Residual Current Device Engineering Guide (RCD, RCCB, RCBO)

What is a residual current device? An RCD is a protective device that continuously compares the current flowing out through the line conductor against the current returning through neutral, and trips within milliseconds once the imbalance — the current leaking to earth — exceeds a rated threshold IΔn, per IEC 61008/61009. A 30 mA device clears a shock-level fault before the current reaches the level associated with cardiac risk at typical contact durations, which is why most electrical codes mandate it on socket-outlet circuits and on TT installations generally. This guide covers the toroidal-core sensing principle, the sensitivity classes from 10 mA to 500 mA, the waveform types (AC, A, F, B, SI), RCCB vs RCBO vs ELCB vs GFCI terminology, earthing-system requirements, selectivity between board levels, EV/PV/VFD special cases, a selection checklist, and how Schneider, ABB and Siemens ranges compare.

How an RCD Detects Earth Leakage: The Toroidal Core Principle

Every RCD works around one component: a toroidal (ring) current transformer with the line and neutral conductors threaded through the core as primaries, and a sense winding wrapped around the ring as the secondary. In a healthy circuit, the current going out through line equals the current coming back through neutral, so the two magnetic fields cancel and the sense winding sees close to zero flux. Earth leakage — current returning to source through an unintended path, a person, damp insulation, a cracked cable sheath — breaks that balance. The core now carries a net flux proportional to the leakage current, which induces a voltage in the sense winding.

On a purely electromechanical RCCB, that induced voltage drives a small trip coil directly; there is no amplifier, no auxiliary supply, and nothing in software between the fault and the mechanism. That is deliberate: an RCD has to trip even during a supply anomaly, so the sensing-to-tripping path stays passive on the base design. Electronic RCDs, common in some low-cost residential ranges, add an amplifier stage ahead of the trip relay, which can improve sensitivity at low IΔn but introduces a dependency on auxiliary power worth checking before specifying — a device marketed as "electronic" rather than "electromechanical" can lose protection if that auxiliary supply fails on some designs.

The full component-level breakdown, including test-button behavior and why it only proves the mechanism rather than the calibrated trip current, is in how an RCD works.

RCD Sensitivity Classes: 10 mA to 500 mA

IΔn sets what the device is protecting. 10 mA and 30 mA devices exist for personal protection: additional protection against direct contact, the scenario where someone touches a live part or a faulty appliance chassis. 30 mA is the sensitivity most electrical codes specify for socket outlets and portable-equipment circuits, sized so it clears within the IEC 61008 time limits well below the current level linked to cardiac risk at typical fault durations. 10 mA devices go where the contact risk is higher or the tolerance for shock current is lower — some medical locations, some bathroom zones, some jurisdictions' childcare wiring rules.

100 mA, 300 mA and 500 mA devices are fire and equipment protection, not shock protection. Leakage at those levels will not necessarily injure a person on contact, but sustained arcing at a loose connection or degraded insulation can ignite the surrounding material well before it reaches a level a 30 mA device would even notice at the main incomer. A 300 mA RCD on the main switchboard, backed by several 30 mA RCBOs downstream, is a common two-tier arrangement: the main device catches slow-building fire risk across the whole installation while each final circuit still gets shock protection.

What we see in the field: panel builders sometimes fit a single 30 mA RCD ahead of an entire distribution board to save one module's width. It works until the cumulative leakage of dozens of small SMPS loads (LED drivers, VFDs, PLCs) pushes the standing leakage current close enough to 30 mA that one additional fault, or just a humid day, trips the whole board. Splitting sensitive circuits across separate 30 mA RCBOs, with a higher-rated device only at the main, avoids that failure mode. The code-specific thresholds are covered in RCD sensitivity classes.

RCD Types by Waveform: AC, A, F, B and SI

IΔn tells you how sensitive an RCD is. Type tells you what it can actually see. IEC 62423 and IEC 61008-1 define the waveform classes because a toroidal core sized only for pure sinusoidal leakage will not detect, or will under-detect, the residual currents that modern electronic loads actually produce.

Type AC senses sinusoidal AC residual current only. It is the oldest class and it is now a poor fit for most commercial and industrial boards, because almost every modern load has a rectifier stage somewhere in it. Type A adds detection of pulsating DC residual current superimposed on the AC waveform, the residual signature of single-phase rectifiers, SMPS, dimmers and most electronic equipment; it is the practical default for general distribution today. Type F extends Type A to cover mixed-frequency residual currents from single-phase variable-frequency drives and similar frequency-converted loads. Type B goes further and detects smooth, pure DC residual current, the kind produced by three-phase VFDs, EV chargers and transformerless PV inverters — currents a Type A device cannot see at all, not partially, not at reduced sensitivity. At all.

Type AC being a poor fit for electronic loads is not a minor caveat: a smooth DC fault current can saturate an AC-only or Type A core so it stops responding correctly to any residual current, AC included. That is why several national codes now restrict or ban Type AC on new circuits feeding electronic equipment. The full type-by-type breakdown, including which standard clause covers each class, is in RCD types explained.

RCCB vs RCBO vs ELCB vs GFCI: Terminology

RCD is the umbrella term covering both. ELCB, an older acronym, historically meant a voltage-operated earth-leakage breaker that sensed the voltage on the earth conductor itself rather than the residual current; that construction is obsolete and largely out of production. Where "ELCB" is still used informally today, it almost always means a current-operated device, an RCD in modern terminology, and the label persists out of habit rather than a different technology. GFCI is the North American equivalent term, referring to a 5-6 mA residual device built into a receptacle or breaker under UL 943 and the NEC, a much tighter sensitivity than the 30 mA standard used elsewhere because U.S. receptacle-level protection targets direct hand contact specifically.

Schneider's Vigi block and Acti9 iDPN Vigi, ABB's DS201/DS202C, and Siemens' 5SU1/5SV1 are all RCBOs, built two different ways: Schneider's Vigi is an add-on earth-leakage block that clips onto an existing iC60 MCB to form the combined function, while ABB and Siemens sell the one-piece integrated RCBO as the base part number. Neither approach changes the protection delivered; it changes board space, spare-parts strategy, and whether earth-leakage protection can be added to an MCB already installed.

The full decision tree for RCCB-plus-MCB against a one-piece RCBO is in the differences between MCB, RCD, RCCB and RCBO.

Earthing Systems: TT, TN and IT — Where RCDs Are Required

Whether an RCD is essential or supplementary depends on the earthing system, not on preference. On a TT system, the installation's earth electrode is independent of the supply earth, so loop impedance back to the source is typically high, often too high for an MCB or fuse to clear a line-to-earth fault within a safe disconnection time on overcurrent alone. An RCD is the primary means of achieving fast disconnection on TT, and the required trip current has to satisfy a touch-voltage limit, not just trip eventually.

Work the formula backward on a real site: a TT electrode measuring 200 Ω limits you to IΔn ≤ 250 mA to stay under the 50 V rule, comfortably met by a 30 mA device. It also shows why an installation with a poor electrode, say 800 Ω, still passes with 30 mA while it would fail outright at 300 mA with no lower-rated RCD downstream.

TN systems (TN-C-S / TN-S) usually let the MCB or MCCB clear a line-to-earth fault on overcurrent alone, because the fault path runs through a low-impedance metallic protective conductor rather than the earth mass. That does not remove the requirement for 30 mA RCDs on socket-outlet circuits and other additional-protection scenarios in most codes; it just means the RCD there is supplementary, guarding against contact faults the overcurrent device would never see — a fault current low enough to never trip the MCB but high enough to be dangerous through a person. IT systems take a different first line of defense: an insulation monitoring device supervises the whole system for a first fault, since IT is designed to keep running through one fault without tripping. RCDs still appear on IT sub-circuits and final distribution, sized so they do not nuisance-trip on the system's normal, small unbalance currents.

Full earthing-system requirements, including where IT sub-circuit sizing differs from TT and TN, are in RCD requirements by earthing system.

RCD Selectivity: S-Type and Time-Delayed Discrimination

Stack two RCDs on the same feeder, an upstream 300 mA at the main board and a downstream 30 mA at a final circuit, and an uncoordinated pair will sometimes both trip on a single fault, killing the whole board when only the faulted circuit needed to drop. Selectivity (discrimination) avoids that. The rule has two parts: the upstream device's IΔn must be at least twice the downstream device's IΔn, and the upstream device must carry a time delay so the downstream instantaneous device gets the chance to clear the fault first.

S-type (selective) RCDs build in that time delay, typically tens to a couple hundred milliseconds depending on the fault current level, plus a slightly higher immunity to transient nuisance tripping. That is why S-type devices also show up at the head of installations with known transient leakage even where selectivity with a downstream RCD isn't the only driver. G-type (general, short time delay) sits between instantaneous and full S-type coordination, used in some three-tier board designs. None of this is optional bookkeeping: an upstream device without the delay, even at double the downstream IΔn, can still trip simultaneously with the downstream device on a fast-rising fault current, because instantaneous devices at any IΔn respond to the same waveform edge.

This depends on the actual let-through current and the manufacturer's time/current curves for the specific model pair. A "2x IΔn plus S-type" rule of thumb gets you most of the way, but full discrimination on a given board should be checked against the supplier's coordination tables for that combination, not assumed from the class labels alone.

Special Applications: EV Chargers, Solar PV and VFDs

Three load types force the Type B question: EV chargers, PV inverters and three-phase VFDs, all of which can produce a smooth DC fault current that a Type A or Type F device cannot detect at all. For an EV charge point, the requirement is protection against that DC fault current, met either with a Type B RCD on the circuit, or with a Type A RCD combined with a 6 mA DC residual current detection device (RDC-DD) built into the charger or a dedicated RCBO under IEC 62955. The RDC-DD route is common in charger hardware because it lets the installation keep a Type A device upstream rather than specify Type B at every charge point, which matters on multi-point installations where Type B devices cost noticeably more per pole.

Transformerless PV inverters can inject a DC component onto the AC side under certain fault conditions, which is why most inverter manufacturers specify Type B on the AC output circuit unless the inverter itself carries an integrated residual current monitoring unit that meets the same detection requirement. Check the inverter's own compliance statement before assuming a Type A device is sufficient just because the installation isn't a charger.

Three-phase VFDs producing smooth DC leakage under certain fault modes need the same Type B treatment on the supply side. Single-phase VFDs are usually Type F territory since their residual signature is a mixed frequency rather than pure DC, but this varies by drive topology and rectifier design, so treat "single-phase equals Type F is enough" as a starting assumption to verify against the specific drive's manual, not a fixed rule.

The full side-by-side on EV charger circuits is in EV charger RCD requirements, and drive and inverter selection is covered in Type A vs Type B RCD selection.

Selecting an RCD: Sensitivity, Type, Poles and Rated Current

Four parameters have to be right, not three: sensitivity (IΔn), type (waveform class), poles (2P for single-phase, 4P for three-phase) and rated current In, matched to the circuit's design load and the upstream protective device.

Work it in that order. Start with sensitivity: 30 mA for any circuit with a direct-contact risk (sockets, portable equipment, most final circuits), 10 mA where the code or the location demands tighter protection, 100-500 mA only at main incomers for fire and equipment duty where a downstream 30 mA device already covers shock protection. Then type: default to Type A unless the load is a known Type F or Type B case (VFDs, EV chargers, PV inverters), and don't specify Type AC on new work feeding any electronic load. Then poles: match the supply, 2P for single-phase final circuits, 4P for three-phase feeders and boards, and check whether the installation needs all four poles switched or a 3P+N arrangement, since that changes fault coverage on a lost-neutral condition. Finally, rated current In: size to the circuit's design current with the standard steps (25/40/63/80/100 A and up), never below it, and check the RCCB's rated conditional short-circuit current Inc against the upstream backup device if it's an RCCB rather than an RCBO — an RCCB with no backup MCB has essentially no short-circuit withstand of its own.

What we see in the field: the most common sizing mistake isn't the sensitivity or type call, it's In. Engineers pick an RCCB rated at exactly the design current with no margin, then find nuisance tripping on motor inrush or transformer energizing current that a slightly higher In rating, at the same IΔn, would have avoided without changing the actual leakage protection at all.

The full selection checklist, worked through with example circuits, is at the RCD selection checklist.

Schneider Acti9, ABB F200 and Siemens SENTRON: RCD Ranges Compared

All three brands converge on the same core offering: Type A as the modern default, Type B available for DC-fault duty, and a super-immunized variant for nuisance-prone sites. The differences sit in construction and range depth rather than sensing technology.

Schneider's residual current devices center on the Acti9 iID RCCB (2P/4P, 25-100 A, Types AC/A/SI/B, IΔn 10/30/100/300/500 mA) and the Vigi iC60 add-on block, which clips onto an existing iC60 MCB to form an RCBO rather than shipping as one integrated part, useful where boards already carry iC60 MCBs and only need earth-leakage protection retrofitted. The compact RCBOs in the iDPN Vigi / Reload line cover final-circuit duty in one module, and the "si" range is Schneider's label for super-immunized.

ABB's F200 series (F202 2P, F204 4P, 25-125 A, Types AC/A/F/B, IΔn 10-500 mA, S-type selective variants) runs a wider current range at the top end than Schneider's iID, and the FH200 sits as an economy tier below it. ABB's RCBOs, DS201 (1P+N) and DS202C/DS203NC, are one-piece integrated modules rather than an add-on block, and the F200 APR variant adds auto-reclosing, useful on remote or unattended boards where a nuisance trip would otherwise mean a site visit.

Siemens' SENTRON 5SV RCCB (5SV3/5SV4/5SV6, 2P/4P, Types AC/A/F, 16-125 A) pairs with the 5SU1 and 5SV1 RCBOs for compact combined protection, and the 5SM3 is Siemens' Type B device for EV, PV and VFD duty. The 5SV "SIGRES" variant targets humid or corrosive environments specifically, and for feeders above roughly 125 A none of the three brands' fixed RCCBs apply. That range calls for a residual current relay (Siemens 5SM2, ABB RD3, Schneider Vigirex) with a separate toroid fitted around the feeder conductors rather than a device rated for the full load current itself.

Testing, Commissioning and Nuisance-Trip Troubleshooting

The front test button proves the mechanism trips: it injects an artificial imbalance and confirms the trip linkage moves. It proves nothing about the calibrated trip current or trip time, which is why commissioning needs an RCD tester: measure trip time at IΔn and at 5x IΔn, and ramp the test current up from zero to find the actual trip point against the rated IΔn. A device that passes the button test but trips at 45 mA instead of a rated 30 mA has failed calibration, not function, and the button test alone will never catch it.

Nuisance tripping, frequent trips with no obvious cause, usually traces to cumulative leakage rather than a hard fault: dozens of small SMPS loads each contributing a few hundred microamps of steady-state leakage, summing across a board until a normal load swing or added humidity pushes the total over IΔn. Splitting a board into more RCBOs each covering fewer circuits, moving to Type A or SI devices, or checking for a genuinely faulty appliance rather than assuming the RCD is over-sensitive, are the standard diagnostic steps, roughly in that order.

Full troubleshooting steps, including how to isolate which circuit is contributing the leakage without de-energizing the whole board, are at troubleshooting nuisance tripping.

Conclusion

Specifying an RCD correctly means getting four independent parameters right at once, sensitivity, type, poles and rated current, against the earthing system and load type in front of you, not picking "30 mA, Type A" by default and hoping it fits. Type A covers most modern electronic loads. Type B is non-negotiable for EV chargers, PV inverters and three-phase VFDs. Selectivity between board levels needs both a 2x IΔn margin and a time delay, not one or the other, and TT systems need the touch-voltage formula checked against actual earth resistance, not assumed. Schneider, ABB and Siemens deliver equivalent protection at a given IΔn and type; the real decision is RCBO form factor and range depth for the board in front of you. Browse Stoklink's residual current devices and RCBOs for in-stock Schneider, ABB and Siemens ranges across the sensitivity and type classes covered above.