How to Find the Circuit Causing RCD Trips
How do you find which circuit is tripping an RCD? An RCD trips on the vector sum of leakage across every circuit sharing its toroidal core, so a trip at 30 mA (or 100 mA on a fire-protection device) tells you the threshold was crossed somewhere downstream, not which cable, appliance, or terminal caused it. Chase the wrong circuit and the fault stays live while the device keeps opening on reconnection, burning site hours on parts that were never at fault. This piece covers sequential isolation, reading mixed RCBO/RCCB boards before touching a breaker, clamp-meter leakage measurement, intermittent faults from moisture and VFDs, and the load types behind most repeat trips.
Why the Board Only Tells You a Threshold Was Crossed
A single RCCB or RCBO measures one number: the vector sum of current entering and leaving through its own toroidal core. If that sum exceeds IΔn for long enough, the device opens. It has no way to report which downstream conductor, joint, or appliance contributed the imbalance — the toroid sees the combined return current from every circuit behind it, not each one separately.
On a board with one main 30 mA RCCB feeding six or eight MCBs, a fault on the water heater circuit and a fault on the outdoor socket circuit look identical at the RCD. The trip is real; the location isn't given for free. Locating the offending circuit is a process of elimination, not a readout.
Sequential Isolation: Split and Test
Switch off every MCB or sub-breaker downstream of the tripped RCD, then reset it. Close one breaker, wait several seconds, and watch. If the RCD holds, move to the next breaker and repeat. The breaker you close immediately before the next trip carries the fault, or at minimum the dominant leakage contribution.
Healthy circuits are not perfectly clean. Cable capacitance, EMI filters in switch-mode supplies, and long conductor runs all contribute standing leakage current, typically a fraction of a milliamp up to a few milliamps per circuit. On a board with several such circuits already closed, cumulative standing leakage can sit close to the trip threshold before the actual fault circuit is even added — so re-test isolated circuits individually rather than assuming the last one closed is always the culprit.
Label each breaker as you go. On a board with unlabeled ways, this step alone often finds the fault faster than any instrument.
Reading the Board Before You Isolate Anything
Check the topology first. A board wired with individual RCBOs per final circuit already tells you which one tripped — no isolation sequence needed, since MCB and RCD function are combined per RCCB vs RCBO differences. A board with one upstream RCCB feeding several separate MCBs requires the full sequential method, because none of those MCBs carry any leakage-detection function on their own.
What we see in the field: many older consumer units run a single 30 mA RCCB across the whole board, with lighting and socket circuits both downstream of it. It saves board space, but one faulty extension lead in a garage circuit takes out lighting on the same RCD. A split-load board — two RCDs, each covering half the ways — narrows the search to one half before touching a single MCB.
Where selectivity is in play, an upstream S-type device backs downstream instantaneous devices per RCD, RCBO, RCCB and MCB differences. A trip at the main confirms the fault current exceeded even the upstream threshold — check whether a downstream device should have opened first and didn't.
Confirm the Earth Path Before Chasing the Load
On a TT installation, an elevated earth electrode resistance doesn't cause nuisance trips by itself, but it changes how much margin the installation has. IEC 60364-4-41 sets the touch-voltage limit that links electrode resistance to sensitivity.
Formula: TT earth-fault touch-voltage limit — Source: IEC 60364-4-41 (TT system disconnection requirement)
RA × IΔn ≤ 50 V
| Symbol | Description | Unit |
|---|---|---|
| RA | Resistance of the earth electrode and protective conductor to the exposed conductive part | Ω |
| IΔn | Rated residual operating current of the RCD protecting the circuit | A |
| 50 V | Maximum permitted touch voltage under fault (conventional limit) | V |
If RA measures high relative to IΔn, the installation is already close to the design limit — a modest additional leakage current from a degraded cable can push touch voltage over 50 V well before the RCD's rated threshold is reached in isolation. Measure RA with an earth-fault-loop or electrode tester before spending an afternoon on sequential isolation; a bad electrode connection sometimes turns out to be worth fixing on its own regardless of which circuit trips.
Measure Leakage Directly With a Clamp Meter
A milliamp-range clamp meter reads net residual current without tripping anything. Clamp around both the line and neutral conductors together on the main tails; the reading is the same vector sum the RCD sees. Then clamp each sub-circuit's line-and-neutral pair in turn, live, without switching anything off.
The circuit with the highest reading is the dominant contributor. This method finds the fault in one pass instead of the repeated trip-and-reset cycle of sequential isolation, and it works on circuits you can't easily de-energize — a running production line, a fridge full of stock, a server room.
When the Fault Only Shows Up Sometimes
An RCD that trips only in wet weather, only at night, or only when a particular machine starts is not malfunctioning — it's reporting a leakage path that isn't always present. Moisture ingress at an outdoor socket, a garden lighting joint, or a cable gland lowers insulation resistance only while it's damp; the same circuit can measure clean on a dry afternoon.
Variable-frequency drives and other switch-mode loads add a different variety of intermittent leakage: their EMI filters route high-frequency noise to earth, and the resulting current depends on motor speed, cable length, and how many drives share the same RCD. This depends on the cumulative leakage from every VFD on the board, not just the one you suspect — a Type A device can pass a single small drive fine and still nuisance-trip once a second drive is added downstream of the same core, a scenario covered in more detail under Type A vs Type B RCD for VFD, EV and solar loads.
If a fault won't reproduce on demand, test under the condition that triggers it: hose down the suspect outdoor circuit (safely, isolated first) and re-test, or run the drive at the speed and duty cycle it normally operates at rather than idling it.
Circuits and Loads That Repeat Offend
Certain circuit types account for a disproportionate share of repeat callouts. Outdoor sockets and garden lighting take the most moisture-related hits. Water heaters and immersion elements develop leakage as the element ages, often gradually enough that the trips start as occasional and become daily. Long cable runs add capacitive leakage proportional to length — a run that was marginal at 20 meters can push a shared RCD over threshold once it's extended.
Length matters on its own. Multiple VFDs, EMI-filtered PSUs, or LED drivers on one RCD sum their standing leakage even when every individual device is within its own limits, which is why RCD sensitivity ratings from 10 mA to 300 mA matter when deciding how many such loads can share one device, rather than assuming a higher-rated RCD alone fixes it.
Once the offending circuit is confirmed, the fix is either repair (reterminate, dry out, replace the failed appliance) or redistribution — moving some loads to a separate RCD or RCBO so no single device carries the combined standing leakage of the whole board. For persistent nuisance trips with no measurable insulation fault, causes and fixes are covered separately in why an RCD keeps tripping, and earthing-system context for TT, TN and IT boards is in RCDs in TT, TN and IT earthing systems. For background on the device itself, see the RCD protection guide, and browse residual current devices or RCBOs for replacement units once the faulty circuit and device are identified.
Frequently Asked Questions
Can the RCD test button tell me which circuit is faulty?
No. The test button injects an artificial imbalance to confirm the trip mechanism still functions; it says nothing about which downstream circuit is leaking. Fault location needs sequential isolation or a clamp-meter leakage check.
Why does my RCD only trip when it rains?
Moisture at an outdoor socket, cable gland, or lighting joint lowers insulation resistance temporarily, letting leakage current rise above IΔn only while the path is wet. The same circuit can test clean on a dry day, so reproduce the wet condition before re-testing.
Can two separate circuits combine to trip one RCD?
Yes. Every circuit downstream of a shared RCCB adds its own standing leakage current to the same toroidal core. Two circuits that are each within limits individually can still sum to more than IΔn on that shared device.
Does a Type B RCD make fault-finding harder than a Type A?
Not fundamentally — the isolation and clamp-meter methods are the same. What changes is the load mix: Type B installations typically cover EV chargers, PV inverters, or VFDs, where smooth DC leakage can be present continuously rather than only under fault, so a healthy reading needs to be taken under normal operating load, not at idle.
Should the RCD be replaced if no fault circuit can be found?
Only after cumulative standing leakage from all downstream circuits has been measured and ruled out as the cause. A device that trips just above the combined normal leakage of several legitimate circuits is not necessarily defective — redistributing loads across two RCDs is often the correct fix, not replacement.
How long should I wait between closing each breaker during isolation?
A few seconds is enough for a hard fault to show. Intermittent faults from moisture or load-dependent leakage may need the circuit left on longer, under the condition that normally triggers the trip, before ruling it out.
Conclusion
Finding the circuit behind an RCD trip is elimination, not diagnosis by readout. Start with the board topology — individual RCBOs point directly at the fault, a shared RCCB doesn't. Sequential isolation and clamp-meter leakage measurement both work; the clamp meter is faster and doesn't require repeated trips on circuits that are hard to de-energize. On TT systems, rule out an elevated earth electrode resistance before assuming the fault is purely downstream. Once the offending circuit is confirmed, decide between repair and redistribution — moving some loads to a second RCD or RCBO is often cheaper than chasing a fault that turns out to be several legitimate circuits summing their leakage on one device.
