Electrical Fault

Hello, I recently had a new consumer unit fitted into an old house and the associated full electrical system check and new earthing, etc. However we now have an annoying fault. Quite randomly, when some light switches are turned off in one area of the house, this trips the new 30 Ma RCD and fails the 13 amp ring main for the entire house. (this may happen once or twice a week) The light circuits in question are mainly flourecent ones. The company that fitted the new consumer board and did the testing said that perhaps a "neutral wire has been crossed over" but said that this would be expensive to track down. The fault can be induced by turning an older type light switch off slowly which causes some arching and then the failure. I would be grateful for any suggestion or ideas about this fault. Would it be worth replacing the flourecent lights for normal ones?

Kind Regards, Dick

Reply to
dick.garwood
Loading thread data ...

I assume from your description of the fault, that this is a split unit with an RCD on the sockets etc. The neutral bars are also split: - (there are 2 of them). You'll probably find that the neutrals for the lighting and/or the sockets are crossed on the neutral bar. If so, the installer requires a kita. Jaymack

Reply to
John McLean

It's the flour in the lights that's causing it! :-)

They did the installation, they did the tests. They should have tested for crossed neutrals. It's their problem.

Reply to
Bob Eager

Am I right in thinking that you had a split board fitted where some circuits are protected by the 30mA RDC and some are not?

OK I can only go by your wording here as I've obviously not seen the installation concerned. But assuming your comments are accurate then I would be concerned by the word "perhaps".

There could be other reasons, but a crossed or "borrowed" neutral would cause the problem you describe. It's not a terribly difficult thing to track down with the right test gear. (An electrician would normally have this with them to test the new consumer unit anyway).

I'm also surprised that the problem didn't show up when they were testing the new consumer unit. They should not have left an installation in a potentially dangerous state. The minimum they should have done is disconnect the faulty lights or circuits.

I would either get them back for another go, or have an independent company give you a second opinion on the state of the installation.

No. You need to get the fault sorted out first as the problem would still be there whatever light fittings you used.

John

Reply to
John White

Can we assume the lighting circuits in question are fed from the non RCD side of the CU?

If my above assumption is correct then I am not sure about the "crossed neutrals" diagnosis. Normally if you return the neutral of a circuit to the wrong bus bar in the CU, then you would expect it to cause a trip the moment you drew *any* power on that circuit. This would also be repeatable and happen every time.

Spotting this fault (if it is in the CU) is usually not that difficult by inspection - look at the MCB and follow the live back to the cable. Now follow the corresponding neutral to a bus bar and check it is the appropriate one. However if the fault is here then it is down to the installers to find and fix at their expense.

A bridged neutral elsewhere in the house that joins the neutral of a RCD protected to a non RCD protected circuit could lead to the problem you are seeing (but even then you would expect more frequent trips - especially at higher power draws (i.e. more lights on)). This would not be the fault of the installers (in the sense of them having created it), although you would expect them to have found it while testing[1].

If the lighting circuit in question was fed from the RCD protected side of the CU then this makes explanation a little easier. The first question however would be is why is this circuit on the 30mA RCD in the first place?

If your lighting circuits are on the RCD then the fact that they are older fluorescent fittings may have some relevance since these may present a moderately inductive load (i.e. they have a poor power factor). This combined with an arcing switch and a fast RCD could give rise to it seeing an imbalance where none really exists.

No, better to find out what is wrong first.

[1] There is one possible scenario I can visualise here however that would be quite tricky to find resulting from switching inductive loads, getting large back voltages, and having insulation breakdown occurring between protected and non protected neutrals, but this is a bit obscure!
Reply to
John Rumm

John,

Thanks for you reply. What you say about a bridged neutral may make sense. There are more trips when the 13 amp circuit is under more load. Would it be possible to find the bridged neutral by turning the light circuit off on the consumer unit then testing each of the 13 amp outlets and appliances on the RCD protected ring main with the assumption that if one of these did not work then it would be the one with the bridged neutral? I am grateful for your time. Regards, Dick

Reply to
dick.garwood

John, Thanks for your reply. The company told me that they had checked the consumer unit for that fault. Regards, Dick

Reply to
dick.garwood

With power off, and MCBs opened, you could put a table lamp into a socket on each power circuit and switch on a tungsten light in each lighting circuit. Open the RCD as well

Then using a meter on resistance range, remove the neutral wires one at a time and test that they go to the correct neutral bus bar.

If the electrician has done a reasonable job, they should run in the same order as the circuit breakers.

Reply to
Andy Hall

classic installation design fault, common. Total rcd circuit(s) leakage is higher than the rcd will tolerate, so any slight increase in unbalance or leakage trips the rcd.

Solution is to get rid of the rcd or replace it with 100mA one, and fit RCBOs to the formerly rcded circuits. In some cases you may not need any rcd to comply with regs.

Arcing a switch creates hf, hf on capacitive cable and other loads creates unbalanced current.

I woud be questioning why lights cause rcd trip, lights are best not on an rcd. Perhaps you have a TT system.

NT

Reply to
meow2222

If you have all things wired correctly (in the sense that the live and neutral for each circuit are on their appropriate side of the CU), but you had a join between neutrals (or lives - but that would be easier to spot!) then you have a parallel return path.

The difficulty here is that with a parallel path it would be hard not

*not* get significant current flow through both of the paths, and you only need some 20mA or so going down the "other" path to trip your RCD. Hence you would expect far more situations would result in a trip. So it does not really fit the symptoms you are seeing.

If you do have a parallel return path problem then a more likely possibility is a high resistance join between circuits. This could result in small enough currents down the "wrong" return to only sensitise the RCD (by using most of its trip limit) rather than actually tripping it. (I alluded to this with my footnote in the last message since a combination of inductive loads with a pulsed current connect / disconnect could generate significant back voltages. These could cause insulation breakdown that would not occur at normal mains voltage).

The more likely possibility is that you actually have a high leakage appliance somewhere, that is pushing you near the limit of the trip. Hence even small disturbances may result in spurious trips.

That would only show up a pretty significant fault - i.e. the wiring of a socket live connection from a lighting circuit! Pretty unlikely to happen you would have thought.

(Make sure the power is off before doing any of these tests)

If you disconnect both ends of a socket circuit at the CU, and then measure resistance between both of the socket circuit neutrals (i.e. neutral at both ends of the ring) and the neutral bus bar (i.e the neutrals of all the other circuits combined) you would expect to see an infinite reading even with the meter on its highest range. If you don't then you may have found your problem. Repeat for each circuit in turn. Generally there should be no detectable connection between circuits. If the multimeter shows nothing, then it is still possible the same test at a higher voltage (i.e. with a megger) would show something.

Reply to
John Rumm

Design fault? All the evidence points to a definite installation fault

- bridged neutrals, a neutral to earth short, or some combination thereof. The electrician seems to have admitted that he knew there was a problem and walked away from it. I wonder what certification the OP got...

Not a relevant remark because the trips are being caused by loads that aren't, or certainly shouldn't be, on the RCD side of the board (viz. lights).

The solution is the find the actual problem and fix it. An RCD protecting the socket circuits (reasonably expected to be used to supply portable equipment outdoors) must be 30 mA to comply with Reg. 471-16-01.

which does tend to suggest that there is just one socket circuit on the RCD side. That is as it should be - the OSG says (sec. 3.6.2, p. 20):

"30 mA RCDs installed to provide protection to socket-outlets likely to feed portable equipment outdoors should protect only those sockets."

Mention of "new earthing" and of only one RCD suggest not.

Reply to
Andy Wade

Dear All,

Thanks to all for your help with this matter. I perhaps should have mentioned that the only thing which causes the trip is when you switch a main room light off (all th eones on a certain circuit). The 13 amp ringmain fails as you physicallly move the light switch to off. The fault has never shown during "switch on" of the lights or for any other reason. I have tried replacing the light switches for new ones but it still happens. You could fail it every time if you held one of the older type switches in the mid position causing arching. This has happened with 3 different room switches which are all on the same circuit area of the house. The fault is really sporadic but appears to happen more as the load is built on the 13 amp.

The testing company told me that the fault was not dangerous, just annoying, would you agree with them?

I hope this provides more detail. Thanks for your time, regards, Dick

Reply to
dick.garwood

I see nothing that tells us its this rather than simply an rcd supplying so many loads that it cant stay on stably due to normal leakage. It might well be either, but nothing confirms it either way, unless I've overlooked something.

I saw it as likely the spark just wanting the op to go away, and floating some bs to try to achieve that. Of course it could be either. It seems fair to say a competent electrician should check for circuit interconnection when fitting an rcd.

which is clearly a design fault.

Having all your sockets on one rcd is also a design error, since its a recipe for unreliability. The fact that its common and legal doesnt make it good design or competent engineering.

And really it is relevant, as its the cause of the problem.

of course. Maybe. Sometimes an attempted fix is the easiest quickest and cheapest way to diagnose a problem, so its not always so clear.

hence rcbos with or without a 100mA rcd. Or in some cases no rcd, eg when a flat is not on the ground floor.

the op also said arcing the light switch trips the rcd, so there is some evidence both ways. Add the 13A ring description of the 30A ring and its clear we really dont know.

yes. OTOH you reckon the spark did a bodge job and walked away leaving it faulty, so I dont think we can automatically assume its not TT. As ever, we dont have the necessary quantity or quality of info to really know.

NT

Reply to
meow2222

It depends what the fault is. Certainly it would worry me that whoever installed the original circuits might have used some unusual wiring arrangements. That said the installation is not working correctly and needs fixing anyway.

Yes - thank you. Just another couple of questions though:

1) When the RCD has tripped, knocking out the sockets, do the lights in question still work without you resetting the trip? 2) If you turn off the circuit breaker for the lights in the consumer unit, can you still trip out the sockets as described above?

It's a good one is this. You're not near Derbyshire by any chance are you?

John

Reply to
John White

John, The answer to your questions:

The lights still work each time without setting the trip.

I am not sure about your second question and I am unable to try as I am away from home for the week. I will try at the weekend and let you know. I am a long way from Derbyshire.

I have just looked back through the original inspection report for the property and one item which I know was not corrected reads: investigate high R1 + R2/Zs for 13 amp socket in kitchen. Not sure if this is relevant!

Thanks for your continued help,

Regards, Dick

Reply to
dick.garwood

Would a 10ma one comply. I have my outside socket on one.

Reply to
<me9

Probably. The relevant section is 412-06-02(ii)

"the residual current device shall have a rated residual operating current (I[delta]n) not exceeding 30mA and an operating time not exceeding 40ms at a residual current of 5 I[delta]n ..."

So assuming it acts fast enough your device complies.

John

Reply to
John White

OK so that rules out a "whole house" RCD and, given that a RCD is a double pole device, a borrowed neutral.

My thoughts are turning to some sort of earth fault here. Perhaps a floating, or high resistance, section that is in some way reacting to the ballast in one of the fluorescent lights and again with a suppressor in some appliance on the ring. This could be enough to push a near-the-limit RCD over the edge, as others have suggested.

Fine. FWIW my guess is that you won't be able to trip the RCD with the breaker for the lights off.

OK. Had you been closer I would have offered to come and have a look for you.

Well R1 is the resistance of a phase conductor (Live in this case) and R2 is the resistance of the CPC (earth) conductor between the point under test and the supply. Zs is the earth fault loop impedance.

This is suggesting that there is something wrong with the earthing arrangements for that particular socket. While this may not be directly relevant, it is certainly something to be looked at with urgency and reinforces meow2222's point about leakage.

It's a pleasure.

John

Reply to
John White

Agreed.

Reply to
Andy Wade

Some of you are mention a TT system. Could you explain what this means? Thanks. Dick

Reply to
dick.garwood

HomeOwnersHub website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.