Every other device in a distribution board protects the wiring. The MCB trips when a cable would overheat; the fuse blows when a fault would melt something. The residual current device is the one device that protects the person: it watches for current that has left the circuit and found another way back to earth, through a damaged cable, a wet appliance or a human being, and it opens before the current has done its damage. In an installation earthed through the consumer's own rod, as most in Pakistan are, it is also the only device that will disconnect an earth fault at all.
How it works: the sum of the currents
The line and neutral conductors of the circuit pass together through a toroidal core, a ring of magnetic material with a sense winding on it. In a healthy circuit every milliamp that goes out on the line comes back on the neutral; the two currents are equal and opposite, their magnetic fields in the core cancel, and the sense winding sees nothing. If some of the current returns another way, through an insulation fault to an earthed case, through a person to the ground, or through a neutral-to-earth fault, the line and neutral currents no longer balance, the difference magnetises the core, and the sense winding produces a signal. When that difference exceeds the device's rated residual current, IΔn, a trip coil releases the mechanism and both poles open. The device is a current comparator; it does not measure the current to earth directly, it measures the imbalance, which amounts to the same thing.
IEC 61008-1 (RCCBs) and IEC 61009-1 (RCBOs) set the speed. A general-type device must trip within 300 ms at its rated residual current, within 150 ms at twice it and within 40 ms at five times it, and must not trip below half of it; so a 30 mA device may trip anywhere between 15 and 30 mA, and at 150 mA it opens in under 40 ms, about two cycles. The test button connects the line on one side of the core to the neutral on the other through a resistor, so the test current passes through the core once and creates a real imbalance; pressing it proves the core, the electronics and the mechanism, but not the earth path, which is why a proper test also needs an instrument.
What 30 mA means for a human
IEC 60479-1 maps what alternating current at 50 Hz does to a person as zones on a current-against-time chart. At about 0.5 mA the current is felt. At about 10 mA, less for children, the muscles of the hand contract and the person cannot let go of the conductor, so the shock lasts as long as the supply does. Above that, for currents that flow for longer than a second, breathing becomes difficult and the probability of ventricular fibrillation, the disordered heart rhythm that kills, begins to rise at a few tens of milliamps and is substantial at 50 mA and beyond; for very short shocks the heart tolerates more, several hundred milliamps if the current is gone within a fraction of a heartbeat.
A 30 mA device is chosen because its trip point sits below the region where fibrillation becomes likely, and its speed at higher currents (40 ms at 150 mA) keeps a heavy shock short enough to stay out of the dangerous zone. What it does not do is prevent the shock. The person still receives the current until the device opens, still feels it and may still fall off a ladder; the RCD makes the shock survivable, not painless. And a person in a bath, with the resistance of the body reduced by water and the contact area large, can be endangered by less than 30 mA, which is why special locations use 10 mA devices or extra-low voltage.
Ratings and types
The rated residual current is chosen by purpose. 10 mA is for locations where a person is wet or cannot escape: baths, pools, some medical areas. 30 mA is additional protection against direct contact, and IEC 60364-4-41 clause 411.3.3 requires it for socket-outlets for general use and for mobile equipment used outdoors; it is the standard for every circuit in a home. 100 mA and 300 mA devices are not for people; they protect against fire from earth-leakage currents in cables and fixed equipment, and they serve as the upstream device in a board where 30 mA units protect each final circuit. A time-delayed (S-type) device waits 130 to 500 ms at its rated current so that a 30 mA device downstream trips first and the fault takes out one circuit rather than the whole building; without that delay two RCDs in series both trip.
The type is about the waveform of the residual current, and it has become the most important letter on the device. Type AC responds to sinusoidal residual current only: it was adequate when loads were heaters and motors, and it is now obsolete for most circuits, because a residual current with a DC component saturates its core and can blind it entirely. Type A also responds to pulsating DC, the waveform produced by the half-wave and phase-controlled rectifiers in almost everything with electronics, and it tolerates up to 6 mA of smooth DC without losing its sensitivity; it is the minimum for any circuit that feeds a computer, a washing machine, an LED driver or a charger. Type F adds the mixed-frequency residual currents produced by single-phase variable-frequency drives, such as those in inverter air conditioners and heat pumps. Type B (IEC 62423) also detects smooth DC residual currents, which come from three-phase rectifiers: three-phase drives, electric vehicle chargers, solar inverters and UPS equipment. Install a type AC upstream of a three-phase drive and a DC earth fault in the drive's motor cable will pass through it undetected, and worse, the DC will blind it to any AC fault on the same circuit.
The form is the last choice. An RCCB is a residual current device alone, with no overcurrent protection, and it protects a group of circuits that each have their own MCB. An RCBO combines the RCD and the MCB for one circuit in one device, so a fault takes out that circuit only and the tripping of one does not tell you which of six circuits caused it. For larger currents, a CBR is a moulded-case circuit breaker with a residual current release built in or added as a module, rated to IEC 60947-2 Annex B, and it is how an incoming supply of several hundred amps gets a 300 mA fire-protection function.
What an RCD does not do
An RCCB provides no overload or short-circuit protection: the current that flows out on the line and back on the neutral is balanced however large it is, and a bolted short between line and neutral will not move the sense winding at all. Every RCCB needs an MCB or fuse in series, and the RCCB itself must be rated to carry the circuit's current and to withstand the fault current the MCB will clear. An RCBO carries its own MCB and needs neither.
An RCD gives no protection against a shock between line and neutral. A person who touches both conductors, or the line while standing on an insulated floor with no path to earth, is a load like any other: the current goes in one hand and out the other and returns on the neutral, balanced, and the device holds. The 30 mA rating is only meaningful when the current finds its way to earth. And a type AC device gives no protection against a fault whose current has a DC component large enough to saturate the core, which is the case for the increasing number of loads described above; if the earth path goes through a person and the core is blinded, the device is a switch.
Finally, an RCD cannot protect what it does not see. A fault upstream of it, in the meter tails or the incoming cable, is outside its loop; a neutral borrowed from another circuit downstream of it bypasses its core; and a device that has never been tested may have a seized mechanism, which is why the test button exists.
Nuisance tripping and its real causes
An RCD that trips "for no reason" is almost always tripping for a good one, and the reasons fall into a short list. Every modern appliance leaks a little current to earth by design, through the capacitors of its EMC filter: the equipment standards allow about 0.5 to 0.75 mA for a portable appliance and up to 3.5 mA for fixed equipment and IT equipment, and a workshop with a dozen machines, a rack of computers or an office of chargers can add up to a standing leakage of 10 or 15 mA before anything is wrong. A 30 mA device may trip at 15 mA, and the usual design limit is a standing leakage of no more than a third of IΔn. Damp does the rest: an outdoor socket after rain, an old immersion heater whose element insulation has absorbed moisture, a buried cable with a damaged sheath, an oven element the first time it heats after months idle. A neutral shared between two circuits on different RCDs, common in installations that have been altered, sends return current through the wrong core and trips both. A variable-frequency drive produces high-frequency leakage through its cable capacitance and its filter that a type AC or A device may trip on or be blinded by, and a solar inverter or EV charger can produce smooth DC leakage for which only type B is designed.
The remedies follow the causes. Split the installation across several RCBOs so that no device carries more than a few milliamps of standing leakage and a trip takes out one circuit. Fit the right type: A as a minimum, F for single-phase drives, B for three-phase drives, chargers and inverters. Use a 100 mA or 300 mA S-type device on the incomer for fire protection and 30 mA on each final circuit, so that selectivity keeps the lights on. Give a drive its own RCD and follow its maker's instructions on leakage and cable length. And find the fault: a device that trips on damp is reporting a cable that will eventually fail, and an installation that "works fine" with the RCD bypassed is one earth fault away from a death.
Why TT installations in Pakistan depend on them
In a TN system, where the supply provides a metallic earth back to the transformer, an earth fault draws hundreds or thousands of amps and the MCB clears it as fast as a short circuit; the RCD there is additional protection, a second layer for sockets and wet places. In Pakistan the service to most premises brings phase and neutral only, and the installation's earth is a rod in the ground: a TT system. An earth fault must then return through the consumer's electrode, the soil and the transformer's electrode, and the current is 230 V divided by the sum of those resistances. With a typical rod at 40 Ω the fault current is about 230 ÷ 40 ≈ 6 A. No MCB will open on 6 A; the faulty appliance simply sits with its case at a couple of hundred volts, and the first person to touch it and something earthed becomes the fuse.
IEC 60364-4-41 clause 411.5 makes the RCD the normal means of fault protection in TT systems for exactly that reason, and sets the condition RA × IΔn ≤ 50 V: the touch voltage during the time the device takes to trip must stay below 50 V. With a 30 mA device that permits an electrode of up to 1 667 Ω, so the requirement is easily met; the earthing guide explains why the electrode is nevertheless specified far lower, and the earth electrode calculator runs the check. The point to hold on to is the asymmetry. In a TN country an installation without RCDs is less safe than it should be. In a TT installation it has no earth-fault protection at all: the earth rod, the green-and-yellow conductor and the bonding are only there to let the RCD do its job.
Choosing and testing
For a home or a small office, an RCBO on each circuit, type A, 30 mA: it costs more than one RCCB for the board but it turns every trip into a labelled circuit instead of a dark building, and it removes the shared-neutral trap. For a workshop or a plant, 30 mA RCBOs on socket circuits and on anything a person handles, type B where a three-phase drive or a charger sits on the circuit, and a 100 or 300 mA S-type CBR on the incomer for fire protection and as a back-up. For tools used outdoors or on a site, an RCD-protected socket or an in-line RCD at the point of use, tested before use. Rate the RCCB's current at or above the MCB or MCBs it serves, and check its short-circuit withstand against the protective device ahead of it. The cable sizing guide covers the overcurrent side of the same board.
Test at commissioning with an RCD tester, not just the button: at half the rated residual current it must hold; at IΔn it must trip within 300 ms; at 5 × IΔn, within 40 ms; on both half-cycles, since a fault can begin on either. Write the times on the test certificate and keep it with the board. After that, press the test button monthly (the label on the device says so, and it is the only part of the mechanism that otherwise never moves) and repeat the instrument test at the periodic inspection. A device that fails the button test is replaced the same day; a circuit with an RCD that "keeps tripping" is investigated, not bypassed. Ahmedonics specifies and tests protection as part of every control panel and installation it builds, and the RCD is the line on the test sheet we look at first.
References
- IEC 61008-1, Residual current operated circuit-breakers without integral overcurrent protection for household and similar uses (RCCBs) — Part 1: General rules — tripping times and non-tripping current
- IEC 61009-1, Residual current operated circuit-breakers with integral overcurrent protection for household and similar uses (RCBOs) — Part 1: General rules
- IEC 60364-4-41:2005, Low-voltage electrical installations — Part 4-41: Protection for safety — Protection against electric shock — clause 411.3.3 additional protection by 30 mA RCD; clause 411.5 TT systems
- IEC 60479-1:2018, Effects of current on human beings and livestock — Part 1: General aspects — current–time zones for AC 15–100 Hz
- IEC 62423, Type F and type B residual current operated circuit-breakers with and without integral overcurrent protection for household and similar uses
- IEC 60947-2, Low-voltage switchgear and controlgear — Part 2: Circuit-breakers — Annex B, circuit-breakers incorporating residual current protection (CBRs)