Extending the ring vs spur

[jacoscar]

Hello, I need to add a couple of extra sockets to my downstairs ring
I thought the best way was to break into the ring and extending it using a pair of 2.5mm
However I’ve read that the maximum length of the ring is 100m and considering that the cables come from the ceiling void, each socket will add at least 7m (2.35+2.35), so I am starting to wonder if I’m getting close to that limit (even if the floor area is about 35-40 sqm, I know the maximum is 100)

How do you make sure you don’t reach the limit?
And if you do, could you just start adding spurs from the ring using single cable (as I think that would not count as extending the ring length)?

 

[John D v2.0]

2.35+2.35 is actually 4.7m not 7m, but anyway you can have a junction box in the ceiling as long as it's certified maintenance free.
You can always measure the end to end resistance to see how long the circuit is

 

[EFLImpudence]

You can always measure the end to end resistance to see how long the circuit is

It's not the length of the circuit which is the restricting factor, but the actual Earth Fault Loop Impedance.

 

[jacoscar]

2.35+2.35 is actually 4.7m, not 7m

I think I need to sleep more

 

[JohnW2]

I think I need to sleep more

... maybe, but don't loose any sleep over it - we are all more than capable of such 'oversights' (be they due to 'senior moments' or whatever), even when we've had plenty of sleep!

Kind Regards, John

 

[AndyPRK]

adding a spur is generally the sensible way.

If you want more than 1 double socket on the spur it's normal to fit a 13A fuse before the new sockets

 

[John D v2.0]

It's not the length of the circuit which is the restricting factor, but the actual Earth Fault Loop Impedance.

That does matter overall, but I always thought it was volt drop that have the surprisingly precise 100m limit, I'm sure there was a very long thread on here about it, and people calculated the actual limit as around 106m give or take

 

[EFLImpudence]

Exactly.

 

[John D v2.0]

So the summary to the op is to check both the resulting earth loop and volt drop, only the latter being covered by the length limit

 

[JohnW2]

That does matter overall, but I always thought it was volt drop that have the surprisingly precise 100m limit, I'm sure there was a very long thread on here about it, and people calculated the actual limit as around 106m give or take

I think you're right that, at least with TN-C-S, VD (rather than EFLI) is usually the length-limiting factor - but that may not necessarily always be true with TN-S.

106m is the figure that eric is always quoting, and I think that may come from the OSG. However, as often discussed here, arriving at such a 'maximum' figure (based on VD) is more of an art than a science, since it relies on assumptions about distribution of the load around the ring (eric always cites the "20A at midpoint, and the other 12A equally distributed around the rest of the ring" - but that is obviously pretty 'arbitrary').

Kind Regards, John

 

[JohnW2]

So the summary to the op is to check both the resulting earth loop and volt drop...

What do you mean by "check the .... voltage drop"?

As I've just written, deciding what it will be (presumably at the furthest socket), with given cables of given resistance/impedance, is dependent upon assumptions about the distribution of the load on the circuit.

EFLI does not present the same problem, since it can be measured directly (at the furthest socket) and is not reliant on any assumptions.

Kind Regards, John

 

[John D v2.0]

What do you mean by "check the .... voltage drop"?

That was the original bit about 100m
I just wanted to make clear to the op that there is both efli's efli point and the original vd point, they are separate but both affect the length. VD is easy just check the overall length, efli is indeed a bit more complicated and affected by the ze well.

 

[ericmark]

There are two independent restrictions as to the length of a circuit, one is volt drop, the other is the prospective short circuit current, the latter is normally taken as the lowest line - neutral or line - earth however with RCD protection the line - earth is less important.

Calculation has to take into account the conductor temperature, so the formula given in BS7671 is rather complex, but unless on the edge one can in the main forget about the correction, tabulated mV/A/m for 2.5 mm² is 18 mV/A/m but after correction as used to get the 106 meter limit it's 16.5 mV/A/m also it is assumed 20 amp mid ring and 12 amp even load so the design current for circuit Ib is 26 amp not 32.

There is nothing official about these figures, so you will not find it in BS7671 it is up to the circuit designer to decide, so if for example setting out a ring final around work benches in a school one could take 13 amp centre and 19 amp even spread, there is no hard and fast rule.

In theory all you need is a low ohm meter, you are permitted to use enquiry, so you should be able to phone or write to the DNO asking what the earth loop impedance and prospective short circuit current is, again in theory that is a better method than measuring, as where the supply can be from more than one cable, they will give you the worst reading.

In practice the DNO will quote 0.35Ω for a TN-C-S and 0.8Ω for a TN-S and with a TT earth loop is up to you, and 660 amp as the prospective short circuit current regardless of real reading, so in practice you need to either measure it, or get it from an installation certificate and the meter needed to measure it is expensive.

For the volt drop a low ohm ohmmeter is all that is required, for the tripping current for the fuse/MCB it needs the expensive meter.

A B type MCB should trip between 3 and 5 times the rated valve with the magnetic part of the trip, and the magnetic part will trip in around 0.01 seconds so to comply it must be able to use the magnetic part, so 5 x 32A = 160A for 160A to flow ohms law 230/160 = 1.4375 Ω working on 95% for a safety margin 1.365625 Ω that is the pass mark. With a type C then 10 times so ½ that figure.

A BS 1362 fuse we get from table at 2.42 Ω with 13A fuse, so if you can't measure it is unlikely you will exceed the 2.42 Ω limit.

So for an electrician who can measure, extending the ring final is better if within limits, but for DIY then may be better to fit a fused spur.

However with RCD protection the question is do we need the circuit to trip that fast? The RCD will trip with a line to earth fault in 40 mS, so faster than a fuse, but for a line to neutral fault it will rely on the thermal part of the trip if it is not drawing the 660 amp.

Now hitting a grey area, fault currents are not the same as short circuit current, and the magnetic part of trip is only good for short circuit current, and the thermal part will open the trip before damage to the cable.

The volt drop is another questionable area, even with a 20 volt drop, well over the limit, it is unlikely to cause a problem. The main problem is when you get the EICR done and it fails.

However I built a program to use the loop impedance readings to work out volt drop, and realised although the meter may say 0.35 Ω in real terms it could be anywhere between 0.33 and 0.37 Ω, the same at the socket looking for 0.94 Ω but could be 0.02 up or down from that figure, so unless the readings are 0.04 Ω out then can't say 100% the volt drop is exceeded, and then you have to assume using 26 amp as design current for circuit Ib.

So I have not seen many even try to measure the volt drop with an EICR, as to earth loop impedance yes that will be reported, but don't think one can even give it a C3 if there is RCD protection.

I am sure others will try and pull all this apart, and it can be debated until some sysop locks the thread, but at the end of the day, you should be filling out the minor works certificate and so your the only one who can decide.

 

[JohnW2]

I just wanted to make clear to the op that there is both efli's efli point and the original vd point, they are separate but both affect the length. VD is easy just check the overall length, efli is indeed a bit more complicated and affected by the ze well.

That's interesting since, as you will realise from my recent posts, it is the opposite of my view ....

... I would say that EFLI (aka Zs) was the one that is "easy" (at least, for someone with the appropriate equipment), since it can be ascertained with just a single measurement (at the most distant socket) which takes just a few seconds. The fact that it is "influenced by Ze" is not a problem, since the measured Zs includes the Ze component.

On the other hand, I would say that VD was the one which was "a bit more complicated", since it cannot really be measured, but can only be estimated on the basis of various assumptions of load distribution. To even make that estimate would, for an existing circuit, require one to estimate the total length of the circuit by measuring the end-to-end resistance of one of the conductors and calculating it from that.

The only way of producing a figure for 'maximum possible VD' which involved no assumptions about load distribution would be to assume a full (e.g. 32A) load applied at the centre of the ring - again requiring measurement of the end-to-end resistance of one of the ring conductors (unless total cable length was actually known). If one does that, using the VD figures for 70°C, one seems to get a maximum ring length of just under 80m. (79.86m).

Kind Regards, John

 

[JohnW2]

Calculation has to take into account the conductor temperature, so the formula given in BS7671 is rather complex, ... tabulated mV/A/m for 2.5 mm² is 18 mV/A/m but after correction as used to get the 106 meter limit it's 16.5 mV/A/m

As you say, it's quite complex. If one wants to calculate the maximum VD, one presumably has to do that in relation to the situation in which the current in the cable is at its 'maximum', and determining the VD at that current requires knowledge of the operating temp the conductors would achieve at that current.

The 18 mV/A/m for 2.5mm² cable (for a conductor temp of 70°C) presumably implies that the conductors are expected to achieve a temp of 70C when fully loaded to its tabulated CCC (27A for Method C). When ambient temp is 20°C and virtually no current flowing (i.e. no cable heating) that would correspond to a 'VD' of about 14.8 mV/A/m.

The 16.5 mV/A/m figure you mention would imply an operating conductor temp of about 48.7°C. Whether that is reasonable for the 'maximum current' flowing in each arm of the ring (13A, if one uses your assumption of an 'effective current' of 26A for a 'fully-loaded' 32A ring - see below), I don't know - but it is probably not an unreasonable assumption for Method C 2.5mm².

... it is assumed 20 amp mid ring and 12 amp even load so the design current for circuit Ib is 26 amp not 32. ... There is nothing official about these figures, so you will not find it in BS7671 it is up to the circuit designer to decide, so if for example setting out a ring final around work benches in a school one could take 13 amp centre and 19 amp even spread, there is no hard and fast rule.

As you say, there's nothing absolute about this, and certainly no regulations, or even associated guidance (that I have seen) about this. As you say, there are certainly situations in which an assumption that the load will be 'concentrated' near the centre of a ring does not really seem to be justified - in fact, I would suspect that this is, in practice, not a particularly common situation. On the other hand, as per my recent post, if one really wants to consider a 'worst-case scenario', one would have to assume that the entire full load (e.g. 32A) was applied at the mid-point of the ring.

The other associated thing which surprises me is that I have never heard of anyone correcting Zs measurements for temperature. Measurements are commonly undertaken on essentially 'cold' installations (i.e. conductors roughly at ambient temp) and, if those measurements are below the prescribed 'maximum Zs' figures, that is considered satisfactory/acceptable. However, if the measured Zs is close to the 'maximum', then it's quite possible that the Zs could be above the maximum permitted (for satisfactory ADS) if the circuit were 'fully loaded' (hence with conductors at their maximum operating temp).

Given that, frankly speaking ('in my opinion'!), Zs is far more important than VD, it seems odd that we do 'adjust' the latter for temp, but not the former!

Kind Regards, John