Random Tripping on new consumer unit

[gw3xjc]

Hi, i'm new to this forum, I am more of an electronic engineer rather than an electrician so will need some guidance please. I wired my house out 30 years ago and the standard installation with a Wylex CU fed with a 100ma rcd. The only time this would trip is if there was a genuine appliances fault or if we had a lightning strike.
I bought a new split load CU last week and fitted it yesterday. I have fed all the lighting off one side and feed ring circuits off the 30 ma RCD side. The whole thing is fed from the same old 100ma rcd. I am awaiting the supplier to fit a PME system which will be in about 3 weeks.
after fitting yesterday I had three random trips over a 24 hr period. Nothing had changed or been switched on, or plugged in. I checked with a meggar this morning and the three ring circuits are reading about 1meg to earth each as indeed the neutral is. BTW this was will all appliances connect, perhaps not recommended but thats what I done anyway.

Can anyone give some guidance as to where to start looking?? I am well used to fault finding in other fields but not this one. When it does trip it is sometime difficult to reset the rcd, it may take two or three times, any clues there or is that normal. It has not tripped for 10 hrs now.

Regards Roger

 

[jj4091]

Fault finding is the same for everything, isolate 1 cct at a time until you find the one that causes the problem, then break it down to the component.

 

[gw3xjc]

Thanks for that, but its of very little use to me, as I say I am well versed in fault finding and if the fault was apparent all the time or if I had an unusual reading on one of the circuits I would be on the case straight away.

What I was looking for is some advice from an electrician who had perhaps experienced this problem before with many hours between tripping and perhaps could offer the benefit of his experience on how to eliminate things.
Thanks Roger.

 

[jj4091]

How do YOU find intermittent faults then?

 

[industryspark]

firstly do the IR tests with all appliances disconnected. leaving them connected can often give miss leading results and besides you dont really want to be putting 500V down your new plasma tv!

do it circuit by circuit one at a time. theres no quick fix, its a process of elimination. if you dont turn up any faults is it possible an old appliance is giving off some earth leakage above its norm or that you have a 'build up' of leakage?

im assuming its the 30ma that is tripping.

 

[gw3xjc]

Intermittent faults are always a devil to find as you know.

If the fault was on a printed circuit board we would see if it's a mechanical problem by firstly taping the board to see if affects the operation. We may also flex the board to find out the same thing. If this causes the fault to come and go we can normally say its a dry soldering joint or break in the copper print. If it is a dry joint it can normally be tracked down to some components which are maybe normally running quite hot.

If it was a thermal fault we would either, or perhaps both, warm the peice of equipment up by covering it with a cloth or towel until the fault appears or we would use a can of freezer spray which would rapidly bring the temperature of the suspect component down.

 

[gw3xjc]

Pressed the wrong key then, hadn't finished, more to follow
Roger

 

[gw3xjc]

Intermittent faults continued------

In all intermitent faults we would try and re create the fault, it is much easier to find that way.
Of course there is always the voice of experience who has come across the problem before or can suggest a quick way of fixing the thing which is why I've posted on here. I am sure there will be someone on here who has experienced this problem before and have a suggestion as to which way to tackle it. My intial thoughts are to restore the 100 ma trip in all circuits and feed one circuit only with the 30 ma rcd and over a period of days perhaps eliminate which of the four 32a and one 40 circuit is causing the problem.

I have heard that the fault need not be in my house even?

Regards Roger

 

[gw3xjc]

Thanks industry spark, What do you think of my idea of re installing the 100ma rcd to every circuit except one and then feed that one with the 30ma trip. And yes it is the 30 ma rcd thats going. It's been fine now since 11pm last night. I suppose them will always be some leakage with some supressors and transient vdrs on equipment. What sort of leakage in your experience is acceptable?
Your right about a little knowledge for sure.
Regards Roger.

 

[Softus]

Have you noticed this button?

 

[securespark]

Is your 100mA RCD S type?

You have TT at the moment?

Have you got an instrument to test the RCD trip times & ramp test?

A ramp test will tell you what the RCD actually trips at. Do both these tests with the RCD disconnected from outgoing circuits, otherwise your results (esp. ramp test) may be inaccurate.

I know it's new, but I've had 2 new straight out of the box & fail on me.

Test both RCD's then if they're OK, and you have not had this problem on the 100mA when your old board was connected, then it could well be something that is connected through the new 30mA RCD.

Is it just ring finals on this RCD or do you have other circuits?

As has been said, you need to narrow it down by isolating circuits (depending on the type of earth fault, switching off the breaker may not suffice. You'll need to isolate the neutral as well) to find the culprit. If it is a ring, then unplug (or switch off if it's hard-wired) everything on that circuit. From then on it's a process of elimination.

 

[RF Lighting]

I'm thinking that the problem has possibly existed for a while, and the old 100mA RCD is voltage operated, so leaves the fault undetected, but the new 30mA RCD is current operated so will see the fault?

 

[gw3xjc]

wow thanks guys. Securespark, Hmm not sure about S type, i do know its a 100 ma rcd current trip, not voltage. Enlighten me on the S type.

I do have TT at the moment but awaiting PME

I do not have any instruments to test trip times and ramp test. However I can probably get access to them.

The fact that it maybe faulty had occurred to me, I was not sure how common it was.

I do have only the rings on the 30 ma trip circuit, the lighting is on the other circuit protected by the 100 ma rcd.

I did disconnect the neutrals as i realised it could be a neutral to earth fault.

Intermittent faults are a pain !!!!!!!!!

RF Lightning You may well be correct the fault, if there is one may have been there for some time.

It always puzzled me why the rcd would trip under lightning conditions, the next door would not be affected????????

Thanks for you help guys, learning by the minute.

Regards Roger

 

[RF Lighting]

An S type RCD has a time delay for operating. This is to provide discrimination between the two RCDs in a standard TT setup.

This is so that if you have an earth fault on a socket circuit (Which is protected by both the 30mA & the 100mA RCDs) then the 30mA RCD will trip before the 100mA, leaving the circuits such as lights etc still with power.

An RCD shuold not operate during an electrical storm. This would suggest a fault somewhere on your system. As your old 100mA RCD is CO then maybe the fault is >30mA but <100mA, or maybe the old RCD is the faulty one.

I would first of all run some tests for trip times and the ramp tests to both RCDs, with the RCDs in isolation. (Having circuits connected can throw up some misleading results).

I find doing a ramp test with the RCD in isolation, and then with the circuits connected and energised will often give an indication of how much leakage to earth is present under normal (with a possible fault) conditions.


This is a Type S RCD

Note the letter S in a box above the test button.

 

[securespark]

I would first of all run some tests for trip times and the ramp tests to both RCDs, with the RCDs in isolation. (Having circuits connected can throw up some misleading results).

Have you got an instrument to test the RCD trip times & ramp test?

A ramp test will tell you what the RCD actually trips at. Do both these tests with the RCD disconnected from outgoing circuits, otherwise your results (esp. ramp test) may be inaccurate.

[emoticon=raspberry]