RCD tripping overnight

[macmini]

I'd ask around, if it's happening at a definite time then there's obviously a reason for it.

Ask neighbours etc if they have anything that starts at that time?

 

[securespark]

Some RCD's do not trip (even with an earth fault present) if no current is flowing.

Disconnect the circuits that are connected to the culprit RCD altogether, then see if it still trips.

If it does, the unit is faulty or there is an external influence.

It really needs testing to see if it is performing as it should.

 

[bernardgreen]

Richard

Can you confirm there are no earth connections to the 100/100 RCD module. I have just seen a recent ( about 3 years ago ) re-wire where the first RCD ( voltage operated one ) has been left in place.

 

[securespark]

Yeah... If it has Chilton on it, that's a dead give-away... plus the dreaded initials V.O.E.L.C.B....

 

[runci21]

if it does have chilton on it just bin it , every chilton item ive had from cef has broken within a month

 

[Richard C]

Hi all, an update for those interested. I managed to borrow some test kit & insulation resistance testing narrowed it down to an earth fault on the downstairs lighting circuit. Disconnecting the circuit at the C/U resulted in our first trip free night for a week & I spent most of today tracing the fault through the circuit ceiling roses & junction boxes; it turned out to be an earth/neutral short on one of the feeds to the kitchen down lights which were installed nearly 3 years ago & have been working quiet happily for the 18 months the new C/U has been installed; I still don’t understand why it should suddenly start throwing wobbles in the middle of the night & this has led me up the garden path a bit! To anyone experiencing similar probs you definitely need to get hold of some test kit or you could be chasing yourself & ‘blue kippers’ around for days /weeks as I was; an hour with the proper kit had it cracked but I expect you sparks out there take it all in your stride!

One question regarding the MCD trip tests; the 100mA RCD checked out OK but the 30mA RCD continually fails the no trip test on half of the sine, reduce it to 20mA it’s OK; it’s obviously slightly out of tolerance but is this significant? Does it mean the RCD is faulty? Should I replace it?

Thanks to everyone for their comments & help; the pain aside, I’ve learned an awful lot in the last week!

 

[RF Lighting]

One question regarding the MCD trip tests; the 100mA RCD checked out OK but the 30mA RCD continually fails the no trip test on half of the sine, reduce it to 20mA it’s OK; it’s obviously slightly out of tolerance but is this significant? Does it mean the RCD is faulty? Should I replace it?

Did you test the RCD with everything disconnected from it?

If for example you have a 30mA RCD and the circuit connected to it has a leakage to earth of 10mA then the RCD will trip at 20mA which could be misinterprited as a faulty RCD



oh and out of interest you can still spend days tracking an intermittent RCD fault with all the proper test equipment!

 

[Thermo]

oh and by the way isnt it a red herring, not a blue kipper???

 

[Richard C]

Did you test the RCD with everything disconnected from it? If for example you have a 30mA RCD and the circuit connected to it has a leakage to earth of 10mA then the RCD will trip at 20mA which could be misinterprited as a faulty RCD

Err, no! but it was a lighting circuit causing the tripping problems & that’s only on the main, 100mA RCD. It's the 30mA RCD (everything except the lighting circuits) that’s tripping on the ½ sine no trip test & which I’m not sure about; I will repeat the test disconnecetd but does this mean I could have some leakage on the one of the other circuits?

oh and out of interest you can still spend days tracking an intermittent RCD fault with all the proper test equipment!

Seems I’ve been lucky & done well then!

 

[Richard C]

oh and by the way isnt it a red herring, not a blue kipper???

Not widely known but the ‘blue kipper’ is a very close cousin of the ‘red herring’ but much more fickle !

 

[ColJack]

30mA RCD continually fails the no trip test on half of the sine, reduce it to 20mA it’s OK

????

you sure you're doing the test right?

The RCD should trip at 30mA on both the 0 and 180 degree cycles..

the RCD should NOT trip on 15mA

 

[securespark]

It's sometimes helpful to run a ramp test on an rcd. This tells you what level of current it's tripping at. If you do it in isolation and then from a socket outlet, it gives you an idea of how much the stuff in circuit is dragging down the true figure.

Much less than 20mA on the true figure & you could say the RCD is on the way out.

 

[Richard C]

Re-did the 30mA test today with all circuits disconnected at the C/U; I repeated the test several times & the results were consistent as follows:
x½ (no trip test) 0 - No trip 180 - No trip
x1 (Trip test) 0- 17ms 180- 27ms

x5 (Fast trip test) 0- 13ms 180- 7ms


I re-ran the tests at a couple of different socket outlets but got quiet widely varying results with the RCD sometimes tripping on the 10mA x1 setting!

The test meter is a Robin KMP 5404 & has no facility for varying the trip current other than selecting 10, 20, 30mA etc & selecting the x½, x1 or x5 scale so I can't ramp it up gradually. I’m no longer suffering from overnight trips on the 100mA RCD but I would appreciate if anyone can throw some interpretation on the results obtained on the 30mA RCD; do I have a problem either with the RCD or on one or more of the circuits connected to it, or both?

 

[Adam_151]

As far as the 30ma rcd giving different readings from a socket outlet than at the board, all I can suggest is you have got a load connected, if you make sure there is no load connected all the results should be the same regardless of where you run the test from

As to the original problem, I can put forward a logical explanation as to the clockwork nature of the problem:

Due to the impedance of the earth electrode, a neutral-earth fault on a TT system is quite likely to not trip an RCD instantly like it may do on a TN system, if you assume a Ze of 95ohms and assume fault of very little impednace, you'd need a neutral, earth difference of 95*0.03 = 2.85v to trip the RCD, they are one and the same at the transformer star point and any voltage difference at your end is going to be down to volt drop in either path, so you can see that if have 0.15ohms of neutral conductor between starpoint and your house that up until 2.85/0.15 = 19A is drawn the RCD does not trip, but if the cable that feeds your house goes onto feed other houses, any load in other houses can increase the voltage drop between star point and your CU, so it can be a load outside your control that is affecting the current through your fault. Also consider volt drop on the earth path, this path not only includes your earth electrode, but the one that connects to the transformer star point to complete the path which may be as high as 21 ohms, which means, say its 20 ohms, if the earth leakage from other houses on the same transformer adds upto 140ma, there is 2.85v between your earth and your neutral and out drops your RCD (and even if the electrode at the transfomer is quite low impedance, someone testing Ze using an ELFI tester on 25A range in another house would still trip your RCD) ... I hope that makes sense, I do sometimes struggle to explain things clearly in words as opposed to diagrams sometimes!

 

[Richard C]

I hope that makes sense, I do sometimes struggle to explain things clearly in words as opposed to diagrams sometimes!

Thanks; yes a little deep but yes I can sort of follow

OK I can see that running the socket test with probably several loads connected has affected the result & would explain why they were all over the place - I forgot to remove all the plugs again! But what about the trip time disparity I got for the readings on the 0 & 180 trip & fast trip tests which were done at the C/U with all the circuits; these were consistent over several test but does this point to a faulty RCD or is it normal to get different readings on each half of the cycle?