RCD tripping - bit different this one

[Ant2110]

Hi,

Just come back from holiday, before I go I like to turn off the non essential circuits. When I come to put them back on again the main RCD trips when I turn on the circuit for the 1st floor sockets.
I did the usual thing of removing all the items plugged in to that circuit, but the RCD still trips when activating that breaker.
I can't understand it. I've measured the resistance on the circuit and the following results;

Live - Neutral (fluctuates between 50kohm - 100kohm)
Live - Earth (fluctuates between 5kohm - 10kohm)

These readings are with the breaker off and noting connected in the circuit.

Do these readings appear normal? I can take more readings to help diagnose, but I am scratching my head with this one.

Thanks for any suggestions,
Ant.

 

[JohnW2]

Live - Neutral (fluctuates between 50kohm - 100kohm)
Live - Earth (fluctuates between 5kohm - 10kohm)
These readings are with the breaker off and noting connected in the circuit.
Do these readings appear normal?

Do I take it that those measurements were taken with a multimeter?

IF the readings are correct and IF there is absolutely nothing connected to the circuit, then they are far too low. Are you absolutely certain that nothing is still connected to the circuit? - it's not just a question of what is 'plugged in' - there might be things 'hard wired' via fused connection units, or just 'overlooked' (TV aerial amplifiers, telephones, routers etc. etc.).

Kind Regards, John

 

[Jackrae]

A neutral to earth fault will (or should) also trip the RCD so you need to do a resistance check on that also.

 

[Ant2110]

I'd ruled out neutral earth as a fault as all of the other circuits are still live and not causing a trip, the neutral and earth on the 'faulty' circuit are also still connected, it's when the breaker is switch on that the main RCD for that side of the CU trips.

I thought that the readings where way to low also, but I am pretty sure nothing else is connected (or hard wired). I did find a couple of extra sockets randomly on another floor that are also on this circuit that I didn't realise but nothing is connected now. I guess for the reading to be that low there 'must' be something still connected..? I find it difficult to imaging a physical wiring fault has occurred whilst away for a week with it being untouched..

Yes readings were with a standard multi meter - I do have a Megger, it's not very accurate above 100ohms (just reads >99.9ohms)

 

[JohnW2]

I thought that the readings where way to low also, but I am pretty sure nothing else is connected (or hard wired). I did find a couple of extra sockets randomly on another floor that are also on this circuit that I didn't realise but nothing is connected now. I guess for the reading to be that low there 'must' be something still connected..?

Maybe, but (whilst it would explain a low L-N resistance reading) even something non-faulty connected would not explain the very low L-E resistance you measured - and that (well, certainly 5k&#937 is low enough to trip an RCD.

I find it difficult to imaging a physical wiring fault has occurred whilst away for a week with it being untouched..

That's certainly improbable (although, for example, rodents don't respect the holidays of the owners of the house!) - but maybe 'something connected' that you have so far overlooked has developed an L-E fault during your absence. Are you sure that nothing 'outdoors' (hence at risk of water ingress) is connected to this circuit?

Kind Regards, John

 

[EFLImpudence]

I take it that you are measuring at a socket with just the relevant MCB open.

As, it seems, the Neutrals and CPCs (earths) are still connected at the board and therefore still connected together (albeit, perhaps at the sub-station (unless TT)), why is there such a difference between the L-N and L-E readings?
It should be no more than a few ohms.

You say you have discounted a N-E fault (that may be reasonable with a heavy load on elsewhere) but a measurement between N & E may still be informative.
It should be no more than a few ohms (unless TT).

 

[bernardgreen]

If you are measuring resistance with a digital multimeter then you need to turn off ALL power ( main switch ) before measuring. The average digital multi-meter resistance measuring system will give false readings if there is any voltage induced in the circuit you are measuring by current / voltage in adjacent cables.

The test current the average multi-meter puts through the circuit when measuring resistance is very small, a few milli-amps or even micro amps and this test current is easily swamped by induced voltage and / or induced currents.

 

[JohnW2]

You say you have discounted a N-E fault (that may be reasonable with a heavy load on elsewhere) but a measurement between N & E may still be informative. It should be no more than a few ohms (unless TT).

Yes, it could be informative, if it could be measured accurately. However, if it's TN-S (or even TN-C-S, if there are significant loads on the installation), the probable small pd between N and E could completely muck up attempts at resistance measurement, particularly low-voltage resistance measurement. I also suspect that, if it were TT, even if there were no such 'pd issues' (which there probably would be!), the measured N-E resistance might well be 'very high' (maybe of the order of figures reported by OP) if measured at multimeter testing voltages.

Given the many uncertainties, it might be better to forget about the OP's resistance measurements and simply approach this as (yet another) case of 'RCD tripping for currently unknown reason'.

Kind Regards, John

 

[Ant2110]

Thanks for the advise all....

I have been measuring at the CU itself, the circuit in question terminates to the MCB as a pair of cables (a single ring), measuring the L-E resistance on both of these showed the same resistance (as you'd expect)
I left these separated and chose a socket which I would expect to be somewhere at the centre of the ring and separated the live's and measured L-E resistance on each side...
One side is open circuit the other is 6k ohm. So I've narrowed the fault (or connected device) to one side now. Guess I can just work my way round and see if I can figure it out. This being a big old Victorian house doesn't help - the wiring is not straight forward most of the time!

 

[ericmark]

For a friend I tested his compressor with a multi-meter and it showed 500kΩ fuse ruptured so retested with 500 volt insulation tester result 50Ω. I will admit rather surprised there was so much difference, but it really pointed out for insulation testing 9 Vdc is just not good enough.

A RCD rated 30 mA is designed to trip between 15 mA and 30 mA with some exceptions like the X-Pole. As at 15 mA around 15 kΩ is your limit. You are below that limit so yes it should trip.

 

[JohnW2]

For a friend I tested his compressor with a multi-meter and it showed 500kΩ fuse ruptured so retested with 500 volt insulation tester result 50Ω. I will admit rather surprised there was so much difference, but it really pointed out for insulation testing 9 Vdc is just not good enough.

Some multimeters use only 1.5V or 3V! ... Yes, it's a problem but, in general (from the point of view of testing voltage alone), if low-voltage testing gives incorrect answers then (as in your example) the indicated resistance will usually be too high, not too low. Of course, as bernard has pointed out, the other issue relates to induced voltages/currents in installations which are not totally dead. The lower the resistance test voltage, the lower the currents it is measuring, hence the more likely are those measurements to be distorted by the induced voltages.

A RCD rated 30 mA is designed to trip between 15 mA and 30 mA with some exceptions like the X-Pole. As at 15 mA around 15 kΩ is your limit. You are below that limit so yes it should trip.

Indeed - as I said, if the OP's L-E measurement is correct, that's low enough to trip a 30mA RCD.

Kind Regards, John

 

[deadshort]

Any neon indicators on FCU's ?

Kind regards,


DS

 

[JohnW2]

Any neon indicators on FCU's ?

They would appear as 'infinite resistance' if tested at multimeter test voltages - you would need something in excess of 100V or so to get any current to flow.

Kind Regards, John

 

[Spark123]

Yes readings were with a standard multi meter - I do have a Megger, it's not very accurate above 100ohms (just reads >99.9ohms)

That can't be an insulation resistance tester, IR testers read Mohms (10^6 ohms).
My MEGGER IR tester reads 299Mohms> @500v when the insulation is above this value.
Thinking about it, it would be pretty useless as an IR tester with low values such as that.

 

[JohnW2]

That can't be an insulation resistance tester ...

As you say, it can't be. However, despite the historical origins (and the sound of the name!) the "Megger" brand now encompasses a lot more than just IR testers. However, I agree that whatever "Megger" the OP has does not sound as if it would be of any use to him in relation to the current issue!

Kind Regards, John