How a Bimetallic Overload Relay Trips: The Differential Principle

How does a bimetallic overload relay trip? Three bimetal strips, one per phase, heat in proportion to I²R losses from the motor current passing through them, bend as the metal warms, and push a common trip bar that releases the relay's 95-96 NC contact once the bend crosses the calibrated set point defined in IEC 60947-4-1. Skip the differential linkage in that trip bar and a lost phase on a running motor waits for the same slow thermal buildup as an ordinary balanced overload, by which time the windings on the two remaining phases have already taken damage. This article covers the three-strip construction, how the trip bar reads the difference between phases rather than just the sum of the heat, the inverse-time trip curve, the auxiliary contact and reset behavior, ambient compensation, and where the differential mechanism gives way to CT-based electronic sensing.

The overload relay is one leg of the classic motor starter: a fuse or motor protection circuit breaker clears short circuits, a contactor switches the motor on and off, and the overload relay protects against sustained running overload — the third leg is what this piece focuses on. For the full range across brands, see the thermal overload relays collection; for the broader construction and terminology, start with what a thermal overload relay is and how it works.

The Three Bimetal Strips and How They Bend

Each bimetal strip pairs two metals with different thermal expansion coefficients — typically a low-expansion invar-type alloy bonded to a higher-expansion nickel-iron or steel alloy. Motor current flows through a heater winding wrapped around or bonded to the strip, and I²R heating in that winding raises the strip's temperature. Because the two metals expand at different rates, the strip bends toward the low-expansion side as it warms. One strip sits in each phase conductor, so a balanced three-phase overload heats and bends all three strips at close to the same rate.

The bend is proportional to temperature rise, and temperature rise is proportional to the square of the current integrated over time, not to current alone. A strip carrying 150% of its rated current does not bend one and a half times as far as a strip at 100%; over the same heating duration it moves closer to 2.25 times as far. That squared relationship is the physical basis for the inverse-time trip curve covered further down.

The Differential Trip Bar: Sensing Imbalance Between Phases

The three strips push against a common trip bar through individual pins or levers. On a balanced overload the bar moves because all three strips bend together; the relay is reading the average heating. Lose one phase — a blown fuse, a loose lug, a failed contactor pole — and the picture changes: current on the two remaining phases rises to roughly 1.7x normal, while the strip on the open phase cools. The bar no longer moves symmetrically. It tilts.

Why the Tilt Trips Faster Than the Sum of the Heat

A phase-loss-sensitive design links the bar so an uneven bend, not just a large average bend, advances the trip mechanism. The two overheating strips push harder than the third pulls back, and the bar reaches release travel sooner than a plain, non-differential linkage would on the same current imbalance. This is why catalog sheets list "phase-loss sensitive" as a separate feature from the base trip curve: without the differential linkage, a single-phasing motor running at 170% of FLC on two legs can take close to the same time to trip as a balanced 170% overload — well after insulation on the overloaded phases starts to cook. See phase-loss and single-phasing protection for how manufacturers rate this sensitivity.

What we see in the field: single-phasing failures show up more often on submersible pump and compressor circuits with long, vibration-exposed cable runs than the failure-mode tables suggest — exactly where a phase-loss-sensitive relay earns its keep over a basic thermal-only design.

The Inverse Thermal (I²t) Trip Curve

Trip time is not fixed. It follows an inverse-time curve set by IEC 60947-4-1: small overloads take minutes to trip, large ones take seconds, and the curve is standardized into trip classes so a Class 10A relay trips in 2-10 s and a Class 30 relay in 9-30 s, both measured at 7.2x the current setting from a cold start.

The curve shape is what lets a bimetal relay pass a motor's 6-8x FLC starting inrush without tripping, then still catch a 110% sustained overload within a reasonable time — minutes, not hours. Set the class too fast for the load's run-up time, a Class 10 relay on a high-inertia fan for example, and the relay reads the acceleration current as a fault it was never meant to catch. See overload relay trip classes 10A, 10, 20 and 30 for how to match class to load.

Trip Point, Reset, and the NC Auxiliary Contact

Once the trip bar reaches its release travel, a snap-action spring — not the slow bimetal movement itself — opens the 95-96 NC auxiliary contact in a fraction of a second. That contact wires in series with the contactor coil circuit, so the trip drops the contactor out regardless of what the main current-carrying contacts are doing; the overload relay never switches the motor current directly. A separate 97-98 NO contact typically closes at the same instant, available for an alarm or PLC input.

Reset is a separate step from the trip itself. HAND reset holds the mechanism latched open until a person presses the reset button — the default for most motors, since nobody wants a faulted motor auto-restarting into the same overload unattended. AUTO reset allows the mechanism to re-close once the strips cool back down, and is appropriate mainly for pumps and similar loads where an unattended restart does not put anyone at risk.

Ambient Compensation and Mounting

A compensating bimetal strip, wired into the linkage but not exposed to the heater winding, bends with panel ambient temperature and offsets the main strips' response so the trip point does not drift much between roughly -5 and +55/60°C. Without that compensation, the same relay in a hot panel interior trips earlier than one in a ventilated enclosure at identical motor current — the strip is reading two heat sources, motor current and cabinet air, as one.

Mount the relay in the same thermal environment as the rest of the motor's control gear, not bolted to a hot busbar or crammed next to a VFD heatsink, and the calibration holds. This depends on how tightly the panel is packed and how much the enclosure ventilates; a compensated relay narrows the error, it does not erase the effect of an unusually hot mounting location.

Bimetallic vs Electronic: Where the Differential Principle Stops

Electronic (solid-state) overload relays replace the three bimetal strips with current transformers or shunts feeding a microcontroller that models motor heating in firmware. The differential-bar mechanics described above do not exist in that design; phase imbalance is instead computed by comparing the three CT readings directly, which is more sensitive and configurable than a mechanical tilt but also less transparent — a bent strip can be inspected, a firmware trip decision cannot. Electronic units add thermal memory that survives a power cycle, a wider 1:3 to 1:4 setting ratio versus roughly 1:1.5 on bimetal, and often ground-fault and stall detection the bimetal principle cannot provide at all. For a fuller comparison of when the extra electronics justify their cost over a bimetal relay, see thermal vs electronic overload relays.

Conclusion

A bimetallic overload relay is, at its core, three heat-sensitive strips and a bar that tells the difference between "all three phases are running hot" and "one phase went missing." That differential reading is what separates a phase-loss-sensitive relay from a plain thermal-only design, and the inverse-time curve behind it is what lets the same relay ride out a motor's starting inrush while still catching a slow overload minutes later. Set the dial to motor FLC, pick the trip class to match the load's run-up time, and mount the relay where its ambient compensation was designed to work — the mechanism does the rest. For sizing, coordination, and reset behavior across the full range, see the thermal overload relay engineering guide.