I think it is the Coronavirus lockdown and everything that goes with it, but whatever it is everyone seems to have lost the will to argue the toss about Amd 3.


It could however be that without any opinions or arguments put forward by people who think the amendments are a good idea it is hard to actually have a discussion.


Andy Betteridge

everyone seems to have lost the will to argue the toss about Amd 3.

Oh, not at all!  Just be a little patient - I'll have something more to debate soon I'm sure (initial slow start because of the awful BSI web site, but now some PDFs have appeared I'm making quicker progress)


  - Andy.

Very reassuring to know that AJJ is on the job. Your post near the start of this topic seems a very good summary of the situation.  I became interested in AFDDs a few years ago, for a variety of reasons I needn't go into here - my feelings about their usefulness for European or UK-specific installations have varied back and forward when I see different results, statistics, etc. For example, 'glowing contacts' (resistive heating) are sometimes seen as a separate non-arcing series fault for which AFDDs would be irrelevant, and yet have also been claimed as typically connected with periods of arcing that an AFDD should supposedly notice (see later link to a thesis).


It seemed likely there would be a gentle move towards pushing AFDDs in the regulations, step by step over many versions. So I was amazed to see - instead - this sudden, wholesale change. The cost/benefit needs clear, open justification before such a major move should be made. Comments here have well pointed out the wider picture also, such as potential negative consequences of putting users off updates, bodging new devices into existing CUs where they don't fit well, etc.


Choice of any such protective device involves some trade-off of cost and benefit. If a study of benefit has been done, as would be hoped for when deciding on this major change, it would seem appropriate to state conditions of cost that make the choice of AFDD worthwhile. For example "if an AFDD is available at less than £30 (including any additional cost due to changes of the enclosure) then it's worth fitting it for ". How well would that go down? (This isn't meant very seriously, as it would make the rules even more complicated. But there does appear to be little regard for the cost side.) The only positive side is in the longer term, if the AFDD market becomes much more competitive and diverse.


It would be interesting to known what's actually in the upstream documents, if the change is mainly from the European level. For example, in perhaps half the countries in Europe the concept of "lighting circuit" scarcely exists, but instead 1.5 mm2 circuits feed sockets and lights together; this makes me wonder whether the draft already contains some UK-specific parts. Likewise, a substantial proportion (half again?) in Europe have three-phase installations, but I don't yet see AFDDs corresponding to 3-phase or 2-phase MCBs that might be used for cookers, heaters etc in those countries.


We know that statistics should be treated very cautiously. One point is the obvious, about poor knowledge of causes of fires and of how many would have been prevented by an AFDD. Another point is differences in electrical installations and building structure, which could make statistics from one country largely irrelevant in another: some examples are given in the following few paragraphs. So applying statistics from north America (the main region with experience of arc detection) to our installations is probably not very fair, even if the statistics are correct ... which they probably aren't (particularly if from interested parties), due e.g. to other factors that change the fire statistics at the same time as the AFCI population is growing.


Parallel arcing. Due to the lower system voltage in the US, the arc voltage is a larger proportion of the source voltage, so the ratio of arcing current to solid short-circuit current is smaller. As a further consequence of lower voltage, the circuit's rated current is higher for a similar load power. Also, circuit-breakers at least used to have higher instantaneous trip levels than the IEC B or C MCB characteristics. A site called Paceforensics had a video from the 1990s or 2000s showing sustained arcing in a cable protected by a US-style breaker; the site has closed but its video links can be found here thanks to the web archive.  I did experiments a few years ago as part of a utility's study of what dangers might increase during island operation of small LV networks with battery storage, confirming the unsurprising point that a 230 V parallel arc in a cable (e.g. 1.5 or 2.5 mm2) with a multi-hundred-amp prospective current will quickly grow its current and operate IEC MCBs or gG fuses in a half-cycle, whereas a weaker source such as an inverter with 100 A limit can sustain arcing, often until a weld happens. Many cases of parallel arcing at 230 V appear therefore to be well covered by existing overcurrent protection.


Series arcing. The use of aluminium wiring in (some) US homes has been causing trouble for decades, particularly from the early generation of poorly selected alloy. US requirements on application of AFCIs have not been limited to these affected installations, but it seems likely that the aluminium conductors helped give a justification for a special detection device.


UK feature: FTE (flat twin-and-earth) cables. For installation cables in most UK domestic situations, parallel arcing in a damaged cable would be caught by an RCD, due to the bare cpc between the other conductors.  I suspect that even series arcing in a cable would have a tough job causing damage to things outside the sheath without bridging over to the cpc. It's rather a UK-versus-most-of-Europe distinction, having this conductor bare and reliably in the middle, in contrast to rounded cables. Cables with earthed screens/armour are another way to pass many cases to RCDs. But this doesn't help for typical flexible cords, junction boxes, appliance internals, etc.


And even within one country's installations there are important differences, as was pointed out earlier about metal conduit versus plastic-sheathed cables.


For those who'd be interested in a bit more technical background about series arcing and AFDDs, there are actually some rather good introductions from the manufacturer that's behind so many of the EU and UK devices ... we know that a little caution is in order, and they're hardly going to try to play down the benefit, but these do give some interesting points and references about the phenomena and the detection methods Primer and, with more detail about 'Series arc faults in low-voltage AC electrical installations' a thesis that notes it was sponsored by the same company.


A year ago I bought 4 different makes of AFDD from Conrad: a Siemens, Hager, Eaton, and Schneider. They've been connected separately or in cascade, on various circuits with a lot of different household loads, and haven't caused any nuisance trip so far, which was almost surprising in view of all the comments on US forums. (This 2017 US article from ECmag gives a good contrast within one page: the article itself is strongly on the supportive side, but the comments from users and electricians are vehemently against.)  The AFCI and AFDD product standards are different.  IEC 62606 (AFDD) includes a list of appliance types that must be supplied without causing a trip - e.g. a vacuum cleaner, thyristor dimmer, switch-mode supply, "hand tool such as electric drill at least 600 W and conditioned by 24h operation", etc.  I'm not sure how similar the corresponding UL standard (AFCIs) is about things that shouldn't cause a trip, although its arcing test-objects look quite similar to the ones later adopted in the IEC standard.


As has often been noted, the AFDDs also aren't overly enthusiastic to trip on things that definitely are arcs. I went through the copper, nickel, carbon etc, attempts when I got them. It would be foolish to dismiss this as a sign of bad function: the series arcs being looked for in real circuits are not the sort that one easily produces without some sort of conditioning of the conductors and possibly of insulation surfaces too. This is one of the main troubles in testing and understanding series arcing faults, that the dangerous conditions arise over a long time and probably in just a small subset of all the cases of loose contacts or broken conductors where the conditions are just right. The interesting question is what realistic dangerous forms of arcing might be overlooked by today's standards and the resulting products.


The power consumption of AFDDs was mentioned here sometime months ago. The four AFDDs I've been testing draw around 2 W for all four together, and vary a bit individually. If there'd be, say, 5 such devices per CU, in each of ~25M homes in the UK (never mind other installations) that's some 60 MW.  Ok - not that much in the big scale of things, and it would take years to reach most of the housing stock, particularly if people avoid CU changes due to the expected cost! But we often hear of the importance of saving a watt here and there in standby modes of equipment in order to avoid "X power stations across Europe", so this is a further little factor when considering all the harms that need to be offset by clear gains in this drastic proposed change. And the devices would probably be able to reduce their consumption a good deal if given incentive.


It's very bright of some manufacturers to have come out with the neat 1-module MCB/RCD/AFDD just now. The multi-module devices such as Eaton's may have an advantage of more clearly indicating what caused a trip, but the 1-module Siemens unit that I have is supposed to use flash sequences and colours of its one lamp to indicate this. If the AMD goes ahead, I suppose we'll all be doing plenty of deciphering flashes. Perhaps the next generation will connect to a cloud service, send a fault-recording of the waveform that tripped them, and try to provide a diagnostic for the user.


Summary related to John's original questions:  I'm very surprised at what seems an inappropriately big step; effects

on installation industry are better handled by others here who've already extensively discussed and doubtless will continue; and the evidence for benefits should be the duty of whoever made the change to the regulations - such a costly change shouldn't get there without pretty strong and country-relevant evidence provided in detail.

I remember hot debates on a fire safety forum when BS5839-6 placed the vast majority of houses in a category that did not require smoke detectors in bedrooms, which seemed odd given the number of fires originating in bedrooms and the obvious desire to warn occupants who were asleep. One of the authors of the document posted that a detailed cost benefit analysis indicated that the additional financial burden that would be placed on householders, housing authorities etc could not be justified. Despite several requests, we never got to see that bit of accounting either!

Well, I still will not be recommending the extra substantial expenditure involved in fitting these things to my customers, no matter what the regs or amd says.

And before everyone jumps in to shout about insurance companies not paying out, I invite them to cite specific instances where an insurer has refused to pay out because of the lack of RCD protection.............

These things smack of a desperate hunt for a solution to a problem which does not really exist in the real world.

Housing Associations don’t have to fit smoke alarms in tenanted homes, so some don’t.

This may be of interest.

https://www.researchgate.net/publication/345910304_Arc_Fault_Detection_Devices_AFDDs_-_Are_they_worth_it