You haven't thought it through. there is ALWAYS some neutral earth
leakage. It occurs every time an oh so rigorously RFI suppressed switch
mode power supply (TV, computer, monitor, LV lighting) or any suppresed
motor is plugged into the circuits. Enough of these and the 30mA RCD is
on a hair trigger. Any voltage surge and over she goes.
Here it was a toss up whether the RCD. MCB, or both would go.
Now with a 100mA RCD, its not gone since.
Whether or not I install RCBO's in the few circuits the regs demand
depends on the building inspector.
Proper standrads of wiring and better plugs, sockets, and particularly
the use of double insulation and plastic cases have done far far more to
reduce shock hazard than RCD's.
I can clearly remember getting shocks of metal fires, whose earthing
wwas poor or nonexistent, and whose rubber coated wires frayed all too
easily. Now we seldom have metal frames, and we don't have users
installing their own plugs nuch either.
I have just explained why.
Actually, I think its quite good, provided it trips only when there is a
significant dangerous fault.
Its the nusiance tripping that is so serious.
Where does this stray energy go?
All LV Transformers I have seen only have L&N, no earth, so where does it
What use is a 100mA RCD?
You die with a 40mA shock, that's why they make them 30mA
Just like a split load would, but it would only take out the faulty section,
not the whole lot.
(If it took out the lighing, then there are still table lamps!)
Well I have NEVER had one single nuisance trip on my lighting RCD, and it
has been in since the 1980's
It has never tripped when a bulb has blown - the only time it did trip was
when some dickhead drilled through a cable
(A builder, NOT me!!)
We did however have a fault develop with our "sockets" RCD, this was then
tested (along with the lighting one) and it was found to be defective (over
Mmm. Ok, I'll take out the LV transformers. Bercause thinking about
mine, they are alos two wrire devices. That is probably why they
interfere so badly with radios.
Well it trips if you get a neutral ertha short, or some other appliance
Whether you die at 40mA is not an established yes and no type thing.
Much less applied across the heart can stop it in a vulnerable subject:
Much more applied across just a finger, will burn, but not kill.
The problem with 30mA is its proneness to false triggering, especially
applied to a whole house.
The other issue is WHAT should be protected. Arguably the most dangerous
circuit in the house is lighting, where you have potentially two user
accesible terminals, and a (now rigorously earthed) metal housing in
close proximity. Grab the housing with one hand, and stck yer other
finger in the socket and Whazza! you are across the floor...
Not in my house. They have theor own lighting sockets on the relevant
I like the idea of split loads, but Peter has made me rethink all of this.
Frankly, I cannot see an ideal solution.
I am very happy that the discussion continue - if a general consensus of
best practice CAN be established, then a fait accompli could be
presented to those who dream up building regs.
Firstly, it usually takes about 200mA to be killed. Secondly, a 100mA RCD is
not for supplementary direct protection. i.e. it is not intended to prevent
electrocution when you touch the live. Its purpose is to rapidly disconnect
the supply in the event of a minor earth fault, rather than relying on there
being a massive earth fault. Circuits likely to require supplementary direct
protection should use dedicated 30mA RCBOs for that circuit only. Whole
house RCDs (except 100mA Type S) and even split load consumer units are a
Mostly - increasingly switched mode PSU's require an earth (and noise
filter) to meet the EMC directive. There is some discussion going on
now as to whether house circuits should be classified as high
protective conductor current circuits (which would effectively stop
30mA RCD's being fitted to these circuits).
No you don't, it's nothing like that simple. The chances of death
from a brief shock caused by touching a conductor carrying below
several thousand volts is quite low. The figures vary depending upon
the conduction path - the most dangerous being through your left arm
to the feet or right arm (hence the advice to keep your left hand in
your pocket when working on live circuits). Whether you grasp the
conductor is also critically important, unless you grasp the
conductor 240V is rarely fatal (always use the back of your hand for
first contact if you really have to touch something which may be
live). Even small currents cause the hand muscles to contract
preventing you from letting go.
There are widely varying figures depending upon where you look for
the simple reason that not a lot of practical work has actually been
carried out on the subject. For some reason they couldn't get many
volunteers. Most of the levels are therefore extrapolations from the
results seen at lower levels.
Critically important is the "Let go" current, the current above which
you cannot release your grasp upon a live conductor. 99% of the
female population have a let go limit above 6 mA, with an average
of 10.5 mA. 99% of the male population have a let go limit above
9mA, with an average of 15.5mA.
Prolonged exposure to 50 Hz currents greater than 20mA across the
chest causes the diaphragm to contract which prevents breathing and
causes the victim to suffocate so it is quite possible (albeit
improbable) to die while your RCD stays firmly in place.
The more immediate cardiac effects depend upon the frequency of the
electrical current as well as its magnitude. Humans are most
susceptible to frequencies at 50 to 60Hz as the internal frequency of
the nerve signals controlling the heart is approximately 60 hertz.
Ventricular fibrillation occurs when 50/60Hz current from the
electric shock interferes with the natural rhythm of the heart. At
100mA the current would need to be maintained for about 3 seconds to
start fibrillation, at 900mA it would take only 3 milliseconds. The
heart loses its ability to pump and death quickly follows.
Ventricular fibrillation can occur at current levels as low as 30 mA
for a two year old child and 60 mA for adults. Most adults will go
into ventricular fibrillation at hand to hand currents of about 100 -
Humans are able to withstand 10 times more current at DC and at 1kHz
hertz than at 50 or 60 Hz. Interestingly the most dangerous current
range is somewhere between 200mA and 4A which causes the heart to
fibrillate which cannot be stopped by first aid methods. Above that
current the heart is paralysed rather than going into fibrillation
and simply stops - it can sometimes be restarted with a blow to the
The following figures are taken from NASA, Stanford and the Royal
Current Physiological reaction
<1 mA None
1 mA Perception threshold
1-3 mA Mild sensation
3-10 mA Painful sensation
10 mA Paralysis threshold of arms
30 mA Respiratory paralysis
75 mA Fibrillation threshold (0.5%)
250 mA Fibrillation threshold (99.5%)
4 A Heart paralysis threshold
>5 A Tissue burning
Sub lethal electrical shocks can often cause permanent nerve damage
which only becomes apparent weeks or months later when the wastage of
muscle tissue is noticed.
50-150 Milliamperes - Extreme pain, respiratory arrest, severe
muscular contractions. Individual cannot let go. Death is possible.
0.2-4A - Ventricular fibrillation. (The rhythmic pumping action of
the heart ceases.) Muscular contraction and nerve damage occur. Death
is most likely.
10A - Cardiac arrest, severe burns and probable death.
Under 1 mA - Not perceptible
1 to 8 mA - Mild sensation
8 to 15 mA - Painful shock but person still has control of muscles
and can let go of source of shock
15 to 20 mA - Painful shock. Unable to let go
20 to 50 mA - Severe shock. Severe muscular contractions. Breathing
50 to 100 mA - Extreme shock. Extreme breathing difficulties
100 to 200 mA - Death by Ventricular Fibrillation. Not reversible
200 mA up - Severe burns. Breathing stops. Chest muscles clamp
and stop the heart but no ventricular fibrillation. Victim may
survive if given immediate resuscitation.
Assuming the skin (which is quite a fair insulator)is intact typical
resistance levels for an adult are:-
Dry skin 100k to 600k Ohms
Wet skin - 500 to 1k Ohms
Internal (broken skin) hand to foot - 400 to 600 Ohms, Ear to ear
Unless you have wet hands it much more difficult than you might think
to dangerously electrocute yourself with 240VAC. Certainly in your
changing a light bulb scenario it would be most improbable.
A 30mA RCD will pass its test if it trips at 16mA.
But it is all a lot more clear now Peter.
Could you possibly tie in with our FAQ chief to lodge the essence of
what has really been a multi-stage tutorial?
It would be much more readily accessible to future enquirers in the FAQ,
than Googling thro a 42 element thread.
How do you feel about 100mA rcd protection of lighting circuits as a
I have a special interest in this, having a house where rodents have
been chewing at the wiring.....
Good idea about the FAQ.
you mention fire prevention & 100ma RCDs
In the US they tackled this a different way with
AFCIs - Arc fault interuptors
see here http://members.tripod.com/~masterslic/afci.html
Should we go this way too ?
Of course there can be, and is. You are forgetting what happens when
a bulb goes and the effects of inductance. You are also forgetting
the effect of sensitisation by unbalanced noise filters. Indeed it
is questionable if simple 30mA protected power circuits have any
future because of this alone.
Because their most common failure mode is fail on power on and they
have poor reliability.
CFL, permanent low level lighting, whatever suits the need.
Any RCD on the lighting circuit is bad, and it is that which you do
appear to be suggesting.
This is not my experience either.
Can you propose a mechanism for this? The only one which comes to mind
would be that you already had a neutral-earth leakage fault, and the rise in
neutral current caused by the bulb-failure arc was sufficient to create a
leak to earth.
It wouldn't matter what kind of lamp illuminated the stairs if the lighting
circuit had tripped.
Yes, but you're a (former?) member of the emergency services, aren't you?
As such, I have every sympathy with the gruesome sights you might have had
to deal with in your chosen profession, but I do believe that members of
such services are uniquely *ill* placed to give general advice about risks
to ordinary people.
To a fireman, a horrible death in a house fire is a common occurence - to
me, as a resident of a non-smoking household, it's extremely unlikely to be
the cause of my, or my family's demise. Personally, if I'm doing the
neurotic parent bit I worry about road safety a darn sight more.
TBH, if you know of a property that has an RCD which trips when light bulbs
blow, then you ought to advice the occupants to get the wiring fixed.
There is always some leakage from both L & N to earth. If nothing else
due to the capacitance between the miles of T & E laid over the house.
And more especially in all the capacitors used to RFI filter electronic
There is always some inductance in the supply, due to cable lengths
and transformer secondaries. High speed switching currents will
work with the capacitances to flip RCDs
You don't need resistive leakage to trip an RCD and the problems of
pre-sensitisation are become more acute as more equipment includes
Indeed not, that is why you chose a lighting type of high reliability
whose failure mode is such that it would not trip protective
circuits. Incandescent bulbs have poor reliability and less than
optimal failure characteristics.
Why? Because they understand the balance of risks better?
Actually it isn't.
Quite reasonable, but risk assessment is about how much work you have
to put into circumventing a risk. Removing (or not fitting) an RCD
on a lighting circuit is trivial and although the risk of fire is
small the consequences are devastating and I suggest the minimal
effort is worth it.
What's to fix? I can set up a demonstration to prove an RCD will
trip on a bulb blowing on a perfectly serviceable circuit with ease.
The point is that an RCD on a lighting circuit does no good at all
but potentially may be very harmful, a situation recognised in the
wiring regulations. What's the agrement for fitting them?
I dont know enough about this to comment, only that as I have said before,
in my installation (With incadecent, CFL and LV switchmodes) I have never
the RCD trip on a bulb failure (Including with 500w halogens) - So in my
I really dont see the added risk. please explain why in my setup it wouldnt
Sorry but I totally disagree with that statment.
It will stop people getting killed by electrocution.
Please explain, in a property where there are never any nucence tips, whay
is it bad?
I'm quite prepared to be shot down in flames
I operate a wet area supplied via 30ma RCDs
..it has 3kw of halogen lighting - failure mode is
water spray on tubes.
it also has two water pumps
many fractional HP motors
a large compressor
several EMI filters
many 240v 3 port valves which are damp
and the only RCD trip we ever had was from
a cable rupture in a pudle of water.
MCBs pop often with incandescant failure
but never the RCD
(it's NICIEC tested regularly)
People don't get killed by electrocution from lighting circuits.
They may sometimes get a belt from them (which an RCD will reduce but
won't prevent) but they do not cause death or injury.
Because in a fire the lights all go out as the RCD trips.
Electrocution in the home kills 25 people a year and causes 2000
minor and serious injuries. I only have access to detailed figures
for 1998 and in that year none of the 23 deaths by electrocution
involved lighting circuits - all bar 6 were people dismantling live
equipment, the remaining 6 were handling faulty supply cords.
In one year 500 people die and 18,000 are seriously injured in fires
in the home.
That's why the wiring regulations don't support fitting of RCDs to
JOOI, can you describe the circuit? As I say, I have many years experience
of non-nuisance-tripping 30mA whole-house RCDs, and always with a room full
of computer equipment somewhere on the circit.
I'm not particularly arguing for the fitting of them against the
regulations, only against the received wisdom that A. they'll give you lots
of nuisance tripping, and B. that they're the major cause of nuisance
tripping on lighting circuits. (Which we can all agree is a bad thing.)
However, my experience of whole-house RCDs is that they've given warning of
incipient problems, both with wiring and with appliances. For a start, they
detect neutral-earth failures (which one would guess represent about 50% of
insulation problems) which are not typically detected by overcurrent
protection at all. They also give early warning of thinks like heater
sheath failure in immersion heaters and electric showers - I don't consider
myself odd in thinking that I'd like to know about an immersed heater
failure while there's 50mA flowing to earth rather than 50A.
And, like I say, I know someone personally who had a bad shock off a fixed
light fitting, and a teenage girl in the village I went to school in was
killed by an electric shock from a light fitting - her parents returned to
the house after a holiday to find her dead on the living room floor.
I'm afraid I interpret your gruesome anecdote as a warning about the way one
locks ones house at night (not being clear that it has any link to lighting
circuit RCDs, anyway), and mine as warnings about earth protection on wiring
. No doubt we both feel rather closer to our own cautionary tales.
Well said Peter. You are probably a voice in the wilderness, like that
chief inspector who categorically stated that 'excess speed is the cause
of less than 7% of accidents' and 'speed cameras do not (statistically)
reduce road deaths'.
I am definitely with you on this. My 30mA RCD tripped about 50% of the
time when any bulb blew, with the MCB about 60%. Sometimes both.
Its bloody dangerous feeling down te saircase to get to the consumer
unit, especialy if you have left the casing off...:-)
Peter Parry wrote:
Reading this, and especially the very sad ending to Peter's post makes
me think that it would not be a bad idea to add emergency lighting as
something that should be included in Building Regulations for new
properties and incentivised for others in some way, especially for the
elderly and others with mobility difficulties..
We already have mains powered smoke detectors as an addition,
basically to get over people taking the batteries out or forgetting to
Non-maintained small fluorescent lights are not very expensive,
although the fittings are pretty ugly. Undoubtedly this could be
improved if there were a consumer premises market for them.
It would seem to me that the populace would be better served with a
lighting measure like this rather than the mandatory low energy lamps
stuff. In fact with a bit of creativity, both objectives could be
met in a single package.
I am not sure what sort of size and power level that they would need
to be to provide sufficient light say at the top and bottom of a
staircase in a fire, but to be able to see adequately during a power
cut, then the little 8w ones are probably enough.
I have my consumer unit at the back of a cupboard in the kitchen and
it would not be easy to see in the dark.
I fitted a small maintained fitting, run from a (non-RCD ptotected)
lighting circuit and operated also by a plunger switch on the cupboard
door. Thus I have a light over the CU for when I want to work on it
with the power deliberately off and also to be able to see the
breakers after a trip.
To email, substitute .nospam with .gl
HomeOwnersHub.com is a website for homeowners and building and maintenance pros. It 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.