Nuisance tripping of an RCD — diagnosing cumulative leakage current

Testing an RCD — trip time and operating current describes how an RCD is periodically tested against the operating current it was designed for. This article covers a different, often confusing problem: an RCD that trips during normal operation without any traceable fault (insulation defect, short circuit) being found — and that, when tested, simply performs within the standard.

Why "no fault found" is not the same as "no leakage current"

Every load with an EMC filter — variable frequency drives, LED drivers, switch-mode power supplies, computer equipment — leaks a small, continuous current to earth via the filter capacitors between phase/neutral and the filter's earth connection. This is not a defect: it is inherent to how the EMC filter works and stays within the leakage current limits stated by the manufacturer for the device. The problem arises when several such devices are connected to the same circuit, or behind the same RCD: their individual leakage currents add up. A single device with 1-2 mA of leakage current is harmless; ten such devices on one circuit can together already cause 10-20 mA of background leakage current — well before anything is actually defective.

The rule of thumb: one third of the rated operating current

In practice, the rule of thumb often applied is that the continuous background leakage current of a circuit should not exceed approximately one third of the rated residual operating current (IΔn) of the RCD. For a standard 30 mA RCD this means a practical margin of around 10 mA of background leakage current, before a small additional switching-peak leakage current (for example when another device is switched on, or a brief mains disturbance) is enough to still trip the RCD. This margin is not a normative value from an IEC standard, but a practical rule of thumb to keep sufficient distance from the operating threshold without there being an actual insulation defect.

Variable frequency drives as a separate source

Variable frequency drives are a particularly relevant source of this kind of leakage current, because the high-frequency switching edges of the output transistors leak to earth via the parasitic capacitance of the motor cable — an effect that increases with cable length and the drive's switching frequency. See also DC bus residual voltage in variable frequency drives for another practical point of attention with the same devices. This high-frequency leakage current is also one reason why a Type B RCD is often required for variable frequency drives instead of an ordinary Type A — see the overview of RCD types AC/A/F/B — though nuisance tripping is primarily about the amount of leakage current, not only its waveform.

How this is diagnosed in practice

Measure the background leakage current of the circuit with a leakage current clamp (a clamp meter placed around all active conductors together) while the load is running normally — not only after a trip has already occurred.
Compare this with the rule of thumb of one third of the IΔn of the RCD protecting that circuit.
Split the circuit if the background leakage current approaches or exceeds the margin — spread the leakage-current-causing devices over several circuits with their own RCDs instead of placing them all behind one RCD.
Give a variable frequency drive with a long motor cable its own circuit with its own (preferably Type B) RCD, rather than sharing it with other sensitive equipment.

Practical relevance

For a complaint about an RCD that trips "for no reason" without the periodic test (trip time and operating current) showing any deviation, measuring the cumulative background leakage current of the circuit — and comparing it with the one-third rule of thumb — is often more informative than replacing the RCD itself, which rarely solves the problem if the cause lies in the summed connected load.

Common mistakes

Replacing the RCD without measuring the background leakage current — if the cause is cumulative leakage current from the connected load, a new RCD of the same type solves nothing.
Connecting all EMC-filtered equipment to one circuit without accounting for the sum of their individual leakage currents.
Installing an RCD with a higher IΔn to stop the tripping, without assessing whether the required maximum touch-voltage disconnection time for that circuit still allows it.
Ignoring variable frequency drives with long motor cables as a separate leakage current source, while their high-frequency capacitive leakage current via the motor cable is precisely one of the larger contributors.