Circuit energised even after RCD off/MCB off?

[choff]

Moved into an old house, and I'm turning things off at the MCB + using a socket tester to see what's on each circuit from the CU (the stickers aren't particularly helpful). I saw something odd: there's a set of sockets (4 doubles) in the basement that leave the socket tester partially lit even after I turn the MCB off for that circuit, and also even after I trip the RCD for all the MCBs on that RCD.

1 LED stays faintly green when the MCB is off; it turns to a faint red when the RCD is off, whether or not the MCB is on or off. This doesn't correspond to anything shown on the tester's labelling. What might be happening?

With both the MCB and RCD on (i.e when the circuit is live), all lights are green - all OK.

The tester is a TIS 100. The faint red/green light is the leftmost one.

I appreciate that I presumably need to get an electrician in, but I'd like to have an informed discussion with him/her based on this site's collective wisdom! Plus, if it's as simple as checking the wiring in the four sockets, or replacing them (they're old), maybe I can try that first.

I have a decent multimeter, so I can do some (safe) further testing if that would be sensible.

 

[ericmark]

Those testers are not good at detecting an earth fault, need one with loop for that, you can test with multi-meter but need to know how to use it. Not sure if really a DIY job, it has done what it is designed to do, it has alerted one there may be some thing wrong.

 

[Murdochcat]

Use a plug in lamp and repeat the exercise. Those testers are NOT to be trusted.

AND if in doubt, turn off the main switch of the consumer unit

 

[ebee]

Use a plug in lamp and repeat the exercise. Those testers are NOT to be trusted.

AND if in doubt, turn off the main switch of the consumer unit

Only if you are using an ALL INSULATED lamp , not a metal one or a plastic one with a metal lampholder.
I strongly advice you get a pro involved and don`t touch anything on that circuit. There might be a serious problem so best not touching anything.
I said might be, it could be some minor problem but treat that unknown as potentially dangerous until you have it expertly comfirmed.

 

[securespark]

AND if in doubt, turn off the main switch of the consumer unit

Something I did without fail on TT fed supplies.

 

[choff]

Could this just be phantom voltage? Using a multimeter across termini in sockets on the circuit (with the MCB off), I'm seeing voltages in the 0-5V range, or 0-3V when the lighting circuit MCB is also turned off.

 

[ericmark]

A TT and TN system are very different when it comes to testing, and we don't know which you have, with a TT we use an earth rod or mat, and the earth and neutral could have up to 50 volt difference although it real terms likely to be no more that 10 volt. With a TN system the earth comes from the supplier, and it is likely only 1 volt or less between neutral and earth.

Just to confuse matters homes like mine with solar panels have an earth rod even with a TN supply for if there is a supply failure.

But the tester can't really test that neutral and earth are not connected together, or swapped over, as they are often the same voltage, or even if the earth is good enough to allow the safety devices to disconnect unless it has a loop test, even then the loop required for a ring final is around 1.38 ohms and most the cheap (£50) testers pass at around 1.9 ohms, clearly not designed for the British ring final. To get a tester which can measure 1.38 ohms we are looking at the £200 range.

So to test we need the isolator off, some RCBO's isolate, but most RCBO's and MCB's only turn off one live wire, which we call the line, the other live wire called the neutral is left connected, so it needs the RCD or Isolator to be turned off, and we can then test a group of circuits, to test one circuit we need to delve into the consumer unit and disconnect wires.

This is why I said not really a DIY job, one it needs expensive test gear, and two you need to know what you are doing.

House hold electrics are getting more and more complex, when I started we refereed to domestic electricians as house bashers, we considered them just a little better to an electricians mate, they did not have to work with three phase etc. And as to programmable logic controls, (PLC's) they likely did not know what they were, but times have moved on, today the home can have rather complex systems, old days the central heating controls were the most complex system in the home, but that is no longer the case.

Through my working life I has to do courses, and training to keep me up to date, so now having retired I am slowly getting out of touch, but it does seem odd that inspection and testing was one of the courses I had to take, what I did in collage in the late 60's early 70's was no longer good enough, and I had to do 24 weeks in night class to take a series of three courses and exams so I knew what was required (the regulations back then 16th edition) could test equipment (PAT testing) both 6 weeks, and then the PIR (periodic installation report) as it was called then now EICR or the electrical installation condition report. Which was 12 weeks, and had rather a high failure rate.

There is no way we can instruct some one on a forum as to what it took 36 hours study to learn for guys who were already electricians. The systems have become far to complex today to guide one through the testing ensuring we have not missed some thing.

 

[bernardgreen]

Could this just be phantom voltage?

Yes it could be, capacitive coupling from a live circuit cable can induce a voltage in an isolated cable. Voltage but very little current,

Testing with a high impedance meter will show the voltage because the meter takes less current than the capacitive coupling can supply.

Testing with a low impedance meter will take more current than the capacitive coupling can supply and the induced voltage will be reduced.

0-3V when the lighting circuit MCB is also turned off.

which reduces the amount of capacitive coupling of Live to the wire being tested and hence a lower voltage on the meter,

 

[opps]

Use a plug in lamp and repeat the exercise. Those testers are NOT to be trusted.

AND if in doubt, turn off the main switch of the consumer unit

I have the Martindale EZ150

Martindale EZ150 Socket Tester - Proactive Technical Training

Martindale's EZ150 socket tester is a reliable and easy-to-use tool for testing the safety of electrical sockets. Identify faults that other checkers can't. Get One Now.

www.proactivetechnicaltraining.co.uk

It has the non-trip earth loop test.

Is that equally un-trustworthy?

 

[flameport]

Is that equally un-trustworthy?

They all are, as certain faults cannot be detected and others will cause inconclusive results

 

[opps]

They all are, as certain faults cannot be detected and others will cause inconclusive results

Thanks for the reply, can you elaborate?

Tnx.

 

[flameport]

N-E reverse can't be detected.
Situations with more than one problem or anything that's not one of the problems listed on the front will give misleading results - all of which is explained in the manual available here: https://martindale-electric.co.uk/product/ez150-eze-check-xtra-earth-loop-impedance-indicator/

Suitable for a quick basic check only, such as if replacing a socket - test first on the old one, replace socket, then test again and results should be identical.
For anything else such as things which did work no longer do, other testing equipment and methods will be required.

 

[choff]

I feel like an obvious test I could do is unbundle the wires coming off my CU - they're ziptied together - and see if the effect goes away.

 

[opps]

N-E reverse can't be detected.
Situations with more than one problem or anything that's not one of the problems listed on the front will give misleading results - all of which is explained in the manual available here: https://martindale-electric.co.uk/product/ez150-eze-check-xtra-earth-loop-impedance-indicator/

Suitable for a quick basic check only, such as if replacing a socket - test first on the old one, replace socket, then test again and results should be identical.
For anything else such as things which did work no longer do, other testing equipment and methods will be required.

I really appreciate your response, but how likely would it be that someone has a socket where the earth and are neutral switched ?

 

[flameport]

No idea, but I have seen it on older installs with no RCD.
With an RCD it should trip as soon as any kind of load was connected.
With 3 wires there are multiple options for getting it wrong.

Plenty of other possibilities such as both N&E missing but socket still live, cable damage elsewhere resulting in N&E shorted, high resistance connections on one or more of the conductors and so on.

Nothing wrong with those socket testers for what they can do, but they are not magic and can only detect the problems they are designed to do.