my parts. TLC will be pleased...

Cheers

Tim