A question for our resident Sparkies

[Cursitor Doom]

Gentlemen,

Some of the 13A power socket installing in our house leaves a bit to be desired. The sockets are fine old English-made ones, but they look to have been installed by a wanker with a blunt penknife. The connection quality is generally poor. This was all done before we bought the place so don't blame me for being a cheapskate!! The plumbing was pretty much up (or down) to the same standard, but I've replaced all that.

I would like to test each individual outlet's earth connection. A) to see what the earth loop impedance is and B) to see if there's an earth connected at all! I've seen a lot of stuff on YT about how to test your external ELI and overall ELI, but nothing on testing individual sockets which is what I'd like to do. Apart from AVO and Megger multimeters, the only thing I've got which *might* be capable of such a measurement is a Robin 4112 ELI tester (an analogue one of uncertain vintage but appears to be in good nick). A young, recently qualified apprentice of my acquaintance professed astonishment at it and couldn't figure out how it worked, but I'm guessing that's just a symptom of our modern education system - and callow youth! So.... how do I proceed, please?

[John Rumm]

Well the basic plan would be to take an ELI reading with it on a socket very close to the CU. That will give you a baseline reading where most of the reading will be accounted for the external loop impedance, and only a small amount for the circuit under test. You can then test other sockets in the same way, and subtract the first reading from them to get a reasonable estimate of their individual loop impedances.

Note that if the circuits have RCD protection you may need to bridge out the RCD to stop it being tripped by the tester.

[Al]

Should do it. That's the one with that weird little fly lead going to a HV probe connected to the meter via a 3.5mm audio jack IIRC! Very odd arrangement. Still, they had a good name overall til they were eventually taken over by Fluke - and I presume still have.

[dennis]

Why would the RCD trip just because you were putting some current through the earth? The earth shouldn't be connected to the RCD in any way assuming all the appliances are unplugged.

[John Rumm]

It won't trip due to passing a current through the earth as such, but obviously that current is coming from the circuit line wire, and not completing a circuit using the corresponding neutral - thus bypassing the RCD. So it will see it as an imbalance.

(most modern ELI testers have a "non trip" feature that allows them to test the ELI typically using very short pulses that the RCD will hopefully not notice. The older ELI testers will often pass a significant current to earth for the duration of the test (and hence can actually give you a better test of the earth performance under "load"))

[Cursitor Doom]

At this installation there is just the one RCD at the mains incoming box aside the meter and company fuse. It's frequently given to nuisance tripping at the slightest provocation (I even have to stifle my sneezes, for example). How do I go about bypassing this RCD - for the duration of the test (at least?) Is there an approved manner of going about it? I'm guessing there will be. ;)

[John Rumm]

Generally its considered better to not bridge them[1]. However, if you are going to bridge the RCD, then the preferred way is usually to use a spare incomer style double pole switch wired across the switched poles of the RCD. That way you can "switch out" the RCD for just long enough for each test. (note that some of the seemingly excess caution here is dealing with the situation where you may be testing an installation that others could be using at the same time).

One other thing to beware is that some RCDs can give a false reading (typically 0.2 to 0.3 ohms higher than they should) when measuring the ELI through them using the "non trip" settings on modern test gear (MK ones have a bit of a reputation here). It does depend on how the test gear does the test though - some use short duration, and others very low test currents.

[1] If you can measure Ze with the tester (i.e. the loop valve before the RCD), then measure the R1 + R2 values[2] on the circuit of interest with it disconnected, using a normal low ohms meter. You can then add them to the Ze measurement to get Zs at each socket. [2] you can follow the procedure in the On site Guide section 10.4 for testing a ring (the bit where you cross connect the line and neutral at the ends of the circuit, and then measure the L to N resistance at each socket). Also described here:

[harry]

Assuming you have all metal cold water piping. The straightforward way is to measure the resistance between the socket earth pin and the nearest bit of pipework. (using your resistance meter and a long bit of wire. Should be 0.5 ohms or less.

This additional to an earth loop test

[Cursitor Doom]

Thanks, John. I'll look into that. As I said it's the condition of the earthing in all the 13A output sockets that concerns me primarily.

[Cursitor Doom]

Interesting suggestion, Harry, thanks!

[John Rumm]

The tests as described will give you a good feel for the quality of the earthing at each socket - they are designed to highlight any problems on an unknown and untrusted socket.

[John Rumm]

Interesting, but not directly useful since it does not tell you what portion of the resistance measured is attributable to the pipe and that to the circuit.

However you could measure resistance of your local TT earth (i.e. the pipe) with your ELI tester, or by following one of the test procedures here:

[Cursitor Doom]

Thanks, John. I'll proceed along those lines.

[John Rumm]

For the avoidance of doubt; I meant to say "untrusted circuit" rather than socket.