Tripping RCD on main disribution board - Help Requested

message from Steve Reeves on 21 May 2004

Hi all

I am wondering if somebody would help me out.

Over the last 3 weeks the main RCD fuse on a Crabtree Starbreaker
domestic distribution board has been tripping out randomly - most often
when we are not home and there is no serious load active. (All of the
other circuits never trip).

I have done some research and I understand that sometimes these fuses
can get over sensitive and I was hoping that I could buy another one and
replace it to see if that helped.

Unfortunately I don't have the instructions for the Starbreaker - and so
I was hoping somebody could explain what I need to do to replace the
main RCD switch (30mA).

A photo of my Starbreaker can be found here:

http://www.stever.demon.co.uk/starbreaker.jpg

I'd be grateful for any advice and guidance. I expect it's a simple
excercise but I'd rather be sure what I was getting into in case I do
something wrong and electrocute myself!!! I had assumed the fuses pulled
out - but there isn't much to grip onto - so I am not at all sure that
this is the case.

Thanks for any help, and regards to all.

Steve Reeves replied to Steve Reeves on 23 May 2004

Hi there, cheers for the reply.

Is it not possible that the supply to the distribution board could cause
this problem?

I see - the other switches then are simply circuit isolating switches
then? Doesn't surprise me - the electrics were done by the original
house builders/contracters - has always been this way since we moved in
5 years ago.

Well, I'll make time and I'll find the cash if absolutely necessary -
I'll see what I can do first.

I hear what you're saying - will do this.

Sorry, what's a RCBO?

cheers

N. Thornton replied to Steve Reeves on 22 May 2004

The main switch on the Consumer Unit in the pic is both a switch and
an RCD. An RCD switches off when it detects earth leakage. You almost
certainly have earth leakage somewhere in your house. It could be
anywhere, in a portable appliance or the fixed wiring.

Dont contact the power company, that is really not an appropriate
course of action unless you want to pay them to come sort it out.

Your install is unsatisfactory, because you have everything on one
RCD, and exactly this problem is almost bound to occur some time.

The relevant question is do you have time or money?

If money, you can get an electrician to PAT test all appliances in the
house, and check the wiring for leakage. Problem will be located and
probably fixed in a day - but it will cost.

The free option is to narrow down which appliance it is (it is >90%
likely an appliance). I'd start by unplugging everything not
essential. Now if it trips again your fault is among the things still
plugged in. If it doesnt, the fault is among what you unplugged. Note
that items must be unplugged, not just switched off. A process of
elimination will probably find the problem. But its not guaranteed, as
sometimes the trip can be due to accumulated fault currents from
several items.

A split load CU would have been a much better choice, and the use of
RCBOs better still.

Regards, NT

stefek.zaba replied to N. Thornton on 24 May 2004

Crucially, they're not *isolating* switches: in a domestic consumer unit
the individual-circuit MCBs are *single*-pole, so they interrrupt the
live ("phase" if we're being all modern-n-pedantic) when operated manually
or on an overload, but they leave the neutral path still connected. Since
your RCD is a current-balance device, roughly "I counted the milliamps out
and I counted them all back in, and they all (well, all bar maybe 20 of 'em)
came back)", if you give enough of those pesky milliamps a different route
back to earth through a low resistance between N and E - yes, even on a
circuit you think you've "isolated", since its N conductor is strapped to
the N conductors of all the still-active circuits through the N busbar
in the consumer unit - the RCD will trip. Hence, when you're
fault-finding, you need to *isolate* appliances and/or whole circuits,
not just operate their on/off switches or the (typically single-pole)
switches at sockets.

As Nick T says, and as your background tells you, you're in for a b****r
of a time tracking down an intermittent/only-just-there threshold fault
such as this one. If you have the house to yourself (or for better safe
working practice, one other person with you who won't want to turn on
kettles, central heating, TVs and the like ;-) you could try "binary chop"
to speed up your fault-finding:

suspect-circuits = all-of-them;
REPEAT
cut all power;
isolate half the suspect-circuits;
reapply power
IF RCD pops
suspect-circuits = the-ones-still-connected
ELSE
suspect-circuits = the-ones-you-just-disconnected
UNTIL number-of-suspected-circuits = 1

(As noted above, to fully do the "isolate" thing means not only operating
the MCB, but disconnecting the circuit's N conductor(s); so you need to
be sure you're competent to fiddle in the guts of the CU, and realise that
the incoming L terminal is STILL LIVE when you've operated the main switch;
and "reapply power" in the above is not best implemented by using the main
switch alone if there are significant loads (immersion, other kitchen kit,
spindle moulder in the garage without no-volt switch ;-) still connected.)

Once you've narrowed it down to a single circuit, you can try repeating the
approach to find the appliance or section of wiring which is causing the
fault.

But Sod and Murphy are ready to give you grief. We already know the condition
is marginal: it's positively likely that there's a a little bit of L-to-E
leakage going on in multiple circuits (interference suppressors have
capacitors running from L to E and N to E, and yer woshdosh, washmosh,
'puter, prolly CH boiler, and other things will all have one, pushing up
the "baseline" leakage to the point where the main culprit is enough to
unbalance the RCD, but isolating these other leakers will reduce the
the "baseline" leakage so that the main culprit no longer makes the RCD
pop reliably. If you're competent to do V = IR and P = IV calcs,
you could knock up a "leak-inducer" to add, say, 10mA deliberate leakage
from L to E as part of faultfinding using a suitably-rated (for both
voltage and power - go overboard on the power rating as the nominal rating
on wirewounds allows for a surface temp of 100 degs C or more!); if not I
certainly won't put you in harm's way by doing those calcs ;-)

Likely sources of increasing leakage are: immersion heater elements;
leccy cooker elements (rings, oven, grill); kettles, irons, dishwashers,
washing machines. Less likely but quite possible is cable damage in some
out-of-the-way place - one "sodit" factor can be wiring whose insulation
has been damaged (rodents, original install cockups, whatever) which are
very-close-but-not-usually-touching but which move enough when warm
(either ambient heat or through passing a load) to cause a fault; once
they've done it a couple of times they accumalate a little layer of carbon
making further faults less spectacular and so (in the case of an L-to-E)
more likely to make the RCD pop than the relevant circuit's MCB.

RCBO = all-in-one device incorporating the functions both of an RCD (the
"current balance" thing) and an MCB (overcurrent and short-circuit protection).
They're still pricey (hard to find retail/undiscounted-trade-counter at
under 30 quid, compared to MCBs at 5-6 quid a go), but are coming down (were
more like 40-45 quid only 3-4 years ago). Since each one protects just one
final circuit, they make nuisance tripping less likely, as there's only
one circuit's worth of "baseline" leakage for each one, and you know right
away which final circuit is faulty. (Some even have a different indication
for an imbalance-trip vs. an overcurrent trip; and most take up twice as
much room side-to-side as an MCB). With a whole-house RCD as
you seem to have, nuisance tripping is maximally likely and maximally
inconvenient: with a split-load CU only selected circuits with a higher
likelihood of safety-problematic leakage are on the RCD, while "boring"
circuits (e.g. lighting) are fed from the non-RCD side - keeping the
lights on contributes materially to overall, practical safety!

I wouldn't spend forever trying to track this one down: if you get stuck,
find a smallish local electrical contractor, be helpful to your sparks,
tell them which circuits seem still to be suspect and so on, bring tea
and non-manky biscuits, and the combination of test-gear-in-hand, and
experience in using it, may well result in 50-80 quid more effectively
spent than similar money on a used (and how well calibrated and
cared-for?) RCD tester wot you get off some guy on eBay and won't use
ever again. Guy I got in in the previous house did this for me as part
of an inspection after I'd done multiple bits of surgery on the house
electrickery over a few years, and quite quickly tracked the nuisance
trips on the then-fashionable whole-house RCD to the combination of
'puters (which I knew), ageing grill *and* one hob element on the cooker
(needed his leakage tester for that), and a just-a-bit-too-sensitive RCD
(again requiring use of his RCD tester; a 30mA nominal RCD is supposed
(a) to let a 15mA imbalance pass for an unbounded time, (b) to trip
within small-number-of-millisecs for a 30mA imbalance - ours was tripping
at 13mA.) And a sparks will be a bit less phased (merely eyebrow-rolling
rather than dangerous-to-self incredulous "they couldn't have done THAT,
could they!?) if the root cause comes down to miswiring in the original
install - sadly not unknown even (or expecially!) in new builds...

HTH, and let us know how you get on - Stefek

Alex Threlfall replied to Steve Reeves on 21 May 2004

as the other poster mentioned, if you don't know how to get it out, then
you don't know enough to be messing with it!

You'd need to unscrew the whole box iirc, and they then unscrew behind
there and unclip, but don't go messing in there as it's dangerous!

Lurch replied to Alex Threlfall on 21 May 2004

Nearly, the new Starbreakers are plug in MCBs so no screwing involved,
apart from the cover. Are you sure you should be offering advice! ;-)

Lurch replied to Lurch on 21 May 2004

Could be, without a bit of testing I couldn't say for certain.

And switching everything unneccesary off, that's about as far as you
can go before calling someone in.

Depends what it is, sounds like your fault is somewhat intermittent.
The type any electrician dreads. It could be found in 1/2 an hour, it
could take hours\days. Until everything is tested we shall never know!

Something like that, I'll leave whoever it is you get in to explain
how he's using it and what it is he's looking for.

Alex Threlfall replied to Lurch on 21 May 2004

i did say iirc - if i recall correctly, i'm not familiar with
starbreakers anyway, but i've seen the boxes/breakers in catalogues so i
can see how they normally go together... ;)

Lurch replied to Alex Threlfall on 22 May 2004

It was tongue in cheek, I usually keep quiet if it looks like I might get it
wrong! Even that plan doesn't work most of the time!

BigWallop replied to Lurch on 21 May 2004

<<<snipped>>>

Sorry if this sounds cheeky but, If you've done some electrical work around the
house, could it be something you've changed yourself in the passed two or three
weeks ?

It would mean disconnecting each circuit supplied by this CU and testing each
one with meters to determine if it is showing fault or not. You can't really
use small DIY multi-meters for this type of thing. if the circuits themselves
are testing OK, then the next things to test are each individual appliance that
you leave connected when you're not home.

Something like a rodent attack on the insulation of cables somewhere can also be
the problem, because you say it is mostly happening when the house isn't
occupied, so rodents could be having a run around in the quiet house and chewing
your cables. This could also be the random part in the timing I suppose.

A few thing to have a look at would be the flexes from the fridge and or
freezer. Any other appliance that is left connected on the circuits should also
be checked for damage to the flex or connection points.

It could also be spiking currents from the mains side of the supply that are
causing your unit to trip out, so it doesn't have to be something on or in your
house that's causing this to happen. The only way to find out if this is the
cause, is to have the supply company fit a recording meter on the system for a
week or so, to see if it picks up on surges in the supply lines.

There are a few other situations that could be causing this to happen, so it's
best to start with your supply company and ask them to check their side of
things, then you can start to eliminate from there.

dave replied to Lurch on 23 May 2004

Well we will know where to come if someone on here starts building a ROV
in the back garden and gets in a fix <G>.

Dave

Steve Reeves replied to Steve Reeves on 21 May 2004

:) Hi guys, thanks for the information. I do have "some" clue about
household electrics but I've never touched a distribution board before.
I don't plan to go messing anywhere that needs a professional!

I had done some research (via google etc.) and had gotten the impression
that the RCDs can get hypersensitive and just wanted (if a safe and
simple job) to see if I could change the trip switch to see if it cures
it. If it didn't then clearly I have a fault somewhere. Clearly I
shouldn't mess!

It's a strange one - only started a couple of weeks back and appears to
happen once a day at random times - particularly when there is no one in
the house. Maybe it's a fridge motor or something kicking in.

Clealry I want to see if I can find an obvious culprit before I spend
money on getting an electrician in - so I guess it's a process of
elimination job - unplug a load of stuff and see what happens. Any
better ideas?

If I did get an electrician in, how easy it is for them to trace these
kind of faults? Used to be a test engineer (marine umbilicals) and we
had Time Delay Reflectometers to test for shielding shorts and stuff.
None of it was AC though - and it was all over much longer cable
distances than used in the home. Does the modern electrician have some
magic gizmo like that?

Thanks for your replies - I would be grateful for any suggestions.

Lurch replied to Steve Reeves on 21 May 2004

There is no 'RCD fuse, it's one or the other. In this case it's an
RCD.
It's not a fuse, and the scenario you describe is highly unlikely.
More likely to be a fault with an appliance somewhere.

Leave it alone.

They aren't fuses, they're not meant to pull out.

Not meaning to sound rude, but you seem to be clueless as far as
electrics go. I'd get someone in who know's what they're doing if I
were you. With the RCD covering all of the circuits the fault could be
anywhere.

Steve Reeves replied to Steve Reeves on 25 May 2004

Hi Stefek,

Many thanks for your very detailed and helpful post. I will indeed let
you know how I get on. Currently eliminating items around the house in
an orderly fashion at the moment to see if I can find a pattern. So far
no joy - but early days. Will review your comments and improve tactics.

To complicate things (like I need that!!) today, again whilst we were
out, we had a power cut (RCD did not trip!) But was back on by the time
I got home.

You know what would be really handy, is an electric analogue clock
without battery backup) that would stop working when the power dies and
start again when it come back on without resetting. This would at least
give me some indication of:

a) When the power went (maybe be a pattern)
b) How long me sausages have spent defrosting in the freezer (to prevent
food poisoining!!).

You'd think there'd be something around an average house that could do
that - but nope everything is electronic and resets after a power
outage!!
Ah well....

Thanks again for your (and everybody elses help).

Pete C replied to Steve Reeves on 23 May 2004

Hi,

One way to help diagnose the problem would be to put a high value
variable resistor across live and earth, adjust it until the breaker
*almost* trips, then switch in the house wiring and plug in appliances
etc to see if they make a difference. This will tell how sensitive
your RCD is and give some idea how much earth leakage there is
elsewhere.

cheers,
Pete.

Jb replied to Pete C on 23 May 2004

Actually asking your supply to monitor the supply for a week is a good idea
and will determine if there are outside sources of interference, (street
lights can trip an RCD etc) and it's free.

Steve Reeves replied to Steve Reeves on 23 May 2004

Hi Z

:) Well I was using it as an example of a magic gizmo - but if you
insist - OK OK - I am CLUELESS!! :))

Good point, will check.

Nope.
Nope.
I'll double check the garage where the supply comes in but I don't think
so.
Nope.
Nope. And, you'll all be glad to hear - never any done my me!!
Nothing new certainly.

Ahh, godammit :) RCD tester is a good idea as other posters have
suggested.

Understood.

A company called Jacques Cable Systems (Littleport, Norfolk) - about 15
years ago. Not sure they still exist now - may have gone or been taken
over. We used to do marine cables, ROV umbilicals etc. for the likes of
the MOD and Snamprogetti (sp?) - An Italian outfit if I recall. We did
one HUGE umbilical - very long and wide for some offshore rig - treble
armoured - huge numbers of comms, power cables and hot/cold water hoses
- that one was a pain to test! Took weeks!

Thanks for your help.

Z replied to Steve Reeves on 23 May 2004

But suggesting a time delay reflectometer.

Have you de-frosted it recently, does it have defrost heaters? Check the
drip tray they often overflow onto the motor protector blocks.
Any water pen to the house or flat?
Heavy rains?
Rodent infestation?
Any joinery work done recently in the house?
Any electrical work done recently in the house?
New computer? PCs and other modern equipment use switched mode power
supplies (SMPSs) which often have high leakage currents.

You'd be cheaper buying an RCD tester than having an Electcian check all
your appliances. Unfortunately the RCD testers wouldn't be much use for
offshore earth monitoring systems for umbilicals like those made by
Bender.

After having confirmed the RCD is not tripping at too low a leakage
current and insulation resistance testing of all circuits protected by
the RCD they'd have to do the same as you - check suspect appliances
which can take hours, days for which they will gladly bill you but IME
not gladly do.

Who were you a test technician with - Simrad, Perry, Hydrovision?

LOL, very unlikely. Most electrical contracting companies do not even
have enough basic test gear (insulation resistance meters, loop testers,
rcd testers) to go round.
An ETDR, OTDR tester as you use in underwater umbilicals would be of
little use anyway as once one find the distance to the fault one would
have to find how the wiring is routed. If you know about domestic wiring
and you know about RF, which the TDRs use you will know a spur or switch
drop will act as a stub.

The test gear you need is

eyes for visual inspection, please do check your appliances.
RCD tester and with every load removed check the RCD does not trip at
too low a current or in too short a time.
Insulation resistance tester - with all loads removed check there is a
high IR between live conductor and earths.

Archived message: Tripping RCD on main disribution board - Help Requested (UK DIY Home Improvement)