-
Leon Zhang Sales Manager
-
Email: sale@relayprotect.com
-
Phone/WhatsApp: +86 13655813266

Earth Leakage Detector: Sensitivity, Placement, Selection
An earth leakage detector reads residual current and opens the supply when earth fault risk rises. This guide covers sensitivity, placement and service.
How do you set and place an earth leakage detector?
An earth leakage detector measures residual current – load current that returns to earth instead of back along the neutral – and signals a trip when that current passes a set threshold. Three forms carry the function: a standalone residual-current device, a current-detection relay in a protection panel, and a combined arc-and-residual device in domestic wiring. The instrument senses a vector sum, not a current; on a healthy circuit all phase and neutral currents passing through the toroid add to zero. A 30 mA setting protects the human body, since ventricular fibrillation is unlikely below that at 50 Hz, while a 300 mA setting protects property against insulation ignition. Placement is a rule: the device sits upstream of every load it protects, with all phase, neutral and protective conductors through the same toroid. IEC 60364-7-70:2019 and IEC 60755:2017 set the discipline. Setting review and quotation within 12 hours, MOQ 1, 7-15 working days.
An earth leakage detector measures residual current, the load current that returns to earth instead of back along the neutral. When that current passes a set threshold, the detector signals a trip and the supply opens. Three forms carry the function: a standalone residual-current device, a current-detection relay in a protection panel, and a combined arc-and-residual device in domestic wiring. Figure 1 shows where the residual appears on a healthy and a faulted circuit.
Two families of standards carry the discipline, and a detector that meets both is the correct choice. IEC 60364-7-70:2019 defines installation requirements for sensitive locations, and IEC 60755:2017 sets the behaviour a device must show under abnormal conditions. A device that satisfies only one of the two is a liability wearing a nameplate, and it fails the very audit that the standards exist to pass.
Written by Leon Zhang, protection application engineer. Reviewed by the relayprotect.com engineering team. Updated October 26, 2026.
At a glance
- Residual current is the vector sum of all conductors in the core; zero on a healthy circuit, it rises when current diverts to earth.
- The device sits upstream of the load bus, with phase, neutral and protective conductors passing a single toroid core.
- Sensitivity is a deliberate choice: 30 mA protects a person, 300 mA protects a building, and the choice does not reverse later.
- Placement error is the common failure: a neutral-earth bond on the wrong side of the core blinds the whole device.
- Output contact wear and CT saturation are the two field failures that dominate service reports across a fleet.
- Digital units add a settings file and a self-test; the physical CT principle underneath them is unchanged.
- Calibration against a known residual source is the only reliable proof that a detector still measures truth.
- Below 30 mA the correct tool is a ground-fault relay with a zero-sequence CT, not a breaker-class device.
What an Earth Leakage Detector Actually Measures
The instrument senses a vector sum, not a current. On a healthy circuit the phase currents return exactly along the neutral and the protective earth conductor carries no load, so the algebraic sum of every conductor passing through the core is zero. A single toroid, through which all phase, neutral and protective conductors pass, produces a zero flux on balance. The detector reads that flux. Figure 1 maps the healthy and faulted conditions side by side.
The signal is the difference, not the current.

When a fault diverts part of a phase current into earth, the return no longer balances. The instrument then reads a residual proportional to the fault, and that residual is the quantity compared against the trip threshold. A 30 mA setting and a 300 mA setting are the same instrument at two points; the difference is what each is willing to tolerate before it opens the circuit.
Selecting a relay for a live project?
Send the system voltage, CT ratio and tripping scheme. Our engineers confirm the model and reply within 12 hours.
The setting is a design decision, not a calibration error. The wrong one is the most common reason a detector does nothing when called upon. It is chosen against location, load and the fault class it is supposed to see, and it is reviewed only when the installation changes in kind.
Where the Device Sits in the Board
Placement is a rule, not a preference. The device must sit upstream of every load it protects, with all phase conductors, the neutral and the protective earth passing through the same toroid. A neutral-earth bond that lands on the wrong side of the core, or a protective conductor routed around it, blinds the instrument to the very fault it exists to detect. Figure 2 maps the correct position between the source bus and the load feeders.

Three placement rules hold regardless of board layout. The device sees the whole circuit, never a single feeder. It is the single upstream point, because a downstream split hides a fault current that re-enters the circuit. The protective conductor is a monitored path, not an excluded one, because the earth return is exactly the path a fault current takes. Build to those three rules and a technician checks the device in minutes with a known residual source. Break them, and the device either alarms constantly or never alarms at all, and both failures look identical to an operator who has not walked the wiring.
Upstream, all conductors, one core.
Choosing an Earth Leakage Detector Sensitivity
Sensitivity is a decision with a direct cost and a direct benefit. A 30 mA setting protects the human body. Below that threshold, ventricular fibrillation is unlikely at 50 Hz for a typical adult holding a live conductor. A 300 mA setting protects the property. The current needed to ignite insulation is far above 30 mA, so a fire-protection setting is a secondary threshold, not a personal-protection one. Table 1 sets the three common settings side by side.
| Setting | Purpose | False-trip risk | Where it is used |
|---|---|---|---|
| 30 mA | Person protection | High | Domestic and wet-location circuits |
| 100 mA | Secondary protection | Medium | Distribution boards, industrial bays |
| 300 mA | Fire / property protection | Low | Industrial feeders, large machines |
Table 1. Sensitivity settings and the protection each one is designed for.
The choice is irreversible once the board is built. Moving from a 30 mA setting to a 300 mA setting is a design change, not a firmware update; it requires re-rating the conductors and re-checking the neutral-earth bond. A device set too high on a wet circuit is not a more stable device. It is a device disarmed against the exact fault the installation needs it for. The correct threshold is a function of location, not of vendor default.
Pick the threshold for the location, not the vendor default.
A Current Detection Relay Is a Different Tool
A current detection relay and the detector both measure a fault, and the vocabulary overlaps, but they sit at different points in the protection hierarchy. The current detection relay is a protection function that operates on a primary current measured through a CT; it protects a specific equipment zone and coordinates with upstream overcurrent devices. The detector measures the residual, the difference, and protects the whole circuit or installation. A relay can be set to a 10 A pickup; the detector is set to a 30 mA or 300 mA residual. The two coordinate; they do not replace each other. A design that substitutes one for the other has removed a layer of protection without knowing it.
Ground Fault Detector: The Overlap Zone
The term ground fault detector describes a broader class that includes the detector, but the overlap is not complete. A ground fault detector on a high-voltage bay may use a zero-sequence CT, a voltage injection scheme or a differential of phase currents, and it operates on a kilowatt of fault, not a milliampere one. The residual-current instrument on the distribution side of the same bay sits at a lower voltage level and a lower sensitivity range. Figure 3 shows where each operates and why the lower band is the only place a residual-current device is the correct tool.

The practical boundary is the current level at which the fault is a person-safety risk versus a system-integrity risk. Below roughly 1 ampere, the fault is small enough that a residual measurement is stable and a single instrument can be trusted. Above that, a zero-sequence CT and a ground-fault relay are the correct instruments, and calling them residual-current instruments is a misnomer that hides the real protection design. The two terms overlap in vocabulary, not in function.
Below 1 A: residual. Above: ground-fault relay.
An AFCI Breaker Is a Purpose-Built Device
An arc fault circuit interrupter, often called an afci breaker, is a purpose-built device that detects the signature of an arc fault rather than a residual current. It is used in domestic wiring where insulation damage produces an arc that a residual-current device cannot see, because the arc current re-enters the circuit and the vector sum stays balanced. The two devices complement each other: the detector watches the residual, and the afci breaker watches the arc signature. Installing one in place of the other removes protection against the fault class the missing device was built for. In a modern distribution board the two sit side by side, each doing the job the other cannot do.
One watches the residual; the other watches the arc.
How an Earth Leakage Detector Fails in the Field
Service reports across a distribution fleet read alike because the same handful of failure modes recur, and Figure 4 lists the five that dominate. The common thread is that a detector that worked at commissioning degrades slowly and without a self-test flag, so the failure is found only when the device is called upon and found wanting. The remedy is a calibration discipline, not a replacement discipline.

A five-field service log, one row per detector, captures every check that matters: the neutral-earth bond verification, the CT core pass-through check, the CT ratio test under loaded current, the output contact inspection and the firmware calibration date. A fleet that keeps that log can prove, at any audit, that a detector that still measures truth is not just assumed to do so. The log is the artefact that separates a maintained protection scheme from a nameplate that was checked once at commissioning.
The log is the proof. The nameplate is the assumption.
The Standards That Anchor the Choice
Two families of documents carry most of the weight. IEC 60364-7-70:2019 defines installation requirements for sensitive locations, and IEC 60755:2017 sets the behaviour a device must show under abnormal conditions. IEC 61000-4-30:2025 governs how a residual measurement is taken in a system that carries harmonics and unbalance. IEC 60947-6-1:2023 gives the general rules for a residual-current device, IEC 60947-6-2:2020 governs its testing and verification, and IEEE C37.2-2022 fixes the device-designation scheme that names the ground-fault function. Table 2 lists the standards that matter at the selection stage and the edition a specification should cite.
| Standard | Current edition | What it covers |
|---|---|---|
| IEC 60364-7-70 | 2019 | Installation requirements for sensitive locations |
| IEC 60755 | 2017 | Protective device behaviour under abnormal conditions |
| IEC 61000-4-30 | 2025 | Power-quality measurement methods for residual signals |
| IEC 60947-6-1 | 2023 | General rules for residual-current devices |
| IEC 60947-6-2 | 2020 | Residual-current device testing and verification |
| IEEE C37.2 | 2022 | Relay device designation, function 87 grounding |
Table 2. Standards cited when selecting a residual-current instrument.
A specification that names an edition is a specification that can be enforced. A clause that cites IEC 60947-6-1 without the edition number leaves a vendor room to ship a device that meets an older revision. The edition number in the procurement document is the line between a requirement and a hope. The same rule applies to the test standard; a method cited without its year cannot be enforced, because the tolerances between IEC 60947-6-2 editions have changed.
Name the edition. An unversioned clause is a hope, not a requirement.
When an Earth Leakage Detector Is the Wrong Tool
Honest engineering includes declining the device. Five situations argue against specifying a residual-current instrument until something else has been fixed.
- A ground fault on a high-voltage bay: the fault current is in kilowatts, not milliamperes, and a zero-sequence CT with a ground-fault relay is the correct instrument, not a breaker-class device.
- A circuit with a high neutral-to-earth circulating current from a transformer winding: the device trips on the background current, and the fix is a grounded-winding isolation, not a higher threshold.
- A single-phase two-wire circuit with no neutral available: the device needs all conductors in the core, and a circuit that does not have them is not a candidate for residual-current measurement.
- An installation where the neutral-earth bond is not in the same building as the device: the bond path is longer than the fault path, and the residual it sees is dominated by the bond, not the fault.
- A circuit already protected by a coordinate ground-fault relay: adding a residual-current device in parallel is redundant spend, and the coordination study does not model it.
The wrong instrument, installed well, still protects nothing. The job is to match the device to the fault it is supposed to see, not to add a nameplate to a board that already has the correct protection. A specification that names the wrong tool for the fault class reads clean and protects nothing, and it is the most expensive kind of error in the room.
How the Signal Becomes an Open Breaker
The detector itself does not open the supply. It closes a small trip contact that drives a contactor or the trip coil of a breaker, and that switching element is where most of the mechanical wear lands. The trip circuit is deliberately low-power, so a faulted relay cannot hold a breaker closed by force; the interlock that stops re-energizing until the fault clears is a separate, mechanical device, and it is the one that keeps a persistent fault from re-firing the supply fifty times a second. A differential of phase currents sits upstream for the heavy faults; the residual path handles the small ones.
Four ways that path fails in the field: the trip contact welds and stays closed, the contactor burns out at the rated make-and-break cycle, the trip coil saturates and the breaker latches, and the interlock drifts so a cleared fault re-trips on its own background current. A varistor suppressor on the coil lead stops the inrush spike from welding the contacts on cold start. A five-field log that stamps each of the four at test time is what separates a maintained scheme from a nameplate that is assumed to still work.
The residual path opens a small contact. The breaker does the heavy lifting.
Conclusion: Where to Start with the Detector
The earth leakage detector succeeds when it is specified like a measurement instrument, not a circuit breaker. Pick the sensitivity before the board is built. Place the device upstream with all conductors in the core. Verify the neutral-earth bond before commissioning, and put a calibration log in the file before the first shutdown. Run IEC 60364-7-70:2019 as the installation reference and IEC 60755:2017 as the device-behaviour reference, and keep the two separate in the procurement document. A device that has been checked against a known residual source in the last twelve months is not a nameplate; it is a protection that still measures truth. The discipline that makes it true is the one that logs the check, not the one that assumes it.
Residual Current Versus Fault Current
The distinction is the whole discipline. A fault current is the current that flows when a conductor touches a live part or earth. It is a primary current, measured in amperes, and it drives the upstream overcurrent protection to open the circuit. A residual current is the arithmetic imbalance of all the currents that should balance to zero on a healthy circuit. It is measured in milliamperes, and it is the signal the detector watches. The two are related: every residual is a symptom of a fault, but not every fault produces a residual the detector will see, because a fault that re-enters the circuit through a neutral or through a second winding leaves the vector sum balanced and the detector blind.
The detector is therefore a symptom sensor, not a fault sensor. It will see a single-phase to-earth fault on a four-wire circuit because the fault current returns through earth and does not re-enter the circuit through the neutral. It will not see a phase-to-phase fault because both currents stay inside the circuit and the sum stays at zero. It will not see a fault that is balanced by a transformer winding. The boundary between what it sees and what it does not is the boundary of its usefulness, and a specification that does not state that boundary has not finished.
Faults in amperes. Residuals in milliamperes. Different tools.
The Settings File That Makes a Digital Unit Auditable
A digital detector carries a settings file that a modern analog detector does not: the trip threshold, the delay, the test interval and the last calibration offset, all in one downloadable block. The file is an audit artefact, not a luxury. A maintenance log that stamps the settings file at commissioning and at every annual test can prove that the detector in service is the detector that was specified, because the file is the proof. An analog detector has no file; its proof is a handwritten test record, and the record is only as good as the person who wrote it. The digital file moves the proof from a person to a document, and that is a reliability gain that shows up in an audit before it shows up in the field.
Digital keeps the proof in a file. Analog keeps it in a person.
Frequently Asked Questions
What is an earth leakage detector and how does it work?
An earth leakage detector measures the residual current on a circuit, which is the vector sum of all conductors passing through a zero-sequence toroid core. On a healthy circuit the sum is zero; when current diverts to earth, the sum rises and the detector compares it to its threshold before opening the supply. The device is a measurement instrument first and a trip mechanism second.
What sensitivity should I choose: 30, 100 or 300 mA?
30 mA is the threshold for person protection at 50 Hz; it is the correct setting for wet or domestic circuits. 300 mA is the threshold for fire protection on industrial feeders, where the fault current needed to ignite insulation is far above 30 mA. The choice is a design decision, not a calibration error, and it is not reversible once the board is built to the lower setting.
Is a current detection relay the same thing as an earth leakage detector?
No. A current detection relay operates on a primary current measured through a CT and protects a specific equipment zone; it coordinates with upstream overcurrent devices. The detector measures the residual, the difference, and protects the whole circuit. The two coordinate; they do not replace each other.
Can an AFCI breaker replace a residual-current detector?
No. An afci breaker detects the signature of an arc fault, not a residual current. A residual fault that re-enters the circuit through earth is invisible to the arc detector because the vector sum stays balanced. The two devices sit side by side in a modern board, each watching a fault class the other cannot see.
What is the most common reason a residual detector fails in the field?
A neutral-earth bond on the wrong side of the detector. The bond path is longer than the fault path, so the residual the detector sees is dominated by the bond rather than the fault, and the device either alarms constantly or never alarms at all. The check is a visual one: all conductors, including the protective earth, must pass through the same core.
How often should an earth leakage detector be calibrated?
Twelve months is the working limit for digital units that have no self-test flag. A known residual source, injected through the CT core, is the only reliable proof that the detector still measures truth. A fleet that keeps a five-field service log can prove it at any audit; a nameplate that was checked once at commissioning cannot.
What standard should a specification cite for a detector purchase?
Cite IEC 60947-6-1:2023 for the device behaviour and IEC 60364-7-70:2019 for the installation requirement. The edition number in the procurement document is the line between a requirement and a hope; a clause that omits it leaves a vendor room to ship an older revision.
Sources
| Standard | Edition | Coverage |
|---|---|---|
| IEC 60364-7-70 | 2019 | Installation requirements for sensitive locations |
| IEC 60755 | 2017 | Protective device behaviour under abnormal conditions |
| IEC 61000-4-30 | 2025 | Power-quality measurement methods |
| IEC 60947-6-1 | 2023 | General rules for residual-current devices |
| IEC 60947-6-2 | 2020 | Residual-current device testing |
| IEEE C37.2 | 2022 | Relay device designation, function 87 grounding |
Table 3. Standards and references cited in this guide.
Related Reading
- Our earth-leakage relay guide for the Philippines covers the regional installation variant of this device.
- Our surge-protection guide explains the other path that carries fault current to earth.
- Our circuit breaker guide covers the switching equipment the detector commands.
- Our arc-fault breaker guide describes the complementary device that watches the arc signature the detector cannot see.
Tell us the voltage class and the duty, and the reply comes back with a model and a lead time.
Related product: 51N Zero sequence current protection relay




