On-Site Guide BS 7671:2018 - 7.2.2 Socket-outlet circuits

As Andy has pointed out, the 106 m max length (all the way around) of a standard T&E RFC is based upon voltage drop assuming that it is fully loaded (26 A) and that the load is distributed, but not all RFCs are going to be fully loaded, especially on the upper floors of a conventional house.

Thank you for responding to my question. I haven’t been thinking about this problem recently but I had hoped to develop a better understanding of the issue (and have, so far, failed). This is the paradox: the ‘1/8 rule’ (as declared in the On-Site Guide, 7.2.2 on page 76), appears to contradict the very idea that every socket-outlet on a ring circuit can have one spur  (H2.4, page 189...’the number of non-fused spurs should not exceed the total number of socket-outlets’...) i.e. if the maximum length at the ‘furthest point’ is zero then you cannot have a spur at that point. I am convinced that my understanding is at fault and I should be happier were it resolved (but, clearly, I need to study more).


For example, if I wished to put a spur at the ‘furthest point’ on a radial circuit, I should, effectively, be simply extending the length of that radial (which would be fine, so long as I did not exceed the maximum length for the cable as identified in Table 7.1, pages 65 onwards). However, because a ring-final circuit operates differently, the electrical load is ‘distributed’ (i.e. the electrical power is drawn further one way or shorter the other way, depending upon which location on the ring circuit the socket-outlet is sited). This is why I doubt the conclusion being drawn from the interpretation of the 1/8 rule which suggests that you cannot place a spur at the furthest point.


Look at it this way, if you have a socket-outlet at the furthest point on a ring final circuit, the electrical power runs half way around the first half of the circuit and half way around the second half of the circuit (i.e. one full circuit); if you have a socket outlet at, or very close to, the consumer unit, the electrical power, effectively, runs all the way around the whole circuit. My conclusion being that, the electrical power on a ring-final circuit will, indeed, run the same distance around that circuit regardless of the location of the socket-outlet.


Therefore, if you had a longer ring-final circuit, there may be implications for the maximum length of any spur; that maximum length will have been determined by the 1/8 rule. Is this the seed of a valid conclusion? I can’t quite fully get my head around this puzzle at this very moment. What a shame that the On-Site Guide isn’t proving its worth as something comprehensively helpful to me as a novice domestic electrical installer.

Look at it this way, if you have a socket-outlet at the furthest point on a ring final circuit, the electrical power runs half way around the first half of the circuit and half way around the second half of the circuit (i.e. one full circuit); if you have a socket outlet at, or very close to, the consumer unit, the electrical power, effectively, runs all the way around the whole circuit. My conclusion being that, the electrical power on a ring-final circuit will, indeed, run the same distance around that circuit regardless of the location of the socket-outlet.

Not quite. The current drawn though a socket will be divided between the two sides (legs) of the ring in proportion to the conductivity (1/resistance) of each leg. So a socket at the exact mid point drawing 13A will see 6.5A flowing through both sides. A socket near to the CU will see almost all its current flow through the shortest leg, and very little go the long way around.


You can verify that simply using Ohm's Law - noting that the voltage at both legs at the CU must be the same (as they're connected together) and likewise the voltage on both legs where they meet at a socket must also be the same (again as they're solidly connected together) - but the longer leg will have a higher resistance so Ohm's law say that it must therefore carry a proportionally lower current if V=IR is to hold true for both legs at the same time.


   - Andy.

Well, I still don’t have a definitive answer but this is what I feel must be somewhere nearer to the truth than me just guessing...


On an 80m ring final circuit, the electricity is ‘evenly distributed’. So, if I plug a radio into a socket, it will draw electricity both ways around the ring. If the socket happens to be near to the supply it will draw the electricity a short way one way and a long way the other way. If the socket happens to be 1/4 of the way around the ring, the radio will draw electricity 1/4 one way and 3/4 the other way around the ring. Again, if the socket is half way around the ring it will draw the electricity 1/2 one way and 1/2 the other way. Effectively, in whichever location the socket happens to be the electricity ‘travels’ one whole circuit of the ring (...).


By induction, I suspect that the 1/8 rule must work on a similar principle. If the socket happens to be at the ‘furthest point‘, it is 0m one way but 80m measuring the other way. Dividing by two gives 40m. Applying the rule of thumb results in a maximum spur length of 5m. Somehow, because a ring final operates differently to a radial - because the electricity is ‘evenly distributed’ - the maximum length of any spur for a given ring is always the same (and the maximum length of the spur relates directly to the maximum length of the ring - up to the safe maximum for a ring)...


Hmm, well it just doesn’t make any useful sense interpreted the other way. The On-Site Guide can’t say that ‘you can have one unfused spur at every socket’ and then give you a rule of thumb that if interpreted literally, precludes the use of spurs anywhere approaching ‘the furthest point’.


Most unsatisfactory...

because the electricity is ‘evenly distributed’

I still say that's a misleading conclusion.


Try a few worked examples using some arbitrary cable resistances and Ohm's Law.


The other option is that one of the authors of the OSG has made something up that isn't of itself entirely logical (unlike many of the other "rules of thumb" for rings that are well known and go back generations, this 1/8th rule seems new has never appeared in the regs or any other guidance as far as I know).


   - Andy.

First, to avoid terminological confusion: the "furthest point" is the point furthest from the CU - if you have an 80m ring, then it's the point 40m from the CU.


It's best to think of the rule as a rough guide to not increasing the maximum resistance (Zs) of the ring beyond what it already is. This is important for two reasons: first to avoid excessive voltage drop across the cables (wastes power and things plugged in might not work properly) and to ensure that a fault will cause enough current to flow to trip the breaker. Given an existing ring, its reasonable to assume that the person who designed it did in such a way that at the socket at the furthest point from the CU will have a Zs which is within spec. If you follow the 1/8 guideline, this guarantees that the new socket will have a Zs not more than the highest Zs of any existing socket in the ring.


If you want the maths, consider a ring with total resistance 2r: so at the furthest point, each arm has resistance r, and the resistance from a socket at that point to the CU is r/2 (rule for resistances in parallel). If you consider a spur tap point in the ring of proportion k between the furthest point and the CU (so k=1 implies at the furthest point, k=0 implies at the CU), then the OSG rule says you can add a spur at that point whose extra resistance is kr/8. If you do the maths, you get that the resistance at the socket at the end of the spur is then


(1-k²)r/2 + kr/8


If you evaluate that for various values of k, you get values between about 0.3r and 0.5r. The latter is the value at the furthest point, so your spur socket is no worse than any existing ring socket.


Since its a rule of thumb, its not precise; in particular, when k = 0.75, you get 0.3r. In this case you could in fact put in a longer spur without exceeding 0.5r.


And remember that this is just a rule of thumb to not make things worse than they already are. If you really need the extra length, you can always measure Zdb and calculate exactly how long the spur can be.

Thank you, I do appreciate all of your responses. During Lockdown I began to learn about ‘domestic electrical installation’ with the idea that I might enrol upon a short course to gain some competence. Perhaps the On-Site Guide and Wiring Regulations were not the best place to start, particularly without the support of experienced teachers. The old adage, ‘A little knowledge is a dangerous thing’, strikes me as pertinent.

wallywombat:

If you evaluate that for various values of k, you get values between about 0.3r and 0.5r. The latter is the value at the furthest point, so your spur socket is no worse than any existing ring socket.

Correcting myself: it's between 0.125r and 0.5r. Once you get near the CU, the Zs of the spur socket drops off rapidly: here the rule-of-thumb is badly underestimating the maximum safe length of spur. But from half-way to full-way to the furthest point the formula works well.