RCD test on all circuits or just once for each RCD?

[John D v2.0]

You have a very odd socket that is not earthed, but has the neutral and earth terminals joined together

Ultimately the only circumstances in which someone could do an EICR without identifying such a blatant defect would be if they were totally incompetent, completely missed most/all of the testing and then completed the report with purely fictional information before riding off into the sunset on their horse.

I was under the impression you didn't have to check every single socket on an eicr, just every circuit, and a sample of accessories. Is that not correct?

 

[eveares]

Which would be the very first test when doing an EICR.

Now I feel silly and a bit stupid. Of course turn off the main isolator, connect your wandering R2 lead to the MET, other end to your MFT and with the short jump/flying lead test all the earths for continuity.

 

[JohnW2]

Of course you could check the continuity of the earth in question against a known earth....

Which would be the very first test when doing an EICR.

A couple of points ....

Firstly, that would presumably only stand any chance of detecting the rather extraordinary wiring defect that has been postulated if one tested for (happened to test for) continuity between 'known earth' (e.g. the MET) and the earth receptacle of the socket in question.

Secondly, with a TN-C-S installation, one would certainly find 'continuity' between that receptacle and the MET. The error would presumably only be detected if one also measured Rn at the same socket and realised the significance of it being the same as (rather than a little less than) the measured "R2". If the socket were close to the CU, such a differentiation might be very difficult, given measurement errors and precision etc.

Kind Regards, John

 

[flameport]

I was under the impression you didn't have to check every single socket on an eicr, just every circuit, and a sample of accessories.

You don't have to do anything, all depends on what the inspection is for and what the person wants - but not checking for earth continuity at each item would be severely negligent, and if that wasn't done, how could someone could then claim the installation was safe / satisfactory ?

 

[flameport]

A couple of points ....

The point is that you and several others have missed the point of testing entirely.

Continuity tests are done with the system isolated, so whether it's TN-C-S, TN-S or any other system is irrelevant.
Any link between the MET and N will only exist when the system is energised.

There are two correct methods of testing CPC continuity.
One is to attach a long lead to the MET, and confirm continuity between that and earth points at sockets etc. This will detect the fault, as there will be no connection between the MET and the earth of the socket.
The other method involves linking the CPC and line conductor at the consumer unit, then checking at every point for continuity between L&E. This will also detect the fault - as with the previous method, there will be no continuity between L&E at the socket.

Other dubious methods of using a plug with a link inside between L&E at the end of the circuit and then testing at the CU are not suitable, as many faults will go unnoticed.

And yes - it is necessary to check at every single socket, switch, light and other item. Just checking at the end of a circuit proves nothing, it's entirely possible to have CPC continuity from one end of the circuit to the other and yet still have items along that circuit which are not connected. A common example is where a plastic drylining box is used behind a metal accessory, the CPCs are linked in a connector, but there is no connection from that to the accessory. Usually occurs when a plastic switch is changed to a metal one. The same can happen when ceiling roses are changed to fancy lights which require separate connections in the ceiling.

If some buffoon decided to just 'test' an installation by ignoring most of the tests and just checking for Zs at the socket, this type of fault would not be found. However neither would most other faults, making the whole procedure totally irrelevant.

If any test was to be ignored, then it should be Zs - if other tests have been done correctly, Zs is only confirmation of an already known fact.
Manufacturers of test equipment do not agree - hence the proliferation of 'no trip' loop testing using only a few mA, and those overpriced adaptor pieces for the inexplicable situation where a 3 wire Zs test is needed at a ceiling rose.

 

[JohnW2]

The point is that you and several others have missed the point of testing entirely. ... Continuity tests are done with the system isolated, so whether it's TN-C-S, TN-S or any other system is irrelevant. Any link between the MET and N will only exist when the system is energised.

That is a good point, given that, in practice, the isolation/de-energisation will nearly always be achieved (in an in-use installation) by operation of a DP device, thereby disconnecting the installation's N from the MET in a TN-C-S installation. I confess that I was thinking of 'SP' isolation which BS7671 appears to regard as acceptable in TN installations (and which would leave the MET-N connection intact).

And yes - it is necessary to check at every single socket, switch, light and other item. ....

That was really my main point since, as others have suggested, my impression was that this is not commonly done. Am I wrong?

How/where (if at all) are the results of all these tests at "every single socket, switch, light and other item" documented?

Kind Regards, John

 

[John D v2.0]

You don't have to document something just because you've done it. Unless you're going to video the whole process it would be very time consuming to do that level of result.
I would have thought the best way to test a circuit would be to disconnect the conductors at the cu and test the bare ends. Then there's less chance of external paths influencing things.

 

[JohnW2]

You don't have to document something just because you've done it. Unless you're going to video the whole process it would be very time consuming to do that level of result.

Whilst that is true, if not documented, one cannot have much confidence that it's been done. Let's face it, even with the limited amount of documentation (per standard 'forms'), it is clearly not unknown for the recorded test results to be fictional.

I would have thought the best way to test a circuit would be to disconnect the conductors at the cu and test the bare ends. Then there's less chance of external paths influencing things.

It still would not be possible to detect problems such as the rather bizarre one hypothesised in this thread without undertaking tests (and/or inspection) at each and every socket. Tests undertaken just at the 'CU ends' of the conductors would certainly not detect such a problem.

Kind Regards, John

 

[John D v2.0]

Tests undertaken just at the 'CU ends' of the conductors would certainly not detect such a problem.

I wasn't meaning just test there,I was just meaning disconnect before starting work.
But yes, in all cases, a specific accessory with incorrect wiiringth would likely need testing directly by some means to find the fault

 

[JohnW2]

I wasn't meaning just test there,I was just meaning disconnect before starting work.

Yes, I realised that (i.e. that it's the most certain way to 'isolate' a circuit) ...

...But yes, in all cases, a specific accessory with incorrect wiiringth would likely need testing directly by some means to find the fault

Kind Regards, John

 

[Adam_151]

I will stick my neck out and suggest that such a fault (locally created TNC earth from neutral instead of the proper cpc) could be missed on the testing part of an EICR of any sizeable installation. R1R2 is not generally done in such an environment, earth continuity is generally verified using a mix of R2 and earth loop measurements, for the R2 test to indicate a problem not only would the installation have to be isolated at the time, but the neutral link would have to be opened... now most installations you can normally find time to isolate and drop the neutral for a quick global IR, but to have it in that state long enough to go and do a dead earth continuity on say every circuit from a 12WAY TPN board, well chances are its going to take longer than the battery duration on the EM lights, so you might be going round an commericial/industrial environment with various hazards in the dark, the IT dept. are sat on their behinds wanting to know when they can reboot the server as its got to be back up by morning

The flip size is such fault is unlikely to occur by accident, its likely to be delibrate, in which it likely one of the following:

1) A DIY addition wired in two core flex or similar and hes thought doing that makes it alright..... such instances usually stand out on the visual ispection and get opened for closer inspection

2) One of the installations which was granted special permission to be wired as TN-C using single core pyro where the core is live and the sheath is CNE. These are rare enough that I have never seen it, But i've heard about it from multiple sources, so I am happy to believe it did happen. This would be all points on the installation, so you'd see it from your sample of accessories opened and from a glance in the DB

 

[JohnW2]

I will stick my neck out and suggest that such a fault (locally created TNC earth from neutral instead of the proper cpc) could be missed on the testing part of an EICR of any sizeable installation.

That is what I was inclined to think.

... for the R2 test to indicate a problem not only would the installation have to be isolated at the time, but the neutral link would have to be opened...

If (as would usually be the case) the installation had been isolated by means of a DP (or TP+N) device upstream of the DB(s), then it wouldn't be necessary to also open the neutral link, would it (since that leak would be upstream of the isolator)? With the link still present/closed, the MET (and hence the installation's CPCs) would still be connected to the incoming neutral, but not to the neutral within the installation.

... A DIY addition wired in two core flex or similar and hes thought doing that makes it alright.....

Maybe, but I would have thought that for someone to even think of doing that would require a degree of understanding that would also make them aware of the fact that it was not "alright". I imagine that a high proportion of DIYers probably regard "neutral" and "earth" as being two totally separate things, and would not be aware of the relationship between them in a TN-C-S installation.

... One of the installations which was granted special permission to be wired as TN-C using single core pyro where the core is live and the sheath is CNE. These are rare enough that I have never seen it ...

As a matter of interest, who would "grant this special permission", and why?

Kind Regards, John

 

[John D v2.0]

CNE is not allowed within an installation, for a start how on earth would you do the bonding? We already have issues with tn-c-s with high currents flowing around the system, but at least the equipotential zone is fairly clear.
Upstream RCD protection would not be possible and a copper pipe (to the gas grill) would beat a 10mm bonding conductor any day for introducing a potential next to an earthed (via the neutral) cooker.

With voltage drop limits maybe that would not be enough for danger, but still it doesn't sound simple.

 

[JohnW2]

CNE is not allowed within an installation ...

That's certainly the general position of BS7671 - and makes sense, for the reasons you describe (although I suspect that the prohibition existed before the days of RCDs).

However, Adam appeared to be saying that somone (who? .. and why?) could "grant special permission" for this to be done, at least with single core pyro - hence my questions.

Kind Regards, John

 

[Adam_151]

I'm told it was the Home Office John, the hows and whys escape me, however the words "crown immunity" often also come up: See https://talk.electricianforum.co.uk/topic/32196-garage-power/?page=3&tab=comments#comment-438538

I can also remember being told about this at college, though it was a particular chain of department stores that were said to be wired that way.

The thing in common is its always heard from those who are retired / semi-retired and they always mention it was a job they were on as an apprentice/not long out their time. So defintaly pre RCD era

Thats all I know John, like I said, I imagine it probably was relativly rare, and most have probably been re-wired by now. I've certainly not come across it, i'll take some pictures if I ever do!