A rotating shaft under torque twists by an angle proportional to the applied torque, the shaft length, and the material's shear modulus. For a solid steel shaft 25 mm in diameter and 500 mm long, a torque of 100 N·m produces a twist angle of roughly 0.15° — barely visible but easily measurable with strain gauges, optical encoders, or magnetic field sensors. The challenge is not measuring the twist; it is getting the measurement signal off a shaft that is rotating at up to 30,000 rpm without a physical electrical connection that wears out. This is the slip ring problem, and it has driven the development of non-contact torque measurement technologies that eliminate the last wearing component in the torque measurement chain.
Slip Ring Torque Sensors: Proven, Affordable, and High-Maintenance
A slip ring torque sensor bonds four strain gauges to the shaft in a Wheatstone bridge configuration. The bridge excitation voltage and the millivolt-level bridge output signal pass through a set of spring-loaded silver-graphite brushes riding on polished brass or silver slip rings. At speeds below 3,000 rpm with clean, dry air, a well-maintained slip ring torque sensor operates reliably for 500–1,000 hours between brush inspections. At speeds above 10,000 rpm, brush bounce generates electrical noise of 1–5% of full-scale output — a 1,000 N·m sensor may show 10–50 N·m of noise, rendering it useless for precise measurements below 10% of full scale. At speeds above 20,000 rpm, centrifugal force lifts the brushes off the rings entirely, breaking the electrical connection. And in any environment with oil mist, metal dust, or high humidity, the brush-ring interface degrades two to five times faster than in clean-room conditions.
The economic case for slip rings is straightforward: a slip ring torque sensor costs $2,000–5,000 for common ranges (10–5,000 N·m), making it the lowest-initial-cost option for test stand applications where the sensor runs 200–1,000 hours per year and can be removed for brush replacement and recalibration during planned downtime.
Non-Contact Magnetic Torque Measurement: How It Eliminates the Slip Ring
Non-contact torque measurement replaces the strain-gauge-plus-slip-ring assembly with one of two physical principles. The first — magnetoelastic — exploits the Villari effect: when a ferromagnetic material is mechanically stressed, its magnetic permeability changes. A shaft made of a magnetoelastic alloy (typically a nickel-iron-molybdenum maraging steel) develops a measurable change in its magnetic anisotropy when torque is applied. A pair of magnetic field sensors — Hall-effect, fluxgate, or magnetoresistive — positioned 1–2 mm from the shaft surface detect this change without physical contact. The sensor head encircles the shaft but does not touch it, and the signal passes through a stationary electronics module via a cable — no brushes, no rings, no wear surfaces.
The second principle — phase-displacement — uses two multi-pole magnetic rings press-fitted or adhesively bonded to the shaft at a known distance apart. Each ring generates a sinusoidal magnetic field pattern as the shaft rotates. Non-contact Hall sensor arrays mounted 1–3 mm from each ring detect the phase of each magnetic pattern. When torque is applied, the shaft twists, the two magnetic rings rotate slightly relative to each other, and the phase difference between the two sensor signals is directly proportional to the torque. This method measures the twist angle of the shaft between the two rings — it does not require a magnetoelastic shaft material, so it can be retrofitted onto an existing steel shaft without replacing it.
When Non-Contact Pays Back
The strongest case for non-contact torque measurement is endurance testing. An automotive transmission test stand running 24/7 accumulates 8,760 hours per year. A slip ring sensor on that stand requires 8–17 brush changes per year, each requiring sensor removal, disassembly, brush replacement, reassembly, and recalibration against a deadweight torque standard — 4–8 hours of downtime per service event. A non-contact magnetic sensor on the same stand runs for the full 8,760 hours with zero contact-wear-related service events. At a burdened test cell cost of $150–300 per hour, recovering 32–136 hours of downtime per year saves $4,800–40,800 annually — typically exceeding the price premium of the non-contact sensor within the first year.
High-speed applications above 15,000 rpm are the second decisive case. A slip ring sensor at 20,000 rpm simply does not work — the brushes lift off. A non-contact sensor at the same speed has no speed-dependent wear mechanism; its maximum operating speed is limited by the shaft's mechanical balance and the bearing's DN limit, not by the measurement principle. For electric motor dynamometer testing, gas turbine accessory gearbox testing, and high-speed spindle torque monitoring above 15,000 rpm, non-contact is not an upgrade option — it is the only option.
For load monitoring and measurement across rotating machinery, see our sensors & instrumentation and vibration sensors categories for complementary condition monitoring technologies.
Can a non-contact torque sensor measure static torque?
Magnetoelastic sensors can measure static torque — the shaft does not need to rotate for the Villari effect to produce a measurable permeability change. Phase-displacement sensors require shaft rotation to generate the sinusoidal magnetic signal that the Hall sensors detect. At zero speed, there is no phase signal. Some phase-displacement sensors include a secondary static measurement method — typically a strain gauge bridge on the stationary housing that measures the reaction torque — but this adds cost and complexity. For test stands that need to measure both static torque (preload on a bolted joint, breakaway torque of a seized bearing) and dynamic torque, magnetoelastic sensing is the more versatile choice.
How often does a non-contact torque sensor need calibration?
Since there are no wearing components in the measurement chain, the primary source of calibration drift in a non-contact torque sensor is mechanical — creep in the adhesive bond of a phase-displacement ring or stress relaxation in the shaft material of a magnetoelastic sensor. In practice, non-contact sensors hold their calibration for 2–5 years versus 6–12 months for slip ring sensors under comparable usage. The calibration procedure is the same for both technologies: apply a known torque with a calibrated deadweight arm and lever, record the sensor output at 5–10 points across the full scale, and verify linearity and hysteresis are within specification. Browse our calibration & maintenance category for torque calibration standards.
