Relying on loads not being able to overload

[mfarrow]

You appear to be totally confusing the concepts of 'fault' and 'overload', at least in the language of BS7671. Overload, as we (and BS7671) understand it, is (or can be) the antithesis of an adiabatic process - so that adiabatic calculations are inapplicable to, and no help in determining the fate of cable in, 'overload' situations.

I can see now why the adiabatic will not work (coming off a train was the wrong time to be thinking about this) however my point about 'fault condition' (in inverted commas, as I had written) was meant in the generic context, i.e. a condition which is not normal, and I've been going through my head considering the examples of such an occurance in fixed wiring which are not protected by another requirement of BS 7671. For the time being I have yet to find one!

I suppose the closest would be IT installations, where e.g. a short to earth on one line can cause a potential rise in the other without causing disruption of the supply, and who's voltage a)can be raised above LV, or the rated voltage of the cable, and b) allow someone to more easily kill themselves by touching that other line, by introduction of an earth reference. However these are controlled installations and I don't think that really counts.

With regards to your shorted heating element, what you're suggesting would introduce an overload to earth, or rather the cpc of the circuit, in many instances. If the majority of the current were to flow through this cpc (as I suspect it would), not the neutral, then the cpc would, in many domestic circuits, be undersized to support the sustained overload. Would you therefore suggest installing T&E with equally-sized conductors?

 

[mfarrow]

Oops

 

[EFLImpudence]

I'm not sure that I even fully understand what you feel is 'accepted' (or acceptable).

Yes you do.

Do I take it that, for example, you are saying that for someone (like yourself) who has adequate equipment and knowledge to verify that the circuit is satisfactory (assuming, of course, that it does prove to be satisfactory), you feel that it is acceptable not 'fuse down' when running lighting (wired in 1.0mm² or 1.5mm² cable) from, say, a 32A ring or radial final circuit?

Another can of ifs and buts.

Well, strictly speaking, 1mm² may have to be fused because it is not allowed if this is considered a power circuit - whatever that is.

1.5mm² would be alright on its own but if a lighting circuit is limited to 6A because of the accessories then, I suppose, the same would apply.


However, taking the conductors alone, is there any point fitting a 3A fuse for a 20A cable with a quarter amp load which cannot cause an overload.


I know we argue about a lot of the regulations but does not the fact that it is allowed mean someone has thought about it before and deemed it safe?

 

[JohnW2]

I can see now why the adiabatic will not work (coming off a train was the wrong time to be thinking about this) however my point about 'fault condition' (in inverted commas, as I had written) was meant in the generic context, i.e. a condition which is not normal ....

Fair enough. However, a distinction between different types of over-current situations is important, in electrical as well as regulatory terms. In particular, 'faults' (again everyday sense) of negligible impedance (i.e. what BS7671 calls "faults"), such that the current is limited only by Zs of the circuit, are a very different kettle of fish from any other type of 'fault' (everyday sense).

I think the we (and BS7671) are really stuck with the fact that it is usually impractical to achieve (or demand) 'satisfactory' disconnection times in response to faults of low, but not 'negligible' impedance.

... and I've been going through my head considering the examples of such an occurance in fixed wiring which are not protected by another requirement of BS 7671. For the time being I have yet to find one!

I'm not totally sure what you're saying here. Are we not simply talking about what BS7671 calls 'overload', and the reason I started this thread - i.e. to stimulate discussion about what sorts of hard-wired loads can, at least under some circumstances, result in such 'overload' (i.e. excessive current, but less than that due to a fault of negligible resistance). I have suggested that equipment/appliances which include, for example, heating elements and/or motors are, at least in theory, capable of producing such 'overloads' - and the only primary protection in such situations is that provided by an OPD.

With regards to your shorted heating element, what you're suggesting would introduce an overload to earth, or rather the cpc of the circuit, in many instances. If the majority of the current were to flow through this cpc (as I suspect it would), not the neutral, then the cpc would, in many domestic circuits, be undersized to support the sustained overload. Would you therefore suggest installing T&E with equally-sized conductors?

That, of course, is another problem, which the regs do not appear to address. They seem to think only in terms of negligible impedance L-CPC faults (which is required to be disconnected very rapidly), in which situation one can undertake an adiabatic calculation to determine the required CPC csa. If a persistent over-current flows primarily through L and CPC then, as you say, the CPC may be inadequate. Indeed, even at a current equal to the CCC of the (L & N) of the cable (let alone an 'over-current'), the CPC would not be theoretically 'adequate' should that current flow through L and CPC, rather than through L and N. I think this is probably another example of the regs not really considering L-CPC faults of more than 'negligible' impedance. Motors, of course, are capable of producing 'overload' currents which flow through only L and N.

More generally, I would be interested to know how you would answer the question I recently posed to EFLI, about not 'fusing down' prior to 1mm² or 1.5mm² lighting wiring fed from a 32A circuit.

Kind Regards, John

 

[ekmdgrf]

The inverse of "what harm could it do?" is "does it need to be there?". There's an old engineering adage that "the perfect solution is attained not when there is nothing more to add but when there is nothing more to take away."

In my mind, if the current should not exceed a given threshold, then that is where the breakers should be set. If the current consumption goes out of spec then the current should be disconnected. It's very well to argue that a situation where an appliance draws over it's specified current but below that of the (oversized) breaker cannot be conceived but in my experience things tend to go horribly wrong because something that was perceived as not possible or completely unforeseen goes and happens.

I recall a documentary I saw about Swiss Air 111, which had an in-flight fire propagated by MPET insulation. It was pointed out that MPET had been extensively tested by the FAA and found not to be able to support a flame. Interestingly the documentary went on to wax lyrical about Kapton which would replace MPET due to it being extensively tested and found to be safe. Strange, as I had a feeling in the back of my mind that I saw another documentary about an aircraft being lost due to its Kapton wiring catching fire... While checking some details of SR111, I also came across "Investigators identified evidence of arcing in wiring of the in-flight entertainment system network, but this did not trip the circuit breakers."

 

[JohnW2]

Do I take it that, for example, you are saying that for someone (like yourself) who has adequate equipment and knowledge to verify that the circuit is satisfactory (assuming, of course, that it does prove to be satisfactory), you feel that it is acceptable not 'fuse down' when running lighting (wired in 1.0mm² or 1.5mm² cable) from, say, a 32A ring or radial final circuit?

Another can of ifs and buts. Well, strictly speaking, 1mm² may have to be fused because it is not allowed if this is considered a power circuit - whatever that is.

Fair enough. Let's forget about 1mm² then - I'm not even sure that 'fusing down' would necessarily be acceptable if the circuit as a whole was deemed to be ' a power circuit'!

1.5mm² would be alright on its own but if a lighting circuit is limited to 6A because of the accessories then, I suppose, the same would apply.

Given the discussion we're having, I'm not sure that this "limited to 6A because of the accessories" concept is valid, is it? If the assumption is that the load is such that the current flowing in this part of the circuit (including through the accessories) can never get anywhere near 6A, then the rating of the accessories becomes an irrelevance, doesn't it?

However, taking the conductors alone, is there any point fitting a 3A fuse for a 20A cable with a quarter amp load which cannot cause an overload.

If one is convinced that the load cannot cause an overload, then, in electrical terms, there clearly is not any point. However, as I've just been discussing with mfarrow, as well as 'overload' in the sense that you probably think about it, one perhaps should also consider the possibility of a low, but not 'negligible', impedance L-CPC fault - which I suppose is always a theoretical possibility, even when the load itself cannot really result in an 'overload'.

I know we argue about a lot of the regulations but does not the fact that it is allowed mean someone has thought about it before and deemed it safe?

I'm not sure you fully understand why I started this discussion. For a start, it is only 'allowed' if one is convinced that 'overload' is impossible, so I wanted to see opinions as to what loads could, and could not, result in 'overloads'.

However, perhaps more importantly, I was trying to gain an understanding of what electricians actually do (which, for reasons you've mentioned, may be different from what you would advise DIYers to do). As you say, it can easily be argued that 'there is no point' in interspersing, say, a 3A fuse when one connects, say, a single light fitting with 1.5mm² cable to a 32A-protected sockets circuit - but would you, as an electrician, actually do it without a fuse (assuming that the circuit checked out OK)?

Kind Regards, John

 

[JohnW2]

In my mind, if the current should not exceed a given threshold, then that is where the breakers should be set. If the current consumption goes out of spec then the current should be disconnected. It's very well to argue that a situation where an appliance draws over it's specified current but below that of the (oversized) breaker cannot be conceived but in my experience things tend to go horribly wrong because something that was perceived as not possible or completely unforeseen goes and happens.

It sounds as if you are agreeing with my 'discomfort' about what is being suggested.

Indeed, you are also really re-iterating the feeling which underlies my 'discomfort'. As you say "...something that was perceived as not possible or completely unforeseen goes and happens". In other words, in context, one can never really be absolutely certain that the overload/over-current situations being discussed could never arise (even if by mechanisms as yet not even considered) - and if one cannot be certain that it is impossible, then to use 'undersized cable' (or an 'oversized OPD') is, as far as I can see, not even 'allowed' by the regs.

I have to say that, despite views being expressed in this thread (and others), I rather doubt that many electricians would actually do the things about which I am uncomfortable!

Kind Regards, John

 

[EFLImpudence]

1.5mm² would be alright on its own but if a lighting circuit is limited to 6A because of the accessories then, I suppose, the same would apply.

Given the discussion we're having, I'm not sure that this "limited to 6A because of the accessories" concept is valid, is it? If the assumption is that the load is such that the current flowing in this part of the circuit (including through the accessories) can never get anywhere near 6A, then the rating of the accessories becomes an irrelevance, doesn't it?

Yes, it does but I thought that was the conclusion we drew when this was discussed.
If that is now the case then lighting circuits may as well be protected by 10A, 16A or 20A MCBs depending on the cable used.

However, as I've just been discussing with mfarrow, as well as 'overload' in the sense that you probably think about it, one perhaps should also consider the possibility of a low, but not 'negligible', impedance L-CPC fault - which I suppose is always a theoretical possibility, even when the load itself cannot really result in an 'overload'.

But how does it happen?
Perhaps it is not considered for a reason.

I'm not sure you fully understand why I started this discussion. For a start, it is only 'allowed' if one is convinced that 'overload' is impossible, so I wanted to see opinions as to what loads could, and could not, result in 'overloads'.

I would only do it when convinced.

However, perhaps more importantly, I was trying to gain an understanding of what electricians actually do (which, for reasons you've mentioned, may be different from what you would advise DIYers to do). As you say, it can easily be argued that 'there is no point' in interspersing, say, a 3A fuse when one connects, say, a single light fitting with 1.5mm² cable to a 32A-protected sockets circuit - but would you, as an electrician, actually do it without a fuse (assuming that the circuit checked out OK)?

Yes. Why not?

It is more likely to be done for oven connections and similar things rather than connecting a light to a ring.


I think the distinction is between designing satisfactory circuits and 'coverall' solutions as with the OSG.

 

[JohnW2]

If the assumption is that the load is such that the current flowing in this part of the circuit (including through the accessories) can never get anywhere near 6A, then the rating of the accessories becomes an irrelevance, doesn't it?

Yes, it does but I thought that was the conclusion we drew when this was discussed.

The two issues are obviously inter-related. The (often 6A) 'rating'of lighting accessories only becomes an irrelevance if one is convinced that the circuit/loads could never result in more than 6A flowing through the accessories.

If that is now the case then lighting circuits may as well be protected by 10A, 16A or 20A MCBs depending on the cable used.

As above - if one is convinced. I don't have the regs to hand, but IIRC, there is actually an explicit maximum of 16A for a lighting circuit.

However, as I've just been discussing with mfarrow, as well as 'overload' in the sense that you probably think about it, one perhaps should also consider the possibility of a low, but not 'negligible', impedance L-CPC fault - which I suppose is always a theoretical possibility, even when the load itself cannot really result in an 'overload'.

But how does it happen? Perhaps it is not considered for a reason.

It surely could easily happen. An L-CPC 'short' due to a fault (everyday sense) is surely a fairly haphazard thing and is therefore not necessarily a 'perfect connection' - I imagine that it could easily have an impedance of, say, 0.5Ω almost as easily as 0.01Ω, couldn't it?

For a start, it is only 'allowed' if one is convinced that 'overload' is impossible, so I wanted to see opinions as to what loads could, and could not, result in 'overloads'.

I would only do it when convinced.

Fair enough. As a matter of interest, which loads are you convinced cannot result in 'overload' currents - lighting (all lighting?), immersions, ovens/cookers, fans ....??

... but would you, as an electrician, actually do it without a fuse (assuming that the circuit checked out OK)?

Yes. Why not? ... I think the distinction is between designing satisfactory circuits and 'coverall' solutions as with the OSG.

Interesting. What you say obviously has a lot of truth in it. As for 'why not?', that obviously becomes difficult to answer if one is convinced that overload current is impossible. However, one issue is potential 'nuisance value' for your customer - since I strongly suspect that a goodly proportion of 'less thoughtful' electricians would, at least initially, find fault (e.g. in an EICR) with what you had done. Of course, they might even be 'right' to do so if they were less convinced than you about the impossibility of an overload current.

Kind Regards, John

 

[EFLImpudence]

As above - if one is convinced. I don't have the regs to hand, but IIRC, there is actually an explicit maximum of 16A for a lighting circuit.

There is but I'm not sure why.

It surely could easily happen. An L-CPC 'short' due to a fault (everyday sense) is surely a fairly haphazard thing and is therefore not necessarily a 'perfect connection' - I imagine that it could easily have an impedance of, say, 0.5Ω almost as easily as 0.01Ω, couldn't it?

Could it?

As a matter of interest, which loads are you convinced cannot result in 'overload' currents - lighting (all lighting?), immersions, ovens/cookers, fans ....??

Not all lighting nor motors but anything with a filament.

However, one issue is potential 'nuisance value' for your customer - since I strongly suspect that a goodly proportion of 'less thoughtful' electricians would, at least initially, find fault (e.g. in an EICR) with what you had done.

I do not have to install with regard to other's incompetence.

Of course, they might even be 'right' to do so if they were less convinced than you about the impossibility of an overload current.

A bit more than 'less convinced' perhaps but if they were correct then I would have been wrong.


The best example is a 13A oven connected to a normal cooker circuit after the removal of the plug - fault current and other requirements satisfactory.
Uneasy?

 

[JohnW2]

It surely could easily happen. An L-CPC 'short' due to a fault (everyday sense) is surely a fairly haphazard thing and is therefore not necessarily a 'perfect connection' - I imagine that it could easily have an impedance of, say, 0.5Ω almost as easily as 0.01Ω, couldn't it?

Could it?

I would have thought pretty easily. A flapping bit of (perhaps rusty) wire happening to touch something earthed could surely result in a a 'less-than-perfect' connection?

Not all lighting nor motors but anything with a filament.

Fair enough. I take it (from what you've been saying) that 'filament' includes any sort of heating element?

Of course, they might even be 'right' to do so if they were less convinced than you about the impossibility of an overload current.

A bit more than 'less convinced' perhaps but if they were correct then I would have been wrong.

That's part of the problem. We are talking about opinions, not absolute rights and wrongs. Opinions will vary.

The best example is a 13A oven connected to a normal cooker circuit after the removal of the plug - fault current and other requirements satisfactory. Uneasy?

I still think you may be somewhat missing my point. Given the seemingly pretty massive 'safety margins' that appear to exist in relation to the cable CCCs we works with, I don't suppose that any of the scenarios we're talking about would often, if ever, result in any harm coming to the 'undersized' cable. In that sense, I'm not particularly 'uneasy'. However, by the same token, the chances of the wiring of a 1.5mm² 32A ring final ever coming to any harm would probably be extremely low. However, we don't do it, and are not allowed to do it. What I'm talking about is what is 'allowed' in terms of the regs - and what I'm a little 'uneasy' about is claiming compliance on the basis of an opinion about what 'can never happen'.

I really would love to hear some other electricians' views about this!

Kind Regards, John

 

[EFLImpudence]

I imagine that it could easily have an impedance of, say, 0.5Ω almost as easily as 0.01Ω, couldn't it?

That would be 480A; what is going to happen?

Shouldn't you be more concerned with a 'fault' of 6Ω ?

 

[JohnW2]

I imagine that it could easily have an impedance of, say, 0.5Ω almost as easily as 0.01Ω, couldn't it?

That would be 480A; what is going to happen?

Well, it would be 480A if Zs of the circuit was zero. More realistically, it's more likely to be about 160A. However ....

Shouldn't you be more concerned with a 'fault' of 6Ω ?

I think you're probably being a bit unfair! As I'm sure you realise, I was just making a general point, and my typing fingers just pulled a 'low but not zero' figure (which happened to be 0.5&#937 out of the air, without stopping to do any calculations Your suggested 6Ω 'fault', plus a bit for Zs, would only result in a current of only around 35A or so - which, although a B32 would allow it to flow 'indefinitely', I probably wouldn't be too worried about with 1.5mm² cable (we agreed to forget about 1mm², for other reasons!). If I had tried to be precise, I probably would have talked about a fault of a couple of ohms or so.

Kind Regards, John

 

[THRIPSTER]

Hi John,

Why don't you pose this question on the IET forum?

I must confess to having the same concerns as you, originally. However, once I understood that most motor circuits have their own overload protection control gear and that the chances of equipment such as showers and immersion heaters would have a very low risk of failing to an overload condition, then I just accepted it. But hey, what do I know?

As a matter of interest, is the W2 in your username because you are located in London, W2?

Regards

 

[ekmdgrf]

So on the one side, one can argue that it's very hard to have an overcurrent fault and one hopefully has smoke detection. One could also say given the odds there are problem statistically more significant issues in the bigger picture that one should address before this.

There again on the flip side, one could easily argue, when one has to select a breaker, why select one greater than one needs.

As the owner (and I stress, not installer) of the installation that precipitated this debate, my guess is the that equestrian installer in question selected it because he'd purchased an off the shelf CU that came with a range of MCBs, so he used that one. As the old saying goes "when all one has is a B32, there's a tendency for everything to look like 4mm t&e".