I would like to add my morsel.
When I did my C&G course on initial testing & inspection, I was taught to use an MFT with plug & lead at a socket to conduct RCD testing. I can see when there perhaps is no other option to test at the RCD terminals, i.e. RCBO protected radial circuit with just FCU at the end. Electricity at Work Regs does permit working live, but only if ‘unreasonable in all circumstances to be dead’ etc.
I have looked at several references on how to carry out the functional test of an RCD.
GN3 states; ‘the test is made on the load side of the RCD between line conductor of the protected circuit and associated cpc’.
OSG goes on to add; ‘as near as practicable to its point of installation’.
I find that somewhat ambiguous. If I was trying to describe the test being conducted at the RCD terminals only, I would not use the phrase ‘load side of the RCD between line conductor of the protected circuit and associated cpc, and as near as practicable to its point of installation’.
I would have written something like ‘on the load side terminals of the device being tested and not on the circuit being protected’. But that’s my opinion.
Unfortunately neither GN3 nor OSG have a pretty set of diagrams (unlike most other tests) to reinforce what they are actually describing.
BEAMA RCD Handbook; All tests must be performed with all loads disconnected, making use of an appropriate calibrated test instrument connected as close to the RCD as possible for convenience.
Advanced Electrical Installation Work publication (by Trevor Linsley); ‘…using standard lead supplied with MFT, disconnect all loads and plug in test lead into socket at the centre of the ring (i.e. socket at the furthest point from the source of supply).
I have spent some time web engine searching to find anything to suggest that the test ‘MUST’ be carried out on the RCD terminals ‘ONLY’, with a negative result.
I’ve emailed some of the RCD manufacturers, putting forward both sides of the argument (citing, poor IR, other leakage currents and possibly cable capacitance reasons for not carrying out test at socket etc), some of them responded;
MK: Unfortunately we are unable to comment on site specific testing of any devices including RCDs; however we would advise that all testing must be carried out strictly in accordance with the requirements of BS7671 and associated guidance notes.
Hager: The correct way to test an rccb is at the terminals with the load disconnected in accordance with the on site guide.
Wylex: The test should be carried out as near as practicable to its point of installation. This may vary from install to install and it may be that a test using a plug in lead at the first socket is that nearest practicable point, just depends upon the install and any accessibility issues that may exist.
Then others;
Elecsa: The issue here is being able to test the RCD in a safe manner. Whilst the OSG and GN3 state “load side of the RCD with loads disconnected” which suggests purely testing the RCD at the board, it is quite an acceptable option to test the RCD by plugging the tester into a socket as long as your tester has that facility to allow you to do that.
Electrical Safety First: Usually we would refer you to the (scheme helpline), but I have scanned in this page (attached in .pdf format) from our “snags and solutions” guide, which will hopefully be helpful. This refers to the NICEIC booklet part 3 Inspection & testing; Snag 49 RCD testing – when current using equipment is connected. Answer: disconnect equipment, measure at any convenient point in the circuit, such as socket outlet. There is also a diagram that shows test being done at the RCD terminals with circuit still connected, but loads disconnected.
I then asked my training establishment: 'It does not matter where on the circuit you carry out the test as the tester uses a set amount of current i.e. for a 30 mA rcd at 1x the tester will use 30 mA thus the resistance of the circuit cable is compensated for. If the circuit resistance was a factor so would the Ze value be. It is convenient and safer to carry out the test at a convenient socket.' (There is a longer response from them, which I haven’t included here).
Kewtech: Both are correct. You have to look at practicality and accessibility. What happens if you test at a socket and the RCD does not trip due to an earth fault and you then go to the board and test the RCD and it trips? You have proved the RCD at the board as opposed to the socket.
Lastly I spoke to my MFT manufacturer (Megger). Who said; neither is the wrong way to carry out the test, but it would be preferable to test on the circuit being tested. Poor I/R should have already been tested and investigated, earth leakage should not be present (as loads disconnect) & the circuit would have to be extremely long for cable capacitance to have an effect (also loads disconnect) (not verbatim conversation). Also said ‘we are aware of an RCBO – MEM MR30 – which will not function test with the tester at its terminals, but will when tested on the circuit being protected (I can provide the citation if required).