9.1 RCD Testing

Purpose of RCD Testing

A residual current device (RCD) disconnects the supply when it detects a difference between the line and neutral currents — current that is leaking to earth through an unintended path. Verifying an RCD confirms that it will actually operate, and operate quickly enough, when that happens.

BS 7671 uses RCDs for two distinct jobs, and the exam expects you to separate them:

• Fault protection — on a TT system, and on some TN circuits, the RCD is the device that achieves automatic disconnection of supply. Its operation is verified as part of Regulation 643.7 (protection by automatic disconnection of supply), specifically Reg 643.7.1.
• Additional protection — a 30 mA RCD fitted under Regulation 415.1 to protect against direct contact, for example on socket-outlet circuits and cables concealed in walls. Its effectiveness is verified under Regulation 643.8.

A third, separate check sits under Regulation 643.10 (functional testing): operating the device’s own test facility — the integral "T" button — to confirm the test facility works.

RCD verification is a live test, performed after the dead tests and after the loop impedance measurements. The device needs a live supply to drive a residual current through its sensing core, so it cannot be verified dead.

RCD Types and Detection Types

Four mechanical forms appear in installations:

• RCCB (residual current circuit breaker) — provides earth-leakage protection only, with no overcurrent protection; it requires a separate MCB or fuse for overcurrent.
• RCBO (residual current circuit breaker with overcurrent) — combines an RCD and an MCB (miniature circuit breaker) in one device, protecting a single circuit for both fault and overcurrent.
• SRCD (socket-integral RCD) — built into a socket-outlet.
• PRCD (portable RCD) — a plug-in portable device used with portable equipment.

Four detection types are classified by the residual current waveform they respond to:

• Type AC — responds to sinusoidal alternating residual current only. BS 7671 restricts its use, and Type A is the practical minimum for general circuits.
• Type A — responds to AC and pulsating DC residual currents.
• Type F — responds to AC, pulsating DC and composite residual currents from single-phase frequency-converter loads such as some washing machines and inverter drives.
• Type B — responds to AC, pulsating DC and smooth DC; used where smooth DC residual current can occur, for example EV (electric vehicle) charging equipment and PV (photovoltaic) installations.

Selecting the correct detection type matters: a Type AC RCD on an EV charger may be blinded by DC leakage and fail to trip. Type A is the baseline expectation for general installations; Type B is required where smooth DC can be present.

What BS 7671:2018+A4:2026 Actually Requires

This is the single most-changed area of RCD testing in recent amendments, and stale revision material is everywhere. Amendment 2:2022 simplified the requirement, and Amendment 4:2026 deleted Table 3A of Appendix 3 altogether.

The verification the current standard requires is:

An alternating-current test at the rated residual operating current (IΔn), whatever the device type (AC, A, F or B).

Connect the tester between line and earth on the load side of the RCD, and test at both the 0° and 180° start points — a residual fault can begin in either half-cycle, and the core response can differ between them. Record the longer (worst-case) time on the schedule of test results, not the average and not the best.

What happened to 0.5× and 5×?

The familiar 0.5× / 1× / 5× ladder came from the 17th Edition and earlier practice:

Most multifunction testers still offer all three settings, and many centres and scheme providers still ask for the full ladder because it is good diagnostic practice. That is fine — but if the exam asks what BS 7671 requires, the answer is the single AC test at IΔn. Answers built on "0.5× must not trip" are testing the superseded rule.

Where the 40 ms figure really comes from

The 40 ms limit is a product-standard characteristic, not a BS 7671 field-test requirement. Under BS EN 61008-1 and BS EN 61009-1, for a general-type device the manufacturer may declare either 5 IΔn or 0.25 A as the residual current at which the 40 ms maximum break time is achieved. Several manufacturers have adopted 0.25 A (250 mA).

The practical consequence: for a 30 mA device, 5 IΔn is 150 mA, but a device declared to the 0.25 A characteristic is only guaranteed to achieve 40 ms at 250 mA. Testing it at 150 mA and rejecting the result as a fail is a real-world error — and a plausible exam distractor. Read the manufacturer’s data before deciding the test current. This applies to general-type devices regardless of whether they are Type AC, A, F or B; it is not specific to Type A.

Additional Protection and the Shock-Protection Rationale

Where a 30 mA RCD provides additional protection under Regulation 415.1 — socket-outlet circuits, cables concealed in walls at a depth of less than 50 mm, circuits in locations containing a bath or shower — Regulation 643.8 requires its effectiveness to be verified. In practice that is the same instrument test at IΔn described above, applied to the device that provides the additional protection.

The reason additional protection exists at all is shock physiology. A 30 mA residual current is below the threshold at which a healthy adult is likely to suffer ventricular fibrillation provided the exposure is brief; the fast disconnection is what makes the protection work. That is the origin of the 40 ms product characteristic — it is designed to clear the fault before the current can span the vulnerable part of the cardiac cycle.

Two exam points follow from this:

• Additional protection is additional. It supplements basic and fault protection; it never replaces them. An RCD does not excuse a missing or high-resistance protective conductor.
• A 30 mA RCD limits the duration of the shock, not the magnitude. You can still receive a painful shock from a circuit protected by a working 30 mA RCD.

Type S (Time-Delayed / Selective) RCDs

Type S RCDs are deliberately time-delayed so that they discriminate with RCDs downstream of them. The delay is a design feature, so a Type S device has both a minimum and a maximum operating time.

At IΔn, a Type S device must operate between 130 ms and 500 ms.

Both ends of that band matter:

• Operating slower than 500 ms is a failure — the device is too slow to give protection.
• Operating faster than 130 ms is also a failure — the device has lost its selectivity delay, so it may trip together with (or instead of) the downstream RCD, disconnecting circuits that were never affected by the fault.

A common inspection error is recording a Type S time of, say, 180 ms at IΔn and rejecting it as "too slow". It is comfortably inside the 130–500 ms band and is compliant. Conversely, a Type S operating in 80 ms is non-compliant even though it looks fast.

Where a Type S device is fitted upstream of general-type RCDs or RCBOs, that arrangement is the whole point: a fault on one final circuit should trip only that circuit’s device, leaving the rest of the installation energised.

What Amendment 4:2026 Changed

BS 7671:2018+A4:2026 was published on 15 April 2026, with the previous A2:2022/A3:2024 text due to be withdrawn on 15 October 2026. For RCD verification the headline change is a deletion:

• Table 3A of Appendix 3 has been removed. That table set out the maximum operating times used for RCD testing. With it gone, the verification requirement rests on the single alternating-current test at IΔn introduced by Amendment 2:2022, together with the product-standard characteristics of the device itself.

Amendment 4 also reworked Part 6 and the Appendix 6 model forms more broadly — the EICR observation-code outcome rules changed as well, which is covered in section 10.3.

For the exam, hold two things in your head at once: what the current standard requires (the IΔn test) and what is still done in practice (the full 0.5× / 1× / 5× ladder at both polarities, because scheme rules, client specifications and habit all keep it alive). A question asking what BS 7671 requires is asking for the first; a question about good practice on site may be asking for the second.

Instrument Use and the Integral Test Button

An RCD tester provides selectable multiples of IΔn (typically 0.5, 1 and 5), a start-polarity selector (0° / 180°), and often a ramp test that increases the residual current gradually to find the actual current at which the device operates. Connect between line and earth on the load side of the device, disconnect loads where practicable, and be aware that standing earth-leakage from connected equipment adds to your injected current and shortens the measured time.

The integral test button (T) is a different thing entirely. Operating it diverts a small current through an internal resistor to prove the mechanism is not seized. It is a functional check of the test facility under Regulation 643.10 — it measures nothing and proves nothing about operating time. A device with a working button can still be far outside its time limits.

The notice required by Reg 514.12.2, instructing the user to operate the test button every six months, is a user-maintenance instruction, not a substitute for the inspector’s instrument test. Note the interval: earlier editions of BS 7671 called for a quarterly user test, and BS 7671:2018 changed it to six months, so a notice printing the old wording is an observation rather than a compliant label.

Recording the results

Record on the schedule of test results:

• the operating time at IΔn, taking the longer of the 0° and 180° results;
• the device type and rating (IΔn, and whether general or Type S), because the pass criterion depends on it;
• any additional diagnostic results your centre or client requires.

An RCD that fails its time limits is a defect. On new work it must be corrected before the Electrical Installation Certificate is issued; at periodic inspection it is normally a C2 observation on the EICR, because additional protection that does not operate in time is potentially dangerous rather than immediately dangerous.