RCD behaviour and loop impedance on a village hall kitchen

What the results mean

• The device is not failing to operate - it is operating when it should not. A 30 mA RCD is permitted to hold at 15 mA; it should not trip. Tripping at ½×IΔn points to the residual current in the circuit, not the device.
• The 11 mA measured on the kitchen circuit is only part of the picture: the same RCD also serves the lighting and the outside socket, and leakage currents add. 11 mA of standing leakage plus the 15 mA test current puts the device at or over its actual threshold.

Sequence before coding

1. Measure total standing leakage at the RCD with all loads connected, using a leakage clamp.
2. Repeat the ½×IΔn test with all outgoing circuits isolated, tester connected directly across the RCD. If it still trips with nothing connected, the device is over-sensitive or faulty - replace it.
3. If it only trips with the circuits connected, the RCD is fine and the problem is cumulative leakage. As a rule keep standing leakage below roughly a third of the rating (about 9–10 mA here), so this installation is over.

Action Redistribute circuits across two RCDs (or fit RCBOs) so that the kitchen equipment is not sharing with the lighting, then re-test and record the outcome. No code against the RCD itself if it passes with circuits isolated.

The technical point A Type AC device responds to sinusoidal alternating residual current only. A variable-speed drive, many LED drivers, PV inverters and some EV charge points can produce pulsating or smooth DC residual current, which can desensitise or blind a Type AC RCD so that it does not operate when a genuine earth fault occurs.

The judgement The deciding factors are what the RCD is doing and what it is protecting:

• Here the installation is TN-C-S, the sockets have MCB fault protection and the RCD provides additional protection. Loss of the RCD does not remove the shock protection for a dead short, so a C3 with a recommendation to replace it with a Type A device is defensible.
• It becomes a C2 where the RCD is the sole means of disconnection (a TT system, or a circuit with an otherwise excessive Zs), where the DC-producing equipment is specified as needing Type A or B, or where there are several such appliances on one device.

Also check against the current edition Type AC is no longer accepted for new work, and the equipment here is exactly the sort that argues against continued use. The recommendation should say so explicitly: replace the shared Type AC RCD with Type A RCBOs, or at minimum fit a Type A device for the kitchen circuits.

Comparison

• Maximum Zs for a 10 A Type B device from the table is 4.37 Ω.
• The measured value is taken with conductors at ambient temperature; for conductors up to 35 mm² the measured value is compared with 0.8 × tabulated: 0.8 × 4.37 = 3.50 Ω as the maximum acceptable measured reading.
• 4.1 Ω exceeds 3.50 Ω. Corrected to operating temperature it is roughly 4.1 × 1.2 ≈ 4.9 Ω, above 4.37 Ω, so the 0.4 s disconnection time for a final circuit in a TN system cannot be relied on.

Conclusion This is a failure of automatic disconnection of supply - the circuit is not compliant and is coded C2. It is not a C3, because a C3 means the circuit is safe but could be improved.

Investigate, then remedy

• Ze is 0.35 Ω, so around 3.75 Ω is in the circuit and cpc - check for a loose connection before assuming it is a long run, then check the actual route and cpc size against the design.
• Remedies: reduce the device to 6 A where the load allows, improve the cpc (1.0 mm² is already unusual for a 10 A circuit in T&E), or fit a 30 mA RCD to provide disconnection - permissible on TN where the RCD is verified to operate within the required time.
• Do not simply re-test and hope; the reading needs to be explained before the report is signed.