RCD will not reset after test until random appliance disconnected

[Applestalks]

Hi,
About once a month I have the main 80A RCD trip. It will never reset immediately, so I turn off major items one by one (cooker, washer, fridge, dryer, etc.) until the RCD will reset.

1. I only have one appliance off at a time until I can rest the RCD.

2. Mostly, the appliance that allows the reset is not in use.

3. Here is the interesting bit, it is a different appliance each time that allows me to reset the RCD.

4. I can then always turn on the offending appliance without tripping the RCD

I then tried pressing the test button on the RCD box. Same result - I can only reset after a seemingly random appliance is turned off. Now I can reproduce the problem but am no closer to diagnosing the problem.

Any suggestions would be gratefully received.


Some detail:
RCD is LeGrand 086 31 80A.
House built in 2003
Of the major appliances, all are less that 2 years old except washer (7 years)

 

[sircerebus666]

Do all of your circuits (lighting , sockets, cooker, shower etc ) go through this RCD?

When the RCD trips do you turn off all the mcbs , reset the RCD , then turn on the mcbs one by one until it trips again? Therefore you can work out which circuit it is that's at fault?

I had a similar problem at a friend's house where I was told the RCD would randomly trip for no reason , when I investigated I found that it happened every time it rained , looked outside and somebody had just cut through the mains cable supplying a security light leaving the cable live!

In your case I would be changing the RCD

 

[crystal ball]

I went to a similar fault, RCD covered all circuits, tripped randomly, would not reset until, fridge, TV, kettle were unplugged, fault was found in the shower, melted cable, even though the cores were separate I think it was tracking across the carbon to earth

 

[ericmark]

With a line - earth fault disconnecting the faulty items with easy show which item is faulty, but with a neutral - earth faulty this is not so easy, the problem is a faulty toaster for example can cause the RCD to trip when kettle is used, this is due to the volt rise in the neutral wire when there is a load.

I had the same problem in last house, the RCD would trip maybe for a month with no fault found, and then would run a year with no trips, they were old type AC RCD's with no electronics in them and were susceptible to being tripped by voltage spikes, and I had too many circuits on one RCD, were fitted early 90's and at that time RCD's were new, and the RCBO was not available for fuse boxes of the time.

I had always intended to replace the fuse box with a modern consumer unit with RCBO's but never got around to it, and did loose a few freezers full of food over the years, so this house all RCBO which reduces the chance of tripping for no reason, but £500 plus for a new consumer unit to stop the odd trip is a lot of money, and it may not be cured then, so first you need some one with the test equipment.

You could buy an insulation tester at around £35 and one by one test every appliance neutral to earth looking for a fault, in essence your doing a part PAT test, but the point is even then it may find no fault, and a multi meter is useless at finding neutral to earth faults, the electrician has three tools he often uses, one the insulation tester, two the low current clamp on ammeter, and three the RCD tester, so he has more chance on finding a problem to you, there may be no problem other than having too large of a circuit.

DIY wise don't leave items plugged in when not in use, not so bad with Class 2 equipment i.e. stuff fed with two core cable which have no earth or charger units where the earth pin is plastic, but any item with 3 core cable that needs an earth, unplug when not in use, you may be lucky.

 

[Applestalks]

Do all of your circuits (lighting , sockets, cooker, shower etc ) go through this RCD?

Yes. All main kitchen appliances plus downstairs sockets.

When the RCD trips do you turn off all the mcbs , reset the RCD , then turn on the mcbs one by one until it trips again? Therefore you can work out which circuit it is that's at fault?

If I do this, they all come on, no problem. There is no single triggering appliance.

If I leave all MCBs on then the RCD cannot be reset.
However, when I do this:

* choose 1 MCB randomly and swich it off
* try to reset RCD
* if RCD does not reset switch the MCB back on
* repeat

If I do this, I will repeat the above 4 or 5 times and then the RCD will reset. It is always a different appliance, when switched off, that allows the RCD to be reset.

There would seem to be no correlation to weather, time of day, etc, as I can provoke this behaviour by simply pressing the test button on the RCD.

 

[sircerebus666]

I'm confused by your terminology , I'm talking about the different circuits that have an MCB attached to them which is looked after by the RCD

Do you mean it is always a different mcb (not appliance ) when turned off allows you to reset the RCD?

If you can recreate this issue by testing the RCD and it won't reset , then that suggests the RCD is probably at fault and needs changing

Any chance of a pic of your cu showing the mcbs and RCD ?

 

[Applestalks]

Sorry, that was a bit confusing.

Below is a picture of the cu. RCD it tripped and 1 MCD is on.

1. I press the test button. RCD trips.
2. Four times out of five, the RCD will not reset with all or any single MCB on. It does not matter which MCB is on, it will not reset.
3. After trying four or five times it will come on.

The confusion with appliances came about as I can leave all MCBs on and turn off, then on again, individual appliances around the house until it resets.

One new piece of information: it is much more likely to reset if I move the switch very fast. Speed of switching might be the variable that affects whether it can be reset. - But I was never switching it hesitantly, always with a good positive movement.

I never used to have this problem every time a light bulb blew and tripped the RCD - so I assume there is a problem more than technique.

 

[SUNRAY]

In your case I would be changing the RCD

I'd get the RCD tested rather than replacing it on a whim, and at the same time test the circuit/appliances to find the real problem.
Additionally I wouldn't be recommending working in a consumer unit until I had established the OP's compency level.

 

[ericmark]

I had exactly the same problem with my old house, two RCD's fed two old wilex fuse boxes with fuse carriers swapped for MCB's and some times resetting one RCD would trip the other, and some times you had to switch off MCB's to get it to reset, and since I had an insulation tester and RCD tester I did check the insulation resistance which was OK, and the RCD which also showed as OK.

Fitted around 1992 the RCD's were old with no or little electronics, and I noted when the welder next door left no more problems, but the problem would come in batches, for maybe two months we would have a problem then there would be some road works and no problem for years.

I assumed but could never prove it was spikes on the line causing the trips, electric storms would some times take them out, so when I moved I had enough with random trips and I fitted all RCBO's to this house, yes tripped, but always with a good reason, like roof leaking. I think there were simply too many circuits on the RCD in old house, mother's house had 6 RCD's 4 RCBO's and two RCD's and no problem with tripping, kitchen was all RCBO's.

Had I had a modern consumer unit then I would have fitted some RCBO's but the old Wilex fuse box did not have the option, I did consider a more modern RCD like the moeller x pole RCD which is claimed not to trip as often and has a warning window to let you know it's getting near the point of tripping, and was claimed to trip between 90% and 100% leakage rather than the required 50% to 100% leakage. However they are the old type AC so never got around to changing it.

Personally I think

(iii) take account of danger that may arise from the failure of a single circuit such as a lighting circuit
(iv) reduce the possibility of unwanted tripping of RCDs due to excessive protective conductor currents produced by equipment in normal operation.

means we should not use just 2 RCD's except for very small homes, it would be impossible to split the circuits in this house so lights and sockets are not on the same RCD, as sockets split front/back and lights split up/down, but the rest of the trade does not seem to agree with me, the valid point is made with a general power cut you will always loose both sockets and lights, except for emergency lights.

So it is a case of compromise, keep installation cost down, and give some options to use alternative sockets, splitting front/back or side/side is not too bad, you can run an extension lead to keep things running, but the up/down split means that extension lead would be on the stairs, I think that is clearly wrong, but seen it done again and again.

You can clearly throw money at the problem and swap the CU, but if anything like my old house it will go away on its own if left, or go to a point where you can find the problem.

 

[securespark]

I'd get the RCD tested.....

The spark ought to know this, but the RCD should always be tested in isolation, as stuff attached to it can cause spurious readings.

 

[flameport]

means we should not use just 2 RCD's except for very small homes,

Perhaps - but for a very small home, the cost of RCBOs will be very similar to dual RCD.

 

[ericmark]

The spark ought to know this, but the RCD should always be tested in isolation, as stuff attached to it can cause spurious readings.

That was my opinion, however with the problems of type AC, A, B, and F now not so sure, maybe better tested in service.

 

[ericmark]

Perhaps - but for a very small home, the cost of RCBOs will be very similar to dual RCD.

I have been told one should test the back ground leakage is less that 1/3 of rated valve, so with a 30 mA that's a 9 mA limit, but I have never owned a clamp on meter able to measure that, so I have never checked. In fact my RCD tester did not have ramp, so could not even test with that, but I was always told to test the RCD with all MCB's off, however I now question that, if tested with all turned on if there is excessive AC leakage then it will show up. The problem is an insulation meter uses DC, so resistance is not the same as impedance.