Basement CU

[JohnW2]

I did not say that, and to me machines spinning down would be the least of my worries if I suddenly found myself in a pitch black room.

As I've indicated, it's not a major one of my worries, either, but it essential is the issue which this discussion is meant to have been about. As you go on to agree:

I have no problem mitigating all the mechanisms which may lead to what I consider to be the hazard here, which it lights going out in the dark, but some are easier to mitigate than others. The mechanisms I can think of, some of which we've discussed, are: <list of 7 possible mechanisms>

Indeed - and, as I've said, the issue of discrimination, which you jumped in to discuss at length, would only be useful (reduce, but not eliminate, the hazard) in relation to just one of those seven mechanisms.

Clearly we have agreed the a reasonable mitigating measure for most of these (save a loose/poor connection on the switch wiring) would be emergency lighting ...

Exactly, which is why I have mentioned it more than once. In fact, AFAICS battery-operated emergency lighting would be the only effective measure (at least, of those reasonably implemented in a domestic setting) that would be useful in relation to 4 or 5 of those possible mechanisms.

With regards to discrimination therefore, which seems to be what you are suggesting cannot be reasonably achieved, I agree that sometimes it is difficult, but it can be done with reasonable due care, a PFCC reading, and a few calculations, most of which you can do in Excel, or even by hand.

[remembering that this relates to just one of the 7 or more possible mechanisms of the hazard] ... One can obviously always achieve a reasonable, perhaps high, probablitiy of achieving discrimination by 'proper' design but, as you say, in a domestic installation, the question relates to what is 'reasonable'. As I've indicated, one hardly needs to do any serious calculations, since the situation/options (domestically) are fairly obvious. If there is a downstream 32A MCB, you are not going to get reliable discrimination from any domestic MCB, since even a 50A MCB upstream will not give reasonable discrimination. Even with a 20A downstream MCB, discrimination probably wouldn't be that assured from a 50A upstream one. Hence, one's only chance of reasonable discrimination would be to have submain protected by a 60A, 80A or 100A fuse, with appropriate cable, (not much point in going above the In of the cutout fuse). However, is it reasonable', in a domestic setting, to go to that length to supply '2 lights and two sockets' in a basement'? Surely it would be far more 'reasonable' to either have emergency lighting or else run the lights from some other circuit? The latter, of course, would still leave several (most) of the possible mechansims of the hazard still present.

Kind Regards, John

 

[JohnW2]

At the end of the day if you beleive that loss of local lighting will create a particular hazard then perhaps the easiest way is to mitigate it with one or more 8w non maintained bulkheads, although in domestic it would have to be pointed out to the client that the batteries do not last for ever (A full BS5266 compliant solution complete with log book, as fitted drawings, etc might be a bit over skill for a garden workshop )

Quite so - particularly given that maintained lighting is really the only solution which would address the majority of possible causes of loss of lighting.

As to discrimination, if you were for example doing a design and build contract on say a secondary school for health centre for the local authority then the spec would generally insist quite rightly on total discrimination between devices in final DBs and submain protection, which isn't actually that big a head ache to prove, there is at least one design package that will run through the design checking discrimination at each point, telling you at what fault levels it is acheived at each point etc, using manufacturer specific data for the family of breakers you have specified

Sure - but, as I keep saying/asking, how far can/should one 'reasonably go' in order to supply 'two lights and two sockets' in a domestic basement?!

Even in the situation of an industrial, commercial or public building, what degree of discrimination can be achieved will presumably depend upon what 'spread of currents' one has got to play with, no matter how much calculation one does - even more so if one considers faults of less than 'negligible impedance' (i.e. 'bolted faults')? As I've been saying, in a domestic installation with a downstream 32A MCB and a cutout fuse of probably 60A or 80A, one hasn't got all that much room for manoeuvre (although admittedly one can probably help things a bit by using an upstream fuse with different characteristics from the cutout one).

Kind Regards, John