On 18/02/2013 18:18, Stephen H wrote:
On 16/02/2013 20:00, John Rumm wrote:

(In reality a 20A plate switch and a single 2.5mm T&E would have been
adequate (i.e. in effect creating a switched unfused spur)).

I did actually sit down and think this through from a good design point
of view...

As stated, there is a 2.5mm2 ring main with a 32A RCBO.

I was not worried about L-E or N-E shorts as there is the 30mA RCD part
in the RCBO that takes care of that.

However, I had two scenarios to think about regarding my outside sockets.

I wanted to put a single external socket in the front porch.

So that meant a single 2.5mm2 spur T&E cable from the 32A ring main in
the garage to a single socket in the front porch via a bedroom.

Given that the max current available on the ring main would be 32A
before the consumer unit's 32A RCBO opens and that the maximum
in-service non-fault conditions continuous current would be 13A at the
porch socket, (limited by the 13A or less fuse in the plug of whatever
appliance plugged in)

I had thought about what could happen if there was a dead L-N short on
the single 2.5mm T&E cable running between the ring main and the single
gang socket.

Obviously this single 2.5mm cable is capable of taking a 13A current
continuously between the ring main and the single socket via the spur
cable, I felt uncomfortable with it potentially carrying up to 32A
before RCBO opens in case of a L-N short on the spur cable.

Ah, right I see the line of thought you are following. In fact you are
over worrying since this exact scenario has already been thought about
in the design of unfused spurs...

As you are probably aware there are two overcurrent categories that need
to be considered with these things - that of overload, and that of a
fault. Each have their own characteristics, and each need (possibly)
separate handling.

The fault current situation is the most important - i.e. stopping the
cable vaporising or bursting into flames on a hard short. The circuit
protective device at the origin of the circuit *must* protect against
fault currents. It has been shown that a single 2.5mm^2 cable (as used
in a typical spur arrangement from a standard ring circuit) will be
adequately protected from the effects of fault current by the B32 MCB at
the origin of the circuit.

For more on this see:

http://wiki.diyfaq.org.uk/index.php?...abatic _Check

The second issue is that of overload. i.e. what happens if a load in
excess of the design current is placed on it for an extended period.
(short term overloads are permissible, and it is even desirable that a
circuit will not trip in response to them). You are correct in assuming
that you can't rely on the B32 MCB/RCBO to offer overload protection for
a single 2.5mm^2 cable. However this class of over current does not
have to be covered by the device at the origin of the circuit. In the
case of fused spurs, its enforced by the downstream fuse. In the case of
unfused spurs, then it is imposed either by the number of available
connections (e.g. a single socket can't take a load large enough) or by
the characteristics of the load itself (e.g. a fixed appliance that has
no mechanism for drawing a sustained overload)

Thus, I put a 13A fused switched neon faceplate at the point where the
spur is connected to the ring main in the garage to protect the spur.

The other scenario was the double gang socket on the front driveway.

Now I could plug in two appliances. Lets say both draw up to 13A before
their individual plug fuses pops. Thats a total load of 26A which is on
a 32A rated 2.5mm ring main.

A cable that if surface wired on in masonry actually has a 27A
capacity... however that's not the point. For the purposes of design, a
double socket is rated at 20A total load. In reality some will do better
than that, but many will not like it if you actually stick two 13A long
term loads on. Chances are the socket will suffer long before the cable.

Now doing 26A continuously on a single 2.5mm2 spur cable between the
ring main and double socket I was not comfortable with, let alone a
potential up to 32A flowing in the case of a dead L-N short.

Some will argue that you can get even more load on it since a 13A BS1361
fuse will carry 20A before blowing... armed with some multi way
extension leads you could have the socket in a molten mess in very short
order ;-) In reality it never happens since actually getting long term
loads (i.e. those that are not intermittent due to usage patterns, or
due to the actions of thermostats etc) is actually quite difficult (a
couple of 3kW fan heaters to warm the garden perhaps?).

Along the same line of thought it is also worth remembering that a 32A
circuit breaker does not trip at 33A. In fact you need 1.45 x nominal
current to ensure they trip, and 5x normal current to insure they trip
"instantly" (i.e. when clearing fault currents)

Have a look at the response curve of a 32A type B MCB:

http://wiki.diyfaq.org.uk/images/d/d...e-MCBTypeB.png

Switch isolators come rated as 20A or 45A. So 20A is less than 26A that
the double gang socket could be going up to. So I went to 45A rating for
the switch neon isolator.

No harm in doing that, except that you pay more for the switch, and the
socket etc is not really up to the job of matching the performance of
the switch.

So I put a 45A rated neon switch on the 2.5mm2 ring main and reasoned
that as current flows both ways round the ring, that 5.0mm2 is the same
as 2 x 2.5mm2. So I opted to run the spur cable as 4.0mm2 from the 45
amp isolator to the double gang socket. Now I am then comfortable with
up to 26A flowing continuously on this 4.0mm cable from isolator to
double gang socket.

Indeed. although normally if one were proposing to have very heavy loads
on sockets in this way, they would ideally be better in the ring and not
on a spur.

For your actual application, I would be entirely comfortable with a
single 2.5mm^2 cable and a double (IP 56) socket on the end.

I fully appreciate the above scenario is highly improbable and that I
have probably over-engineered my circuit but hopefully at least a sparky
can't find fault with my reasoning!

There is nothing dangerous about what you propose, although its over
egging things a tad. Lots of design effort over the years has gone into
make sure that these things work well in reality, and while there are
theoretical weaknesses in the designs of unfused spurs, they prove time
and again in practice to be very robust and reliable.


--
Cheers,

John.

/================================================== ===============\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\================================================= ================/