What Is an RCD and How Does It Work

What is an RCD? A residual current device (RCD) is a protective switching device that continuously compares the current going out through the line conductor against the current coming back through the neutral, per IEC 61008-1, and opens the circuit the instant the difference — the residual or earth-leakage current — reaches its rated threshold IΔn. Left undetected, that imbalance can pass through a person's body or smolder in damaged insulation long before a standard MCB notices anything wrong, because a 30 mA leakage fault draws nowhere near the current an overcurrent device is built to see. This article covers the toroidal-core sensing principle, the IΔn thresholds that separate shock protection from fire protection, how RCD/RCCB/RCBO/ELCB/GFCI terminology actually maps to different devices, why the earthing system changes what's required, and how the device gets tested and commissioned in the field.

How the Toroidal Core Detects a Fault

Inside every RCD sits a ring-shaped (toroidal) current transformer. The line and neutral conductors of the protected circuit both pass through the center of this ring. In a healthy circuit, every ampere that leaves through the line conductor returns through the neutral — the two currents are equal and opposite, so their combined magnetic field in the core cancels to near zero.

An earth fault breaks that symmetry. If current finds a path to earth — through a person touching a live part, through damp insulation, through a nicked cable — less current comes back through the neutral than went out through the line. That imbalance produces a net magnetic flux in the toroidal core, which induces a current in a secondary sense winding wrapped around the same core. Once that induced current reaches a threshold, it triggers the trip mechanism, usually a solenoid-actuated latch that opens the contacts mechanically. No external power supply is required for this trip path on a standard electromechanical RCCB.

What we see in the field: a "nuisance trip" complaint is often not nuisance at all — it's cumulative leakage from several loads (long cable runs, EMI filters, multiple SMPS devices) adding up under one RCD until the total crosses IΔn. Splitting the load onto two RCDs frequently fixes it without touching a single fault.

Rated Residual Operating Current (IΔn): Shock Protection vs Fire Protection

The IΔn rating tells you what the device is designed to protect against, not just how sensitive it is. 10 mA and 30 mA units target personal protection — additional protection against electric shock from direct or indirect contact — because 30 mA is the threshold most electrical codes treat as the safe upper limit for the trip time involved. 100 mA, 300 mA and 500 mA units exist for fire and equipment protection instead: leakage current in that range can heat insulation to ignition before it ever reaches a level that would stop a heart.

A 30 mA RCD must trip within the time limits set by IEC 61008 at rated current, and faster still at 5x IΔn. That speed matters as much as the threshold. A device that senses the fault correctly but opens too slowly provides no real protection.

See the full breakdown of RCD sensitivity levels for where each IΔn value is actually specified by code.

RCD, RCCB, RCBO, ELCB and GFCI: Same Family, Different Jobs

RCD is the umbrella term — any device whose protection function is based on sensing residual current. Under that umbrella, an RCCB (residual current circuit breaker) senses earth leakage only and carries no overload or short-circuit protection of its own. An RCBO adds an MCB's thermal-magnetic trip into the same module, so one device covers earth leakage, overload and short circuit together.

Two older or regional terms cause most of the confusion. ELCB historically meant a voltage-operated earth-leakage breaker, a design that sensed voltage on the earth conductor rather than current imbalance, now obsolete and effectively unavailable new. When someone says "current-operated ELCB" today, they mean an RCD. GFCI is the North American label for the same current-sensing principle at a much lower threshold, typically 5-6 mA, built into a receptacle or breaker under UL 943 and the NEC, rather than the 30 mA IEC devices this article describes.

For the full comparison of trip functions, standards and where each fits a panel, see RCCB vs RCBO differences and the related MCB vs RCBO vs RCD vs RCCB differences comparison.

Why an RCCB Needs a Backup MCB — and Why RCBOs Exist

An RCCB will not clear a short circuit. Its internal contacts and trip mechanism are sized for earth-fault current, not the thousands of amps a bolted short can produce, so every RCCB installation needs an upstream or downstream MCB rated for the circuit's prospective fault current. That's a two-device, two-module solution per circuit.

An RCBO collapses both functions into one module wired to a single circuit. It costs more per circuit than a shared RCCB feeding several MCBs, but it isolates a fault to one circuit instead of tripping every load under a shared RCCB, the classic complaint of "the whole board goes dark when one appliance leaks." For final circuits where nuisance-trip isolation matters, such as freezers, IT equipment, or individually metered tenancies, RCBOs are usually the better trade.

Browse RCBOs or the broader residual current devices range for both approaches.

Earthing System Changes the Requirement

How critical the RCD is depends on the earthing system behind it. On a TT system, the installation's earth electrode has a high loop impedance compared with a solidly earthed neutral, so an overcurrent device alone often can't clear a line-to-earth fault fast enough; an RCD is not optional there, it's the primary means of disconnection for earth faults. On a TN system, the MCB or MCCB can usually clear a line-earth fault on its own because the loop impedance is low, but codes still require 30 mA RCDs on socket outlets and other circuits with direct contact risk as additional protection. IT systems take a different approach entirely: insulation monitoring detects the first fault, since the system is designed to ride through it, with RCDs applied at sub-circuit level rather than as the primary protection.

This isn't a detail to skip past when specifying a panel for export. A design built around TN assumptions doesn't transfer cleanly to a site running TT, and vice versa. See the full breakdown of earthing systems and RCD requirements for the loop-impedance numbers behind each case.

Testing an RCD in the Field

The test button on the front of the device does not test a real fault path. It injects a small artificial imbalance internally to confirm the trip mechanism moves; it proves the mechanical linkage works, nothing about the actual IΔn threshold or trip time. Proper commissioning uses a dedicated RCD tester that ramps current from zero and records the exact trip current, then separately confirms trip time at IΔn and at 5x IΔn against the limits in IEC 61008 (RCCB) or IEC 61009 (RCBO).

Skipping instrument testing at commissioning is common, and it's also how a marginal device — one that trips at 45 mA instead of a rated 30 mA — goes unnoticed until an incident report asks why. The button will have passed every time.

Conclusion

An RCD's entire function comes down to one comparison: current out through the line conductor against current back through the neutral, sensed through a toroidal core, tripped once the difference reaches IΔn. Everything else — RCCB vs RCBO, Type AC vs Type A vs Type B, 30 mA vs 300 mA, TT vs TN — is a variation on how that core measurement gets applied to a specific circuit and a specific fault risk. Get the sensitivity or the device type wrong for the load behind it, and the toroidal core can be sensing correctly while the protection still fails to do its job. For the full engineering picture across sensitivity, type, earthing and selection criteria, see the RCD protection guide, or move on to how to select an RCD for sensitivity, type and pole count together.