Does BS7671 require short circuit protection

[plugwash]

I had assumed previously that "fault protection" included both protection against earth faults, and protection against short circuits but a comment in another thread lead me to do some searching and it seems that the BS7671 definition of fault protection only covers earth faults.

Most of the time on a TN install, even if a RCD is present we still design circuits such that MCBs will provide earth fault protection. Since the PN loop is normally lower than the PE loop this means they will also provide short-circuit protection, but there seem to be a few cases where this would not be the case.

1. If multiple circuits of different sizes are running in the same containment system and a short circuit develops between conductors of different circuits.
2. On a TT install (which was the comment that lead me to ask this) we rely on RCDs for earth fault protection since the MCBs wouldn't have a hope of providing it.
3. If the earth is unusually heavy for some reason (say a shared earth between multiple circuits of different sizes, or an earth that is enlarged for bonding reasons) then the PN loop may in-fact be higher than the PE loop.


So what if-anyting are the rules on short-circuit protection.

 

[ericmark]

Cables will clearly melt if not short circuit L - N or L - L protection, so we always fit fuses or MCB or MCCB or RCBO in the circuit, however it takes some time to melt cables, so the thermal part of the trip is often good enough. But we do need to look at the overload device rating and the prospective short circuit current. Only once have I found the PSCC was too high, tower crane supply and wanted to use a RCBO to feed a 13A socket to charge the 2 way radio. And it was rated 6kA and the PSCC was over that limit.

I read about the let through value of fuses, about as clear as mud, but it was clear it needed a fuse before the RCBO, so I fitted a fuse, and then found the resistance of the fuse was enough to drop the PSCC to permitted levels, so never did get my head around let through values.

But on the lower side I would agree if the loop impedance is such that the MCB will not trip on the magnetic part, then there could be an extended time for the overload. But in real terms only with 110 volt site systems has there really ever been a problem, and this is really due to where the MCB or other overload device is placed.

A 16 amp socket should have an overload which will in fullness of time open an overload device at 20 amp, however I have seen yellow bricks with the overload and fuse on incomer only, so with a 10 amp overload a line to earth fault needs to draw 10x230/55 = 42 amps, well will not trip at 10 so likely getting to 50 amp to trip the overload. So have seem melted 110 volt cables, the damage is most evident at plug or socket, so always open them when doing PAT testing.

So does the yellow brick comply with BS 7671? So 433 says you can have up to 3 meters installed in such a manner as to reduce the risk of fault to a minimum, and it is installed in such a manner as to reduce to a minimum the risk of fire or danger to persons, I would say the yellow brick is not therefore permitted under BS 7671 as the cables do not normally run in a safe zone, that's why we use 110 volt site supplies.

However I have argued the fact with safety officers and pointed out 230 volt with RCD protection would be safer, but to no avail, still ended up with two transformers one down to 110 and one built into the equipment being used to step it up again.

 

[JohnW2]

I had assumed previously that "fault protection" included both protection against earth faults, and protection against short circuits but a comment in another thread lead me to do some searching and it seems that the BS7671 definition of fault protection only covers earth faults.

Indeed, but I think that's because (probably quite rightly) they look at the two things in totally different ways.

With "true BS7671 {L-E} faults", they are not really concerned about the potentially very high fault currents but, rather, the risk to persons that arises from 'live' exposed-conductive parts, and hence is dealt with in Chapter 41 of BS7671 ("Protection against electric shock"). On the other hand, with 'short circuits' (L-N faults), the concern is specifically about the risk (primarily to cables, hence possible fire risks etc.) due to the (potentially very high) 'fault' current, per se, and is hence dealt with in Chapter 43 ("Protection against overcurrent").

Most of the time on a TN install, even if a RCD is present we still design circuits such that MCBs will provide earth fault protection. Since the PN loop is normally lower than the PE loop this means they will also provide short-circuit protection, but there seem to be a few cases where this would not be the case.

I think you are getting concerned about a problem which doesn't exist.

For a start, with TN-C-S the L-N and L-E loop resistances are obviously the same thing and (although not really relevant) with TT the L-N loop impedance is inevitably a lot lower than the L-E impedance. Hence, if it could occur, P-N impedance greater than P-E impedance could only occur with TN-S, and I struggle to think of any credible situations (other than a faulty neutral distribution cable) that could result in that.

However, much more to the point, loop impedance is not, in practice, really relevant to 'short circuit protection'. The requirements of Chapter 43 already require that circuits (cables) be adequately protected against even very modest 'over-currents', essentially requiring (in the 'worst-case, of a circuit in which Iz = In) an OPD (assuming MCB) to operate (eventually) if the current exceeds 1.13 times the CCC of the cable and to operate 'immediately' for currents greater than, say (for a Type B MCB), 3-5 x CCC.

The tabulated values of CCC (Iz) we use are such that, together with the known characteristics of OPDs, if the In of the OPD is appropriate (i.e. In≤Iz) the cable will come to no harm - and that equally true of a massive current due to a short circuit as it is of a modest overload (>1.13 x Iz) due to excessive loads having been connected to the circuit. In other words, if In≤Iz, then the OPD will operate at currents far less than would result from a short circuit.

So what if-anyting are the rules on short-circuit protection.

See above. There is no need for specific rules about 'short-circuit protection', since satisfaction of the requirements of Chapter 43 (essentially In≤Iz) means that the OPD will operate at 'overload' currents far less than those that would result from a short-circuit - and, in particular, that satisfying those requirements will ensure that the cable is not damaged, no matter what the current (for as long as the OPD allows the current to flow).

Put another way, if satisfying In≤Iz did not give adequate protection against short circuits, that could only be because the tabulated values of Iz (CCC) were incorrect.

Does that help?

Kind Regards, John

 

[JohnW2]

I suppose I should have talked about (since you mentioned it ...

1. If multiple circuits of different sizes are running in the same containment system and a short circuit develops between conductors of different circuits.

That hypothetical scenario could, indeed, present a problem, but I don't think that (realistically) there is anything one can do about it.

If, say, the L of a 45A or 50A shower circuit 'shorted' to the N of, say, a lighting circuit, then the N of that lighting circuit would, indeed, find itself carrying a potentially 'dangerously high' current, which could damage that conductor and maybe even result in a fire.

However, unless you are going to require that every circuit in an installation uses cable with the CSA required by the highest current circuit in the installation (e.g. requiring 6mm² or 10mm² for lighting circuits), I can't see that one can realistically do anything to protect against that (presumably incredibly unlikley) situation

Edit: I suppose that the message raised by your concern is that one should not put cables of widely differing sizes (hence widely different CCCs) in situations in which there could conceivably be a 'short' between two of them (e.g. avoid pointing cables of widely different sizes in the same 'containment system').

Kind Regards, John

 

[ericmark]

That hypothetical scenario could, indeed, present a problem, but I don't think that (realistically) there is anything one can do about it.

I would expect the RCBO would trip, however I suppose unless double pole then could have excessive neutral currents, but seems unlikely. So I would say you can do something about it, the question is would you do anything?

 

[JohnW2]

I would expect the RCBO would trip, however I suppose unless double pole then could have excessive neutral currents, but seems unlikely. So I would say you can do something about it, the question is would you do anything?

If, per plugwash's scenario, the L of one circuit 'shorted to' the N of a different circuit, then if one (or both) were protected by an RCD that was not also protecting the other one, then at least one RCD would immediately operate. "RCDs" (but not RCBOs) are (I think) invariably DP, so one operating would clear the fault. I must confess that I hadn't considered the possibility of a SP RCBO - but, as you imply, that could leave a vert high 'short-circuit' current flowing if 'the other' circuit were not RCD or RCBO protected,

However, plugwash presumably knows that so the fact that he raised the issue I guess he was thinking of the situation in which neither circuit was RCD protected OR the situation in which both circuits were protected by the same RCD - when, as I said, there IS theoretically a 'potential problem'.

However, the scenario he postulated in one in which the L conductor of one cable and N conductor both lose their insulation (and also their sheaths, if they have them AND the exposed conductors come into contact AND (if there is RCD protection) neither of the conductors come into contract with anything earthed. Other than if the whole show was 'engulfed in flames' (in which case I would think there would be greater things to worry about than the 'short circuit'!), I would think we are in the territory of the annual risk of being struck by lightning (or being killed by the AZ Covid-19 vaccine!) and may even be heading towards the probability of winning the Lottery - so I wouldn't personally loose any sleep over this 'risk'!

Anyway, as I've said, for the ultra-cautious amongst us, the (tiny) risk can be essentially eliminated by keeping cables of appreciably different CSAs which are protected by the same RCD well away from one another!

As for his actual question ("So what if-anyting are the rules on short-circuit protection.?"), as I replied to him, there is no need for the regs to have any requirements/'rules' specifically about 'short-circuit protection', since a short-circuit is merely an extreme case of the much more modest (L-N) 'overload' currents that the regs already require there to be protection against.

Kind Regards, John

 

[333rocky333]

Most short circuits i assume are caused by the Electricians themselves cutting the wrong cable

 

[JohnW2]

Most short circuits i assume are caused by the Electricians themselves cutting the wrong cable

Maybe! As I said, the sort of short-circuit plugwash postulated, between (insulated, and maybe also sheathed) conductors within two cables, simply because they are in proximity to one one another, must be incredibly improbable. However, faults within equipment, or even accessories etc., can (and do) cause 'short-circuits') - do you remember my daughter's RCD which went bang and took out the cutout fuse when she pressed the test button?!

Returning to my reply to plugwash's question, it's occurred to me that there is another bit of 'explanation' I could have added...

The reason why fault loop impedance is important (to the regs) for L-E faults ('earth faults'), but not L-N ones ('short-circuits') is because (since it is about reducing the risk of electric shock), in the former case (but not the latter) very rapid disconnection is required.

In the case of a ('negligible impedance') L-E fault, BS7671 require disconnection within 0.4 seconds (with TN - 0.2 secs for TT, for whatever reason!) (i.e. in the case of MCB protection, 'magnetic tripping'), which is why low Zs ('EFLI') figures are needed. However, in the case of an (L-E) 'overload (of which a 'short-circuit' is the most extreme example) all that it is required is that the fault is cleared before the cable comes to any harm - and, as I've said, the tabulated CCC figures we work with assume that (with knowledge of characteristics of the protective device), the cable will come to no harm with a current above its CCC provided only that it is protected by a device with an In no greater than the Iz ('CCC') of the cable.

So, for example, although reqs would require disconnection within 0.4 seconds (with TN) in the event of an L-E fault, in the case of an 'overload' the tables are telling us that the protective device will operate before any harm comes to the cable, no matter how long that may take - e.g. if it were cable with a CCC of 20A protected by a 20A MCB, then the MCB would allow 29A to flow for about an hour, since the cable is deemed to be able to carry that current for that long without coming to harm.

Kind Regards, John