Talk:Residual-current device

From Wikipedia, the free encyclopedia
Jump to navigation Jump to search

How is the calculation done[edit]

How is the difference in the currents calculated. If we have a resistor and some simple electrical circuit the current is maximum in the input (something like 220 volts / 16 ampers) but when it gets to the circuit it is lowered to lets say 1 amper (16 ohms resistance) and on the neutral wire we get a maximum of 1 amper current. So how do we calculate 16 and 1 ampers of currents to match zero-0?! My point is the current in the RCD measuring the live/phase(L) wire will always measure a big current no matter how much amperes the electrical circuits in a house use because they have resistance which the RCD doesn't know and the current drop happens later in the electrical devices in the house due to lower voltage/current consumption.--Leonardo Da Vinci (talk) 09:25, 1 March 2011 (UTC)

RCDs operate by measuring the current balance between two conductors using a differential current transformer. This measures the difference between the current flowing out the live conductor and that returning through the neutral conductor. If these do not sum to zero, there is a leakage of current to somewhere else (to earth/ground, or to another circuit), and the device will open its contacts.

An RCD or GFI doesn't "calculate" anything, it only responds to the imbalance between live and neutral conductors. --Wtshymanski (talk) 14:16, 1 March 2011 (UTC)

I've read that but still the current will always be different in the Line and Neutral wires. Having the same current means the two wires are shorted... It seems the answer to my misunderstanding lies in "differential current transformer" and the method of calculation. Could it be that the current which is coming with the 220/110 volts be the needed by the consumption (computer, tv, freezer etc.)? And no matter the ammount consumption the returning current is the same as the incoming current?--Leonardo Da Vinci (talk) 14:40, 1 March 2011 (UTC)

Why would you expect the current going out on one wire to be different from the current coming back on the return wire? It's got nothing to do with the voltage, or the type of's more a principle of continuity. --Wtshymanski (talk) 14:57, 1 March 2011 (UTC)

In normal mode all circuits working by first Kirchoff's law: how much amperes go to the load by hot wire, the same amperes return back by neutral wire. But if you touch hot wire by one hand and metal pipe by other hand (don't do it!), hot wire will carry 20...30mA more, than neutral. If we write it from Kirchoff law, it will be: I(hw) = I(nw) + I(gw); RCD/GFCI will disenergize circuit, as I(hw) ≠ I(nw). Work can be explained by Gauss theoremas, when I(hw) = I(nw), magnetic fields will equate each other; when some curren leaks to ground wire, magnetic fields won't equate each other and RCD/GFCI trips (talk) 21:18, 22 March 2011 (UTC)

For example, your washing machine consumps 10 Amperes. In normal mode 10A goes to washing machine by hot wire and 10A goes back. Some nanoamperes leaks through insulation, but it is not danger for human. RCD/GFCI working normally. By the years, rust made hole in storage tank and water begin to flood energized wires during washing. Circuit breaker will not trip, cause short-circuit current value is not enough to beat out usual circuit breaker. If grounding electrodes are missing or far from sewerage pipes, taking shower or bathing might be lethal during using faulty washing machine. But if you have RCD/GFCI, current transformer will feel, that current in hot wire is some more, than current in neutral - that way magnetic field from hot wire will overbalance magnetic field from neutral wire and RCD/GFCI will trip. Widespreadly, those devices have diode inside and when leakage current will overlimit safe value, diode "open" relay's circuit and relay separate contacts between power source and load sides. As the contacts are open, current doesn't gain and protectable circuit is dead.

RCD/GFCI will be useless, if you ground metal covers to neutral; neutral should be separated to ground and working neutral in main panel or in distribution substation. Also, RCD/GFCI will protect if ground wire is broken, but only after someone close circuit to grounded area (it mean shock for short time).

Difference between RCD and GFCI:

  • in Europe I can use one RCD to protect as much as I want circuits (electrical codes regulates less than 8 circuits to be protected by one RCD, cause each circuit normally gives 3...5 mA to earth due to dust)
  • in US and Canada I should waste lot of money, cause one GFCI can protect only one circuit (and I need to find GFCI with proper values, cause I can't use circuit breaker separately from GFCI).

Better to use GFCI sockets in US, but it will also take lot of money and I can't protect 4-5 circuits by one GFCI. Or install several subpanels additionaly to main panel, but it will also waste lot of money. Well, in America it is unfortunate engineering solution :-) (talk) 22:06, 22 March 2011 (UTC)

Well, on this continent, we enjoy the convenience of not having all the circuits in the house go dark if there's a GFCI trip. Not a problem if you're living in Soviet-style worker's housing with only one plug for the whole apartment, I suppose. Plus the average GFCI receptacle costs about an hour's wages - or less. One GFCI receptacle can be used to protect a number of donwstream receptacles, if you're building a summer cottage or a doghouse or something and want to save $8 or so. And if you need to protect a permanently-wired apparatus, you can of course buy a GFCI circuit breaker. There's even a thing called an AFCI. --Wtshymanski (talk) 01:14, 23 March 2011 (UTC)
I suppose, you'll not protect ceiling lights by GFCI in dry rooms; even ground wire does nothing on ceiling despite inspectors require it. But Americans use very powerful devices - distributing 15A to 5 receptacles from junction boxes won't give desired result. In old Soviet apartment houses (till 1960's) main fuse was 6A to one apartment and all thick aluminium circuits were energized from one junction box. It was enough those time, but lot of times homeowners were scared by arc when they were need to replace main fuse or disenergize the circuits. Only in the end of 1960's there begin to appear high-rise fully-concrete apartment blocks with electric stoves and energy-waste elevators in Soviet Union. There were one 25A circuit breaker for stove and two 16A circuit breakers for circuits in rooms (chandeliers and receptacles were energized from one breaker, but junction boxes have been hidden in concrete).
Nowadays I install subpanel inside apartment with 8-20 independent circuits and remove group circuit breakers from panel in stairs. After such reconstruction, rusty panel in stairs contains only main breaker and electricity meter, all group breakers located inside apartment. If apartment contains few circuits, there is enough 2 or 3 RCDs. But if apartment is very big, there lives big family and 20 circuits are engaged (2 or 3 circuits on ceiling for chandeliers) - nobody will agree to buy 17-18 RCDs and spend 480 EUR only for RCDs. In such case I install one RCBO for stove, one RCBO for washing machine, one RCBO for refrigerator - and one RCD to all circuit breakers for kitchen, one RCD to all circuit breakers for room one, one RCD to all circuit breakers for room two, one RCD to all circuit breakers for room three and so one. Such way all receptacles in certain room will be disenergized in case of leakage to ground, but economically available for homeowners. (talk) 20:34, 23 March 2011 (UTC)
6 amps for an apartment? Even at 240 volts, that's not a lot (not enough for electric cooking on any significant scale, I expect). Surely that must be very old construction, dating back to the 1920's or something - I've seen (Manitoba) farmhouses from 100 years ago that had more like a 30 amp or 60 amp (120/240) service and I don't know if 6 amps would ever have been acceptable for residences built to US or Canadian code.
I don't see what difference you're trying to show us; in US and Canadian electrical codes, we don't get to pick and choose which circuits get GFCI or AFCI protection, that's set by regulation. For example, my rather small house has a 42 pole panelboard (only about 20 poles installed, though; 240 volt appliances get 2-pole circruit breakers), 200 ampere main circuit breaker, with no GFCI circuit breakers mounted and only one AFCI panel mounted circuit breaker for the bedroom receptacles; there are GFCI receptacles for bathroom, kitchen and outdoor receptacles. So my house has perhaps 5 or 6 GFCI receptacles protecting a dozen or so receptacles, and one AFCI circuit breaker protecting a few bedroom receptacles - not greatly different than what you describe.
Why would you have an RCBO for the refrigerator? Do European 'fridges short out that often that tne need additional protection? One false trip and you've ruined a lot of food, plus possibly never being able to get the smell out of the 'fridge; N. American codes don't require GFCIs on refrigerator receptacles. --Wtshymanski (talk) 21:43, 23 March 2011 (UTC)
Few years ago code required RCD only to stove and bathroom circuits, nowadays code reqires all receptacles been protected by RCD. But not of all homeowners have money to buy 20 or 30 or more RCDs, that is why one RCD protecting approximately 5 circuits.
As for energy consumption, old Soviet apartment houses widely used gas for cooking and heating. Electricity was used only for chandeliers and watching TV or hearing radio. Only in 1970 gas was forbidden and replaced by electric stoves. As for heating, pump-heating stations were more widespread, than independent heating (even nowadays). Nowadays there appeared electric kettles and washing machines, but 30-years-old thick aluminium wiring can't handle such loads. Private houses have 3-phase power, the same about big apartments. If homeowner wants to use electric stove or sauna in his apartment in new house, we give him 3-phases, like in private house. As for small apartments, main breakers (230V) are 25A without stove and 32A or 40A with stove and it is enough.
But as the main problem is missing of normal ground. 30 years ago stoves reqiured to be grounded to neutral wire (and aluminium wires weren't fat those time) - nowadays lot of people use powerful kettles and washing machines with the same old wiring - lot of times neutral wires burned out from main busbar in main distribution board in basement and all apartments had lethal voltage on "grounded" applieneces. And even neutral wires weren't normally grounded, sometimes it was usual neutralling instead of grounding.
As for refrigerators, it is not good idea, when food is decomposing with related smell cause somewhere (like in holey kettle) was leakage into ground wire. And RCD with cicuit breaker will occupy much space in fuse box, that is why I install RCBO there. But in American fuse boxes I have no choice cause of constantly attaced busbars, even I can't attach relays there if I want to use some automatics (like smart ventilation or time-controlled outdoor lamps). (talk) 08:37, 24 March 2011 (UTC)

Interesting. The induction cooker we put in last year required a 40 amp breaker all by itself. We have a lot more circuits in a home today than 70 years ago. Generally home automation goes downstream of the panelboard, although there are some manufacturers who make (expensive) controllable circuit breakers; this is kind of off-topic for this article. --Wtshymanski (talk) 13:43, 24 March 2011 (UTC)

I suppose, induction stove is connected between 2 hot wires on 220V and it gives 8800W. In Europe 13A circuit breaker will be enough for the same power stoves. But European stove manufacturers usually produce stoves for 16A breakers (in America it will be 3x16x230/220=50A). As for amperage question, Americans use more electric power; Soviet Union used gas power, electric stoves and electric boilers appeared there only in 1970's. Then imagine - I should find 40A RCBO with calibrated leakage. What will be, if the same RCBO is missing in store? But in Europe I can choose: RCBO or RCD + circuit breaker. In America I have only GFCI, even if it is missing in store, I can't find replacing to it. (talk) 20:13, 24 March 2011 (UTC)
Not quite. Even with 3 phases, 13 amperes x 1.732 * 240 V = 5400 watts, and an induction cooktop usually runs around 7500 watts. Preferred sizes in North American practice are 30 amp or 40 amp circuit breakers, and 7500 watts is too much to rely on a 30 amp circuit breaker, so 40 is the next choice. We have similar options in North America, you can buy a (moderately expensive) GFCI or AFCI circuit breaker for the panelboard, or, for outlets, you can buy a GFCI receptacle that can be used to protect downstream receptacles and which costs less than the electrician's labor to install it. I don't know why European wiring devices seem so costly - this must make for great compromises as you describe, where you have many rooms tied to one device. I suppose it's not a problem as long as faults are rare, but you'd get tired of all the power going off every time you plugged in a vacuum cleaner. --Wtshymanski (talk) 21:47, 24 March 2011 (UTC)
You're wrong - 240 x 1.732 is voltage between hot wires, but each hot wire to neutral gives 240V.
European 3-phase explanation.jpg

Correct way is:

  • 1) 240V x 3 phases x 13A = 9360W if the load is connected by star (hot to neutral)
  • 2) 240V x 1.732 x 3 phases x 13A = 16211W if the load is connected by triangle (neutral is not used)
As for receptacles with built-in RCD, they were produced during 1990's in Norway, but nowadays they are missing in store (probably, they are not produced nowadays). As they are missing, nobody can buy them. You can find only power strips with built-in RCD, which are 4.5 times cheaper than RCD for fuse box, but inspectors don't accept them cause they can't be constantly installed.
Faults are rare, cause European RCD have leakage 30mA and more (16mA goes only to electric boilers), lot of appliences have dielectic cover and rules regulate to use 12 Volts in wet rooms. Appliences with metal cover in Europe you can find only in kitchen. But american GFCI have leakage <10mA, it's calibrated by minimum and using power strip on protected circuit will be accompanied by often GFCI faults (talk) 08:44, 25 March 2011 (UTC)

Your calculation is wrong too. You cant pull more power through the same circuit breakers if you use the delta instead of the star configuration.

Correct is 240V * 1,732 * 1,732 * 13A = 9360W = 240V * 3 phases * 13A
When you connect a device between two phases and pull 13A (240V * sqrt(3) * 13A) two of the three circuit breakers or two thirds of the 3-phase-MCB are already at their limit. You then can't do the same on the two remaining 2-phase combinations. --2003:D4:8BD0:ED00:114:CD16:E940:937 (talk) 19:07, 12 December 2018 (UTC)

Thanks for the explanations guys :). Really helpful. My problem in the first place was because of Ohm's Law. You have a neutral wire on which 10 amperes pass for example but the voltage is lower than on the hot wire. And this is my confusion. How is I=U/R accomplished (for the neutral wire current) if we can change the voltage with different resistors in the consumption circuits any time :)? — Preceding unsigned comment added by Leonardo Da Vinci (talkcontribs) 19:27, 25 March 2011 (UTC)

Current goes from biggest voltage value to lowest. It is taught in the school, so remind physics to yourself. I can touch grounded neutral wire barehandly and nothing will happen despite 10A are flowing through neutral wire. But if electric installation hasn't earthing, I can get lehtal shock touching ground wire, cause voltage will "jump" in neutral wire and connected to it ground wire in main panel will also have "jumping" voltage. But when ground and neutral wires are grounded near building, voltage between ground wire and ground will be near 0V, which is safe for people; the same about neutral wire.
You can do next experiment - make circuits (use extra low voltage):
*1) torch battery "+" → ampermeter → torch light bulb → torch battery "-"
*2) torch battery "+" → torch light bulb → ampermeter → torch battery "-"
See ampermeter's statements in both events and make correct conlusion ;-) (talk) 21:54, 25 March 2011 (UTC)


From the article:

The terms ground fault circuit interrupter (GFCI) or ground fault interrupter (GFI) are used in the United States and Canada, but these terms are not, strictly speaking, correct (or at least complete), because a RCD will trip if current leaks anywhere, not just to ground.

In NEC terminology, power flows between the "ungrounded conductor" and the "grounded conductor" (with the load in the middle) . If more power is flowing through the ungrounded conductor then the grounded conductor (unbalanced), then there is a ground fault (a failure to ground properly), and so the circuit should be interrupted. Looking at it this way, the terms GFCI and GFI make more sense, and is how I always assumed they came about.

Notes: The "ungrounded conductor" is colloquially called "hot" or "line". The "grounded conductor" is also called the "identified conductor", and is colloquially (and sometimes incorrectly) called "neutral". Note that none of these are the "equipment grounding conductor", which is colloquially called "safety ground" or just "ground". Yes, NEC terminology is confusing.

--DragonHawk 15:23, 19 Jun 2005 (UTC)

But ground isn't the only place the errant current can go; it might return on the other hot leg of a split-phase system, the neutral leg of another branch circuit, or it might even leak unto a de-energized part of another circuit (the switched leg going to a lamp or a circuit de-energized by opening its breaker). I think it was this thought that provoked the original author to speak of "ground fault" being something of a misnomer. And it's certainly true that the devices sense "balanced current" and not a "ground fault" in particular. Compare this with certain ungrounded systems (for example, railway traction motor circuits) where there ctually are devices that are purely sensitive to ground faults; any grounding of the ungrounded system trips the detectors and removes power. Such devices would be properly called GFCIs.
Atlant 11:55, 20 Jun 2005 (UTC)

Yes, and also in many renewable energy systems, some of them are wired up in a balanced way with neither of the two wires connected to ground, in which case, of course, there must be a breaker on both sides. In this case there is no ground, and no concept of ground fault. Thus what's really being measured and detected is a balance fault, i.e. basically what would be a violation of Kirchoff's current law, were it not for the unknown existence of some other leakage current somewhere else (not necessarily to or from ground). Also the concept of a balance fault nicely generalizes to three phase, where, for example, a three-phase BFI could trip if there was any imbalance of total current (i.e. if the total current didn't sum to exactly zero). Glogger 23:02, 30 December 2005 (UTC)

I have (re-?) inserted a brief explanation that GFCI is the usual name in the U.S. and Canada, which it certainly is, regardless of arguments as to whether the term is strictly defensible. On a related note, is "Balance Fault Interrupter" used anywhere? If it does appear in documentation or industry literature, a cite is in order; if it's new coinage proposed by a Wikipedia contributor, it really doesn't belong. Sharkford 19:27, 1 February 2006 (UTC)

Strictly speaking, a GFCI is different from a RCD in that a GFCI will trip in the presence of a neutral-earth fault (even though there is no voltage difference). It does this by having a second coil around the supply conductors (as well as the sensor coil). This provides a positive feedback loop when the neutral-earth circuit is made (ground-fault) and the GFCI trips. A RCD will not do this.-- 23:14, 17 October 2006 (UTC)


From the article:

Manufacturers are now designing such devices that will disconnect their circuit, even if they don't dectect electrical leaking, if they are wired improperly.

Really? How? Under all possible miswirings, or only some? -- The Anome 08:50, 26 May 2004 (UTC)

I don't understand this either. It's just a guess on my part, but since the most common wiring error is reversal of line and load connections on receptacle devices, the manufacturers may be doing something to detect that particular condition and trip. UninvitedCompany 14:44, 26 May 2004 (UTC)
Some types of RCD incorporate an additional device which trips if the voltage between the earth (ground) and neutral conductors rises above a certain potential. This would also detect a miswiring where the live (hot) and neutral conductors were reversed. 13:20, 17 Jul 2004 (UTC)

mixing mains and signal wiring in us[edit]

I'm concerned by the comments at the bottom that suggest that, in the US, you can intermix line voltage wiring with Ethernet, etc. I'm not aware of *ANY* US code-approved wiring solution that allows that, although I suppose a partitioned mounting box might allow the use of a single Decora faceplate with, say, line voltage components on one side and low-voltage components on the other side.

Any NEC experts here to say yea or nay?

Atlant 18:25, 13 Dec 2004 (UTC)

Hubbell makes a variety of products that put ethernet, phone, and power into the same faceplate. Of course it's nicely designed to keep things separated by metal dividers. From an aesthetic and design point of view, it's fantastic, the way they got everything lined up, so when you install all the sockets you have "all your ducks in a row" so to speak. Glogger 23:07, 30 December 2005 (UTC)

Yes, the low-voltage stuff is either on the other side of a divider or, commonly, outside the box altogether, but sharing a common (large) faceplate. I have edited this section and simply deleted the refs to low-voltage wiring; it's not particularly common and has no special relevance to this article. Sharkford 19:12, 1 February 2006 (UTC)

RCD vs RCCB[edit]

whilst the term RCCB is more descriptive it seems to have fallen out of favor at least here in the UK. Nearly every manufacturer and most standards and guides use the term RCD nowadays. Should we change the article to reflect this (if we do we should probablly also do a pagemove). Plugwash 01:51, 30 Mar 2005 (UTC)

My taste would be to:
  1. Create a redirect from RCD to RCCB
  2. Modify the lede of the article to mention the new term and how it is replacing the old term
But I don't feel strongly about this; if you'd rather move the article, flip all the references from RCCB to RCD, and add a bit to the lede describing the now-archaic ( :-) ) term RCCB, feel free! But don't forget to clean-up all the resulting double-redirects of RCCB that may exist out in Wikispace. (It's for this reason that I generally don't rush to rename an article whose name is still within the bounds of "accurate enough".)
I assume that the lede also mentions GFI, which is the universal, casual term in the United States. If it doesn't, I'll edit it in right now. (It's okay; GFI and GFCI are both there in the second 'graph.)
Atlant 12:16, 30 Mar 2005 (UTC)

Arc-Fault Interrupter[edit]

Here in the U.S., we're starting to see circuit breakers that are able to sense arcing current. (I don't know exactly how they do this, but I'd assume it depends on sensing high-frequency noise in the mains current waveform.) I think this capability is often combined with Balanced-Current fault detection.

Does anyone feel ambitious enough to investigate this and either 1) create a AFI redirect and modify this article to include the concept or 2) create a free-standing article with appropriate cross-references to here and back?

Atlant 12:18, 30 Mar 2005 (UTC)

Aha! I found the pre-existing article and linked it as a "See also".

Atlant 19:03, 30 Mar 2005 (UTC)

Actually, the North America section of this main article is a mess, confusing Arc Fault Circuit Interrupters (AFCI's) with Ground Fault Circuit Interrupters (GFCI's), and also whether Tamper Resistant (TR) outlets are required. They are not the same and they do not exhibit behavior. For example, and AFCI does not normally provide for Ground Fault protection. AFCI's are usually installed at the main panel, while GFCI's come in both outlet and breaker form factors. There are major cost implications involved here and just because the NEC (NFPA-70) recommends it does not make it the law of the land everywhere. So this article needs work to manage the mass confusion. -- (talk) 02:15, 2 January 2015 (UTC)

How do RCDs actually trip?[edit]

Using the example of the illustrated "outlet-style" RCD, the description used to speak about the solenoid being continuously energized and that providing the latch that kept the RCD "reset". This description is clearly inconsistent with at least US-style RCDs as it would imply that a power failure would also lead to the dropping-out of all of the RCD solenoids, their subsequent tripping, and the need to manually reset all the RCDs after power was restored. This ain't the way it works in practice.

I conclude from this that the mechanism operates in the opposite fashion: The latch is mechanically held and the RCD activates the solenoid when it wants to "trip" the RCD by unlatching the latch. I've changed the text to reflect this view of the world. But if Europe has "outlet-style" RCDS and they operate as originally described, please feel free to extend the text to explain this.

Meanwhile, I guess I'll take an American RCD apart this week-end and check for certain. :-)

Atlant 19:27, 3 Jun 2005 (UTC)

In Europe you can get both passive rcds (which don't trip out on power failure) and active rcds (which do). Generally most rcds in consumer units etc are the passive type but the active type is considered advisible for supplying equipment which can pose a danger on unexpected re-energistation. Plugwash 20:57, 3 Jun 2005 (UTC)

Thanks! Atlant 23:22, 3 Jun 2005 (UTC)

Leakage limits: real safety or imaginary?[edit]

"Residual current devices (RCD) or residual current circuit breakers (RCCB) are circuit breakers that operate to disconnect their circuit whenever they detect that current leaking out of the circuit (such as current leaking to earth through a ground fault) exceeds safety limits."

dd —Preceding unsigned comment added by (talk) 15:15, 3 October 2007 (UTC)

Safety limits are not really what its about. A typical example that shows this is a damaged immersion heater. It is very common for the element casings to split open allowing electrical power to flow directly through the hot water. This is a normal part of the life cycle of an immersion heater. This will trip any RCD immediately, although (counterintuitively for those without much electical knowledge) there is no danger, ie people are not getting electrocuted because of it. RCDs tripping due to split elements of fixed appliances does not in reality give any added safety, thus is more correctly described as a nuisance trip. I realise this may sound counterintuitive, but bear in mind safety is removal of danger, and danger equals number of deaths, and split immersion elements are simply a non issue from a safety POV. They arent electrocuting anyone.

How is the 30mA figure reached? It is primarily a question of what will generally stay working. Anything more sensitive will be troublesome, 30mA is mostly ok... though not always. Some installs are horrors from the point of ongoing nuisance trips.

I hope RCDs turn out to reduce deaths, I really do. But... the usual wondrous descriptions of them are I fear significantly inaccurate.

Tabby 21:46, 8 Aug 2005 (UTC)

If you have water energized by an exposed heating element in an immersion heater, and it is in an ungrounded metal container, like grandma's old Maytag washer out on the back porch, or a bucket, and you happen to touch the bucket and a faucet, you will get a painful or dangerous shock. (been there, done that). It is basically no different from the numerous deaths which resulted from someone in a bathtub being electrocuted by a plugged in appliance which fell into the tub. Edison 22:36, 15 March 2007 (UTC)

Are they really safer?[edit]

The article says RCDs increase safety: to be honest we have yet to see. Although they offer protection against known issues, they also introduce a number of dangers.

Firstly users often take full advantage of the protection they offer, using equipment they would never dare before they got their RCD. Examples often seen are plugging wet electrical goods in, using power tools outdoors in the rain, using or installing goods that are known to be problematic safetywise, on the basis the RCD will protect them, and so on.

Also RCDs will kill lighting ciruits quickly in a fire, due to smoke and carbon caused leakage, and deaths are known to have occurred because of this. I curently recommend it is best not to put lighting on an RCD, except with TT installs where they are necessary for basic safety.

Also users tend to believe RCDs will proetct them against risks they dont, and become over confident as a result, exposing themselves to danger. A classic example is that of fitting an RCD plug to an electric drill, in the belief that the power will be cut if they drill into a live wire. In reality no such protection is given by this arangement. First the RCD only detects current imbalance _in the drill's supply_, so it wont even notice the user getting fried, as the shock current is coming from the wire buried in the wall. Second, if by some miracle it did manage to receive a heavenly missive telling it Jo's getting fried, cutting power to the drill, which is all it can do, would make no difference, since the power to the drill case is coming from the wire in the wall, not the drill's own lead. The problem here is simply that the user thinks the RCD will protect them against drilling live cables, ceases worrying about it, and does things that are a threat to life.

I'm not taking an anti-RCD stand here, just being realistic enough to say that the safety benefits are not what is being widely claimed. Or even close.

Only time will tell where the balance lies between the hazards they create and the hazards they protect against. Lets just hope theyre the right move.

Tabby 21:46, 8 Aug 2005 (UTC)

There is discussion about anti-lock brakes, and that people might drive different, knowing that they will be protected. I suspect that most, even those who understand anti-lock brakes, don't understand and rely on RCDs. As well as I understand it, European RCD (where they call them that) trip at 30ma, where US GFIs trip at 5ma. Probably best not to trust your life to a 30ma device. Gah4 (talk) 21:01, 7 March 2021 (UTC)

RCDs and immersion heaters[edit]

Tabby writes:

Safety limits are not really what its about. A typical example that shows this is a damaged immersion heater. It is very common for the element casings to split open allowing electrical power to flow directly through the hot water. This is a normal part of the life cycle of an immersion heater. This will trip any RCD immediately, although (counterintuitively for those without much electical knowledge) there is no danger, ie people are not getting electrocuted because of it. RCDs tripping due to split elements of fixed appliances does not in reality give any added safety, thus is more correctly described as a nuisance trip. I realise this may sound counterintuitive, but bear in mind safety is removal of danger, and danger equals number of deaths, and split immersion elements are simply a non issue from a safety POV. They arent electrocuting anyone.
I'm afraid you're not considering all the possibilities. Let's take, for example, my hot tub. It has this nice immersion heater that is in intimate contact with water that's just loaded with ions that make the water very conductive. When the immersion heater eventiually fails, the water is connected to the live windings of the immersion heater and becomes charged to some arbitrary voltage (depending on where the leak in the immersion heater occurs. Typically, they fail near the top end so in the US, you could almost count on the water being connected to a full 120 vac at the end of the split phase-wired immersion heater. (And if you don't believe me, I'll show you a heater that failed exactly as I'm describing!)
Now, step into my hot tub. Without an RCD, YOU probably form the lowest-impedance path from ground to the energized water. You've just used up all nine of your lives, Tabby. But with an RCD, the water conducts enough leakage current to the grounded pumps and such that the RCD trips and de-energizes the water.
The same sorts of situations can happen in houses with electric hot water heaters and one or more runs of plastic insulative piping. Without an RCD, you can get energized water (although the situation may be less severe than the hot tub because the poable water may contain fewer conductive ions). (By the way, are you familiar with the fact that hot water heaters usually install dielectric joints in the copper piping? So even with copper piping, it's not impossible that the copper somewhere in the house won't be well-grounded.)
RCDs save lives, and properly designed RCD circuits typically don't put emergency egress lighting on the same RCD device (or even any RCD device) as the circuits that pose shock hazards.
Atlant 00:16, 9 August 2005 (UTC)
That would be true if it wasn't grounded, but I assume that it is. If the heater shell is grounded, then it will trip on that, and not current through a person. Gah4 (talk) 21:03, 7 March 2021 (UTC)
On the other hand, for a handheld immersion heater, like for a single cup of coffee, you might get your hand in the water, and the heater element shell is normally not grounded. Gah4 (talk) 21:05, 7 March 2021 (UTC)

How to Evaluate Risk[edit]


I'm afraid you're not considering all the possibilities. Let's take, for example, my hot tub.

Every imaginable combination of faults and design errors is possible. We could sit here and think up scenarios by the score - but sooner or later one notices that some of these scenarios are common, some rare, and some simply dont happen IRL. Thus what scenario we can think of is not what determines level of risk.

What does determine level of risk is how many people are dying in real life from these imagined scenarios. The one you pose here is not one that is hapenning here in the UK, thus, by elementary logical deduction, it is not a significant risk.

Real world risk is best determined by looking at the facts, not by imagining what scenarios might or might not happen. Its a common mistake.


Tabby 22:50, 15 Sep 2005 (UTC)

Re RCDs saving lives[edit]

There is no question that RCDs have saved lives. It is also clear that RCDs have cost lives, and will continue to. The questions are: 1. which number will be bigger over the many years to come, will they save more or kill more? 2. could the cost of millions of RCDs be better spent elsewhere, eg on non slip stair treads? Stairs injure and kill lots of people, electrics dont (in UK).


PS I agree with your comment that escape lighting should not be on RCD, but here in UK is routinely is. We have an ongoing history of mandated whole building RCDs too.

Tabby 22:51, 15 Sep 2005 (UTC)

The only time it is mandated by the electrical regs for everything to be on a RCD is a TT system and in a TT system it really is important to have everything on RCD (or voltage operated ELCB but those had big problems of thier own) but even then it is considered very bad practice to just have one RCD in the system. The reason the horrible practice of one rcd covering everything became common was simply cost. Plugwash 20:16, 11 September 2007 (UTC)

What is the "residual" in RCD?[edit]

I accept the points of those who say that ELCB is not technically accurate but at least it is descriptive. I haven't been able to figure out what "residual" means. Any ideas? Harmonitron

THe residual current is the difference between the phase current and the neutral current. There shouldnt be any diff if you have no leakage to earth. But if there is a leakage it means some normally earthed part is live, and this will be sensed by the balanced coil and trip out.--Light current 23:51, 23 April 2006 (UTC)

How does earth leakage circuit breaker differs from residual current circuit breaker[edit]

IMO it doesnt--Light current 23:48, 23 April 2006 (UTC)

ELCBs don't measure current imbalance, instead they measure current in the earth wire. This gives them much less shock protection. Tabby (talk) 23:06, 7 December 2007 (UTC)

Disadvantages of RCDs?[edit]

I understand from the comments above why RCDs shouldn't be used in lighting. Why are they not used everywhere else? I suppose they cost slightly more and that the differential transformer uses up some extra energy. Are there other downsides? --Stereo 12:42, 6 April 2006 (UTC)

Nuisance tripping!--Light current 23:52, 23 April 2006 (UTC)
Nearly all appliances leak some current to earth and this can seriously add up. For this reason and due to the huge disruption of a trip covering a lot of equipment its inadvisable to put too much kit on the same RCD. However RCDs suitable for fixed installation are still quite expensive (it's a chicken and egg thing, RCDs suitable for fixed wiring won't get cheaper until more people use them and more people won't use them until they get cheaper) so the typical home (at least here in the uk) ends up with a split load board (some circuits only on the main switch the rest all on a single RCD). So a comprimise ends up being made between the desire to have RCD protection and the desire not to load too much on that one RCD. Plugwash 14:23, 5 May 2006 (UTC)
Because of the existence of the combination RCD/duplex-outlet, current American practice seems to be tending towards lots of RCDs protecting individual portions of branch circuits. So, for example, in my house, I have the following RCDs:
  • One in the garage, protecting all four garage outlets and one associated exterior outlet
  • One in one bathroom, protecting just that bathroom
  • One in another bathroom, protecting that bathroom and two others
  • One protecting a portion of the basement workrooms, the central vac, and two exterior outlets
  • One protecting one exterior outlet
  • Three protecting kitchen convenience outlets.
  • One serving the kitchen garbage disposal
  • One serving the dishwasher
With this arrangement, nuisance tripping is minized as is the scope of the failure if an RCD trips. With the exception of one kitchen appliance that trips one RCD (which is apparently sensitive to induced noise; there's no ground fault in the appliance.), we hardly ever see a nuisance trip although we've seen some genuine trips.
Atlant 15:25, 5 May 2006 (UTC)
Yeah you can get RCD outlets in the UK too but unlike your american ones they only protect thier own outlets. Also they are pretty expensive. Plugwash 17:32, 5 May 2006 (UTC)
Here, the going price is about US$12-$13, and they're frequently put "on sale" at $10 or so.
Atlant 19:17, 5 May 2006 (UTC)

about £20 in UK, which is ~$40 a pop, in a slightly less wealthy country than US. And they don't daisy chain, one rcd socket only protects itself. Tabby (talk) 23:08, 7 December 2007 (UTC)

"Whole installations on a single RCD, common in the UK, are prone to nuisance trips that can cause safety problems with loss of lighting and defrosting of food." This statement sounds like trips happen all the time and are normal. In practice, I've never come across trips that weren't caused by a real electrical fault. Has anyone got figures for how often food is spoiled due to nuisance trips? Whole installation RCDs do make fault-finding difficult. pjb 06:47, 25 January 2009 (UTC) —Preceding unsigned comment added by (talk)

RCD's and MEN connections[edit]

What are the implications of using RCD protection where MEN connections are present?

What are MEN conns? Is it the same as PME Protective_multiple_earthing (protective multiple earthing in UK?)--Light current 08:36, 5 May 2006 (UTC)
Assuming you mean a neutral thats tied to real earth at multiple points and/or used as a protective conductor as well as a neutral there are two reasons why a RCD shouldn't be used in such a core.
  1. Paralell paths bypassing the RCD will cause it to trip when it shouldn't
  2. Isolating a combined neutral and earth core is a BAD idea.
However its perfectly fine to feed a RCD from such a system provided neutral and earth remain seperate after the RCD and the only path from the outgoing neutral to the transformer neutral is to go through the RCD. Plugwash 09:09, 5 May 2006 (UTC)
There is a subtle issue with PME and this type of breaker. If you have a neutral to earth fault, you get a low impedance circuit through the PME connection and the fault, so you have a short-circuit around one of the sensing coils. The effect of this is to desensitize the breaker, esp. if the neutral to earth fault is close to the breaker. pjb 07:12, 25 January 2010 (UTC) —Preceding unsigned comment added by (talk)

alternative names for the device[edit]

at this forum thread lots of different names for RCDs in English have been listed. TERdON 22:36, 8 May 2006 (UTC)

Category: Electrical safety[edit]

I added this article to the category Electrical safety because I belive people looking for information on GFCI/RCDs are apt to look in that category.Gerry Ashton 14:34, 25 May 2006 (UTC)

Surge protector into RCD[edit]

Forgive my ignorance, but what's the typical result of plugging a surge protector into a RCD? Will both still function correctly? Nullbit 01:02, 25 September 2006 (UTC)

Yes it will Blaab
Thanks Nullbit
It isn't so obvious that a surge protector can't generate hot to ground currents, but usually not enough. I have an outlet tester, which has three neon lamps, hot-neutral, hot-ground, and ground-neutral. Commonly, neon lamps run 1ma or so, not enough to trip. There might be some current through the MOV, usual in surge protectors, but again should be enough less than 5ma. Gah4 (talk) 21:12, 7 March 2021 (UTC)

RCDs in Australia[edit]

The article doesn't seem to mention anything about Australia's regulations, so the short version:

RCDs must trip at no more that 30mA within 20ms (this is because anything over 35mA is likely to kill a person if it goes through their heart). RCDs must be installed for any new circuits added to a domestic situation (this includes lighting circuits). The only exception to this is where it would be deemed to add greater risk (such as emergency lighting or smoke alarms.) Also, if there is only one appliance wired into a circuit, it does not require an RCD (eg - an oven). It is common practice for electricians to run a dedicated circuit for older fridges that do not pose a risk. Also, RCD's are usually refered to as Safety Switches by the general public.

Hope that added to the info. Blaab 10:34, 28 September 2006 (UTC)

I have difficulty believing this. I live in Victoria, and we had three new circuits installed in September 2007, a year after the previous article, and none of them had RCDs on them. Two were for air conditioners, and one was a dedicated 20 A circuit for a dish washer. My understanding is that RCDs are only required where human contact with the circuit is possible. Groogle (talk) 06:05, 23 December 2008 (UTC)

Did you not note that "Blaab" wrote "if there is only one appliance wired into a circuit, it does not require an RCD (eg - an oven)"? In your case, you had "three new circuits installed" for (two) air conditioners and one dishwasher. This IS a case of each appliance being wired into its own separate circuit. It is almost certain that the air conditioners would have been directly "wired" into the (16 A) circuits concerned (without a socket outlet) and the dish-washer would have been connected to its own dedicated 20 A circuit, with which a socket outlet may or may not have been involved. (I am surprised to hear of a dish-washer which required 20 A at 230 V. This could be over 4,500 Watts, when most domestic dish-washers would draw less than 2,000 Watts. Hence, I presume that yours is an an industrial or restaurant situation.)Fredquint (talk) 15:19, 13 May 2014 (UTC)

Cups of Rice?[edit]

These values were set by tests at Underwriters Laboratories during which volunteers holding cups of rice were subjected to shocks of known amperage and voltage.

What's the purpose of the rice? Is this possibly a joke? 22:50, 29 November 2006 (UTC)

Rename article to more commonly used term Ground fault circuit interrupter[edit]

I wonder if more of these devices are in use in the UK or in the US? What are they called in other English speaking countries? We should title the article with whatever is the more common term. Checking Google, I see for "ground fault interrupter" 96,100 hits, for "ground fault circuit interrupter" 167,000 hits, and for "Residual-current device" only 67,300 which reduces to 65,000 when Wikipedia and its mirror sites are removed. I propose renaming the article to "Ground fault circuit interrupter" unless there is a good reason to use the present term. Edison 22:30, 15 March 2007 (UTC)

Residual current devices are called as such in New Zealand, and presumably, Australia as well because we routinely purchase each other country's brands for such devices. --JNZ 21:21, 19 July 2007 (UTC)

Object Tabby (talk) 23:09, 7 December 2007 (UTC)

Naming on the basis of google results is a dubious approach in any case. Not only does it bias Wikipedia towards the nomenclature of dominant countries on the internet (i.e. the USA), but in cases like this, there are abbreviations which are used far more often than the spelled-out terms. e.g. a google for "RCD electric" turns up 710,000 results. (talk) 01:12, 15 February 2011 (UTC)

Did you notice the dates on the thread? How about ELCB? What do you non-North-Americans call these things again? --Wtshymanski (talk) 01:32, 15 February 2011 (UTC)

A GFI is a different device to an RCD. It has a second coil attached to the main sensor coil around the supply conductors. This enables a GFI to detect and trip on detecting a neutral (grounded wire)-to-earth fault. GFIs are covered by a different standard (UL943) which incorporate tests for the neutral-to-earth fault. Strictly speaking, there should be a separate entry on GFIs. --Kiwigpz (talk) 09:45, 27 December 2007 (UTC)

Do you have a ref for what standard RCDs are covered by in that same set of standards? And for the different arrangement of current sensing coils? GFI and RCD seem to be the same thing. Edison (talk) 06:15, 28 December 2007 (UTC)
There are no US, Canadian or Mexican standards that I know of for RCDs, only GFIs - I stand to be corrected on that as I have worked only on GFCIs for the US market. Internationally for RCDs, there is IEC61540, IEC61008, and IEC60755 Australia and NZ use AS/NZS3190 as well as local adoptions of IEC 61008. The UK uses BS 7288 and BS 7071. All these standards refer to RCDs - not GFIs. It's interesting to note that 120V countries seem to use GFIs and never RCDs and 230V countries use RCDs and never GFIs - don't know why. If you follow this link to the datasheet for the chip that is used by most electronic GFIs and RCDs (my opinion) you will see on page 4 a nice description of the differences in the circuit to give the different operation. Figures 2 & 3 show the standard circuits - you will see two coils in the GFI and one in the RCD.--Kiwigpz (talk) 10:23, 28 December 2007 (UTC)
I am trying to determin if the RCD and GFI are really different devices, or if it is just regional terminology difference. This afects how the article should be set up, or if there should be two articles. The standards referred to by Kiwigpz are only viewable in a preview format without paying a large sum, so the test standards are not viewable to determine the trip and nontrip conditions the RCD and GFI must satisfy. The chip data sheet shows first a circuit for a 2 wire ungrounded outlet which would trip if the neutral and phase currents differ sufficiently, as would be the case for a phase to ground leak, or a leak to the protected circuit's neutral from a different source (less of a concern). The arrangement for a three wire grounded outlet apparently provides for tripping if there is a low resistance connection between the neutral and the ground prongs of the outlet. Are you saying that the first trip condition is also characteristic of a RCD but it lacks the second trip condition? I am looking for the trip/notrip conditions for the two devices (if they are in fact different). Edison (talk) 17:31, 28 December 2007 (UTC)
Yes, to my knowledge, no RCD will trip on a neutral-to-earth fault. Sorry but I have no other references to back that up - just the standards listed. UL943 contains the test for the neutral-to-earth fault that GFCIs will pass, none of the RCD standards do. I will try to find a reference that confirms that.--Kiwigpz (talk) 09:52, 29 December 2007 (UTC)
I don't understand from your link to the RV4145A how it works when there's a neutral-to-earth fault. What is a "magnetic path between the sense transformer and the grounded neutral transformer"? How does this result in an "AC coupling" that closes a positive feedback loop around the op-amp? Eric Kvaalen (talk) 17:32, 2 August 2008 (UTC)
As well as I know, it is usual for European RCD to trip at 30ma, and US GFIs to trip at 5ma. That doesn't mean that the names are actually indicative of that, though. I don't know what the WP:MOS rule is regarding US vs. other countries and work usage. Gah4 (talk) 21:17, 7 March 2021 (UTC)

Trip without load attached[edit]

What would cause a RCD to trip without a load (or an ungrounded load such as a cell phone charger) attached? --D3matt 03:42, 23 March 2007 (UTC)

Well, the obvious cause would be something bumping the "test" button. Older units (of the 1970's ) frequently tripped when there was a lightning storm, and today's might also if a sufficient spike of voltage affected the amplifier circuit which monitors the residual current. A power quality problem in the building could also cause spikes in the supply voltage which might afect a unit. Then there is the possibility of one simply being defective or too sensitive. Or the ungrounded load could conceivably have had a short to some nearby ground, or an extension cord or adaptor could have leaked current to ground. It does not take much current. Edison 15:30, 23 March 2007 (UTC)

I actually had two bad experience's with GFCI outlets - both with no load at the time. A) One GFCI outlet was live but I was stapling a wire to a stud and the vibration caused something internal to the GFCI to arc and smoke. I tested the GFCI and it still had live voltage, but my external outlet tester could not cause it to trip. I threw out the old outlet and replaced it with a new one. B) For a different GFCI outlet, I pushed the TEST button (there was no load) and it behaved fine. But when I plugged in my external outlet tester and pushed the test button (more than once), it failed to trip. Perhaps the trip thresholds were off a bit, but I've used this external outlet tester many times before without issue. The outlet was wired correctly. Again, I disposed of the defective outlet. In summary, I recommend using an external test device for GFCI testing rather than rely on the TEST button on the outlet. — Preceding unsigned comment added by (talk) 01:57, 2 January 2015 (UTC)

It is fairly easy to hit the test button when plugging something in. The test button connects to neutral before the transformer, so doesn't need a ground wire. External outlet testers connect to ground, which isn't required for a GFI. (Especially in older wiring, replacing a two prong outlet.) Gah4 (talk) 21:21, 7 March 2021 (UTC)

Trip on cooker circuit[edit]

I have a current fault of a cooker that trips my RCD. The trip may not come immediately on turning the cooker supply on but it will normally trip after a given amount of time. For example the trip may occur when I turn the oven on or after a few minutes with no action or change occuring (no load change). Another example was when the cooker was on for 59 minutes and I went to take a shower. As soon as I switched the electrical shower on the trip occurred.

This RCD unit is on all of the house supply (separate breakers obviously)and none of the circuits are split. I have recently installed new electrical circuits in the kitchen but the cooker circuit was only modifed (change of position). I cannot work out if...

1. The cooker is faulty (has been in the garage not used for 3 years).

2. The wiring is faulty or there is another faulty circuit somewhere.

3. The RCD trip device is faulty (i.e. over sensitive).

The RCD has some history of nuisance tripping and previously with the original electric cooker it used to trip if sometimes if both ovens were used or if the cooker was not used for a week or so (like returning from holiday).

I could bypass the RCD but I would lose all protection for all circuits.

I hesitate in bypassing a safety device.

Are RCDs suitable for cookers?

A.Wheeler 10th Sept 07. —Preceding unsigned comment added by (talk) 12:03, 11 September 2007 (UTC) Tabby (talk) 23:27, 7 December 2007 (UTC)
No liability is assumed by providing the following comments. Help out your friends across the pond with terminology: the "cooker" is the kitchen range, with burners on top and an oven, yes? Is it powered by 240 volts, with two wires, or 240 volts plus a neutral to operate the controls on 120 v? The requirements for whether the range requires a GFI (ground fault interrupter, or RCD if you prefer) will depend on your applicable electrical code. Outlets in the kitchen generally (in the US) are required to be on ground fault protection because of the presence of the sink, of moisture, and of grounded metal appliances. Here the range is often on a separate circuit from the outlets, and I have certainly seen installations by electricians which were approved by inspectors where the range was not on a RCD. But if it has a convenience outlet as part of it, then that should be on a RCD. It would be appropriate to have a licensed electrician check it out. Safety from shocks is important. At least 3 things are possible: 1) The cooker could certainly have some part of the wiring which at times shorts to the grounded metal part of the appliance, tripping the RCD. This could be part of the heating element, or ebven the clock, RCDs are generally built to trip in milliseconds on milliamps of such leakage current, not like an overload protection device (fuse or breaker) with a long time curve, which might trip on a moderate overload after several minutes. 2) A gadget like an RCD can trip falsely. They make millions of them, and a few are bound to be defective. 3) Something else could be on the same circuit as the cooker which is tripping the RCD. Isolate other appliances. You mentioned the "electrical shower." Is that a demand heater near the shower which runs when hot water is needed, as opposed to a storage heater in a utility area? Is it on the same circuit as the oven? Are you overloading a breaker? Edison (talk) 23:43, 7 December 2007 (UTC)
The original poster appears to be in the UK both from a whois on his IP and from the terminology he uses. I will make that assumption in my reply.
First of all I will answer edisons questions and clear up some differences between UK and american practice
Domestic properties in the UK normally use 240V single ended single phase. Some larger houses and most commercial properties get 240V/415V three phase. 120V is only used in specialist situations.
Cooker generally reffers to a device with both oven and hob (probablly what you reffer to as burners but we tend to reseve the term burner for things that actually burn ;) ) integrated into a single unit though it can also sometimes be used to reffer to a seperate oven or hob.
Afaict american "GFCI breakers" combine the functions of leakage detection and overcurrent protection. We have such devices (we call them RCBOs) but they are expensive and not normally used in domestic installations. More common here is to have a single RCD (which only provides earth fault protection) covering either the entire installation (RCD incomer) or all the circuits that are to be RCD protected (known as a split load arrangement) then seperate MCBs (overcurrent breakers) for the individual circuits.
An electric shower is a unit fixed on the wall in a shower cubical or above a bath that takes cold water in, heats it and lets it out through a shower head. Controls on the front of the electric shower allow the user to turn the water on and off and have some control over flow and power (which indirectly control temperature).
It sounds like his electric shower is on a seperate MCB but on the same RCD as everything else.
Now back to the orginal posters problem
Cookers are prone to becoming leaky devices and it is not that unusual for them to trip RCDs, most electricians therefore prefer to keep them on NON-rcd circuits. However if there is a socket on the cooker circuit and this socket could be "reasonablly expected to supply equipment outside the equipotential zone (interepretations of this reg vary widely) then it either needs to stay on RCD or the socket needs to be removed.
Bypassing the RCD for the whole installation is not a good idea. Really you need the help of an electrican who knows what they are doing and has proper test equipment (no a multimeter is not much use here sorry) to test the RCD and the cooker and the wiring and any other suspect appliances and decide where to go from there.
-- Plugwash (talk) 00:40, 8 December 2007 (UTC)

Very well explained "Plugwash". I congratulate you on you knowledge of electrical systems outside of your "domestic" environment. Unfortunately, such knowledge is not common and many contributors often have marked jingoistic tendencies. Fredquint (talk) 15:19, 13 May 2014 (UTC)


Our home was struck by lightning last night and I found it very interesting that every GFI unit in the home tripped and appears to have saved many of the items they powered. Our 2001 home has several GFI’s. Things that were not using a GFI receptacle were damaged unless they had a surge arrestor built-in. I believe the lightning struck the powered attic vent or the gas water heater’s metal smokestack. The family could feel the electricity.. —Preceding unsigned comment added by (talk) 16:52, 24 April 2008 (UTC)

If the company, who wired your home, done everything honestly, they have put dischargers, which accept all lighting to themself and redirect it to the earth. As there was too high current in ground wire during lighthing, CFI felt it and disenergized circuits (talk) 17:38, 13 December 2010 (UTC)

Picture of GFCI receptacle?[edit]

Should we have a picture of a GFCI receptacle for this section? In the US, these devices are quite common. (talk) 20:46, 19 July 2008 (UTC)

Surge resistance[edit]

I do not understand the section entitled "Surge current resistance [in A]":

Peak current an RCD is designed to withstand (8/20 µs impulse). The IEC 61008 and IEC 61009 standards impose the use of a 0.5 µs/ 100 kHz damped oscillator wave (ring wave) to test the ability of residual current protection devices to withstand operational discharges with a peak current equal to 200 A. With regard to atmospheric discharges, IEC 61008 and 61009 standards establish the 8/20 µs surge current test with 3000 A peak current but limit the requirement to RCDs classified as Selective.

What is an 8/20 µs impulse? What is a "0.5 µs/ 100 kHz damped oscillator wave (ring wave)"? (Note that 100 kHz has a period of 10 μs.) What atmospheric discharge is it talking about? What exactly do IEC 61008 and 61009 concern, respectively?

Eric Kvaalen (talk) 17:32, 2 August 2008 (UTC)

I've rephrased that first sentence, it was unforgiveably jargon-y for a general reference encyclopedia. The 8/20 us impulse refers to a common shape of pulse used for electrical testing. --Wtshymanski (talk) 19:45, 2 August 2008 (UTC)

Bigger picture of internals of RCD needed[edit]

The current picture of the internals of the RCD is quite small, a larger picture would be nice. —Preceding unsigned comment added by (talk) 13:20, 14 April 2009 (UTC)

3-phase picture[edit]

in the picture test button winding is incorrectly connected only to one side of torroid (both are up)

-supposed to connect on one to up - other to the lower side —Preceding unsigned comment added by (talk) 22:27, 16 June 2010 (UTC)

3-phase picture(corrected one)[edit]

3-phase.JPG —Preceding unsigned comment added by (talk) 22:58, 16 June 2010 (UTC)

Great write ups regarding residual current devices however sensitivity can be higher than 300mA free standing electronic rcds with seperate cts for used for higher currents (usually up to 800A)have adustable sensitivity from as little as 10mA up to 2 amps (depending on the manufacturer).The protective circuit breaker (mccb ar acb) is tripped by means of an undervoltage/shunt release.Small modular rcbs for use with miniature circuit breakers have integral cts (or toroids)and are available with a sensitivity of up to 500mA. Dave Smith Maidstone UK. —Preceding unsigned comment added by (talk) 15:32, 30 August 2010 (UTC)

Passive/active/latching/non-latching - cutting the Gordian knot[edit]

We've had a CN tag on that section for almost a year. There was a copy of one manufacturer's brochure for an RCD offered as an example, but strangely, none of the books I've found so far on Google Books talks about two types of RCDs. This makes me suspicious that it may be a peculiar offering by one manufacturer, and not necessarily standard practice. (North American GFCIs, to the limit of my knowledge, don't pay any attention to power interruptions and only trip for ground fault current. Perhaps North American practcies put the responsiblity for not restarting on power restoration in the standards for the connected equipment, and not on the GFCI's mandate? Seems a little weird to me to combine these two functions in the RCD). I've taken the section out. If someone finds a good authoritative reference that classifies two types, please restore that section with a good citation. --Wtshymanski (talk) 14:02, 19 August 2011 (UTC)

As well as I know, this is a European thing. I am not so sure why it is connected, but certainly in some cases being surprised by power restoration is bad. One might turn on some device, forget that it is on, and then leave it. As well as I know, rare in the US. Gah4 (talk) 21:28, 7 March 2021 (UTC)

Confusion in top paragraph[edit]

For revision on Feb 13 2012 The first paragraph on this article lists so many names and abbreviations that it is hard to read it without misunderstanding something. The information may be correct and possibly worth noting but with this many names I would have put information on alternate names in its own section, not in the first paragraph. Even if it deserves to be in the top paragraph it perhaps ought to have a better layout than it has now. — Preceding unsigned comment added by (talk) 14:02, 13 February 2012 (UTC)

2 errors[edit]

Whole installations on a single RCD, common in the UK, are prone to nuisance trips that can cause safety problems with loss of lighting and defrosting of food. RCDs also cause nuisance trips with appliances where earth leakage is common and not a cause of injury or mortality, such as water heaters. However such nuisance trips are easily corrected by removing the problem (such as a blown light bulb) and re-setting the RCD.

The last sentence is quite incorrect

  1. blown lightbulbs are open circuit loads, removing them has nothing to do with resolving earth leakage issues
  2. problems causing RCD trips are known for being not easy to correct, generally speaking
  3. In the case of water heater leakage, it would be unrealistic to describe the solution as easily corrected and reset the rcd.

So I removed the faulty sentence.

This RCD contains a resistor with an improper value in the test circuit, which gives incorrect results for testing. Under certain circumstances it may prove lethal because it will not trip when expected. That is why RCDs should be tested by the socket tester with calibrated leakage

This makes little sense. I've not corrected it, as I didn't know what point the author. was trying to make. Tabby (talk)

As well as I know, tradition is for 5ma in the US and 30ma in Europe. Though with the different line voltage and plugs, people won't be using testers in the wrong system. Gah4 (talk) 21:31, 7 March 2021 (UTC)

What is an MCB?[edit]

"MCB" is not a universal explanation of a circuit breaker. Professional electricians across North America will never run into that terminology unless they speak with an electrician outside of that area. "MCB" should be linked to a correct page or explained as to what it is. — Preceding unsigned comment added by (talk) 21:36, 28 February 2013 (UTC)

The term MCB (for Main Circuit Breaker) is widely used here in California in electrical power engineering. (talk) 23:42, 4 December 2019 (UTC)

Although that's not what it means (in this article, at least). Andy Dingley (talk) 19:56, 16 December 2019 (UTC)

History Section Query[edit]

In the History section it is stated "In the early 1970s most North American GFCI devices were of the circuit breaker type. GFCIs built into the outlet receptacle became commonplace beginning in the 1980s. The circuit breaker type, installed into a distribution panel, suffered from accidental trips mainly caused by poor or inconsistent insulation on the wiring. False trips were frequent when insulation problems were compounded by long circuit lengths. So much current leaked along the length of the conductors' insulation that the breaker might trip with the slightest increase of current imbalance. The migration to outlet receptacle based protection in North American installations reduced the accidental trips and provided obvious verification that wet areas were under electrical code-required protection."

I find it difficult to believe that the insulation on the wiring in North American premises was so poor that GFCI circuit breakers installed at the distribution panel were "falsely" tripped by the leakage and that it was necessary to resort to the installation of (multiple) CFGI devices at the outlet receptacles to get over this problem. (Yet in Europe, and elsewhere, RCDs and RCBOs are normally installed at the distribution board without problems.)

If the above "History" true, this would seem to be an indictment of the quality of the insulation of North American cabling. I submit that in most countries, if it were found that when a GFCI, RCD or RCBO was installed it detected leakage of sufficient magnitude to cause it to "trip", the authority concerned would deem it to be a fault condition that required the replacement of the wiring concerned. Such wiring would not pass the normal (Megger) Insulation Resistance Tests. Fredquint (talk) 15:19, 13 May 2014 (UTC)

I remember stories about the early GFCI trip set at 1ma, and that caused too many problem. It was then changed to 5ma. Circuit breaker type are and were always much more expensive. I believe commonly used for bathrooms and showers in commercial settings, but rarely for home use. PVC wiring should have small enough leakage, but some older wiring might be leakier. Gah4 (talk) 21:35, 7 March 2021 (UTC)

History section correct?[edit]

The history section in the german version of this article ("Fehlerstromschutzschalte") goes back quite a bit further than to south africa 1955:

Der Fehlerstromschutzschalter wurde bereits 1903 von Schuckert unter der Bezeichnung Summenstromschaltung zur Erdschlusserfassung patentiert (DRP-Nr. 160.069).[20] Kuhlmann beschreibt bei AEG eine Methode zur Messung der Erdschlussströme im Berliner Netz. Weiterentwickelt wird die Technik, auf der auch heutige Fehlerstromschutzschalter basieren, von Nicholsen (1908, USA-Pat-Nr. 959.787).[21]

Maybe this should be revised in the english article? — Preceding unsigned comment added by (talk) 10:36, 8 April 2013 (UTC)

A translation from the current article is given below "The residual current circuit breaker was already 1903 patented by Schuckert under the name Total Current for Earth Fault Detection Circuit ( DRP No.. 160 069). [23] Kuhlmann describes AEG a method for measuring the earth fault currents in the Berlin network. Is further developed the technology, on which also present residual current circuit breakers are based, by Nicholson (1908, U.S. Pat no. 959 787). [24] In the early 1950s, after countless suggestions and technical studies on the basic applicability of the circuit for the first time presented as a protective device a mature residual current circuit breakers for wide use in the current client. [25] Busy is in it for 1951 a residual current circuit breaker of the company protection devices Society & Co. mbH. KG, Schalksmuehle / Westphalia. (Schupa) with the trade name cobweb , [26] which was designed in two-, three-and four-pole version for a nominal current of 25 A and voltages up to 380 V with a tripping fault current of 0.3A. A lower trigger threshold was discussed but rejected as economically unsound. The time allowable leakage in heating devices have resulted in a lower trigger threshold also in frequent false alarms. In the year 1957 by Gottfried Meier Biegel in Austria at Felten & Guilleaume a residual current circuit breaker. These were prescribed in Austria in 1980 in private households by law, the trip current was gradually reduced from the original 100 mA at 70, 65 and 30 mA. Since early 1985, this applies to the entry into force of the provision SEV 1000-1.1985 in Switzerland."

Fredquint (talk) 16:09, 13 May 2014 (UTC)

Variation with frequency?[edit]

I have a basic understanding of how these devices operate but have a query as the parameters will vary with supply frequency. Trying to decide if a device designed and specified for use on 50Hz would work at 400Hz. SpikeUK (talk) 08:41, 28 November 2013 (UTC)

It is a current transformer, and that should work independently of frequency, as long as it isn't too low. (Won't work at DC.) Otherwise, I don't know about code rules regarding which devices can be used where. Gah4 (talk) 21:37, 7 March 2021 (UTC)

Use of RCDs for protection against shorts to earth[edit]

Undid revision 588089932 by TinyMark (talk) RCDs are used for earth fault protection. See talk.

RCDs are very much used for protection against short circuits to earth. See below link for reference. In the case of TT installations, where there is no metallic earth return to the supply transformer, the earth loop impedance may be too high to trip a circuit breaker, which is why RCDs are used for additional protection, typically with a 100 or 500 mA trip current, as compared to the 30 mA typically used for protection against electric shock.$file/2CSC420004B0201_RCDs%20EN.pdf

I'm having problems with the terminology here. I must admit that I never had anything to do with these things when I lived in the UK. But surely, in a TT system, the circuit cannot be shorted to earth because a the circuit is not shorted, i.e. the current does not return directly to where it came from????? TINYMARK 12:14, 29 December 2013 (UTC)
If I recall the regulations do not use the term short circuit; it is simply referred to as a fault. But short circuit defines the term as current flowing along an unintended path, which is what happens - in the event of such a fault the current flows directly through the mass of earth to the supply transformer (usually via the earth electrodes). It's not relevant that that path isn't a metallic path. The RCD is used because depending on the local conditions the fault current may be only a few amps, which means a normal circuit breaker could take a long time to trip, or not trip at all. The circuit breaker is still relevant and used to provide overload protection and protection from phase-phase or phase-neutral faults. Alistair1978 (talk) 12:28, 29 December 2013 (UTC)
Well I have been living in Germany for ages now. The circuit breakers are marked with LS (Leitungsschutz) meaning that they are there to protect the power cable - from burning out or causing a fire, I assume. This will not happen with a higher resistance short as you described as the current is too low. Although a fault is present and the RCD will trip - but what is being protected??? TINYMARK 12:52, 29 December 2013 (UTC)
It's for preventing fire at the location of the fault (see page 16 of the PDF linked above). Otherwise, a fault to earth could deliver say 10 amps continuously, which might not melt the supply cabling but could easily cause a fire at the point of fault, as well as causing rise of earth potential which could cause a risk of electric shock. Note that such TT supplies are very common in the UK in rural installations with overhead supplies. Alistair1978 (talk) 14:19, 29 December 2013 (UTC)
Thanks Alistair. My mum's flat is on the first (top) floor of a house and I (2m) could reach up and touch the lines coming into the house! Obviously a TT system, but with very iffy plastic sleeves on the lines. In Germany TT systems are usually only in factories etc. TINYMARK 15:36, 29 December 2013 (UTC)

Lead opening sentence is iffy[edit]

A residual-current device (RCD), or residual-current circuit breaker (RCCB), is a device that instantly breaks an electric circuit to prevent serious harm from an ongoing electric shock. A fuse or circuit breaker also does that, 'instantly' of course, is subject to qualification. How fast a fuse or breaker reacts depends on its profile and the degree of overcurrent in the circuit. They can be made as fact as they need to be. An RCD also has a reaction time that can vary. Actually, it needn't be any faster than a breaker. That's not how it protects against electrocution. That sentence is bad; it misses the issue. Sbalfour (talk) 23:51, 14 November 2017 (UTC)

RCCB full form is Resudual case circuit breaker which are used in industry for prevent from electrical shock to human or animals.Its working sensing by current.Its woking principal of RCCB is its opeated when current is unbalanced from phase or netural . — Preceding unsigned comment added by (talk) 09:21, 13 July 2018 (UTC)

Article Title[edit]

What is with this article title? Why does GFI/GFCI redirect here? As far as I know, RCD is only used in the UK and Australia. The US and Canadian population vastly outnumbers these countries. Having GFCI redirect here might confuse more people than if it were the other way around. (talk) 05:38, 16 May 2019 (UTC)

I don't know the WP:MOS rule, but I don't think it is meant to give greater weight to US usage. Gah4 (talk) 21:40, 7 March 2021 (UTC)


It seems that there is now LCDI, Leakage Current Detector Interrupter, designed to detect faulty power cords. They are required for some portable air conditioners. My understanding is that they put some wire mesh inside the power cord, and outside the actual conductors. If current goes into it, then there is a fault in the cord. That seems somewhat different from RCD or GFI, but maybe it could redirect here. It seems like they can also use Arc Fault Detectors instead, which I think are not well described here, but might be, or might have their own article. Gah4 (talk) 09:57, 20 March 2021 (UTC)

Is there an official definition of an LCDI somewhere, perhaps in the requirements for its presence in certain appliances such as air conditioners? My current understanding of an LCDI is that it is a dedicated GFCI installed together with a mesh-shielded power cord. Please correct or refine this impression if necessary. Reify-tech (talk) 15:07, 23 September 2021 (UTC)