Whilst in some cases surge protection may be useful Graham, the regulations take this to being a requirement and specifically state that this is lightning protection. The specification for its fitting is entirely based on lightning strikes.
Absolute nonsense. BS 7671 says nothing of the sort. In fact, see 2nd para of 443.1.1
This section does not specify requirements for protection against transient overvoltages due to direct or nearby lightning strokes on the structure.
The overvoltages covered by BS 7671 are atmospheric origin or due to switching. As I inferred in my previous response, the former being either cloud-cloud or remote (in which case impedance comes into play).
Were your devices suffering from such spikes or other causes? I know and have measured many mains spikes, but the cause is usually other things, so why not come clean and simply say surge protection may be useful in some circumstances?
Definitely the types of "spikes" inferred in 443.1.1, and in the places concerned, it's a mixture of both ... in reality, not in a dream or a speculation.
The cost-benefit analysis should be up to the consumer, not some remote manufacturer or regulation committee who have no idea of the likelihood or cost of any damage. It might well be a good idea for a data centre, but for Mrs Bloggs who has two LED lights, probably not worthwhile. The energy from switching spikes and similar is very low compared with lightning, and may often be entirely satisfactorily removed with very simple and cheap VDRs. Live conductor to Earth protection is rather different to L-N protection, simpler and effective for preventing damage to connected semiconductors, particularly in with the mains supply impedance being raised with some small series inductors, except in extreme cases. The regulations seem to have a fascination with the line-Earth voltage which is only seen by any attached EMC suppression components. I don't doubt that some equipment is occasionally damaged, or at least fails, but deciding the exact cause is often very difficult.
We are now at a point where the cost-benefit analysis is perhaps moving in favour of fitting them, as I said in my previous e-mail. I agree that, say 15 years ago, my view would have been wholly different, but I'm totally sold on the cost-benefit based solely on the cost of equipment I've had damaged by cloud-cloud strikes alone over the past 20 years, in an urban area (although by the coast). And I know I'm not alone.
We are now at a point where the cost-benefit analysis is perhaps moving in favour of fitting them, as I said in my previous e-mail. I agree that, say 15 years ago, my view would have been wholly different, but I'm totally sold on the cost-benefit based solely on the cost of equipment I've had damaged by cloud-cloud strikes alone over the past 20 years, in an urban area (although by the coast). And I know I'm not alone.
I thought that this thread was about AFDD, but as for SPD, I agree. FWIW, I think that the whole business of lightning stroke/strike in BS7671 is far too complicated and I find it difficult to see why a new build would not have SPD. I have never had any damage personally, but when I install a new CU it will have it. It isn't that the benefit has increased, it is that the cost has decreased. ?
We are now at a point where the cost-benefit analysis is perhaps moving in favour of fitting them, as I said in my previous e-mail. I agree that, say 15 years ago, my view would have been wholly different, but I'm totally sold on the cost-benefit based solely on the cost of equipment I've had damaged by cloud-cloud strikes alone over the past 20 years, in an urban area (although by the coast). And I know I'm not alone.
I thought that this thread was about AFDD, but as for SPD, I agree. FWIW, I think that the whole business of lightning stroke/strike in BS7671 is far too complicated and I find it difficult to see why a new build would not have SPD. I have never had any damage personally, but when I install a new CU it will have it. It isn't that the benefit has increased, it is that the cost has decreased. ?
I agree with you on the new build ... part of the difficulty with BS 7671 is that it applies not only to homes, offices and shops, but other installations, some of which (say if it's just resistive heating loads, or contains loads with in-built surge protection) may not require SPDs.
In addition, there are existing properties with which it may be difficult to adopt a blanket mandate, although those difficulties are there already. For example:
a new circuit is added to existing house, say for a traditional resistive element oven. The oven's circuit is not SPD protected - the choice needs to be there as in this case, perhaps the omission of the SPD won't do too much damage.
an existing building has LPS to BS 6651. You are asked to rewire a floor on the building for an office as part of a refit. There is no surge protection ... what do you do? In this case, the cost of not having the surge protection could well be large, BUT the cost of providing the surge protection might include a re-assessment (and perhaps improvement) of the existing lightning protection system to BS EN 62305 series, and then provision of a hierarchy of SPDs to BS EN 62305-4.
What about if the load were an electric vehicle at say £30K+ - would you recommend SPDs are fitted? I'm sure most charging points are installed without them.
Graham the 443 requirement to fit SPDs is based on the lightning chart and a "magic" formula. True it adds "transients due to switching" as an extra justification, but I am afraid I am far from convinced that they are necessary to protect electronic devices. Induced lightning current in an underground cable is very small, as the conductor spacing is tiny, so why is its length in the formula?
There is an enormous quantity of electronics connected directly to the Mains supply at this time and I am not convinced that there is a significant failure rate due to "Mains transients". Take all the VFDs and Solar inverters in the country, do these regularly fail during thunderstorms? You might say that an extra £50-100 for a bit of insurance is good value, as a cost-benefit exercise, but the question is then"does this guarantee against failure"? I think there is a problem with the theory behind VDR protection here, which I expressed above, and to work properly some series impedance is required unless the device dV/dI slope is essentially zero, like a Zener diode. The clipping action is extremely dependant on the transient energy, and I have yet to see a graph of the protection offered by the various devices on sale, which is curious.
I have just been looking at some data sheets. It then struck me that these devices have not been designed for PME supplies! They fit a neutral to MET suppressor when these are solidly bonded together a couple of feet away. For single phase PME we need a single suppressor, between L and N, but can I buy one. Interesting question.
For SPDs, in terms of preventing damage from induced voltages, the hard part is not voltage spikes between the cores of the incoming supply cables, but the voltages that exist between the incoming supply cable considered as a group, and other things, such as telephone lines, plumbing, structural steels, and if you have one, the lightning protection system- equipment that straddle pairs of such things (phones, cable TVs, fax machines etc ) are especially at risk.
If you like imagine the pick-up working rather like a crude dipole antenna, with the mains wiring as one limb, and the phone lines or the plumbing or whatever as the other.
Cheers Wikipedia for the image ... Note that a surge is more of one or two isolated cycles, not a continuous wave, and the sort of SPD we can buy only work for a few microseconds at their rated current - if there really was a continuous over-voltage, they would simply catch fire without ADS.
A real radio dipole is laid out with the limbs carefully balanced in a straight line and away from obstructions, but the same pick up effect still occurs if neither condition is met, just less efficiently and with more interference picked up- the older ones among us will have seen a coat hanger and a car body used as as the two parts of a medium wave and VHF radio antenna and also working - though that is so unbalanced you may prefer to think of that as a monopole and a pseudo 'earth', the pick-up mechanism is the same.
In more rambling installations, the 'other limb' of the "diplole antenna" may well just be the wiring of the rest of the installation.
However, any device connected L-N cannot do much about that, unless L and N part company and are forming the antenna, but usually they are really the same limb, having stayed side by side for most of the route,
I suspect a lot of SPD will not quite work as expected, as they are not across the actual pair of things that form the dominant pick-up mechanism.
I wonder if in some ways PME makes things worse - similar spikes are imposed on L & CNE nearer the lightning, but then all the additional electrodes pull the CNE, but not the L, a bit closer to true earth before it gets to the consumer. Not that old fashioned TN-S with lead covered cables and rotted hessian would be any better.
I agree about the phone lines - that's where I had most problems at my old Dad's place - usually with equipment that was connected to both the mains and the phone line. (out in the sticks, miles of overhead phone lines, LV underground but only for about 40m to a pole transformer then overhead HV for miles). My worry is that adding SPDs to the electrical installation (but not the phone lines) could actually make things worse as the LV side gets pulled down to something closer to true Earth while the phone line still bounces about as it always did - do the equipment between the two might actually be stretched more than before.
I am sure that you are right Mike. However, the point you are making is important. The only voltage we can "control" (hopefully rather better than the Government attempts to control Covid) with PME (which is, in reality, all we have) is the L-N voltage. We do not have a low impedance Earth which really is the bulk of the planet, so we cannot divert the surge elsewhere whatever we do. The voltage between a telephone cable and the mains Neutral, or any other service cannot sensibly be controlled. The items which can be damaged (in my case over the years Modems and ADSL filters on telephone lines in the exposed country) are not protected by mains surge suppressors because the failure is in the line interface. I have never had a failure on the same site on the mains part of any equipment, despite a local pole TX, TT supply, and a very exposed 11kV line of miles length. The site got quite a lot of lightning and several local trees were hit over 15 years. All the equipment I have seen for sale assumes a TN-S supply and a good local Earth, probably not to a cable armour with the sheath insulating it from real ground. This does not match the reality of the UK, the earth connection is often not local, bonded service pipes are often plastic and not available, and so the Earth connection has significant resistance and series inductance. Perhaps this is why there was pressure for foundation Earthing, to provide a local Earth for the surge suppressors. Unfortunately, I was not party to the discussions!
Following on, the lightning flash can only induce a similar voltage, as a series addition to both or all live conductors, raising the end voltage to real Earth. However real cables have conductors both twisted and very close together so these voltages must be pretty much identical. Overhead HV lines are slightly further apart, but still very close in inductive terms, similar to an open-wire transmission line. Other switching transients may well mainly affect one line, giving a L-N or P-P voltage which could be clipped by a surge suppressor, but it is not clear why this has recently become a serious problem to consumers because that is not my experience.
Graham the 443 requirement to fit SPDs is based on the lightning chart and a "magic" formula. True it adds "transients due to switching" as an extra justification, but I am afraid I am far from convinced that they are necessary to protect electronic devices. Induced lightning current in an underground cable is very small, as the conductor spacing is tiny, so why is its length in the formula?
Of course the risk of issues with non-direct phenomena of atmospheric origin depends on the frequency of occurrence ???
Length is there for "common mode" as Mike points out.
Presumably you can get events which are induced as common mode, but if one of the line conductors flashes over to earth, becomes a differential mode transient.
