RCD Keeps Tripping? A Circuit-by-Circuit Leakage Troubleshooting Workflow

An RCD that trips once after a thunderstorm is inconvenient. An RCD that trips every few days, usually at 02:00 and never while the electrician is standing beside the panel, is a fault-finding job. Replacing the device may appear to solve it for a week, but unless the protective device is proven defective, that approach simply removes evidence. The faster route is to treat the circuit as a current-balance problem and measure where the residual current is actually flowing.

This guide sets out a field workflow for nuisance RCD and RCBO tripping using a leakage current clamp meter. It focuses on measurements that can be repeated, compared, and handed to another technician—not on chasing a single unstable number. The EK-G668A high-accuracy clamp leakage current meter is used as the practical reference because it combines fine leakage-current resolution, a large rectangular jaw, true-RMS measurement, filtering, harmonic analysis, and Bluetooth-assisted recording. The same logic can be applied with other suitable leakage clamps.

Safety comes first: most clamp measurements in this workflow are made on energized circuits. Only qualified personnel should open equipment or work near exposed live parts. Follow the site’s electrical safety rules, approach boundaries, PPE requirements, and lockout procedure. Verify the measurement category and voltage rating of every instrument used. A current clamp is not an approved method for proving that a conductor is de-energized.

Identify the RCD type and expected residual-current waveform before choosing the test instrument. The EK-G668A is an AC clamp: its published bandwidth is 40–70 Hz with the filter on and 40 Hz–1 kHz with the filter off. Smooth DC, components below 40 Hz, and waveforms relevant to some Type F or Type B applications may not be represented by this measurement. The clamp workflow in this article covers compatible AC residual current within the instrument bandwidth; it does not replace a Type B/DC-capable test method where the installation requires one.

What the RCD Is Responding To

An RCD compares the current leaving through the intended live conductors with the current returning through them. On a single-phase circuit, line and neutral currents should balance. On a three-phase circuit, the vector sum of the phase conductors and neutral, where fitted, should be close to zero. Current returning by protective earth, structural metal, cable screens, pipework, or another unintended path creates a residual difference.

The protective device does not know whether that residual current comes from damaged insulation, wet equipment, electromagnetic compatibility filters, long cable capacitance, several electronic loads, incorrect neutral connections, or a person. It responds to magnitude and time according to its design. The technician’s job is to separate normal standing leakage from a new or intermittent contribution, then locate the branch or operating state responsible.

Why Nuisance Tripping Is Often a System Problem

Many modern loads intentionally place small capacitive paths between their internal circuitry and protective earth. Variable-speed drives, UPS systems, LED drivers, IT power supplies, filters, surge protective devices, induction equipment, and long shielded cables can all contribute. One item may be entirely normal; thirty items on one protective device may leave very little margin for humidity, switching transients, or a deteriorating heater.

This is why the word “nuisance” can be misleading. The RCD may be operating correctly in a badly coordinated system. A leakage survey can reveal whether the issue is one failing asset, cumulative standing current, a wiring error, or the wrong protective arrangement for the connected equipment. The final remedy might be repair, circuit separation, filter replacement, neutral correction, or an engineering review of device selection—not simply a new RCD.

Use the Right Clamp for the Current You Need to See

A standard load-current clamp is designed to read amperes on one conductor. Residual-current work may require stable readings in milliamps while large balanced load currents pass through the jaw. The instrument therefore needs adequate low-current resolution, external-field rejection, consistent jaw closure, and enough opening to enclose the complete conductor group.

The EK-G668A measures AC leakage and load current across a wide range and provides 0.01 mA display resolution at the low end. Its 80 mm × 80 mm rectangular jaw is useful when several conductors must be grouped. True-RMS response matters on distorted waveforms, while the selectable fundamental-frequency filter helps distinguish 50/60 Hz leakage from higher-frequency components. These functions answer different questions and should be recorded with every result.

The Conductor Grouping Rule

For a single-phase residual measurement, clamp line and neutral together. Their intended currents flow in opposite directions and cancel magnetically. Keep the protective-earth conductor outside the jaw. If earth is included, some or all of the leakage current may cancel and the result can look artificially low.

For a three-phase three-wire circuit, enclose all three phases. For a three-phase four-wire circuit, enclose all phases and the neutral. Again, exclude protective earth. If the cable construction prevents access to the individual live conductors, measuring the protective conductor alone may provide useful information, but parallel bonding paths can carry part of the current elsewhere. Record exactly where the clamp was placed.

Before Opening the Panel: Build a Trip History

Ask when the trips occur and what changes at that time. “Random” faults often follow a schedule that nobody has connected to the event: defrost heaters at night, washdown at shift change, compressor starts, battery charging, exterior lighting, irrigation, cleaning equipment, or a production recipe. Check whether the trip follows rain, high humidity, temperature change, maintenance, or a specific machine state.

Record the RCD type, rated residual operating current, time delay where applicable, protected circuits, recent alterations, and whether the device can be reset immediately. Do not treat this history as proof, but use it to plan the measurement window. A ten-minute survey at noon will not catch a fault that appears during a two-hour heater cycle before dawn.

Confirm the Protective Device and Circuit Arrangement

Identify the exact RCD or RCBO, its poles, upstream and downstream protection, neutral arrangement, and the circuits it protects. Compare the drawing with the installed wiring. Shared neutrals, borrowed neutrals, neutral-to-earth connections downstream of the device, and incorrectly routed conductors can create residual current even when every appliance is healthy.

Check that every intended live conductor passes through the RCD sensing core and that no unintended conductor does. Inspect for loose or overheated terminals, damaged insulation, water paths, contamination, and modifications. If the device includes a test button, use it only as directed; a successful button test confirms part of the trip mechanism, not the quality of the entire installation. When trip current and operating time must be verified, use a purpose-built instrument such as the EK-S864 leakage protector/RCD tester only after confirming that the system voltage, RCD type and test-current range match its published application. A leakage clamp does not replace a formal trip test.

A Four-Stage Measurement Sequence

Stage 1: Measure the Complete Protected Circuit

Place the leakage clamp around all live conductors leaving the RCD. Center the group, close the jaw completely, and keep the clamp away from adjacent high-current conductors as far as practical. Start with the fundamental-frequency filter enabled and allow the reading to settle. Record the actual load state, filtered value, and time.

Then repeat with the filter disabled. A much higher unfiltered value points to a significant high-frequency contribution or local electromagnetic interference. It does not automatically mean the RCD should ignore that current; device response depends on design and waveform. The comparison tells you that a simple 50/60 Hz insulation model is incomplete and that electronic loads deserve attention.

Do Not Judge the Circuit From One Snapshot

Residual current often changes with load. Use minimum, maximum, and average functions or logging while the equipment completes a representative cycle. Capture start, run, idle, heating, cooling, charging, and shutdown states. If the installation trips intermittently, a recorded maximum and timestamp are far more useful than the value displayed after the system has been restarted.

Note that a short transient and sustained leakage are different problems. A logging interval designed for slow trends may miss a brief event, while a peak function may provide no context. Match the capture mode to the suspected event. Where necessary, correlate the leakage record with PLC states, building-management logs, weather, or operator notes.

Stage 2: Divide the Distribution Tree

If the complete circuit reading is elevated or unstable, move downstream one level. Measure each outgoing branch using the same conductor-grouping method, filter state, and operating condition. The sum of branch readings may not equal the incomer perfectly because of timing, phase relationships, parallel paths, and measurement uncertainty, but the comparison usually identifies where most residual current originates.

Work from large branches to small ones. There is little value in testing thirty individual power supplies before checking which of three subpanels carries the problem. Mark measured branches on a one-line drawing. A divide-and-locate sequence is faster, safer, and easier to audit than disconnecting loads without a plan.

Stage 3: Separate Standing Leakage From the Fault Contribution

Once the suspect branch is identified, measure it with loads in controlled states. Record all loads operating, then switch off or isolate groups according to an approved procedure. Observe the change after each action. The difference between states is often more informative than the absolute reading because it shows the contribution of a specific load group.

Do not confuse “off” with electrically isolated. Electronic equipment can retain connected filters or standby supplies when its front-panel switch is off. Contactors may remove phases but leave a neutral connected. Where safe and permitted, isolate the equipment at the correct point, then repeat the residual measurement. Document the switching state so another technician can reproduce it.

Stage 4: Check the Protective-Earth Path

Clamping an accessible protective-earth conductor can confirm that current is flowing on that path. This is particularly useful for a single machine with a dedicated earth conductor. Compare the direct earth-current reading with the residual reading around the live conductors. Similar values support a straightforward return path; a large difference suggests parallel bonding or an incomplete measurement boundary.

Current can divide through cable armor, conduit, mounting frames, process pipework, data shields, and bonding straps. Never open a protective conductor to improve the measurement. Instead, identify parallel paths and measure them individually where safe. The goal is to understand the return network without compromising protection.

Neutral-to-Earth Faults: The Quiet Source of Confusing Results

A downstream neutral-to-earth connection can divert part of the normal neutral current around the RCD. The resulting residual current varies with load, so the device may trip only when several circuits are busy. This often looks like an insulation problem even though the affected appliances test normally one at a time.

Clues include residual current that follows total load closely, unexpected current on bonding conductors, and changes when unrelated circuits operate. Locating the connection usually requires isolation and resistance or continuity testing under a controlled de-energized procedure. Do not perform resistance measurements on an energized circuit, and disconnect electronics as required by the test method.

Electronic Filters and High-Frequency Leakage

EMC filters commonly use capacitors from line conductors to protective earth. Their current depends on voltage, frequency, circuit arrangement, and the number of connected devices. Drives add common-mode current through motor cables and winding capacitance. Longer shielded cables can increase the effect. An installation may therefore have acceptable individual equipment yet excessive combined leakage.

Compare filtered and wide-band true-RMS results at the incomer and suspect branches. If available and appropriate, harmonic analysis can show whether the waveform is dominated by the fundamental or contains strong higher-order components. Use these observations to guide coordination and equipment investigation. Do not impose a generic “safe” high-frequency value without reference to the actual protective device and installation design.

Load Patterns That Change Leakage

Wet Heaters, Refrigeration and Outdoor Loads

Heating elements can absorb moisture during idle periods. Leakage may be highest at cold start, then fall as the element warms and dries. Refrigeration equipment can show the opposite pattern if condensation develops during a cycle. Outdoor lighting, heat trace, pumps, and junction boxes may track rainfall or dew point rather than production load.

For these assets, record current from energization through the full cycle. Note ambient temperature, humidity, recent weather, and time since last operation. A thermal or visual inspection can support the diagnosis, but do not open wet or energized equipment casually. If offline insulation testing is required, isolate sensitive electronics and follow the manufacturer’s voltage limits.

Variable-Speed Drives and Motor Circuits

Drive systems produce common-mode voltage that can drive current through motor winding capacitance, cable screens, bearings, and earth. The result depends on motor cable length, shielding, filter configuration, carrier frequency, grounding quality, and motor construction. Leakage can change with drive state and output frequency even when load current is steady.

Measure the drive input residual current, the motor-cable conductor group where accessible, and relevant protective paths with instruments approved for those measurement points. Keep the filter state consistent. If the trip follows acceleration or a specific speed, log through that period. For upstream voltage, current, harmonics, unbalance, or event analysis on the drive input or another suitable AC point, use a dedicated instrument such as the EK-F523A power quality analyzer. Do not assume it can be connected to a PWM motor output; use the drive and instrument manuals plus the correct probes and method for output-side work.

UPS, IT and LED Loads

Large numbers of switch-mode power supplies can create substantial cumulative leakage and high-frequency content. A data cabinet, office floor, or LED installation may operate reliably until one more group is added. The new equipment is blamed because it precedes the first trip, but the real issue may be that the entire protected group was already close to its practical margin.

Measure leakage by distribution unit, lighting group, rack, or floor. Record connected-device count and operating state. If one unit contributes far more than its peers, inspect it. If many similar units contribute modest amounts, circuit division or protective coordination may be the engineering answer. Avoid randomly removing protective-earth connections; they exist for safety and EMC performance.

Surge Protective Devices

Surge protective devices can deteriorate after repeated events or a severe surge. Some failure modes increase leakage to earth without producing an obvious short circuit. Inspect status indicators and follow the manufacturer’s test or replacement procedure. A leakage change after a storm is a useful clue, not proof by itself.

When isolating an SPD for diagnosis, use an approved de-energized method and restore protection correctly. Record the before-and-after residual current. Also inspect downstream equipment because the same event may have affected filters or insulation elsewhere.

Using Harmonic Information Without Overinterpreting It

Harmonic readings can help distinguish a predominantly fundamental leakage path from distorted current associated with electronic loads. They can also reveal that the waveform changes with equipment state. Treat harmonic content as a fingerprint for comparison, not a standalone fault code. The same harmonic can arise from several circuit mechanisms.

Record the instrument range, filter state, fundamental frequency, total leakage, and operating state alongside the harmonic values. Compare similar machines under similar conditions. A repeatable difference is actionable; an isolated spectrum with no baseline is mostly a description.

How Much Standing Leakage Is Too Much?

There is no universal field threshold that applies to every RCD type, waveform, installation, and jurisdiction. The protective device’s rated residual operating current is not a design target for continuous standing leakage. Good practice is to maintain margin for tolerances, transients, environmental change, and future additions. The required margin must come from the applicable standard, device data, and site engineering rules.

Use the measured current to answer practical questions: Is the value stable? Which branch contributes most? How close is the system to its historic level? Does a specific load produce a step change? Does the waveform or filter comparison point to electronics? Has the current risen after maintenance or weather? Those answers lead to a repair or design decision more reliably than an unsupported percentage rule.

A Composite Example: The 03:00 Bakery Trip

The following hypothetical scenario combines several common field patterns. A small production bakery reports an RCBO trip on the preparation area two or three nights each week. Daytime residual current is stable and low. The trip is initially blamed on a refrigerator because it is the only visible load running overnight. Replacing the refrigerator would be expensive and unlikely to help.

A leakage logger was left around the outgoing line and neutral. The record showed a sharp rise shortly after 03:00, followed by a trip. The building schedule revealed that a floor-warming circuit and wash-water heater started together at that time. Branch measurements showed the heater contribution rose during damp weather, while the warming cable contributed steady background leakage. Isolating and testing the heater located moisture at a terminal enclosure. Repairing the enclosure and separating the scheduled starts restored margin without changing the RCBO.

The useful evidence was not a single milliamp value. It was the timestamp, the step change, the branch comparison, and the operating schedule. This is the pattern a repeatable workflow is designed to find.

Common Mistakes That Waste Hours

Clamping Only One Live Conductor

This measures load current, not residual current. For leakage, enclose all intended live conductors and exclude protective earth.

Including the Protective Conductor

This can cancel the leakage path you are trying to measure. Draw the conductor group before testing if the cable arrangement is unclear.

Comparing Filtered and Unfiltered Readings as If They Were Identical

The two readings answer different questions. Record the filter state every time and compare like with like during a trend.

Opening the Jaw Near a High-Current Bus

External magnetic fields and poor jaw closure can disturb a low-level reading. Center the conductor group, close the jaw fully, and move away from neighboring conductors when possible.

Resetting Before Capturing the Event

A reset changes the system state and may erase the only clue. If safety permits, record device status, time, active loads, weather, and alarms before restoration.

Replacing the RCD Without Testing the Circuit

Protective devices can fail, but replacement should follow evidence. Test the device with suitable equipment and investigate the residual current rather than using a new device as the diagnostic method.

What to Put in the Test Record

• Panel, protective device, circuit, and test-point identification.
• Date, time, technician, weather, humidity, and relevant process state.
• Exact conductor grouping and whether protective earth was excluded.
• Clamp model, serial number, calibration status, range, and filter state.
• Instantaneous, minimum, maximum, average, and logging interval where used.
• Filtered and unfiltered true-RMS values.
• Branch and individual-load contributions.
• Switching actions, isolations, and before-and-after results.
• Inspection findings, photographs, and corrective actions.
• Final verification after repair and the new baseline.

When to Move to Offline Insulation Testing

Leakage clamping is an energized screening method. It shows residual current under operating conditions but cannot by itself determine insulation resistance or prove that an isolated asset is safe. If a branch or item shows abnormal leakage, an offline insulation test may help localize deterioration after the equipment is safely isolated.

Choose the test voltage to suit the equipment and disconnect sensitive electronics. For motors, cables, generators, and higher-voltage assets, an insulation tester such as the EK-C352C high-voltage insulation meter provides timed resistance and diagnostic modes. Energized leakage and offline insulation resistance are complementary measurements, not interchangeable substitutes.

A Compact Troubleshooting Sequence

1. Collect the trip history and identify timed, weather-related, or process-related patterns.
2. Verify the protective-device type, circuit arrangement, and visible wiring condition.
3. Measure total residual current around every live conductor downstream of the device.
4. Compare filtered and wide-band readings and record the load state.
5. Log long enough to capture the suspected operating cycle.
6. Divide the distribution tree and identify the dominant branch.
7. Separate load groups and quantify each contribution.
8. Check protective-earth and parallel bonding paths without disconnecting protection.
9. Investigate neutral-to-earth faults, filters, drives, heaters, moisture, and SPDs as indicated.
10. Repair or redesign, then repeat the original measurement and store the new baseline.

Where the EK-G668A Fits

Technicians looking up the G668A usually need more than another specification list. They need to know whether the instrument can resolve small, changing residual currents in a live installation and whether its records will help them locate a fault without interrupting every load. That is the job this workflow is designed to demonstrate.

The EK-G668A’s combination of low-level resolution, large jaw, magnetic shielding, true-RMS response, selectable filtering, min/max/average, harmonic functions, and Bluetooth-assisted records supports the workflow described here. Use the EK-G668A leakage current testing guide for a broader explanation of baselines and harmonic interpretation, then apply this article when the specific symptom is an RCD that keeps tripping.

Final Takeaway

Intermittent RCD tripping stops being mysterious when every measurement has a boundary, operating state, filter setting, and timestamp. Start at the protective device, measure the complete residual current, log the actual cycle, then divide the system until one branch, load group, or wiring condition explains the change. Keep protective conductors intact, respect parallel paths, and move to offline testing only under a controlled isolation procedure.

The repair is complete only when the original measurement is repeated and the margin is demonstrably restored. Browse the EK Instruments product range for leakage, insulation, power-quality, and general electrical test equipment that can support the full investigation from first symptom to documented handover.