Is the designer to blame? Should people claim? RCD trips.

[ericmark]

Since the 17th Edition we have had “Every installation shall be divided into circuits, as necessary, to reduce the possibility of unwanted tripping of RCDs due to excessive protective conductor currents produced by equipment in normal operation.” However the active word is normal.


Under normal operation how big can a circuit be? 3.5 mA is the limit where a plug and socket can be used, so normal earth leakage is less than 3.5 mA the 30 mA RCD can trip between 15 and 30 mA so with no fault, i.e. normal operation 10 sockets is within the rules, however clearly 100 sockets is over egging it.


So it would seem 25 sockets is likely considered as reasonable, but as fixed appliances are added which do not have the 3.5 mA limit like cookers, immersion heaters, and lights then we it seems are over stepping the mark.


My old house had just two RCD’s and they would have a batch of tripping, could go a year without a trip, then trip once a week for a month then another year without tripping, since been fitted for some 25 years now, it is clear they did not comply, they did trip in normal operation, two RCD’s were not enough. That is in a 3 bedroom house.


Mothers house had 6 RCD’s two for whole of house except for kitchen, and 4 RCBO’s for kitchen, and I seem to remember only one unwanted trip, so it would seem likely 3 RCD’s with the kitchen on it’s own RCD would have been enough.


This house when we moved in had one RCD with most of house unprotected, it now has 14 RCBO’s as experience with last house had shown 2 RCD’s are not enough, and consumer units may have three neutral bars, but not really designed for 3 RCD’s and also 3 RCD’s take 6 slots, so the consumer unit becomes much larger. However 4 slots are taken up with surge protection and an isolator, and since must be able to isolate else where under HSE rules, never really worked out why we have an independent isolators to isolate else where plus one in the consumer unit that is I suppose going OTT, no real need for an isolator in the consumer unit.


So from my own experience it does seem a few RCBO’s for kitchen and then 2 RCD’s for rest of house is enough for normal operation, however then we have “Take account of danger that may arise from the failure of a single circuit such as a lighting circuit.” We it seems don’t put lights and sockets in the same room on the same RCD, for this house that forced my hand, sockets split front/back and lights split up/down so could not possibly get away with two RCD’s, maybe putting lights on RCBO would have been enough.


However this is all about design, and the electrician is trying to design the installation when in real terms he has no idea what will be plugged in, so where it seems the requirements of division of installation has not been met, can he be called on to correct and add more RCD’s? We sign to say we have designed, not simply followed tradition, so at what point could the house holder blame the designer and say I want to claim for the loss of a freezer of food as you got the design wrong? And if one person is successful claiming from a designer would that open the flood gates?


I would say clearly a caravan has all on one RCD (although it does have battery back up) and a two up, two down clearly two RCD’s are enough, but as the person who designs the installation at what point do we need more that 2 RCD’s?

 

[JohnW2]

Since the 17th Edition we have had “Every installation shall be divided into circuits, as necessary, to reduce the possibility of unwanted tripping of RCDs due to excessive protective conductor currents produced by equipment in normal operation.” However the active word is normal.

Almost. The current wording is actually...

314.1 Every installation shall be divided into circuits, as necessary, to:
.....(iv) reduce the possibility of unwanted tripping of RCDs due to excessive protective conductor (PE) currents not due to a fault

However, "not due to fault" means essentially the same as the "in normal operation" which you quote - so that doesn't make any difference.

....however then we have “Take account of danger that may arise from the failure of a single circuit such as a lighting circuit.”

Indeed - although "taking into account" does not imply any compulsion/requirement to do anything.

As I've said many times before, I think that is probably a rather daft regulation - since, if they regard the loss of a lighting circuit as being an important safety concern, they should insist on emergency lighting. As I've often said, in my house, which has several lighting circuits, all of which are protected by RCDs that also protect other (sockets etc.) circuits, it is FAR more common for loss of a lighting circuit to be due an (often very brief) power cut than to an RCD tripping.

Kind Regards, John

 

[flameport]

The limit is 30% of the RCD trip current in normal operation, which is 9mA for a 30mA RCD.

 

[JohnW2]

The limit is 30% of the RCD trip current in normal operation, which is 9mA for a 30mA RCD.

Maybe, but with sockets circuits usually having many outlets (into any of which anything could be plugged), there's not really any sensible way in which one can design on that basis, is there?

Kind Regards, John

 

[ericmark]

Maybe, but with sockets circuits usually having many outlets (into any of which anything could be plugged), there's not really any sensible way in which one can design on that basis, is there?

Kind Regards, John

Since we can have up to 3.5 mA without special arrangements we can expect up to 3.5 mA per socket circuit however not every socket circuit, so the point is how many sockets is reasonable on a single RCD, and cookers can of course have a larger leakage, and one would hope very much less on the lights.

As to “Take account of danger that may arise from the failure of a single circuit such as a lighting circuit.” I think most consider that is not so much as simply loosing lights due to RCD trip, but as to loosing lights when another circuit in the same room causes the RCD to trip, if some thing goes wrong which could likely give the occupant a shock you don't also want them plunged into darkness.

I consider running a temporary extension lead from one room to another although clearly has some danger, is not a great danger, so it is reasonable to expect some one to run and extension from the living room to kitchen to keep the freezer going when they can't find why the RCD has tripped, however it is not safe to run an extension between floors. Also where the cables are distributed between the floors, splitting the home side to side or front to back reduces cable used, so reduces the ELI so it was common in North Wales at least to split sockets that way rather than up/down, however it was also common to split lights up/down, so the only way to arrange things so sockets would not cause lights to trip was either put sockets or lights on a RCBO so two RCBO's and two RCD's seemed minimum sensible split.

OK my own house now is all RCBO but I realise that is expensive and not everyone wants to do that. But in a reasonable sized house, (three bedrooms) the question is if the electrician has not fitted a board with three neutral bars (high integrity) is it really fit for purpose? I know it is done, and it seems all too common to have just two RCD's but we sign to say we have designed, not the builder, or the owner, or the foreman, we sign to say we designed it, and is it reasonable to design a three bedroom house with only two RCD's?

 

[JohnW2]

Since we can have up to 3.5 mA without special arrangements we can expect up to 3.5 mA per socket circuit ....

I'm not sure what you are suggesting here. 3.5mA earth leakage at each of several outlets on one sockets circuit would not be acceptable, even if that were the only circuit protected by an RCD (or RCBO) - and most sockets circuits have a good few 'double outlets'.

As to “Take account of danger that may arise from the failure of a single circuit such as a lighting circuit.” I think most consider that is not so much as simply loosing lights due to RCD trip, but as to loosing lights when another circuit in the same room causes the RCD to trip, if some thing goes wrong which could likely give the occupant a shock you don't also want them plunged into darkness.

Hmmm - I don't think I would qualify of one of those "most" who would think that.

Other than when someone accidentally touches an N & E together when working on a circuit, it must be very rare for someone to be in a room when 'something goes wrong' in that room which causes an RCD to trip - and, even if that (very rarely) did happen, the fact that the RCD had tripped (as a result of that 'something going wrong') would mean that it was no long "likely" (in fact, no longer possible!) that it could "give the occupant a shock", anyway

Rather than what you suggest, when 'most people' are talking about the dangers of 'being plunged into darkness', they usually seem to mention things like falling off stepladders and dropping pots of boiling oil - which, as I said, are no less dangerous when lighting is lost because of a power cut than when it is due to an RCD trip.

Kind Regards, John

 

[bernardgreen]

which, as I said, are no less dangerous when lighting is lost because of a power cut than when it is due to an RCD trip.

Hence the sense of having emergency lighting in areas where power cuts are a problem. Sudden un-expected darkness due to trips caused by faults / mis-use in the house wiring / appliances is likely to happen more often than darkness due to a power cut.

 

[ericmark]

Have had RCD protection for some 25 years I know it is very possible to get a shock, OK maybe only 40 mS but you still get a shock, last time I went through a horizontal cable which should not have been there, it went around 4 corners, and I was floored by the shock.

It was in day light, but it would have tripped the lights. I use rechargeable torches in stands as stand-by lighting, and with a general power cut they work well, however since plugged in, if lights trip I have not light unless I trigger the PIR built into the torch, I have considered dedicated 13A sockets connected to lighting circuit to power them, but not done it as yet.

The 3.5 mA is the limit after which you can't use a plug and socket, however normally we consider 1MΩ as the limit, so 0.23 mA which would mean 65 sockets would be the limit. This of course includes any extension leads which means one fixed socket outlet could have 7 items plugged in, so what we are talking about is a reasonable leakage, what is a reasonable leakage on a circuit?

You have buy plug in units which draw 7 mA line to earth the Kewtech 107 is an example, it considers 7 mA used to measure ELI should not trip the RCD and it is common for the testers to take 7 mA, a RCD rated at 30 mA must not trip below 15 mA so in real terms we have 8 mA to play with, so all the circuits connected to the RCD together should not be below 30kΩ, if it is then it can't be tested for ELI.

But even when I knew really I had too many circuits on a single RCD I have left it in place for over 25 years, and since fitting the RCD's I know one has never had the ELI tested since it would always trip the RCD, my son has bought that house and is about to fit all RCBO's same as this house, but having two RCD's supplying two old Wilex fuse boxes was not really good enough, fuses were swapped for MCB's. It has never been good enough, but when fitted RCD's were rather expensive and large, the two fitted took up 8 slots in the adaptable box, in a modern consumer unit they would only take up half that room, but my point is my set up I know was not really designed, it was a knee jerk reaction when I realised my son was studying to become a radio ham, at 13 years old, although could not get licence until 14 years old, and that would mean he would likely play with mains supplies, and I wanted to protect him, this was around 1990 and at that point we were really only using 100 mA RCD's with TT supplies, the 30 mA was still quite rare, maybe used for a socket supplying garden but still loads of ELCB-v around at that time.

What has been a surprise is how we gets runs of the RCD tripping, it may trip once a week for 6 weeks, then not trip again for a couple of years, but I have lost a freezer full of food, so I know my design was not good enough. But what is good enough, where is the demarcation line where one must say it needs three RCD's?

 

[plugwash]

The 3.5 mA is the limit after which you can't use a plug and socket, however normally we consider 1MΩ as the limit

Don't confuse DC resistance (what an IR tester measures) with AC leakage. The AC leakage can be much higher than a DC resistance test would indicate.

Getting back to the broader topic of the thread it seems that a major problem is there is no good way to track down the cause of RCD trips, IR and leakage tests on the connected equipment may help, but if the issue in the appliance is intermittent then they may not reveal anything. The only real options at that point are to either live with it or rebuild the CU with RCBOs.

 

[ericmark]

Yes I know impedance and resistance are not the same, however what is the main question is what is normal? If the RCD trips because of damp bread in the toaster that is a fault, so not really concerned about that, there is nothing to say you can't lose whole supply with a fault, what is more the question is random trips which then go away, so not a pre-warning of some thing failing, be it due to spikes on the supply or simply additions of many small leakages, to my mind if the roof leaks and a socket gets wet so the RCD trips that's it doing its job. But if it trip once a week for 6 weeks then does not trip again for a year, then likely the circuit is too large.

I normally test earth to neutral and most switches and relays are on the line, so even a de-frost element on a freezer which only switches on twice a day, will be found with a neutral to earth test. I know my outside lights are only line switched and if my lights did trip, that is what I would look at first, however when a RCBO trips you don't know if over load or earth leakage, so a tungsten bulb which ionises when it ruptures has the same effect as a little water, both trip the RCBO.

But I go to so many homes where so much trips at the same time, and one is really scratching ones head to isolate it, only way is often to open the consumer unit and remove neutral wires which is not some thing the occupier could really do, so is it really good design to have so much on each RCD, and when it causes problems because there is so much, should the designer be blamed and should he be required to modify the installation when time has shown he has made an error of judgement.

When my veranda leaked I called on an expert to fix it, when it still leaked I expected that expert to correct it FOC. I did not tell him what to use or how to do his job, that's why I employed an expert, the same with electrics, if I was not an electrician and the RCD keeps tripping, then one of two things should happen, either the electrician finds a fault, or he corrects some error FOC. When my veranda leaked the builder did not say Oh it's likely a freak wind blowing in wrong direction, he just fixed it. So if a RCD trips and the electrician can't find a fault, and it is likely due to far too much being protected by one RCD, should he be required to fix it, including if required adding RCBO's to some circuits to reduce how much is on each RCD.

If the builder says look can't fix that properly at the moment, but I can do this temporary fix, it may work, it may not, if it doesn't then I will need to do a proper job latter then fair enough if it fails. Same with an electrician if he says really you need RCBO's but I can do a cheap job and see how it goes, then again fair enough, but unless the occupier knows it's being done on the cheap, if it keeps tripping then down to electrician.

 

[JohnW2]

Hence the sense of having emergency lighting in areas where power cuts are a problem.

That's what I said, and was my point. As I said, if the authors of BS7671 are cornered about that potential (albeit very small/rare) danger, then they should 'require' emergency lighting. That will deal with ANY cause of 'being plunged into darkness', whereas protecting lighting circuits from faults etc. on other circuits will deal with only some of them.

Sudden un-expected darkness due to trips caused by faults / mis-use in the house wiring / appliances is likely to happen more often than darkness due to a power cut.

As I said, and have often said, experiences vary. In my case, power cuts (usually very brief) are common (many times per year), and very common (can be several times per hour) if there is lightning with ~10 miles of here! On the other hand, it is many years since an RCD tripping has taken out one of my (many) lighting circuits - even though all of the lighting circuits are protected by RCDs which also protect other circuits.

Kind Regards, John

 

[JohnW2]

The 3.5 mA is the limit after which you can't use a plug and socket, however normally we consider 1MΩ as the limit, so 0.23 mA which would mean 65 sockets would be the limit.

This seems rather confused/confusing.

For a start, although it may just be down to my ignorance, I don't know where this "3.5 mA" comes from - is it perhaps related to PATesting?

Similarly, what is this "1MΩ" which you "usually consider to be the limit" - is that, again, perhaps related to PATesting? Whatever, as has been said, that presumably relates to "Insulation Resistance" when measured with dc, and has nothing to do with the AC leakage current 'during normal use' - which is nearly always down to capacitance/capacitors.

As flameport has indicated, BS7671 says that designers should avoid situations in which total L-E leakage on a circuit is more than 30% of the rated current (IΔn) of an RCD protecting it [ 531.3.2(ii) ] i.e. 9mA total for a 30 mA RCD (which would mean that 3 x 3.5 mA leaks would be 'too much')

However, no matter what figures one uses, my point still remains - that in a 'standard sockets circuits', which will usually have many socket outlets, a designer cannot usually have any clue as to what loads, with what total leakage current, will be plugged into those sockets. In a few cases (e.g. rooms full of computers IT equipment) it will be clear that there is likely to be a high total leakage but in most domestic (and many other) situations, the designer 'will not have a clue'.

Kind Regards, John

 

[scousespark]

Does this logic apply if a customer plugs in an adapter and overloads the socket, or trips over an extension lead?

 

[JohnW2]

Does this logic apply if a customer plugs in an adapter and overloads the socket, or trips over an extension lead?

Good to see you - I hope all is well with you and yours.

What 'logic' are you referring to? - I cannot work out whether you are responding to something I have written, or something that eric (or someone else) has written.

Kind Regards, John

 

[scousespark]

Good to see you - I hope all is well with you and yours.

What 'logic' are you referring to? - I cannot work out whether you are responding to something I have written, or something that eric (or someone else) has written.

Kind Regards, John

Hi John. All good here thank yo. Hope all is well with you and yours. I meant the logic of the thread. I fit 5 sockets and there is a trip due to the equipment used. I don't see how I would be expected to sort this at my expense. I just wonder if Eric thinks I should replace a socket if the customer overloads it, or trips over a lead. I don't have the gift of precognition.