Cable size for sockets?

[HuwT]

Good morning all.

I want to add some extra sockets in a 1930s two-bed semi that I'm renovating down in Bideford and had a query regarding cable size.

At the moment, the consumer unit connects to two mains circuits - one labelled "downstairs sockets" and the other "upstairs sockets". These both are radial circuits (with some spurs) using 4mm cable, protected with 32A breakers.

However, the "upstairs" circuit only powers one socket in the main bedroom and two in the lounge downstairs!

The downstairs circuit powers one socket in the main bedroom, one in the second bedroom, one on the landing, one in the dining room and three in the kitchen, plus spurs for washing machine, fridge and freezer. (9 doubles in total)

My idea is to change the "upstairs" circuit so it supplies the two bedrooms, lounge and dining room, adding in a couple of extra sockets, and use the "downstairs" circuit to supply the kitchen.

From my limited research it seems it's more normal to use 2.5mm cable for mains circuits. I don't want to remove all of the old cabling as a lot is chased into walls. Obviously, I don't want to mix sizes but is there any reason not to use 4mm other than it being more expensive?

Also, is 32A too big for the circuit breaker on a radial mains circuit?

There are two lighting circuits on 6A MCBs so that seems normal. However, the boiler is running off a single cable with a 20A MCB. Is this too big?

The only other circuit is for the cooker - with another 32A MCB.

All advice welcome. Many thanks.

 

[Taylortwocities]

From my limited research it seems it's more normal to use 2.5mm cable for mains circuits. I don't want to remove all of the old cabling as a lot is chased into walls. Obviously, I don't want to mix sizes but is there any reason not to use 4mm other than it being more expensive?

The size of the MCB is selected to protect the cable attached to it.
There are a number of "standardised" circuits

RING FINAL circuit, wired in 2.5mm² cable protected by a 32A MCB
RADIAL circuits
For sockets the standard choices are
wired in 2.5mm² cable protected by a 16 or 20A MCB (it can be up to 25A but not many manufacturers make a 25A MCB)
wired in 4mm² cable protected by a 32A MCB.
other radial circuits are configured depending on load, so you may see a 6mm² circuit protected by a 32A or 40A MCB, or a 10mm² circuit with 45A MCB*

*NOTE that cable current carrying is dependent on the installation environment, the above is a simple guide only. That is why new circuits are notifiable and need proper design, test and certification.

To return to your question. 32A is not too big for a radial if it is wired in 4mm². It is too big if teh wiring is in 2.5mm². Why? 2.5mm cable, in ideal conditions has a max current of 27A. In overload conditions, the MCB needs to trip before the cable overheats and your house burns down.

SO, if the circuit you want to extend starts out in 4mm you need to use 4mm for any extensions.

 

[securespark]

Is the 20A device the only one connected to the boiler? Or is there a fused connection unit local to the boiler with a 3A fuse in?

As for the sockets, I would try and ensure an even loading across the circuits. Make sure the labelling of the MCBs is accurate, with a list if necessary.
32A radials are run in 4 milli. 20A radials in 2.5. Although you can use 2,5 on a 32A radial if only feeding one socket.

 

[HuwT]

Yes, the boiler is the only device on the 20A connection - but there is a separate fuse local to the boiler.

I take your point about even loading - which is why I was thinking of changing the circuits. If I put two bedrooms, lounge and dining room on one circuit and the kitchen on the other does that seem reasonable?

I know things like kettle and toaster draw a lot of power - also have a microwave and breadmaker, as well as washing machine and fridge in the kitchen.

The other circuit will have TV etc but not a lot of heavy load items.

 

[flameport]

4mm cable, protected with 32A breakers.

How have you determined that the existing cable is 4mm²?
Does the consumer unit have RCD(s) in it?

There is no concept of a spur on radial circuits.

 

[ericmark]

With 4 mm² cable if the load is near end of cable to stay within limits for volt drop maximum length is 31 meters, but with a ring final using 2.5 mm² you can have 106 meters of cable, there are a number of reasons for the difference, one is 2 x 2.5 mm² = 5 mm² and also it is assumed you will draw 20 amp in centre of ring and 12 amp even spread so the calculation is for 26 amp not 32 amp. You will use more 2.5 mm² in a ring final because it returns to start, but the first job is to work out how much cable is already used, so for example if the incoming loop impedance is 0.35Ω the highest reading at end of cable calculating at 32 amp is 0.86Ω you can clearly if you know heavy loads not near the end use other design current figures, so using 26 amp same as ring final you could use 40 meters and would expect no more than 0.97Ω loop impedance at end of cable.

However be it 31 meters or 40 meters it is not that much, and it would be easy to exceed the volt drop limits. As to if volt drop is that important is another question, and I have not seen anyone measure volt drop when doing an EICR, however you do need to consider the volt drop even if you don't keep within it, the main items to suffer with volt drop and fridge/freezer and fluorescent lights, so if kitchen is close to consumer unit not so bad.

The other limit is the earth loop impedance, since the earth wire is 2.5 mm² even with 4 mm² cable not sure on the limit, but unlike the volt drop this includes supply loop impedance not just the difference between supply and internal sockets, it was 1.44Ω for a 32A MCB but when the volts were dropped from 240 to 230 it was noted supply volt drop could affect it, so it was reduced by 5% so now 1.37Ω and this is checked with an EICR so you do need to comply.

The problem is a loop impedance meter is not cheap, also the RCD tester is not cheap either, so to test costs money, most DIY people trust to the lord, or inshallah (god willing) if the circuit is protected by a RCD then should the ELI be too high then likely the RCD will protect you, however we are told not to rely on the RCD.

In the main the DIY man uses FCU or RCD FCU to extend the ring final or radial circuit, and the ELI for 13A fuse is 2.30Ω so you have much more leeway, also with a MCB once the device will not trip on the magnetic part the time to trip increases a lot before the thermal part will trip, with a fuse there is no magnetic part, so being slightly over the ELI does not matter as much.

In theory I can assume the incoming with a TN-C-S supply is 0.35Ω so I can use a low ohm ohmmeter with at least 200 mA used to measure to work out the ELI, enquiry is permitted, TN-S normally taken as 0.8Ω, and if you have the installation certificate then the incoming ELI will be on that, but in practice you can't really rely on those figures, often electricians in the past did not disconnect gas and water pipes before testing, and so if water or gas pipes have been replaced with plastic the ELI will have changed.

I personally like ring finals, and would rely on the RCD if the ELI is over 1.37Ω but that is not what the regulations say, so next EICR and it will show as unsatisfactory. So you need to do your own risk assessment, is it worth doing, and if it is worth doing, is it worth doing as a DIY job.

 

[JohnW2]

There is no concept of a spur on radial circuits.

It appears that there is (such a concept) in the minds of some people, including the authors of Appendix 15 of BS7671.

Some people talk about 'branches' of radial circuits when all cables have the same CSA (in which case it would obviously be impossible to define which was 'the circuit' and which was the 'spur',if one tried to use such terminology) BUT refer to a 'branch' which has a smaller CSA than the 'main' part of the radial circuit as a 'spur' (fused or un-fused, with the same implications about what may be supplied by it as would be the case with a spur from a ring final circuit) - and it seems that the authors of Appendix 15 subscribe to that thinking/convention.

Kind Regards, John

 

[EFLImpudence]

With 4 mm² cable if the load is near end of cable to stay within limits for volt drop maximum length is 31 meters, but with a ring final using 2.5 mm² you can have 106 meters of cable,

Maybe, but that is not the only option.

Taking your figures:

on a 31m. radial, at 16m. you could have a 15m. branch - or any number of branches making 31m.

and

on a 86m. ring, you can have any number of 5m. spurs.

 

[ericmark]

1930's is before the ring final was invented, whole idea of the 13A socket is with a fuse in the plug you can have 32 amp supplying 13 amp sockets, before the 13A socket the 15A socket had to come from a 15A fuse so if you wanted more than one 3 kW appliance in a room you needed multi feeds into the room, so pre-war fuse boxes had loads of fuses, one for each circuit.

Be it a ring final or a radial the idea is the same, one 32A breaker can feed many sockets.

But the big problem is although some sockets seem to remember MK being one will allow you to get three 4 mm² wires in one terminal, many simply don't have enough room for 3 wires, best you can often get is 2 x 4 mm² plus 1 x 2.5 mm² the 13A socket was invented to work with the ring final, and once you try to change that it does not work so well.

Theory having a ring final feeding a 4 mm² spur, or having 2 x 13A single sockets on a 2.5 mm² spur would work OK, however it is not allowed, there seems to be some thought given to what one should expect to find. Again see no reason why you should not have a figure of 8 can't even find anything in regulations to say you should not use one, but we were taught for the inspection and testing exam to check there was no figure of 8's. I assume as one could test and assume there was a full ring when one loop of the figure of 8 had a break?

But the point of last post was to point out need for inspection and testing before adding to the circuit, exceeding 106 meters it needs to be a reasonable sized house, exceeding 31 meters however is rather easy. There was an IET Addendum or Corrigendum bs7671:2008 which shows the appendix 15 circle and the options, but I can't find it.

 

[EFLImpudence]

1930's is before the ring final was invented

Also before 32A MCBs so ...

But the big problem is although some sockets seem to remember MK being one will allow you to get three 4 mm² wires in one terminal, many simply don't have enough room for 3 wires, best you can often get is 2 x 4 mm² plus 1 x 2.5 mm²

As there are sockets with 12mm² capacity terminals - obviously not the best you can get.

Theory having a ring final feeding a 4 mm² spur, or having 2 x 13A single sockets on a 2.5 mm² spur would work OK, however it is not allowed,

Of course it is allowed.

But the point of last post was to point out need for inspection and testing before adding to the circuit, exceeding 106 meters it needs to be a reasonable sized house,

Likely not 1930s two-bed semi, then.

exceeding 31 meters however is rather easy. There was an IET Addendum or Corrigendum bs7671:2008 which shows the appendix 15 circle and the options, but I can't find it.

Appendix 15 shows some options.