Replacing a fused CU ?

[Andrew Gabriel]

I don't believe building trades people use on-line ordering much. I bought them from trade counters; IIRC, a combination of ERD (who stocked most things but wouldn't order anything else except in box quantities), and Gibbs & Dandy (who didn't stock much of the range, but would order anything even singularly).

The RCBOs you see there are all made up by me from MCBs and RCBO pods. The RCBO pods were under £50 each (might have been more if I was buying just one), and the MCBs well under £10 each (I tend to ignore the price of those).

Because I made them up from separate MCBs and RCBO pods, all the RCBOs in that picture are actually Type B. There would have been no issue with using Type C because by definition they're all RCD protected. If they weren't RCD protected, you'd have to double-check the earth fault loop impedance against more stringent values for Type C to ensure they still meet disconnect times.

The 6A lighting MCBs are all Type C, as is a 20A radial for things I didn't want RCD protected.

For the outdoor sockets circuit, which is a 20A radial Type C, I used a C20A 10mA RCBO prebuilt. That had to be ordered and didn't arrive until after I took the photo, but it looks exactly the same as the ones I assembled myself. At £60 IIRC, it was the most expensive component. I went for a Type C there because I could imagine it being used for large motor appliances. I went for 10mA because it's generally only used for one or two appliances so cumulative appliance leakage isn't an issue, and earth leakage paths through you outdoors can be subject to additional external resistances which might serve to limit the current below 30mA. That circuit is also TT, where as the rest are all TN-C-S.

Also bear in mind my prices are ~5 years ago.

[Andrew Gabriel]

That reminds me -- the RCBOs do get warm, as they contain functional electronics which runs continuously[1]. I rearranged them slightly after that picture was taken so they weren't all next to each other, so they would stay a bit cooler, not that I have any reason to think that might be a problem. Typical life of electronic components does halve which each

10C temperature rise though.

There are also requirements on not putting the more sensitive RCBOs next to high current circuits (40A or above, IIRC) due to magnetic interference.

[1] I wonder if anyone's worked out how many more power stations we'll need to bring all installations up to 17th Ed regs (or later) over the next 30 years or so?

[Andy Hall]

Is this because of their RCBO properties or through dealing with 10mA detection?

Spacers for this or changing the order?

I suppose if the RCBOs are getting warm to the touch, they are using a couple of watts each?

[Andrew Gabriel]

Yes. The position where the tails came out of the wall was a significant influence on the positioning of the CU. I did move that a couple of inches to make it all line up vertically, which wasn't difficult on a thermal block wall where you can almost use your finger nails to scrape out a larger hole.

[Tim S]

Andrew Gabriel coughed up some electrons that declared:

True enough - just price gauging at the moment. My local TLC are very helpful, but not very good at getting unusual kit (eg a Wylex REC2 DP 100A isolator had them flummoxed, but it's the optimum device for running tails in an out of as it's designed to do exactly that, and the load terminals are accessible by a separate cover so it's possible to work on a live but "off" one safely). MEM are "unusual" for them too. I'll give them a chance to source because they are very helpful, but I don't hold too much hope.

Is that ERD in Luton? Found Gibbs and Dandy - branch in Slough. I'm near T Wells in Kent so it's not a million miles and worth knowing they can source stuff. Thanks for that.

Ah. Just looked at memonline.com harder. I found the RCD for DP MCBs, the MR30. Just re-read it and the MR30 pod seems to do SP too - is that the one you used?

That was part of the attraction with MEM - being able to stick in Type B and Type C according to the requirements of each circuit without having to fudge the issue.

Cheers

Tim

[Andrew Gabriel]

It applies to all RCDs. The 10mA rating doesn't make any difference.

I changed the order.

The spacers were only temporary because I didn't have all the components when the photo was taken. I always fill a CU with MCBs even if they aren't all used. They're cheap, and they might no longer be easily obtainable when you want some more later on. (I got caught out with this once in the past when MK stopped doing their original single way RCBOs.)

Probably less than a watt, but when squashed together, there's not a lot of surface area to dissipate that from the middle ones.

I have a couple of spares -- I could rig something up on the bench to measure the power consumption of the RCBOs.

[Andrew Gabriel]

I thought ERD were all over the place (at least in the South East - I think they are also known as Edmundson Electrical), but strangely enough, it was one of the Luton ones, the one on North Luton Industrial Estate. The other Luton one (near Screwfix) specialises in other makes, but has some more common MEM bits. It was also the Luton Gibbs & Dandy.

Yes.

Be careful with the MCBs though. MEM do more than one range and it's easy to get them mixed up. One is for these industrial boards (and has a 10kA breaking capacity) and the other range is for domestic boards and has a lower breaking capacity (probably 6kA). I don't think the pods are intended to fit to the domestic breakers, but I never tried.

[Tim S]

Andrew Gabriel coughed up some electrons that declared:

Well, nadger me doobries - now I have their full name, have located branches in T Wells and Tonbridge. Never noticed them so they must be raund the back of all the places I normal frequent. Thanks - will look in on them.

Cheers - I noticed that too.

Thanks

Tim

[meow2222]

There is no required disconnect time for existing installations. Required times apply to new wiring. Whether a domestic socket circuit takes 0.45s or 0.35s to fuse makes absolutely blank all difference to real world death rates. And /that/ is whats relevant here, nowt else.

I think the biggest factor is the ratio of safety benefit to expense. Since the safety factor is about zero, and there are mass killers in the average house, twiddling with the fusebox is both pointless and counterproductive.

I agree theyre much more significant than fuses/mcbs, but. Twenty something deaths a year, of which some are down to appliance abuse. RCDs would indeed stop some of those deaths. But they also encourage people to do things they never would have before, causing more of the situations, and they cause loss of sair lighting, loss of lighting in fires etc, which we know kills people. So the real world safety cost or benefit is anything but clear. We simply don't have the figures yet.

But once again... if you want to save lives in the home, there are much greater results to be had by spending much less on the real risk issues.

NT

[John Rumm]

Its a good question, and I don't know how easy it would be to get figures in the exact form required... Until recently[1] deaths from electrocution have fallen year on year in absolute terms for quite a number of years, even though usage of appliances has risen. This won't be attributable to just one improvement in safety, but a collective effect of which RCDs have been a part quite a number of years now.

[1] In the last couple of years this trend seems to be showing signs of reversing - strangely coincident with the introduction of Part P

Not sure I follow the question - do you mean what will be the change in user experience? or the actual process of making the change?

[Dave Plowman (News)]

One thing with fuses is it's all too easy to fit a larger one to an overloaded circuit - I've seen this many times. Not many who do this sort of thing would know how to change an MCB.

[Andy Wade]

That's not the case at all if we're talking about the 30 A rewireable fuse (BS 3036). If you superimpose the adiabatic line for a 1 mm^2 conductor with a k value of 115 over the relevant fuse characteristic you discover that the former lies wholly below the latter. In other words the fuse can't be relied on to protect the 1 mm^2 CPC at any fault current. Hence unfused spurs in ring circuits in many old installations were never compliant in respect of thermal protection of the CPC. That's why the cable standard (BS 6004) was changed in 1984 to require a

1.5 mm^2 CPC.

Common sense might also hint of a problem in the above case: the fuse wire is only about 10 thou' smaller than the conductor it's trying to protect.

That's more a general design matter (coordination), rather than something related specifically to the fuse or cable. Of course the 0.4 s disconnection time requirement hasn't changed for TN systems (but does now have wider scope). What has changed is that the old 'alternative method' (limitation of fault voltage drop) has gone from the 17th ed., replaced by a requirement for local supplementary bonding where 0.4 s disconnection can't be achieved. 30 A rings protected by a rewireable fuse in TN-S earthed systems do rely on the alternative method if the Ze value is high, and hence might not comply with the 17th ed.

In general you can say that although rewireable fuses are still permitted it's now harder than before to ensure full compliance when using them. Standard circuits - per Table 7.1 in the previous OSGs - are not necessarily compliant with the 17th ed. It will be interesting to see how this table looks in the new edition, when it finally comes out.

[John Rumm]

Disconnect times have been specified in the regs for many years. The difference of .1 sec etc may not have a huge impact, however when disconnect times start creeping up into the multiple of seconds then the risks do become more significant (not only due to prolonged exposure to high voltage earthed metalwork, but also that the CPC in the cable will fail, or the cable sustain damage before the fuse operates).

The improving stats for safety of electrical installations in general would seem to disprove that the safety factor is zero, also blanket statements about the pros and cons of changing a CU are not really much use since they concentrate on too small a part of the picture. There will be times when swapping out a working BS3036 CU for a modern one will bring no improvement in safety, and there are times when it will bring about a dramatic and significant improvement.

IIUC the RCDs would indeed stop some of those deaths. But they also encourage

This is probably true - but only in a very small subset of cases where the user of the appliance has a good technical appreciation for what the RCD is and does. IME the vast majority of people have little grasp of what is in the CU or what it does, other than there being lots of "little switches" that they need to reset from time to time, and never necessarily even associate the need to reset them with any other event.

So I would be very surprised if having a CU with RCDs prefitted is going to influence behaviour for many people. (you only need look at the group of accidents that occur with extension leads in the garden - its not usually mowing over it that causes the shock, its the dopey pillock picking up the severed ends to look at them!)

Well this has been known about for some time (most of the reign of the

16th edition) and remedies put in place to address the problem.

Again, it depends on circumstances. In situations where students are likely to overload circuits in their lodging house, and replace fuses with tinfoil etc, or where there are lots of ankle biters roaming about trying to push bits of metal into sockets to see what happens, you have a real risk that needs addressing as a mater of urgency. In other cases where a couple of adults share a flat, never need to mow a lawn, and do nothing more exotic than plug in the TV and the phone charger there is relatively little risk with 3036 fuses and no RCD.

[John W.]

Yes, when I moved into this place some years ago each fuse was wired with fuse wire that went from one screw, round the other and back to the first...

Our place, when we lived in in France, had cartridge fuses and, following that experience, I now much prefer rewireable ones.

[Geo]

I always find it difficult soldering the fuse wire to the cartridge ends...

Geo

[meow2222]

Yes :( FWIW In some cases changing from 5A to 10A can remain withint the cable ratings and be a deliberate upgrade.

NT

[meow2222]

Yes, but there is no requirement for existing installs to meet them. And whats relevant is the real world safety implications.

those are due to many factors, so cant prove or disprove anything. The slow disappearance of 1950s and earlier installations is doubtless one factor.

yes.

I thought it was total electrocutions, but I dont have a reference to hand.

True enough, but I think there's more to it. An illustration we can hopefully all agree on...

Decades ago lots of schookids got a shock at some point or another. We all learnt from that how nasty and scary it was, either directly or by seeing how shaken our friend was. That taught people about electricity and changed peoples' behaviour, everyone's.

As safety practices advanced, most kids grew up without learning this.

Now today its the other way round. Kids muck about, short things out and get a shock, but the RCD trips, and the whole experience is a lot less nasty. The lesson the rest of the kids learn from this is different: if you piss about with it, it'll bite very briefly, and the thingy will trip. That's it.

This influence's almost everyone's behaviour, not just those with some understanding of the technicalities. Its one of several reasons why safety innovations so often fail to deliver the benefit that early promoters claim.

Then here are the many that put themselvs at risk using RCDs. Typical case: RCD on drill plug. It does nothing to protec the user from drilling into a live wire. Nothing at all, but they think it does so are a lot less careful.

:)

What remedies have you in mind? AIUI the 17th's requirement for RCDs on all circuits has merely ensured these safety problems will apply to every house newly wired.

In both cases there are much bigger risks. If you look up the death rates you'll immediately see that. Kids have been putting things in sockets for over half a decade, but the resultant deaths are few. The big risks to them lie elsewhere. The deaths from heart disease and cancer are several orders of magnitude larger, and we know that what kids do does affect their later life outcomes - just for one example.

Its not that bad. There are many millions of houses with rewirables and no RCD, and the electrocution death rate is tiny.

For various reasons a sense of proportion is too often missing from safety discussions.

NT

[meow2222]

AFAIK the data isnt collected.

NT

[John Rumm]

"Existing installs" must include the large number done during which time these requirements have been in force. Note that disconnection times also have a big impact on a secondary risk with electrical installations

- i.e. fire. (as I recall someone was often keen to point out, there are lots of these each year, and some are electrical in origin ;-)

I think that is partly the point I was making - you can't look at just one part of the installation and say safe/not safe based on that.

Yup in fact, I am not sure anymore either... looking at some docs that figure does include both sources, but then others (like those originally cited in in the Part P RIA spoke of several K deaths/year (before later acknowledging they had misunderstood their own stats)).

Yup, I understand how many safety improvements can become counter productive and alluded to that in my comments above.

I don't really have any data to support a general reducing attitude to shock risk though (or any to the contrary either)

To be honest I have never seen anyone use an RCD[1] on a drill. Mowers and hedge trimmers yes, but never a drill.

[1] as in explicitly make a decision to use one, rather than just plugging it into a RCD protected circuit since that is what was close.

With modern (class II drills) its unlikely to be an issue either way.

The same remedies apply in the 17th as earlier editions - namely the requirement for discrimination between faults and circuit protective devices. Split load CUs, multiple RCDs and RCBOs should ensure that you don't lose lighting as the result of a single fault elsewhere. The only time loss of lights as a result of a fault is a notable problem is with "whole house" RCDs.

I was not suggesting that one should focus all efforts on reduction of risk with electrical installations, there are of course plenty of others that would benefit from attention. That is not a reason to ignore potentially dangerous electrics either.

Note we are not only talking about electrocution risk here, but also fire risk, and accident risk directly and indirectly associated with shocks in general - which is a much bigger number than actual deaths from electrocution.

Quite agree - the big killers tend to be things that are no longer even news.

[Terry Fields]

That was my impression too, based on comments on the ng; but I was wondering if any officcial figures were available.

It seems to be common for 'the authorities' to set a policy (i.e Part P) but to fail to put in to place any monitoring of the effects of that policy. Perhaps it's a question of 'if you don't want to know the answer, then don't ask the question'.

It was just a convoluted way of trying to link the technicalities such as the release time for an RCD, or better earthing arrangements, with reduction or otherwise in casualties, but if there's no figures collected, then there's no way to link the cause with the effect. It all seems to run on a 'feel-good' factor.