"David" wrote in message
om...
Hi
Probably bad practice, but when I'm working on the wiring in my house,
rather than switching off the entire supply and incurring the wrath of
family who apparently can't function without electricity, I tend to
just switch off the circuit I'm working on.

What puzzles me, however, is that if I happen to short out the
earth/live/neutral wires (not sure which) in the dead circuit I'm
working on, the RCD trips the whole house off. Why is this? Is there
some infinitesimal residual current in the earth? Does it mean
there's a fault in my system somewhere?

By "switch off the circuit" you will mean removing the fuse/switching off
the live feed only. In this situation the neutral is still connected. The
neutral line will have a few volts on it which is a natural feature of the
generation and distribution system. Shorting neutral to earth (whether or
not the live is isolated) will cause a mismatch between the current in the
live side and the neutral side of the supply to the house. This mismatch
(which can only be accounted for by a leak to earth (either direct or via a
higher resistance path, such as you) is what the RCD detects. Yes, it is
very annoying when you are wiring in the loft! You can only get round this
by isolating the live and disconnecting the neutral feed to that circuit as
well.


--
Bob Mannix
(anti-spam is as easy as 1-2-3 - not)

In uk.d-i-y, David wrote:
What puzzles me, however, is that if I happen to short out the
earth/live/neutral wires (not sure which) in the dead circuit I'm
working on, the RCD trips the whole house off. Why is this? Is there
some infinitesimal residual current in the earth? Does it mean
there's a fault in my system somewhere?

No, it means the RCD is working fine. Your MCB/fuse only isolated the
live ("phase") wire. E and N are still connected back with all the Es and
all the Ns on the load side of your RCD. The RCD's job is to make sure
that (to within 30mA or whatever it's rating is) all of the current which
goes out its L side is coming back up its N side. When you join the N to
the E, you provide an alternative path back for some of the N current, and
the RCD pops because you're out of balance.

Probably bad practice, but when I'm working on the wiring in my house,
rather than switching off the entire supply and incurring the wrath of
family who apparently can't function without electricity, I tend to
just switch off the circuit I'm working on.

The above explanation shows you one reason why relying on the MCB alone
for isolation is less than best practice. You'd do better to switch off
the whole supply, physically remove the L and N conductors from the MCB
and the N busbar respectively - put the bare ends out of harm's way into
a bit of connector block, or for best paranoia into spare ways in the
earth block. Then you can restore power to the rest of the installation,
and work at leisure on the properly-isolated (or even earthed) final circuit
you just disconnected.

The cautious among us when first working on a circuit we haven't
installed ourselves might even measure the voltage from the allegedly
isolated N and L conductors to earth before grounding them, and/or do
the temporary-connection-to-earth through a low-value glass fuse (so
you could see it blow) or an incandescent bulb (so you could see it
glow) Just In Case some eejut has cross-wired the final circuits; but
that's taking caution a bit too far, IMHO. Me, I isolate the Ns and Ls,
put them into free-waving terminal block, and then test AGAIN at the
working position that there is no voltage between earth and the allegedly
isolated Ls and Ns. 'Course, if you use a digital multimeter, you can
be misled by very-low-current induced voltages - connecting a suitable
resistance in parallel with the meter will soon sort out the difference
between a tiny capactively-and-or-inductively coupled current, and a
Nasty cross-connection.

HTH, Stefek

wrote:

In uk.d-i-y, wrote:

................................................. When you join the N to
the E, you provide an alternative path back for some of the N current, ...

(in case it wasn't clear) I should have written "... an alternative path
back for some of the N current *in the other circuits on the RCD which
are still live".

The N voltage to earth may only be a 100mV (hardly measurable - or less)
but the path is usually less than 1 Ohm so you get the 30mA you need to
trip the breaker.

--
Ed Sirett - Property maintainer and registered gas fitter.
The FAQ for uk.diy is at www.diyfaq.org.uk
Gas fitting FAQ http://www.makewrite.demon.co.uk/GasFitting.html
Sealed CH FAQ http://www.makewrite.demon.co.uk/SealedCH.html

wrote in message ...

Probably bad practice, but when I'm working on the wiring in my house,
rather than switching off the entire supply and incurring the wrath of
family who apparently can't function without electricity, I tend to
just switch off the circuit I'm working on.

The above explanation shows you one reason why relying on the MCB alone
for isolation is less than best practice. You'd do better to switch off
the whole supply, physically remove the L and N conductors from the MCB
and the N busbar respectively - put the bare ends out of harm's way into
a bit of connector block, or for best paranoia into spare ways in the
earth block. Then you can restore power to the rest of the installation,
and work at leisure on the properly-isolated (or even earthed) final circuit
you just disconnected.

The cautious among us when first working on a circuit we haven't
installed ourselves might even measure the voltage from the allegedly
isolated N and L conductors to earth before grounding them, and/or do
the temporary-connection-to-earth through a low-value glass fuse (so
you could see it blow) or an incandescent bulb (so you could see it
glow) Just In Case some eejut has cross-wired the final circuits; but
that's taking caution a bit too far, IMHO. Me, I isolate the Ns and Ls,
put them into free-waving terminal block, and then test AGAIN at the
working position that there is no voltage between earth and the allegedly
isolated Ls and Ns. 'Course, if you use a digital multimeter, you can
be misled by very-low-current induced voltages - connecting a suitable
resistance in parallel with the meter will soon sort out the difference
between a tiny capactively-and-or-inductively coupled current, and a
Nasty cross-connection.

Thanks to everyone for the explanation.

So... if I was to switch off just the circuit I want to work on, THEN
confirm with a meter that there is no voltage across the N, L and E
conductors... under what circumstances could that still put me at
risk?

David

wrote in message
...
In uk.d-i-y, David wrote:

So... if I was to switch off just the circuit I want to work on, THEN
confirm with a meter that there is no voltage across the N, L and E
conductors... under what circumstances could that still put me at
risk?

There's little likelihood of things going awry in a normal, properly
installed, domestic installation. But unlucky things can happen if
a bodger has been in the house before you. (Bodgers - not the woodworking
sort! - come in both d-i-y and nominally professional flavours, even on
new builds. Rare, but most unfunny.) There'd have to be a fault in the
installation to put you at risk - for instance, if some wazzock has
"picked up" the N from your circuit "because it was convenient", it
could float up to the live potential - and it might be switched off
when you test, then when some other member of the family turns on the
switch for the outside security light/understairs-cupboard/whatever -
oops, ouch, or worse.

And, yes, it does happen - it happened to me!

The wazzock was the previous owner who fitted a two-way switch for the
landing light. Clearly unaware of the existence of 3-core +E, he had a bit
of a challenge when it came to wiring it up using 2-core+E. Then he had a
bright idea. "I'll use the live from the upstairs lighting circuit for the
upstairs switch and the one from the downstairs circuit (which was present
in the downstaitrs switch box) for the downstairs switch and connect them
together with the 2-core and earth". Hmmmmm. Muggins comes along and wants
to work on a bedroom light. Pulls fuse, all OK, happy working away. SWMBO,
in all innocence, turns on landing light. Disconnected neutral I am hanging
onto rises to live potential through landing bulb which glows fitfully due
to current through my sweaty fingers, while brain thinks WTF!? Guilty cable
also buried in plaster with no conduit - hooray! Yes, it is wired correctly
now. As they say in the X-files, "trust no-one".


--
Bob Mannix
(anti-spam is as easy as 1-2-3 - not)