I would say that it is neither a "should" or "must".
eg if you have a 4mm 32A radial circuit what sized cable are you allowed
to used for an unfused spur powering one single or one double socket?
(app 15 of the regs is your clue).
There's two overlapping requirements covering this.
You need to have a fused connection where cable sizes change, and a
circuit has to be fused to protect the smallest cable in the circuit.
So, if your whole circuit is fused to protect 1mm cable I would
accept a larger cable as a main feed to 1mm spurs. I have done this
where I typically have a larger cable running from the consumer unit
to a convient breakout point, where 1.5mm cable goes off to
individual lighting circuits.
I've used fuse-at-cable-change typically when spuring from a 32A ring
to provide a loft light where the lighting circuits are inaccessible
or there's a need to have a non-lighting light so you can see when
the lighting circuit is switched off. Similarly, I make sure there's
an upstairs socket within extension lead reach of downstairs and
On 30/12/2017 08:43, email@example.com wrote:
While true for things like your example of taking a feed to run a light
fitting from a 32A socket circuit, its not universally true... For
example, an unfused 2.5mm^2 spur from a 4mm^2 radial protected by a B32
There is also some nuance in "protect" here. The fusing etc must provide
fault protection to the whole circuit, but may not be required to
provide overload protection to it all. (a spur from a ring circuit being
one example - the 32A MCB will provide fault protection for the single
run of 2.5mm^2 cable on the spur, but not overload protection. That job
is delegated to the design limitation on the spur of only feeding one
single or double socket).
With 6A lighting circuits, the cable sizes are over specified anyway
with regards to maximum operating current (for reasons of mechanical
durability, and limiting voltage drop).
I'm now more puzzled than before. I could look at the book but its not
Earlier it was said "No, cable should be the same size throughout the
circuit. 1mm is actually adequate for most 6A lighting circuits in
houses where the cable length isn't too long."
I thought this wasn't true at the time. Can we confirm that as long as
the smallest conductor is still fault and overload protected by a MCB,
then all is well?
It changed in about 1983-ish from 1mm to 1.5mm AFAIK, but
ARW or John Rumm will no doubt confirm this. Something
to do with guaranteeing that even a non-rcd protected
circuit would trip or blow its MCB or fuse in the
required minimum period (?40 msec).
So you'll probably be ok with an RCD/RCBO or maybe even just
an MCB, but not an old-fashioned fuse arrangement.
The RCD protection will be a requirement for any new sockets and for
all T&E cables for any circuit that are buried in a wall behind plaster.
In 6 months time the regs will have changed and will require all new
domestic lighting to be RCD protected.
To what value? My board has 100mA to lights and a few other things, then
30mA to sockets, cooker etc.
I've always liked this because if the socket is tripped the lights will
probably stay on - if I've just had a shock I'd rather not be in darkness as
I would expect 30mA (i.e. for shock protection rather than
infrastructure / fire protection).
However "all RCBO" installs are much more cost viable these days, and
even if just using the normal 17th style CU you would normally ensure
the power circuits in a given place are on the "other" RCD to the one
doing the lights.
On Sat, 30 Dec 2017 00:51:54 +0000, John Rumm wrote:
RCBOs are OK for a new installation but probably wouldn't fit on the Wylex
board fitted c1992, judging by the space occupied by the MCBs.
I'd definitely go for 100mA/30mA RCDs and then RCBOs in a new installation.
If you have a TT installation that is about right (assuming that the
100mA RCD is time delayed if in series with the 30mA RCD).
However it will soon be a requirement for lighting to have 30mA RCD
protection on new installations. Although this has pretty much been the
case since the 17th edition came out.
That's a pity - I've had an experience of enough of a shock to make me jump
back - in darkness (assuming that the lighting RCD tripped) I could have
suffered an injury on various edges etc.
I can't recall if my 2 RCDs are in series or not; I suppose if 2x30mA are in
parallel, then the lighting should stay on if there's a fault on the ring.
The rule change was a bit before than I believe, but there could well
have meant some installations still being later than that.
400 ms usually (TN systems)
There is a problem on the old 1mm^2 CPC cable when you have a spur on a
ring protected in particular by a BS 3036 re-wireable fuse.
The single 1mm^2 CPC is undersized to ensure a fault to earth is cleared
(you need about 1.25mm^2 - the ring itself with 2 x 1.00mm^2 is ok, but
the spur is vulnerable)
The "instant" trip current required for a BS3036 fuse is something like
450A compared with 160A for the B32 MCB.
When I moved into my present house, the vendors had taken all the fancy
brass sockets and switches (it was a while ago) and put back the
original 1976 white contract stuff.
But they had managed to snap the 1mm cpc in two of the sockets, which
were on a single ring (house less than 100 sq m), and badly connected
the live to a third, so that it had no cpc at all and the ring was
actually two spurs running from a 30 amp wylex rewireable fuse.
They also fitted a new kitchen and drilled right through the cooker
live wire, and then repaired it with a bit of choc block which
was hidden behind a fitted cabinet. I noticed the burn mark above
the fuse wire carrier, that had actually melted part of the metalwork
but it was 13 year before I removed the kitchen cabinet and found
HomeOwnersHub.com is a website for homeowners and building and maintenance pros. It is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.