Hysteresis, Trip Delay and Latching in Monitoring Relays
What do hysteresis, trip delay and latching actually do inside a monitoring relay? Hysteresis is the gap, set as a percentage of the threshold, between the value that trips the output and the value that resets it; trip delay is the time the measured quantity has to stay out of window before the output actually switches, both defined under IEC 60255 measuring-relay practice. Get either one wrong and the relay either chatters on every load swing or lets a real fault ride through. This article covers how the two settings interact, how to calculate a reset threshold, the difference between automatic and latched (manual) reset, and typical starting points for phase, voltage, current, level and temperature monitoring.
What Hysteresis Actually Does
A monitoring relay does not switch its output the instant a measured value crosses the setpoint and switch back the instant it recrosses it. If it did, a voltage sitting exactly at 253 V — the trip point for a 230 V supply with a 10% overvoltage setting — would flicker the output relay on and off dozens of times a second as line noise pushed the reading a fraction of a volt either side of the line. Hysteresis fixes this by creating two separate thresholds: one to trip, one to reset, with the reset threshold set back inside the safe window.
For an overvoltage function, the reset threshold sits below the trip threshold. For undervoltage, it sits above. A relay set to trip at 253 V with 5% hysteresis will not reset the output until the voltage drops back to roughly 240 V, not the moment it dips below 253 V again. That gap is the whole point — it stops single-cycle noise from re-triggering the output the instant the fault clears.
How Trip Delay Rides Through Transients
Trip delay (sometimes labeled "on-delay" or "response time") is a separate setting from hysteresis, and confusing the two is a common commissioning error. Delay is a timer: the monitored quantity has to stay past the threshold continuously for the set duration before the output switches. Typical adjustable ranges run from about 0.1 s to 10 s, with some phase-monitoring functions defaulting near 0.5-1 s.
Motor starting is the textbook case. Direct-on-line starting pulls a voltage sag on the supply for a few hundred milliseconds. A relay with no delay, or a delay set too short, sees that sag as an undervoltage fault and trips before the motor is even up to speed. Set the delay past the typical sag duration and the relay ignores it; a sustained undervoltage — an actual brownout, a failed contactor, a broken conductor — still trips once it outlasts the delay window.
What we see in the field: panel builders often leave the factory default delay in place without checking it against the actual load. On a site with long motor start times or a soft starter ramping over several seconds, the default delay is too short and the relay nuisance-trips on every start. Match the delay to the measured start transient, not to the datasheet default.
Calculating the Reset Threshold
Hysteresis is normally programmed as a percentage, and the reset point is derived from the trip point, not set independently. For an overvoltage or over-quantity function the reset threshold is below the trip threshold; for an undervoltage or under-quantity function it is above.
Formula: Reset Threshold (Over-Function) — Source: IEC 60255-1, hysteresis definition
Xreset = Xtrip × (1 − H%)
| Symbol | Description | Unit |
|---|---|---|
| Xreset | Value at which the output relay re-energizes after a trip | same as measured quantity (V, A, °C, kOhm) |
| Xtrip | Programmed threshold that switches the output | same as measured quantity |
| H% | Hysteresis, set as a percentage of the trip threshold | % |
For an under-function (undervoltage, undercurrent, low level), flip the sign: Xreset = Xtrip × (1 + H%). A dry-run pump-protection relay set to trip at a low-level threshold with 8% hysteresis will not re-energize the pump output until the level rises 8% above that trip point — enough margin that a single splash on the probe does not restart the pump into a still-low tank.
Latching (Manual Reset) vs Automatic Reset
Independent of hysteresis and delay, every monitoring relay has a reset mode: automatic or latched. Automatic reset re-energizes the output the moment the measured value crosses back over the reset threshold, no operator action needed. Latched (manual) reset holds the output in the tripped state even after the fault clears, until someone presses a reset button or cycles the auxiliary supply.
The choice is an application decision, not a technical default. Automatic reset suits unattended, self-correcting loads — an HVAC fan that should restart on its own once the supply stabilizes. Latching suits anything where an unattended restart is a safety or process risk: a pump that dry-ran and needs a human to check for damage before it runs again, or a motor that tripped on thermistor overtemperature and should not restart hot. Some relays also offer a fault-memory option that latches only specific fault types (phase loss, say) while letting others (a brief voltage sag) reset automatically.
Output logic compounds this decision. A normally-energized (fail-safe) output de-energizes on fault or on loss of the relay's own auxiliary supply, so a tripped panel with a latched, fail-safe relay stays safely open even if control power is cycled — the operator still has to press reset. A normally-de-energized configuration behaves the opposite way and is far less common for safety-relevant monitoring.
Typical Hysteresis and Delay Settings by Monitored Quantity
Three-Phase Supply and Voltage
Phase-sequence and phase-loss functions in relays like the Schneider Zelio Control RM17TE or the ABB CM-MPS typically run a short, largely fixed delay (a fraction of a second) since phase loss is unambiguous — there is no useful transient to ride through. Over/undervoltage functions on the same devices, and on dedicated three-phase voltage monitoring relays, take an adjustable hysteresis (often 1-10%) and a longer delay to clear starting sags. See the phase sequence and phase failure relay guide for how sequence detection interacts with these same settings, and the how to set a voltage monitoring relay article for a worked commissioning example.
Current
Undercurrent protection (broken belt, lost load, dry-running pump on amp draw) generally wants a longer delay than overcurrent, because load current on many machines dips briefly during normal operation — an intermittent conveyor, a reciprocating compressor. Overcurrent (jam detection) can run a shorter delay since a genuine jam current rise is sustained, not momentary.
Level
Level relays add a third dimension beyond hysteresis and delay: probe placement itself sets a physical hysteresis band, since the "on" and "off" electrodes sit at different heights in the tank. The electronic hysteresis and delay then handle surface turbulence at the electrodes — a pump kicking up waves that would otherwise make the reading bounce across the electrode.
Temperature
PTC thermistor functions (per IEC 60947-8) are largely binary — the sensor resistance jumps sharply near the reference point (roughly 3.3 kOhm), so hysteresis is mostly built into the sensor's own resistance curve rather than user-adjustable. PT100/PT1000 relays reading an analog temperature behave more like voltage relays: adjustable threshold, adjustable hysteresis, and a delay to avoid tripping on a brief thermal spike. See the PTC thermistor and PT100 temperature monitoring guide for sensor-specific detail.
Common Field Mistakes
Hysteresis set to zero, or to the tightest available step, is the most common cause of an output relay that chatters and eventually fails from mechanical wear on the contacts. This shows up as intermittent nuisance trips that look random until someone puts a data logger on the supply and sees the reading oscillating right at the threshold.
The second common mistake is the reverse: hysteresis or delay set too wide "to stop the nuisance trips," which quietly disables the protection the relay was installed for. A phase-loss relay with an 8-second delay added to stop a real fault from tripping is not solving the nuisance-trip problem, it is hiding it — and it delays the very protection the panel needed. This depends heavily on how noisy the supply actually is; measuring the transient before adjusting the setting beats guessing at a number and hoping.
Third: mixing up reset mode with hysteresis. A relay that keeps re-tripping after a reset button press usually has a fault condition that has not actually cleared, not a latching problem — check the measured value against the reset threshold before assuming the relay itself is faulty. For related contactor and thermal-overload wiring context, see how to select the right contactor and the thermal overload relay engineering guide, both of which cover the output side these monitoring relays typically drive.
Frequently Asked Questions
What is the difference between hysteresis and trip delay?
Hysteresis is a value gap between the trip point and the reset point, preventing output chatter around the threshold. Trip delay is a time duration the fault must persist before the output switches at all. They solve different problems and are set independently.
What hysteresis percentage should I use for a voltage monitoring relay?
There is no single correct figure — it depends on how much the supply naturally oscillates. A stable utility feed can run tight hysteresis (1-3%); a generator-backed or heavily loaded supply usually needs more (5-10%) to avoid nuisance trips from normal voltage ripple.
Why does my relay keep tripping even after I reset it?
Most often the fault condition has not actually cleared — the measured value is still past the reset threshold, not just past the trip threshold. Check the live reading against both thresholds before assuming the relay itself has failed.
Should I use automatic or manual (latched) reset?
Use latched reset for anything where an unattended restart could damage equipment or create a hazard — motor thermal trips, dry-run pump protection. Use automatic reset for self-correcting loads where a human does not need to intervene once the supply or condition is back in range.
Does trip delay affect protection speed for a real fault?
Yes. Any delay added to ride through transients also delays genuine fault clearance by the same amount. The setting is a trade-off — set the delay just past the longest normal transient the load produces, not longer, to keep real-fault response as fast as the application allows.
Is hysteresis the same on every monitoring relay function?
No. Phase-sequence and phase-loss detection is largely binary with little need for hysteresis. Voltage, current and PT100 temperature functions use adjustable percentage hysteresis. PTC thermistor functions get their hysteresis mostly from the sensor's own resistance curve rather than a user setting.
Conclusion
Hysteresis and trip delay are the two settings that decide whether a monitoring relay protects the load or annoys the operator. Neither one is a "set and forget" default — both need to be matched to the actual electrical behavior of the supply and the load: how noisy the reading is at the threshold, and how long a normal transient runs before it should stop being normal. Latching mode is a third, independent decision about whether a cleared fault should restart the load on its own or wait for a person. Get all three right and the relay trips on real faults, resets predictably, and stays out of the way otherwise. For the wider function set these settings apply across, see the monitoring relay engineering guide, and browse the current monitoring and control relays range for Schneider Zelio Control and ABB CM series stock.
