20 amp RFC, ok or not?!

[timhoward]

We ran into this question on another thread. A couple of people I respect a lot have suggested it's ok. I'd always believed it wasn't ok.
I'm sure it's actually safe, but does the letter of the regs allow it?

The only regulation that I know of specific to the RFC is 433.1.204 which covers the case of the OCPD being above the cable rating, but due to the ring structure with fused sockets along it, that is acceptable.

That's the only reg I know of for an RFC too.
It does actually specifically say "protected by a 30A or 32A protective device" in the reg, and says "Such circuits are deemed to meet the requirements of 433.1.1 if the CCC of the cable is not less than 20A". Given that method 103 problems with RFC's are commonly spoken of I'd always assumed that the 20A was "as installed" not "in free air".
So is 2.5 sq mm T+E installed to method 103 hooked up to a 20A OCPD arranged as an RFC allowed?
And is any installation method with a 20A OCPD arranged as an RFC allowed?
Discuss! And let's keep this civil - it's Christmas!

 

[pc1966]

I guess this is more about "is a 20A RFC acceptable" rather than "would you design with it" as the original context was thermal insulation being added later making it 103 method (so 13.5A CCC). I guess the key points for this (any any other unusual option) are:

• Is it safe?
• Does it contravene any regulation?
• Is it too bizarre that a future spark will get in to difficulties without access to the design plans?

My own take on the 433.1.204 regulation is that the two legs of the RFC give you 50% higher CCC, not the 100% higher from true parallel paths with balanced current sharing. So in that context a CCC of 13.5A would by adequately protected by a 20A OCPD.

However, in the other corner, the 13A sockets do potentially present a proportionally higher point-load at each place (as I try to us up my spare 'p's before Xmas). Is that enough of a concern?

Often a badly mangled RFC gets split and dropped to 20A as two radials as a means of quickly resolving the issue without a lot of cost / disruption to find and fix the break, so it is not uncommon to have a whole floor, etc, off 20A in that situation. Would it be too bizarre for it to be restored to a RFC and the same 20A retained?

Over to you ladies and gentlemen...

 

[LastManOnline]

I installed my second RFC in 30 years recently. It is protected by a 20 amp mcb.The regs give a max OCPD but not a minimum. Bearing in mind that the OCPD is there to protect the wiring, reducing its size cannot cause any harm. However there is a question of possible electrical inconvenience due to the reduction in circuit size , but that is a separate matter.

 

[westward10]

This is a genuine 1.0/1.0 ring final in a factory.

 

[pc1966]

This is a genuine 1.0/1.0 ring final in a factory.

WTF?

I could understand that for a very large light circuit to keep VD down for a given cable size, but absolutely not on 32A MCB!

 

[Julie.]

This is a genuine 1.0/1.0 ring final in a factory.View attachment 93318

Ffs.

 

[timhoward]

I installed my second RFC in 30 years recently. It is protected by a 20 amp mcb.The regs give a max OCPD but not a minimum. Bearing in mind that the OCPD is there to protect the wiring, reducing its size cannot cause any harm.

For an 433.1.204 the protective device is actually quite specific, see below. Here's the reg nearly in full.

433.1.204
Accessories to BS1363 may be supplied through a ring final circuit, with or without unfused spurs protected by a 30A or 32A protective device complying with BS88 / BS3036/ BS EN 60898 / BS EN 61009 etc.
The circuit shall be wired with copper conductors having line and neutral conductions with a minimum cross sectional area of 2.5mm sq. except for two-core mineral insulated cables complying with BS EN 6072-1 for which the minimum cross sectional area is 1.5 mm sq. Such circuits are deemed to meet the requirements of 433.1.1 if the current carrying capacity (Iz) of the cable is not less than 20A and if, under the intended conditions of use the load current in any part of the circuit is unlikely to exceed for long periods the current carrying capacity (Iz) of the cable.

In general, I'm in no doubt that a 20 amp RFC it's a safe setup.
To think about it completely differently a moment, if every point on the circuit has a total of 5 mm csa copper on live conductors and 3mm copper on cpc (considering both directions at once) , then it exceeds the csa of a 4 mm T+E radial circuit which would be be allowed to be protected by a 20A breaker in all circumstances except method 103 (17.5 amp max)
So without any reference to the RFC reg one might argue that apart from method 103 it's safe.

For me, from a regs point of view this boils down to two questions
1) In all other cases that aren't method 103, does it therefore matter that the RFC reg seems quite specific about what can be a RFC when it 'feels ok' on every other count
2) Is there any basis to say that it's still ok with method 103

 

[pc1966]

For me, from a regs point of view this boils down to two questions
1) In all other cases that aren't method 103, does it therefore matter that the RFC reg seems quite specific about what can be a RFC when it 'feels ok' on every other count

My suspicion is the regs are quite specific here to keep the RFC to one very specific "standard circuit" and not to give free reign to ring designs for all sorts of other possible reasons.

2) Is there any basis to say that it's still ok with method 103

My feeling is it is safe based on the relative ratings of cable Iz and OCPD.

But the one point that does come up though is "under the intended conditions of use the load current in any part of the circuit is unlikely to exceed for long periods the current carrying capacity (Iz) of the cable" which is kind of the general point about not putting large fixed loads on to the normal RFC anyway.

That is my minor concern about the method 103 & 2.5mm design with its 13.5A cable rating: how much margin is there for, say, a pair of 13A loads on the ring at some place? For any very long time the OCPD might trip, and if the legs are mostly balanced then no concern about the rating. But also the regs do say "long periods" and then we have the typical diversity of a domestic setup where even if you have a couple of high power white good running they are not likely to both be on heat cycles for long overlapping periods.

Here it is more about cable life-time, as we know it won't burst in to flames at 14A, or even 20A method 103, just you might not get 25+ years life if sustained. Again, for domestic that might be quite reasonable as you would not expect near-continuous loads as commercial/industrial systems might have, especially off 13A sockets.

 

[Pretty Mouth]

My own take on the 433.1.204 regulation is that the two legs of the RFC give you 50% higher CCC, not the 100% higher from true parallel paths with balanced current sharing.

This tallies with regs of old (from 14th ed am3):

For ring final sub circuits, the current rating of the conductors forming the ring is not less than 0.67 times the rating of the fuse or circuit-breaker protecting the final sub-circuit.

 

[pc1966]

This tallies with regs of old (from 14th ed am3):

Makes me wish I had kept my old books!

(At the time I did not expect to need them much, it is only recently and out of necessity that I am doing power work instead of just electronics)

 

[timhoward]

All I can say is that it's gone by 16th edition (oldest one I have on the shelf!)

 

[nicebutdim]

Makes me wish I had kept my old books!

(At the time I did not expect to need them much, it is only recently and out of necessity that I am doing power work instead of just electronics)

You can never have too much information or reference material.

 

[brianmoooore]

14th, 15th, 16th and 17th all on my bookshelf.
14th was the first book I bought in my university's bookshop. Price is written inside the front cover. £1.25.

 

[Lucien Nunes]

if every point on the circuit has a total of 5 mm csa copper on live conductors and 3mm copper on cpc (considering both directions at once) , then it exceeds the csa of a 4 mm T+E radial circuit which would be be allowed to be protected by a 20A breaker in all circumstances except method 103 (17.5 amp max)

This is a distorted view of the CSA of a ring. On a 4mm² radial, it doesn't matter where the load is, there will always be 4mm² of copper to carry it. On a 2.5mm² ring there are only 5mm² of copper for loads in the middle. Loads right near one end flow almost entirely through 2.5mm². E.g. 20A of load 10% of the way round the ring will divde as 18A in one leg and 2A in the other. The heating is proportional to the square of the current so the reduction to 18A in the short leg would reduce the temperature rise to 81% compared with the full 20A in one 2.5mm², but it's still a higher temp rise than the same load at 10% (or anywhere) along a 4mm² radial.

Because one configuration relies on a distributed load and the other does not, they cannot be directly compared in terms of CSA, Iz and Ib alone.

 

[timhoward]

This is a distorted view of the CSA of a ring. On a 4mm² radial, it doesn't matter where the load is, there will always be 4mm² of copper to carry it. On a 2.5mm² ring there are only 5mm² of copper for loads in the middle. Loads right near one end flow almost entirely through 2.5mm². E.g. 20A of load 10% of the way round the ring will divde as 18A in one leg and 2A in the other. The heating is proportional to the square of the current so the reduction to 18A in the short leg would reduce the temperature rise to 81% compared with the full 20A in one 2.5mm², but it's still a higher temp rise than the same load at 10% (or anywhere) along a 4mm² radial.

Because one configuration relies on a distributed load and the other does not, they cannot be directly compared in terms of CSA, Iz and Ib alone.

Thanks - makes sense.