RCBO tripping on an LED lighting circuit: cumulative leakage versus insulation fault
A retail unit is protected by a 30 mA type AC RCBO on each final circuit and a 300 mA S-type RCD at the origin. The RCBO feeding the shop-floor lighting trips about twice a week, always between 17:00 and 19:00, shortly after the lighting is switched on. The circuit supplies 24 LED high-bay luminaires with integral drivers, and there is a three-stage surge protective device at the sub-board. Insulation resistance tests give more than 200 megohms line-to-earth and neutral-to-earth at 500 V DC with the luminaires isolated. A portable tester shows the RCBO tripping at 21 mA in 25 ms. A leakage clamp on the circuit protective conductor reads 14 mA steady with all luminaires on, rising to 19 mA for a few seconds when the chillers start. The shop owner wants the RCBO replaced with a 100 mA device like the one at the main board.
1. Is there an insulation fault? Explain the measurements and the cause of the tripping.
Read the worked answer
There is no insulation fault. Insulation resistance above 200 megohms at 500 V rules out a resistive breakdown to earth, and a 1 megohm result would have been enough to pass. The tripping is caused by accumulated protective-conductor current, not by a fault.
Every LED driver contains an EMI filter with capacitors from line and neutral to the earthed case. Each driver bleeds a small current, typically 0.3 to 0.8 mA. Twenty-four luminaires at an average of 0.6 mA give about 14 mA, which matches the clamp reading exactly. Add the surge protective device, which itself has a continuous leakage component, and the steady total moves towards 17-19 mA.
A 30 mA RCBO must not operate below 15 mA and must operate by 30 mA, so a standing leakage of 14-19 mA sits inside its operating band. The chiller start adds switching transients and a few extra milliamps through driver filters, and the evening switch-on adds inrush through the driver input capacitors. The combination pushes the residual current past the RCBO's threshold. The device is not faulty - it is correctly responding to a current it cannot distinguish from a shock risk - and the 21 mA / 25 ms test result simply confirms it is within specification.
2. Evaluate the owner's proposal to replace the RCBO with a 100 mA device, and state what determines whether 30 mA protection is required on this circuit.
Read the worked answer
The proposal should be refused as presented, but for reasons that must be explained rather than asserted.
- A 30 mA RCD is required for socket-outlets and for circuits supplying portable equipment outdoors; here the important question is whether the circuit is lighting only or whether the lighting circuit also feeds shop-floor socket-outlets, display equipment or anything a person can touch while earthed. Check the as-built drawing and physically verify the circuit before deciding.
- Replacing 30 mA with 100 mA increases the current a person can receive before disconnection by a factor of more than three, and increases the let-go risk at 230 V. It also raises the touch-voltage during an earth fault, since that voltage is roughly the residual operating current multiplied by the earth path resistance.
- Masking the symptom is not fault-finding. The leakage will continue to grow as drivers age and as more LED fittings are added; a 100 mA device merely buys time before the same nuisance tripping returns, at a higher human risk.
- The 300 mA S-type at the origin is there for discrimination and for the fixed installation, not as an acceptable level of final-circuit protection.
If the circuit genuinely supplies only fixed lighting in a location where 30 mA is not mandated, a 100 mA device could be considered, but only after the leakage has been reduced and the decision documented and accepted by the responsible person. That is an engineering decision, not a convenience.
3. Give a repair strategy that keeps the 30 mA protection, with the numbers that justify each step.
Read the worked answer
Target: standing leakage on any one 30 mA device should not exceed about 30 percent of its rating, i.e. 9 mA, leaving margin for transients.
- Quantify. Clamp the CPC of each luminaire group separately with all fittings on. Identify the drivers with the highest leakage; a few failing or low-quality drivers often account for a disproportionate share.
- Split the circuit. Divide the 24 fittings into three lighting circuits of eight. At roughly 0.6 mA each, that is about 5 mA per circuit, inside the 9 mA target, with three 30 mA RCBOs retaining full protection.
- Move the SPD. A three-stage SPD at the sub-board can contribute continuous leakage and a large impulse current. Confirm its leakage contribution and, if significant, connect it upstream of the RCBO or use a type with lower standby leakage, while keeping the required protection level.
- Replace type AC with type A. LED driver and electronic ballast faults can produce pulsating DC residual current that a type AC device may not detect reliably; a type A RCBO is the appropriate choice and also improves immunity to some nuisance sources.
- Rule out the classic false trails. Check for a borrowed neutral between lighting circuits sharing the same neutral bar, a neutral-to-earth short downstream, and moisture in an outdoor fitting. These produce intermittent tripping that mimics leakage.
- Verify and record. Re-test each RCBO at rated and five-times rated residual current, record the leakage readings per circuit, and leave the schedule and a note explaining why the circuit was split so the next person does not simply fit a bigger device.