Disagree.

In the table the typical context that the word "should" should be in is in an informative element.

If the word "should" was used in a normative element, then you should do it.

In the case of 421.1.7 the word "recommended" is used in the normative element of the regulation, it is not within an informative element (the notes). So to substitute it with the word "should" would be out of context. If JPEL 64 wanted to use the word "should" in a normative element, they would have done so, and AFDDs would have to be fitted, just as if they used the word "shall". They used "recommended" because they only, simply recommend that they be fitted.

It does seem a bit convoluted though

Edit; The table is based on this document Microsoft Word - Rules for structure and drafting of UK standards 170420 FOR PUBLICATION (bsigroup.com) Which states "The auxiliary verb “should” is used to express recommendations. In a specification or test method, the auxiliary verb “should” is used only in informative text (i.e. notes, commentary or informative annexes). In a specification, recommendations may be used to indicate that among several possibilities one is recommended as particularly suitable, without mentioning or excluding others, or that a certain course of action is preferred but not necessarily required, or that (in the negative form) a certain possibility or course of action is deprecated but not prohibited." and has been re-worded a bit in BS7671.

I am not permitted to provide an interpretation, but I can provide some ideas which may help.

A single Regulation cannot be read in insolation, to be honest. In this case, it's under Regulation Group 421.

If you look at the 2018 version of the Regulation:

421.1.7 Arc fault detection devices conforming to BS EN 62606 are recommended as a means of providing additional protection against fire caused by arc faults in AC final circuits.

Viewed on its own, with the proposed interpretation which in terms of the word 'recommended' ought NOT to change, because the source BSI document for drafting standards is the same, if you didn't fit them to AC circuits to BS 7671:2018, what did you do?

Surely the other "several possibilities" alluded to are in Regulations 421.1 to 421.6, and other parts of Chapter 42 as appropriate? It's the only way that Regulation 421.1.7 in BS 7671:2018 makes sense not being worded "shall".

Put another way, if we are not considering an installation in any of the bulleted premises, we can effectively grey out the related words for those premises, and view Regulation 421.1.7 in BS 7671:2018+A2:2022 as only a slightly modified version of the one in BS 7671:2018, which is a heck of a lot clearer and accedes to the limitations in the scope of BS EN 62606 in terms of ratings etc.:

421.1.7 Arc fault detection devices (AFDD) conforming to BS EN 62606 shall be provided for single-phase AC final circuits supplying socket-outlets with a rated current not exceeding 32 A in:

• Higher Risk Residential Buildings (HRRB)
• Houses in Multiple Occupation (HMO)
• Purpose-built student accommodation
• Care homes.

NOTE 1: Higher Risk Residential Buildings are assumed to be residential buildings over 18 m in height or in excess of six storeys, whichever is met first. It is anticipated that in many areas higher risk residential buildings will be defined in legislation which can be subject to change over time, as well as in risk management procedures adopted by fire and rescue services. Current legislation should be applied.

For all other premises, the use of AFDDs conforming to BS EN 62606 is recommended for single-phase AC final circuits supplying socket-outlets not exceeding 32 A.

Where used, AFDDs shall be placed at the origin of the circuit to be protected. The use of AFDDs does not obviate the need to apply one or more measures provided in other clauses in BS 7671.

So, if we thought it was optional in 2018, perhaps it's optional now (or was it actually optional all along ...)?

If needed to, what would your defense be for not installing them per the recommendation?
Cost? Questionable efficacy? Space constraints in the DB cupboard?

I know it is not a "requirement" but I'm leaning towards it being a standard item unless there is a documented statement from the client whereby they accept that not installing them is outwith the recommendation of the Regs.

The whole thing about recommendations being just one of several possible ways to skin a cat only holds water when there actually are other ways of skinning it. There just isn't when it comes to AFDDs.

(Assuming they actually work) Nothing else you can readily do will provide the same type of protection against the risk of fire, therefore the recommendation holds.

Whilst a recommendation is not a requirement, the legal implications should something go wrong down the line are potentially rather expensive. Certainly worth the cost of drafting a statement for clients to sign that: explains the clause, the possible implications of not installing AFDDs, and their acceptance of that.

The fact there's debate about this, and the new escape route clause, suggests that there's work to be done on making the language and intent of the regulations clearer.

pww235 said:

The fact there's debate about this, and the new escape route clause, suggests that there's work to be done on making the language and intent of the regulations clearer.

I think the fact that the "helpful" Guidance on the language used in BS 7671 has now been included on Page 18 of BS 7671:2018+A2:2022 has led to the debate; as I pointed out in my earlier response on this thread, the language of the "recommendation" hasn't changed in 2022 ...

Such language is used throughout British (and CENELEC and IEC standards), so moving away from that isn't really an option for JPEL/64.

The whole thing about recommendations being just one of several possible ways to skin a cat only holds water when there actually are other ways of skinning it. There just isn't when it comes to AFDDs.

I can see some possible arguments. If we use wiring systems where damage to a line conductor that would result in arcing would very likely also cause an earth fault (BS 8436 cables, or singles in steel conduit or even simple T&E) then ordinary RCD protection will likely provide very similar levels of protection to the fixed wiring as as AFDD. Possibly ditto for accessories with earthed metallic enclosures.

Then there's an argument that if you allow an arc to start a fire, an acceptable level of safety might still be maintained by other means - e.g. suitable fireproofing in the surrounding environment - or just that the environment itself naturally isn't as risky as those listed (e.g. due to the limited number of vulnerable people that could be affected by a single event). AFDDs aren't going to stop all fires after all - they can't detect simple resistive heating - so there's the issue of diminishing returns and there soon comes a point where the additional safety just isn't worth the cost (if there's money to spare, better to spend it on more effective safety measures).

i read a "recommendation" very similarly to "consideration shall be given" (e.g. in 559.5.1.208 - for N at light switches) - i.e. have a think about it, but there's no compulsion for the moment.

    - Andy.

It is a play on words which provides the grist for lawyers. In the Real World, the likelihood of a AFDD stopping a fire is remote.

Another thing I was wondering about recently was how robust these things will actually turn out to be in the field, especially the multi function variants. All those sensitive electronic bits in close proximity to an arc chute designed to extinguish extremely high fault current when functioning in over current mode.

Then we get to the fact that they won't protect anything downstream of a transformer.

However, no matter how fancy and tortuous a language they wish to dress it up in, BS7671 is non-statutory, and good luck to any lawyer who tries to prove that the absence of such devices enabled a fire to start.

Recommend to me means that the default is to install unless it can be risk assessed out. 

so yes, we should default to installing AFDDs unless we can bring in other control measures that we are happy with in terms of safety. 

No Neil, you are making a serious mistake. You have misunderstood the way the regulations actually work, and the video referred to above is absolutely wrong. You need to refer first to 421.1.7 (Brown book). This is the regulation which uses shall, which means you must, with the conditions.

Then you get the the next part of 421.1.7 (after note 1 which is the small print ONLY)

This then says that it is recommended to fit them in all other premises, this from BS7671 is that it is a choice, which should be made by the CLIENT, not the electrician or someone else. It is not a case of "Risk Assessment" or anything else (you will note those words are missing), it is a simple choice. You are seeing a difficulty where there is none, my definition is "If you want to pay you can fit them which is permitted in all cases, but you can choose to spend the money or not". Simple enough.

I am not going to talk about any technical reasons why they may or may not be useful, but I suggest you study the American fire statistics since AFDDs became mandatory about 12 years ago in their code, and tell the rest of us if there is a statistical difference, I make no comment.

GK cannot say this as he is a member of JPEL/64, but I will as I am not, if you had any idea how many man-hours had been wasted on getting those words in the text as written and not something else you would probably have died a sad death from old age whilst waiting for the next agenda item!

Hi David,

I do not subscribe to allowing unqualified persons a say on items that have been recommended to be installed by those that are qualified.  We can all choose where we wish to fall on the line of compliance and I try and be at the end of 'best practice' as often as I can.  For me, if a Standard 'recommends' I install something then I need a good reason not to.  For me, 'because the client said no' is not a good enough reason to keep me where I want to be on my imaginary line of compliance.

We do not need to go into the science behind AFDDs and if they work on stranded cables etc as I have had many discussions on the subject with those far smarter than I will ever be and to be honest it does not really change how we deal with the term 'recommended'.  

I agree it is a choice to fit them in other examples but it is how we inform that choice which is important.  When I carry out a design risk assessment I will look at things like installation method, past history of EICRs, past history of electrical fires, how the building is maintained etc.  these are all things that could reduce the level of risk of an electrical issue caused by arcs and could mean that I, as the designer deems that AFDDs are not required.  I have then read the Standard, considered the risk, documented the risk and issued the DRA with my design for total transparency.  

This is how I believe we should approach the design process as competent electrical engineers.

Hi Neil

I see that you believe that the wording of this regulation was inadequate, and therefore should have been simply to require AFDDs on every circuit of every installation, because this would have been "safer", whatever that means, and it means many different things to many groups of people. Perhaps you would care to comment on the prevalence of arc faults which are not cleared by other circuit protective devices, and the number of such that cause fires? I would point out that unless you have this data it is impossible to carry out a competent risk assessment, because you have no idea of the actual risk, only a vague idea that there must be some.

The greatest risks to an installation are loose or resistive connections and so far AFDDs have not been shown that they can detect either, and in fact no one can without an inspection process that is thorough.Most installations have loose connections or screws when properly inspected, yet few of them catch fire. It takes a loose tails cable and a large load current in most instances, and these can all be seen to be resistive loss by the spread of the heat along busbars and similar parts.

I suspect that you have not considered the physics of an arc, it is very much more complex than often imagined. The fundamental part is the distance in air through which electricity can jump cold, it is a small fraction of a millimetre per thousand volts, air being a very good insulator. The mechanism is that the air must be ionised, and is then somewhat conductive. Once the conductive path is established, the next mechanism is heating of the air and ends of the electrodes, which produces many more ions at a higher temperature and the arc current can increase significantly. The ends of the electrodes then burn away, increasing the arc length, which may well then be extinguished, or possibly cause ignition of something nearby. Note that copper electrodes need to reach more than 1083C to start to produce ions, so most arcs extinguish long before this temperature is reached unless between tiny points, the heat loss with copper cables being very high due to the thermal conductivity. There is a fairly low temperature source of ions and this is heated plastic which carbonises at a fairly low temperature and has poor thermal conductivity. This is what is simulated in the AFDD demonstration equipment, but is very difficult to simulate with a real length of cable.

I wonder which part of this reasoning you feel makes AFDDs mandatory? Certainly the manufacturers do, but that is to be expected as profits would be huge, there are around 40 Million installations in the country and perhaps an average of 6 or 10 circuits each, so perhaps 400 Million units over perhaps 10 years, and £100 each? Fantastic! Yet very few of those installations would have an arc based fire, and there are many other causes of fires. Are we sure that this is the most important thing to change? Perhaps not.