Relying on loads not being able to overload

[JohnW2]

In recent times, I have on several occasions had to qualify my comments about OPDs which were were rated too high to give adequate overload protection to the cable concerned with something like "... unless you argue that the (hard-wired) load in question is not able to result in an overload situation". The most recent example was 2.5mm² (and some 1mm² !) cable serving a dedicated immersion heater circuit but protected by a B32.

In such situations, if the circuit is such that the OPD can provide adequate fault protection (or if it's a TT installation in which fault protection is reliant on RCDs, anyway), if one argues that overload is not possible, one can seemingly have a compliant set-up even if the In of the OPD considerably exceeds the CCC of the cable.

I really am not very comfortable with that approach, and certainly would not want it taken in my home. The nearest to a rational electrical argument for my discomfort that I can come up with is that, in fairly exceptional circumstances, many fixed loads (e.g. those which include heating elements or motors) can theoretically result in 'overload' situations - i.e. drawing excessive currents, but not high enough to invoke 'fault protection' measures.

What do others think?

Kind Regards, John

 

[EFLImpudence]

Firstly, I would say that as in the recent thread there is no reason to fit a 32A OPD on a circuit with 2.5mm² and 1mm² cables unlike situations where the omission of overload protection could be convenient and satisfactory.

However

The nearest to a rational electrical argument for my discomfort that I can come up with is that, in fairly exceptional circumstances, many fixed loads (e.g. those which include heating elements or motors) can theoretically result in 'overload' situations - i.e. drawing excessive currents, but not high enough to invoke 'fault protection' measures.

How ? Without exceptional heating elements.

I would not include motors in the argument.

 

[JohnW2]

Firstly, I would say that as in the recent thread there is no reason to fit a 32A OPD on a circuit with 2.5mm² and 1mm² cables unlike situations where the omission of overload protection could be convenient and satisfactory.

Agreed - in such a situation I can see absolutely no justification for not having an appropriate size of OPD.

However ...

The nearest to a rational electrical argument for my discomfort that I can come up with is that, in fairly exceptional circumstances, many fixed loads (e.g. those which include heating elements or motors) can theoretically result in 'overload' situations - i.e. drawing excessive currents, but not high enough to invoke 'fault protection' measures.

How ? Without exceptional heating elements. I would not include motors in the argument.

A short between the element and earthed casing 'a little way' from the L end could result in a low, but not 'negligible', impedance current path - resulting in a current not large enough to operate the magnetic part of an MCB but potentially considerably in excess of the CCC of the cable, couldn't it?

The same can happen with winding-earth and winding-winding shorts in motors - and, of course, a motor can also stall/jam and then present an overload.

Kind Regards, John

 

[EFLImpudence]

A short between the element and earthed casing 'a little way' from the L end could result in a low, but not 'negligible', impedance current path - resulting in a current not large enough to operate the magnetic part of an MCB but potentially considerably in excess of the CCC of the cable, couldn't it?

I assume you are thinking of something over 35A (for a fair time) to damage the 2.5mm² in the usual circuit (I'll discount the 1mm² in the thread).

If it is indeed possible for this to happen I presume that the filament will act as its own fuse before nearly three times its design current flows.

 

[JohnW2]

I assume you are thinking of something over 35A (for a fair time) to damage the 2.5mm² in the usual circuit ...

Indeed. There's no reason why it could not be a lot more than 35A (if the short to earth were fairly close to the L end of the element). You will probably argue that even a B32 would then disconnect 'fairly quickly' (say, within a few minutes), and that the cable might not be harmed - but that is surely not within either the word or spirit of the regs, is it?

(I'll discount the 1mm² in the thread).

I don't really think that you can 'discount it', since the general argument (about a load 'which cannot produce an overload') can, provided fault protection requirements are satisfied, be applied as much to a 1mm² cable protected by a B32 as it can to a 2.5mm² cable protected by a B32.

If it is indeed possible for this to happen I presume that the filament will act as its own fuse before nearly three times its design current flows.

You may well be right (at least, in many cases) , but I don't see any reason why that is inevitable - so can/should one rely on it?

Kind Regards, John

 

[mfarrow]

if one argues that overload is not possible, one can seemingly have a compliant set-up even if the In of the OPD considerably exceeds the CCC of the cable.

I really am not very comfortable with that approach, and certainly would not want it taken in my home.

So you don't have any ring mains? Or unfused spurs off them? Or lighting pendants? And you're re-wired your cooker internals with 4mm cable?

 

[JohnW2]

if one argues that overload is not possible, one can seemingly have a compliant set-up even if the In of the OPD considerably exceeds the CCC of the cable. I really am not very comfortable with that approach, and certainly would not want it taken in my home.

So you don't have any ring mains? Or unfused spurs off them?

As I thought was apparent, I was talking specifically about dedicated circuits serving a single hard-wired load (which allegedly could not produce an overload).

Socket circuits of any description obviously present different issues since, whether ring finals or radials, overload is always possible because the designer has no control over what (and how many) loads are 'plugged in'; overload protection per the usual formulae is therefore essential. Ring finals, with or without spurs, obviously present particular issues (making the concept of such circuits 'controversial' in some people's minds), since they cannot be guaranteed never to overload any of the cable.

Are you implying that you would be totally comfortable with a dedicated immersion heater circuit wired in 1mm² (clipped direct) and protected by a B32 (or maybe even B40/B45), if fault protection requirements could be satisfied (and if 1mm² for a 'power' circuit were allowed, which IIRC it is not - otherwise rephrase my question in terms of 1,5mm²!)?

Kind Regards, John

 

[mfarrow]

I live in a house with 253V on the supply, persistently, so to me overload is not uncommon. However, that is mainly on circuits where the overload protection rating is exceeded by the CCC of the cable.

For overload on circuits without suitable protection, ie rated above the cable's CCC, then I suppose we need to understand what else is in place to protect against 'fault conditions'. I believe I need to go back to the adiabatic and work out whether there is a risk of damage/fire, or not, for certain sets of boundary conditions, and that would probably give me the answer I, and you, are seeking.

 

[EFLImpudence]

Indeed. There's no reason why it could not be a lot more than 35A (if the short to earth were fairly close to the L end of the element). You will probably argue that even a B32 would then disconnect 'fairly quickly' (say, within a few minutes), and that the cable might not be harmed - but that is surely not within either the word or spirit of the regs, is it?

No, but I am sure the filament would just burn out bearing in mind that the contact will only be a touch.

(I'll discount the 1mm² in the thread).

I don't really think that you can 'discount it', since the general argument (about a load 'which cannot produce an overload') can, provided fault protection requirements are satisfied, be applied as much to a 1mm² cable protected by a B32 as it can to a 2.5mm² cable protected by a B32.

That's true but then we only have to consider current up to say 25A.
This is still almost twice the filament design current and whilst I don't have any confirmation would the filament not still just melt?

You may well be right (at least, in many cases) , but I don't see any reason why that is inevitable - so can/should one rely on it?

In the accepted belief that an element cannot cause an overload I suppose we can.

 

[JohnW2]

For overload on circuits without suitable protection, ie rated above the cable's CCC, then I suppose we need to understand what else is in place to protect against 'fault conditions'. I believe I need to go back to the adiabatic and work out whether there is a risk of damage/fire, or not, for certain sets of boundary conditions ....

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.

That is, I would imagine, the main reason why the regs generally require than the In of an OPD should not exceed the CCC of the cable it protects - since that is a relatively 'fail safe' approach to take (particularly, as I assume has been shown to be the case, the operating characteristics of OPDs accommodate the true tolerance of cables to overloads beyond their "CCCs").

Kind Regards, John

 

[JohnW2]

... This is still almost twice the filament design current and whilst I don't have any confirmation would the filament not still just melt?

You may well be right (at least, in many cases) , but I don't see any reason why that is inevitable - so can/should one rely on it?

In the accepted belief that an element cannot cause an overload I suppose we can.

You seem to just be asserting a 'traditional belief' in order to create a circular argument - i.e. you are using an assumption than an element cannot produce an overload to 'prove' that it must melt under persistent over-current conditions!

I think we have agreed that there is no real justification for protecting a 2.5mm² dedicated immersion circuit with a 32A MCB. However, are you saying that, even if there is no 'justification', it would be electrically totally satisfactory, and that you would be no less 'comfortable' about it than you would with a 16A or 20A MCB?

Kind Regards, John

 

[JohnW2]

I suppose one of my concerns is the confusion this 'approach' could create for some of the 'less thoughtful' of electricians (not to mention DIYers) around. To teach them that it is never acceptable to have an OPD with an In greater than that of the tabulated CCC of the cable (with the installation method employed) is a pretty simple, 'fail safe' rule for them to apply. If they start hearing/seeing people suggest that there are situations in which they are allowed to break this 'rule', there is clearly scope for mistakes due to less than full understanding.

Kind Regards, John

 

[EFLImpudence]

I may be doing as you say (circular Argument) because it is accepted much the same as diversity with cooking appliances with which you seem equally uncomfortable and I cannot prove the diversity figures either but it has been accepted and worked for a very long time.
I could say you were doing an opposite circular argument in accepting it is compliant but you are uneasy about it.

I never tell DIYers to install 'undersized' cables because they do not have the equipment to verify it is satisfactory.

 

[JohnW2]

I may be doing as you say (circular Argument) because it is accepted ...

I suppose my main reason for starting this thread was in order to ascertain 'how accepted' it actually was, but I haven'y yet got many views!

... much the same as diversity with cooking appliances with which you seem equally uncomfortable and I cannot prove the diversity figures either but it has been accepted and worked for a very long time.

Particularly given that, as we know, there appear to be very generous 'safety margins' built into the CCC figure we work with, I wouldn't say that I am at all uncomfortable with the concept of diversity, which is a perfectly valid statistical concept. Indeed, if I were uncomfortable with that concept, I'd probably be very unhappy to see CUs containing MCBs which 'added up' to 150A+ with a 100A main switch and probably a 60A or 80A service fuse! As for the detail of diversity of cooking appliances, as you say, we have been using the current figures, without change, for many years, during which period both cooking appliances and cooking practices have probably changed a lot - so I do think that someone should probably revisit the situation and attempt to confirm that the diversity calculations we are using are still 'appropriate'.

I could say you were doing an opposite circular argument in accepting it is compliant but you are uneasy about it.

That's not quite what I say. I accept that the 'undersized cables' can be compliant IF (amongst other things) it can be successfully argued that the load in question 'cannot produce an overload'. However, as must be apparent from this whole thread, I am not necessarily convinced that some of the loads in question are all that guaranteed never be able to produce an overload. That brings me back to the reason for this thread - to see if people are convinced about the impossibility of an overload.

I never tell DIYers to install 'undersized' cables because they do not have the equipment to verify it is satisfactory.

That's fair enough. Of course, they don't have the equipment to verify that the circuit is satisfactory even if they don't use 'undersized' cable - but I agree that it's more likely to be satisfactory if their cable is not 'undersized'!

Kind Regards, John

 

[JohnW2]

I may be doing as you say (circular Argument) because it is accepted .... I never tell DIYers to install 'undersized' cables because they do not have the equipment to verify it is satisfactory.

It's a pity that more people have not got involved in this discussion, since I really did want to try to find out what the 'general opinion' was.

In the meantime, I'm not sure that I even fully understand what you feel is 'accepted' (or acceptable). 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?

Kind Regards, John