Maximum length of ring circuit
Graham Gibson
The On Site Guide contains figures for the maximum length of a ring
circuit, limited by voltage drop, then earth loop impedance (in
addition to the requirement for a ring main not serving more than
100m2). IIRC, the figures are something like 88m for a 32A fuse and
91m for a 30A fuse.
Does anybody know how these figures are arrived at? It doesn't appear
to be a simple case of assuming a worst case scenario where each leg
of the ring is assumed to be carrying a maximum 15/16A to a point
furthest from the consumer unit, then applying the voltage drop calcs
in the usual way.
Although my upstairs ring main does not serve an area of more than
100m2, the overall length of the ring is in excess of 110m (using a
rough measure) as the circuit cables run from the consumer unit into
the loft and drop down to the sockets from the loft. With a 2m drop
to 15 socket outlets, this gives a length of 60m just for the drops,
plus at least 10m to and from the consumer unit, plus the length
around the ring in the loft of at least 40m - it all adds up (it's a
big house!) This is exactly as the house was built in 1984. To make
things worse, I added a couple of extra sockets by entending the ring
(thinking this was ultimately better than using spurs and before
getting paranoid about total length).
I am planning to replace my CU soon and get the whole system tested at
the same time - is the length of this circuit likely to cause a
problem? I'm pretty sure I'll be ok from an earth loop impedance
point of view, as I have a T-N-C-S system, and I'll be using RCBOs on
the ring mains when I replace the CU anyway, but voltage drop may be
an issue. It surprises me that the original circuit was designed in
this way.
Many thanks in advance for any comments.
Cheers,
Graham
Christian McArdle
>100m2, the overall length of the ring is in excess of 110m (using a
>rough measure) as the circuit cables run from the consumer unit into
>the loft and drop down to the sockets from the loft.
If access is good, you may be able to split the ring into two. You'd need to
run two extra feeds to a spare way in the consumer unit and find two parts
of the ring about halfway round apart that run close to each other. You can
then reorganise the feeds to suit. On the other hand, if you don't use any
electrical heating, you are very unlikely to run into a practical problems.
An upstairs ring without heating will only ever use a tiny fraction of its
capacity. Also, excessive voltage drop is not dangerous, it will just reduce
the voltage at the end to be similar to the nominal 230V instead of the
actual 240V feed, and only while the maximum current is drawn. A limit due
to earth loop impedance or raw cable capacity is far more important from a
safety perspective.
Of course, are you sure the circuit isn't run in 4mm cable, instead of
2.5mm? This may well bring the voltage drop characteristics into spec.
Christian.
Graham Gibson
> Of course, are you sure the circuit isn't run in 4mm cable, instead of
> 2.5mm? This may well bring the voltage drop characteristics into spec.
Unfortunately, it's definitely 2.5mm2.
Cheers,
Graham
Andy Wade
<1781e401.03031...@posting.google.com>...
> The On Site Guide contains figures for the maximum length of a ring
> circuit, limited by voltage drop, then earth loop impedance (in
> addition to the requirement for a ring main not serving more than
> 100m2). IIRC, the figures are something like 88m for a 32A fuse and
> 91m for a 30A fuse.
The voltage drop limited figures are 90m for a 30A BS 1361 cartridge
fuse, 91m for a 30A rewireable fuse and 84m for a 32A Type B MCB (or old
Type 1 or 2). As you hint, earth fault loop impedance considerations can
reduce the max. length considerably, although with a PME supply this will
only apply if you're using a Type C (or old Type 3) MCB.
> Does anybody know how these figures are arrived at? It doesn't appear
> to be a simple case of assuming a worst case scenario where each leg
> of the ring is assumed to be carrying a maximum 15/16A to a point
> furthest from the consumer unit, then applying the voltage drop calcs
> in the usual way.
The assumption made is a 'point' load of 20A half-way round the ring,
with the remainder of the current, up to the device rating, uniformly
distributed around the circuit. This is equivalent to assuming a 'design
load' of 25 or 26A (for 30 and 32A devices) at the mid-point. In
addition a small temperature correction factor is applied to the
tabulated cable voltage drop values in recognition of the fact that the
conductor temperatures will be less than the 70 degC assumed for the
voltage drop tables.
> Although my upstairs ring main does not serve an area of more than
> 100m2, the overall length of the ring is in excess of 110m [...]
> This is exactly as the house was built in 1984.
It certainly wouldn't have complied with the 15th Edition in 1984, when
the allowable voltage drop was 2.5%!
> things worse, I added a couple of extra sockets by entending the ring
> (thinking this was ultimately better than using spurs and before
> getting paranoid about total length).
Yes, sometimes spurs are the better option.
> I am planning to replace my CU soon and get the whole system tested at
> the same time - is the length of this circuit likely to cause a
> problem? I'm pretty sure I'll be ok from an earth loop impedance
> point of view, as I have a T-N-C-S system, and I'll be using RCBOs on
> the ring mains when I replace the CU anyway, but voltage drop may be
> an issue. It surprises me that the original circuit was designed in
> this way.
Assuming the RCBOs have a Type B overcurrent trip characteristic, I
wouldn't worry too much. Upstairs rings are rarely heavily loaded, and
probably almost never by 20A at the mid-point. A sharp-eyed tester might
query the rather high resistance readings on the continuity test though,
but I'd be inclined to leave that as a bridge to cross.
HTH
--
Andy
Graham Gibson
> > things worse, I added a couple of extra sockets by entending the ring
> > (thinking this was ultimately better than using spurs and before
> > getting paranoid about total length).
>
> Yes, sometimes spurs are the better option.
Andy,
Many thanks for your comments.
Would you recommend that I convert the 'extended ring' sockets back to
spurs, thereby reducing the length of the ring by 15 metres or so?
Cheers,
Graham
Andy Wade
<1781e401.03031...@posting.google.com>...
> Would you recommend that I convert the 'extended ring' sockets back to
> spurs, thereby reducing the length of the ring by 15 metres or so?
Difficult to say, without more information. What's the likely loading
pattern of the circuit, and how long would the spurs be?
--
Andy
Graham Gibson
> pattern of the circuit, and how long would the spurs be?
Loading is very light and distributed fairly evenly around the ring -
bedrooms and upper hall only, so table lamps, portable TVs, one 5A
shower pump etc.
I did some more measuring last night - the ring is about 110m long,
from CU right round and back to CU (without the 'extensions'). I
extended the ring by about 10m, when I added a couple of extra sockets
back to back. I discovered last night that the previous owner had
also extended the ring when he fitted a couple of extra sockets and
that the total length of this extension is around 15m.
Clearly, even the 110m basic ring would not have complied with the
regs in 1984 in terms of voltage drop and possibly not in terms of
earth fault loop impedance (BS3036 fuses, no RCDs).
I'm planning to replace the CU in a few weeks with one containing MCBs
and RCBOs for each ring (downstairs, kitchen and upstairs). This will
take care of any earth fault loop impedance concerns and the volt drop
shouldn't be too much of an issue given the light loading.
However, I'm concerned that this may cause a problem when I get the
system tested before the new CU is connected to the mains i.e.
although the upstairs ring main is perfectly safe, strictly speaking
it does not comply with the regs due to possibly excessive voltage
drop under heavy loading.
What I don't want to do is to go to the expense and effort of
improving the safety of my system to be told that I can't be
reconnected due to non compliance with the regs!
Cheers,
Graham
Andy Wade
<1781e401.03031...@posting.google.com>...
> Loading is very light and distributed fairly evenly around the ring -
> bedrooms and upper hall only, so table lamps, portable TVs, one 5A
> shower pump etc.
Nothing to worry about there then.
> Clearly, even the 110m basic ring would not have complied with the
> regs in 1984 in terms of voltage drop and possibly not in terms of
> earth fault loop impedance (BS3036 fuses, no RCDs).
I reckon that for that length and fusing it will be right on the edge for
earth fault loop impedance, assuming a TN-C-S supply (also assuming it's
not wired with the old 2.5mm^2 T&E which only had a 1mm^2 CPC). For TN-S
it would fail.
> However, I'm concerned that this may cause a problem when I get the
> system tested before the new CU is connected to the mains i.e.
> although the upstairs ring main is perfectly safe, strictly speaking
> it does not comply with the regs due to possibly excessive voltage
> drop under heavy loading.
"Strictly speaking" testing is to verify compliance with the _design_,
not the advice in the OSG. In view of the nature of use of the circuit
you could argue that it's designed for 32A distributed load and therefore
complies. The measured earth fault loop impedance should come in at 0.9
ohm max. (0.54 ohm for R1+R2 and 0.35 ohm max. for Ze) which is well
under the 1.2 ohm measured value allowed with a Type B MCB (OSG Table
2D). Therefore I don't think alarm bells will ring during testing. The
cable will be fully protected against fault and overload currents (even
without the RCD), which is what really matters.
> What I don't want to do is to go to the expense and effort of
> improving the safety of my system to be told that I can't be
> reconnected due to non compliance with the regs!
It's pretty unlikely that the issue would even be noticed, let alone be a
reason for refusal to give a certificate. Save these worries for the
kitchen/utility ring, where the loading pattern is far more likely to
lead to higher voltage drops.
--
Andy
Christian McArdle
>>bedrooms and upper hall only, so table lamps, portable TVs, one 5A
>>shower pump etc.
>
>Nothing to worry about there then.
Out of interest, due to the expected low loading, is it permitted to drop
the ring circuit breaker to 20A? If so, will this be enough to bring the
voltage drop into spec?
Christian.
Andy Wade
<3e70562d$0$4851$ed9e...@reading.news.pipex.net>...
> Out of interest, due to the expected low loading, is it permitted to
drop
> the ring circuit breaker to 20A? If so, will this be enough to bring
the
> voltage drop into spec?
Well you _could_ do that - BS 7671 itself doesn't prescribe any
particular design or topology for final circuits. Any design which
complies with all the requirements is "permitted". If you did it though,
I'd suggest that the floor area served shouldn't exceed 75m^2 (as per the
20A radial circuit) and that might be a problem in the OP's big house.
I don't recommend it though; for domestic applications it's much better
to stick to the standard arrangements that everybody understands and
expects to find.
And who says the voltage drop is going to be out of spec.? The most
likely way for this upstairs ring to be heavily loaded would be the use
of a few electric heaters in the event of a central heating breakdown.
In that case the load is likely to be well-distributed and not have 5 out
of the 7 or so kW plugged into one double socket exactly in the middle of
the ring.
--
Andy