RCD Testing Procedure: How to Test RCDs per BS 7671

RCD (Residual Current Device) testing verifies that every RCD in an electrical installation — whether RCCB (Residual Current Circuit Breaker), RCBO (Residual Current Breaker with Overcurrent protection), or socket-outlet RCD — operates correctly at the specified current levels and within the required trip times. RCDs provide additional protection against electric shock by detecting small leakage currents to earth and disconnecting the supply before a lethal shock can occur.

RCD testing is test number seven in the GN3 testing sequence, performed after all dead tests and the other live tests (earth fault loop impedance and prospective fault current). It is a live test that requires the circuit to be energised and deliberately injects fault current through the earth path.

The test applies the rated residual operating current (IΔn) as an alternating current and verifies that the device operates within the required time. It is performed on both the positive and negative half-cycles of the supply waveform, because an RCD can behave differently depending on the polarity of the residual current when it appears.

A4:2026 introduced an important change to Regs 643.7.1 and 643.8: regardless of RCD type (AC, A, F, B etc.), an alternating current test at rated residual operating current (IΔn) is used to verify the effectiveness of the RCD. This means the 1x IΔn effectiveness test is always performed as an AC sinusoidal test, even for Type A, F, or B devices. Where your MFT has a type selector, select Type AC for the 1x IΔn effectiveness test; instruments without a type selector can only perform Type AC tests and this remains compliant for the effectiveness verification.

Got a reading already? Check it against the limit

Amendment 4 made this shorter than most electricians expect. Table 3A of Appendix 3 — the time/current performance criteria table that drove the old half-rated, rated and five-times routine — has been deleted. Regulations 643.7.1 and 643.8 now verify an RCD with a single alternating current test at the rated residual operating current, IΔn, whatever the device type. Type AC, A, F and B are all verified the same way.

The worst-case (longest) trip time from the 0 degree and 180 degree tests is the value recorded on the schedule of test results. If the test gives 28 ms at 0 degrees and 34 ms at 180 degrees, you record 34 ms.

The half-rated and five-times tests have not become wrong — they have stopped being required. Both remain useful when you are investigating a specific problem, and the familiar 40 ms at five times IΔn is still a real figure from BS EN 61008/61009. It describes how the device must be built, though, not what BS 7671 asks you to measure. The practical consequence on site: record the trip time at IΔn as the certified value, and do not fail a device against a criterion the standard no longer applies.

The test applies the full rated residual operating current — 30 mA for a 30 mA device — as an alternating current, on both the positive and negative half-cycles. A general (non-delay) RCD must operate within 300 milliseconds. Since Amendment 4 this is the whole of the required verification, and it applies identically to Type AC, A, F and B devices.

The 300 ms figure is a limit, not a target. Most healthy RCDs operate far faster — typically 15 to 30 ms at IΔn. A device passing at 250 ms is compliant but worth a second look: trip times drift upward as a device ages, and a reading that high on a modern installation usually means something is wrong with the device rather than the circuit.

Both half-cycles matter because an RCD's sensitivity can differ with the polarity of the residual current at the instant it appears. A device can pass at 0 degrees and fail at 180. The worst-case (longest) trip time from either half-cycle is the one recorded on the certificate.

In the app

Auto-validated RCD trip times in the schedule of tests

Enter the trip time from your instrument and Elec-Mate validates it against the current BS 7671 requirement — 300 ms at IΔn for a general device, 130 to 500 ms for a Type S.

Try it free for 7 days

Free to use, no sign-up — enter the RCD rating and the trip time your instrument gave, and it tells you whether the device passes.

Use the worst-case (longest) of your 0 degree and 180 degree readings, and check it at IΔn — that is the value that goes on the certificate. The five-times option is kept for reference only, since Amendment 4 deleted Table 3A and that test is now a diagnostic rather than part of the required verification.

These three tests used to be part of the routine. Amendment 4 deleted Table 3A of Appendix 3, which is what required them, so they are no longer part of verifying an RCD. They have not become useless — they have become diagnostic tools you reach for when investigating something specific.

Half-rated (0.5x IΔn — 15 mA on a 30 mA device). The device should not trip. This is the test to run when chasing nuisance tripping, because it distinguishes an over-sensitive device from a circuit with genuine earth leakage. BS EN 61008/61009 requires an RCD to operate between 50% and 100% of IΔn, and BS 7671 makes the same point in NOTE 2 to Regulation 531.3.3 — an RCD may operate at any value of residual current in excess of 50% of the rated residual current — so a device tripping below 15 mA is out of specification. Bear in mind that existing leakage downstream — electronic equipment, long runs, damp — adds to the test current, so a trip here is not automatically the device's fault.

Five times (5x IΔn — 150 mA on a 30 mA device). The familiar 40 ms figure comes from BS EN 61008/61009 and describes how the device must be built. The reasoning behind it is sound and worth knowing: currents of that magnitude flowing through the body for more than roughly 40 ms risk inducing ventricular fibrillation, which is why a device intended for additional protection is built to clear a heavy residual current that quickly. What has changed is that BS 7671 no longer asks you to measure it. Run it when you want the extra evidence; record the IΔn result as the certified value.

Ramp test. Rather than applying a fixed current and timing the trip, a ramp test raises the current gradually from zero and reports the current at which the device actually operates — a 30 mA RCD might let go at 22 mA. It is the most informative of the three when a device is behaving oddly, and it has never been part of the required sequence.

The AC supply waveform is a continuous sine wave that oscillates between positive and negative peaks. When a fault occurs, it can happen at any point in this waveform. The magnitude of the fault current at the instant of the fault depends on where in the waveform cycle the fault occurs — at the peak (maximum voltage, maximum current) or at the zero crossing (zero voltage, zero current).

Testing at 0 degrees applies the test current as the waveform passes through zero going positive. Testing at 180 degrees applies it as the waveform passes through zero going negative. These represent the two worst-case starting conditions for the RCD — the points where the fault current starts from zero and must build up before the RCD can detect it.

Your multifunction tester has a setting to select 0 degree or 180 degree phase angle for RCD testing. Both tests must be performed at each current level, and the longer trip time is the one recorded. Some electricians abbreviate this by only testing on one half-cycle — this is not compliant with BS 7671 and could miss a device that fails on the other half-cycle.

The ramp test is a diagnostic test that gradually increases the leakage current until the RCD trips. Unlike the standard tests which apply a fixed current and measure time, the ramp test measures the actual trip current — the precise current level at which the device operates. GN3 describes the instrument behaviour: the ramp starts from less than 50% of the RCD's rating and rises to 110% of it, and the current at which the device trips is recorded as the tripping-current threshold.

BS EN 61008/61009 requires that a general-type RCD must trip between 50% and 100% of its rated residual operating current. For a 30 mA device, this means it must trip between 15 mA and 30 mA during a ramp test. A device that trips below 15 mA is overly sensitive and a nuisance tripping risk. A device that does not trip until above 30 mA is not providing adequate protection and must be replaced.

The ramp test is not always required on the schedule of test results, but it is a valuable diagnostic tool. If you are investigating nuisance tripping, the ramp test can confirm whether the RCD itself is overly sensitive or whether the issue is background leakage current on the circuit.

Every RCD has a test button on its front panel marked "T" or "Test." Pressing this button creates an artificial imbalance in the sensing toroid by passing a small current through an internal resistor, simulating a fault. The RCD should trip immediately.

The push-button test confirms three things: (1) the trip mechanism is free and operates correctly, (2) the contacts separate cleanly, and (3) the mechanical reset mechanism works. It does NOT confirm that the RCD will trip at the correct current level or within the required time — for that, you need the full instrument test sequence.

The push-button test should be performed first, before any instrument testing. If the button test fails (the RCD does not trip when the button is pressed), the device is faulty and must be replaced — there is no point proceeding with instrument tests. Occupants should be advised to press the test button six-monthly as a routine maintenance check between periodic inspections — this is the interval specified in the RCD instruction notice required by BS 7671 Reg 514.12.2. (A4:2026 added an exception to that regulation for domestic premises where certification or an EICR complete with the Appendix 6 guidance for recipients has been issued.) Note also that Regulation 643.10 requires the effectiveness of any test facility incorporated in an RCD to be verified, so the button check is itself a required part of initial verification — it is simply not the trip-time test.

Time-delayed RCDs, designated Type S (selective), have an intentional time delay built into their operation. They are designed to achieve discrimination with downstream general-type RCDs — when a fault occurs, the downstream device should trip first, without the upstream Type S device also tripping.

Note the difference that catches people out: a Type S device has both a maximum and a minimum. Operating faster than 130 ms at IΔn is a failure, not a good result — the device is not providing the intended delay, and the downstream general-type RCD will no longer discriminate against it. Operating slower than 500 ms is also a failure. Both are recorded as such.

RCD discrimination ensures that when a fault occurs, only the RCD nearest to the fault trips — the upstream RCD remains closed, maintaining supply to other circuits. This is essential in modern consumer units where a main switch RCCB protects multiple circuits that also have individual RCBOs.

To achieve discrimination, the upstream device must be a Type S (time-delayed) RCD and the downstream device must be a general-type RCD. The time delay in the Type S device gives the downstream device time to operate first.

Discrimination between two general-type (non-time-delayed) RCDs cannot be guaranteed because both devices will attempt to trip at the same speed. One will trip first, but which one trips is essentially random and depends on the manufacturing tolerances of each device.

RCD test results are recorded on the schedule of test results attached to the EICR or EIC. For each RCD, you must record the type (RCCB, RCBO, socket-outlet), the rated residual operating current (typically 30 mA), and the operating time measured at IΔn.

The worst-case (longest) trip time from the 0 degree and 180 degree half-cycle tests is the value recorded. If the test at IΔn gives 25 ms at 0 degrees and 32 ms at 180 degrees, you record 32 ms. Since Amendment 4 there is no second figure to enter alongside it — if you ran a five-times test as a diagnostic, that result is not the certified value.

In the app

Voice to test results — speak trip times hands-free

On site with your instrument in one hand? Just speak: 'RCBO 1, trip time 28 milliseconds.' Elec-Mate fills in the schedule of test results and validates it.

Try it free for 7 days

If the test fails, record the failure and classify the observation on what the device actually protects. An RCD that does not operate within 300 ms at IΔn on a circuit relying on it for additional protection is normally C2 (potentially dangerous). A device that proves over-sensitive on a diagnostic half-rated test may be C3 (improvement recommended) where it is merely causing nuisance tripping, or C2 where the nuisance is likely to lead occupants to bypass the device altogether.

Several common issues can affect RCD test results or cause confusion during testing.

In the app

EICR and EIC forms capture all RCD test data

Elec-Mate's digital EICR and EIC forms have dedicated fields for RCD results — the operating time at IΔn on both half-cycles, device type and rating.

Try it free for 7 days