Ground Fault Protection Performance Testing: Pickup and Delay

What it is

Ground fault protection performance testing verifies that a ground-fault scheme on low-voltage equipment really clears a fault to ground. The scheme is made of a current sensor, a relay or an electronic trip unit that evaluates the signal, and a breaker or switch that is tripped. The sensor may be a residual connection of the phase and neutral current transformers, a single zero-sequence sensor around all the conductors, or a sensor on the main bonding connection between neutral and ground. The test injects a known current to simulate a ground fault and confirms that the whole chain operates: the pickup level at which it responds, the time delay before it trips, and the actual opening of the breaker or switch. It is performed on molded-case, insulated-case and power circuit breakers with integral ground-fault trip functions, on stand-alone ground-fault relays and on switches or switchboards protected by such a scheme.

Why it is done

Arcing ground faults on solidly grounded low-voltage systems can burn at a current well below the setting of the phase protection, so the phase devices may never see them. Left alone, an arcing fault destroys switchgear and busway quickly and starts fires. Ground fault protection exists to catch that low-level fault, and it is a scheme with many places to go wrong: a sensor installed backward, a neutral bonded in a second place, a relay left at the wrong setting, wiring that was never completed, or an electronic trip unit that was never enabled. A scheme can look correct on the drawings and still not trip. Because it is rarely called on, nothing in normal service reveals the problem, and a documented performance test at installation and on a periodic basis is the only way to know that it works. The same test also exposes the opposite fault, nuisance tripping caused by neutral currents that follow the wrong path.

The principle

In a healthy circuit the phase and neutral currents add to zero, so a sensor that sums them sees nothing. When current leaks to ground and returns through the earth, the bonding conductors or the equipment frame, the sum is no longer zero and the sensor produces a signal proportional to the ground current. The relay compares that signal with a pickup setting and, where the scheme has a time delay, waits out the delay before it commands a trip. The delay allows downstream protection to clear a fault first and keeps the scheme selective, and some devices follow an inverse relationship in which more current trips sooner. The scheme depends on the neutral being bonded to ground at one point only, on the supply side of the sensor, so that neutral current cannot return through a path that bypasses the sensor or is mistaken for ground current. The test relies on those same rules: it injects current where a real ground fault would flow and checks that the response matches the setting.

How it is performed

The equipment is de-energized, isolated and proven dead under the owner's lockout and work-permit procedures. Before any injection the installation is inspected: the sensor location, its direction relative to the load, the single neutral-to-ground bond and the neutral connections are confirmed against the drawings, and the scheme is set to the condition it will be in when energized. A primary current injection set capable of the required current is connected so that the test current passes through the sensor in the same sense as a real ground fault. In schemes with a built-in test feature, that feature is used only as a supplement and not as a substitute. The current is raised slowly until the device responds to establish the pickup, and a separate run applies a current above pickup and measures the interval to the trip output, with the time-delay setting noted. The breaker is confirmed to open, and a reduced control voltage trip check is made where the procedure calls for one, to show that the shunt trip operates when the supply is weak. The delay and pickup are checked against the manufacturer's published curve and the field settings. Afterward the test connections are removed, the scheme is returned to normal and the breaker is reset.

What is recorded

The record shows the equipment and scheme identification, the type and location of the sensor, the pickup and time-delay settings found and left, the measured pickup current and measured delay, whether the breaker or switch opened, the result of any reduced control voltage trip check, and the injection set and timer used with their calibration status. The as-found and as-left conditions are kept separately, and the result is signed and dated as a performance test so that it can be produced for inspection.

How results are judged

The measured pickup and delay are judged against the settings specified for the scheme, within the tolerance the manufacturer publishes, and the breaker must open on every run. Comparison between phases or runs is not used here; instead the response must be repeatable and the delay should sit on the published curve. The scheme must not trip on unrelated neutral current, and a fault that produces no trip is a failure regardless of the numbers. ANSI/NETA ATS-2025 gives the acceptance criteria for these schemes. For periodic testing see ANSI/NETA MTS-2023. Refer to your own licensed copies for the values and tolerances that apply to your equipment.

Variations

Acceptance testing proves a new installation, including the wiring and the single neutral bond, and its record is the baseline. Maintenance testing confirms that the scheme still works, that its settings have not been changed and that no extra neutral-to-ground bond has been introduced by later work. On circuit breakers with an integral electronic trip unit the ground-fault function is exercised together with the long-time and short-time functions and may be tested by primary injection or by secondary injection into the trip unit. On a stand-alone relay with an external sensor the loop includes the sensor and its wiring, so the whole path is injected. On a switch or switchboard with a ground-fault relay and a shunt trip, the trip output and the opening of the switch are verified. Where zone selective interlocking is used, the restraint signal is also exercised.