RCD trip when circuit off

[RoyG]

I was replacing a ceiling rose. Obviously I flipped the MCB off at the consumer unit first _and_ double checked that the power was off to the lights before I started.

Then, out of the blue, the main RCD popped out.

Scratched head. Checked again that the power to the lights was absolutely, definitely off and reset the RCD. It stayed in for a little while and then popped out again.

Long story short, I eventually realised that it was going out when I inadvertantly shorted the neutral to earth in the rose. Presumably, current drain on other circuits in the house was pulling the neutral a little bit away from the earth potential causing an unmatched current to flow when they were connected. Just wondered if this was normal or a possible fault. How far should the neutral get away from the earth?

[ericmark]

There is only one way to isolate a supply and that is to disconnect both the line and neutral. The MCB and even some RCBO's often only switch the line supply and not the neutral.

Although the neutral should remain within the 50 volts of earth considered as safe this is only so long as nothing else goes wrong. Although one hopes one has not got any shared neutrals one can never be sure. So in most houses the only way to isolate at consumer unit is the main switch (Isolator).

We are told time and time again the neutral is classed as a "Live" wire. The Phase wire in a single phase system is called "Line" both Line and Neutral are called "Live".

If the bond between earth and neutral is close to the house and no power is being used than neutral earth faults may not trip the RCD but once the power being used starts to increase so does the voltage difference between neutral and earth. This causes many a person to blame the wrong item when the RCD trips. A neutral earth fault on a fridge may not trip the RCD until one uses the kettle. The power used by kettle increases the voltage difference between neutral and earth over the critical level that will trip the RCD.

This is why it says "537.1.4 A main linked switch or linked circuit-breaker shall be provided as near as practicable to the origin of every installation as a means of switching the supply on load and as a means of isolation.
A main switch intended for operation by ordinary persons, e.g. of a household or similar installation. shall interrupt both live conductors of a single-phase supply."

It does say however "537.2.1.1 Every circuit shall be capable of being isolated from each of the live supply conductors. In a TN-S or TN-C-S system, it is not necessary to isolate or switch the neutral conductor where it is regarded as being reliable connected to Earth by a suitably low impedance.
Provision may be made for isolation of a group of circuits by a common means. if the service conditions allow this."

But one must consider "reliable connected to Earth" and unless you have tested the supply with an earth loop impedance tester then how do you know if it is reliable? The big danger is two way switching on the stairs. All too common this has shared neutrals and with a shared neutral 230v can easy be found on the neutral wire once disconnected.

[RoyG]

Hmm. Interesting. I hadn't realised they could get so far apart. I have to change the other rose in that room next; I'll have a look and see what I get between N and E, although I suppose it will vary depending on what else is running in the house. Last time I think the washing machine or tumble dryer were on

Thanks

[russell fields]

if your supply is TN-CS neutral and earth are connected together within your distributors cut-out normally less than 3m from your CU
3m of 25mm cable has a resistance of approx 0.0022 ohms
to obtain a voltage drop of 50volts from CU to cut-out neutral current must be:
i=v/r
i=50/0.0022
i=22925A (not likely)

if 100A flows (rating of distributors cut-out fuse)
VD=100 x 0.0022
VD=0.22v

[sparx]

I think whilst all comments/calcs are valid the point is not one of voltage but of milli-Amps!
Quote"it's the volts that 'jolts' it the 'mils' that kills".
If virtually any load is connected to an RCD A device which works of course on current imbalance between L&N, then touching N-Earth will provide a by-pass route for N current thus unbalancing the RCD.
As an aside this means it is almost impossible to do a loop impedance test with only L-E connected to test instrument, since the test current trips RCD.
Some testers will do so ie latest Megger @£1500! so 3 wire low current testers are the norm.
So to answer the OP's question this is a perfectly normal occurence, one of many pains leckie have to contend with!