Online assistance for electrical trade people in New Zealand and Australia Login  |  Register  |   Forgot Password
Assistance for electrical trade people




Click here to send Forum Admin a pdf document for publication on this Topic

Documents must be less than 200k in pdf format

Posted By Topic: RCD Tripping, induced voltages and other horseplay

Jan 24 2014 23:22

So, today I have had a curly one, im in need of a few other opinions on what is going on.

A house extension has been done, all new cables from scratch, the switchboard/meterbox is brand new with RCDs protecting three MCBs each, power points, new exterior HWC, lights and switchgear, all brand new.
Cables are in a bit of a bunch under the house with saddles holding up the bunch for around 12m until they feed up a section of wall into the new extension.

The customer called and the RCD tripped this morning with the jug and a juicer going, also a drier which is on another RCD that tripped the day before.
I attended site and unplugged all the appliances on the new extension, tried using various appliances on various circuits and its all good.
I pulled the laundry earth off to megger the circuit and it sparked off the E busbar, 108VAC to earth when that mcb is on, 10VAC when the MCB is off (rcd on). Meggered out fine between A-E N-E and A-N, still has voltage present when powered on.
Pulled another earth from another circuit and the same, except 150Vac between earth of circuit cable to earth busbar when that circuit is on.
Took out the MEN link, E-N read 0V, A-E 230V A-N 230V.
Ran a lead out the window to the E peg and had a circuit between the main earth cable and my lead, 150VAC between the circuit earth and my lead too.
By this point I was getting a little annoyed so touched the circuit E cable, nothing, not even a tingle, measured 150VAC to earth.
ELI was .45 Ohm with the MEN link in and 54 Ohm with it out, is this normal, I have never tested ELI with the link out?
I have clamped both the MEN link and the main earth conductor and have had 0Amp readings.

The N and E busbars are not live as in a reverse polarity situation but im fast running out of ideas.

In brief, subcircuit wiring has 100VAC to 150VAC non loaded voltages kicking around, enough to spark on the busbar but not to feel any shock. RCD is tripping.

Any suggestions or ideas on what to test are most welcome.


Jan 25 2014 08:19

Are you using a digital voltmeter without a 2000to 3000 ohm resistor across the measuring leads?

If so, repeat all voltmeter readings, and I think you will find very different voltage readings.

AS/NZS 3017 has a section covering a digital meter being used to read leakage voltages.

Jan 25 2014 09:16

For RCDs to be tripping, either you have imbalanced current in circuit or a fault in RCD.

Anything related to RCDs,you really need a milli-amp clamp. Reading 0 A means nothing, readings in mA are the only ones worth getting

Disconnect load side of each RCD, and test them with no circuit attached.

Then megger between A+N and E of each subcircuit.

Then put it all back together and tong each subcircuit; first A+N then E.
do this with various appliances on load, including combinations.

You may well have a few mA leakage from some appliances, adding up to enough to cause a trip. I\'ve seen 7mA on a brand new 500 W halogen floodlight.

Remember a 30 mA RCD typically trips at around 24 mA, some more some less. Should be none over 30 mA, and none under 15 A.

Your link out EFLI reading of 54 ohm seems not unreasonably high, but this varies depending on local conditions. So it\'s a good idea to do his test every chance you get, to build an idea of what to expect in
your area.

With bunched cables there\'s always the chance of capacitive coupling, and while the resulting voltages may vanish when using a high-impedance meter instead of an average DMM they can be \"real\".
But they are likely to be unrelated to your RCD tripping issues as the currents involved are typically miniscule.

concentrate on the currents rather than the voltages, \'cos it\'s currents that RCDs work on.

Jan 25 2014 10:39

Thanks for the fast replies so far.
My DMM/Clamp is OK, it reads down to 0.00A in the clamp, might be time to upgrade as my fluke died a horrible death.
Also im borrowing a PAT to individually test a few appliances which might show if some have a leakage.


Jan 25 2014 11:49

The error on your DMM clamp will be of 3 or 4 digits (maybe even 5 digits). Check the specs.
that means take what the last digit means (2 decimal places in Amps means last digit is reading to a resolution of 10 mA) and multiply by the error factor +/- so many digits).
So on a decent DMM with 3-digit error, a reading in figures of 0.01 A (10 mA) has a error of +/- 30 mA (ie can be anywhere from -20 to +50 mA).
These readings are of NO use when hunting for leakages of say 5 to 25 mA.
You need a clamp that has a milliamp scale reading to one or peferably 2 decimal places, so the error will be +/- 3 or 4 tenths or hundredths of a milliamp instead of +/- 30 or 40 mA.

The PAT may help with individual appliances, buy you need to see what the total in a group of subcircuits fed from one RCD is.