Hysteresis error is one of the most decisive parameters governing the accuracy, repeatability, and reliability of smart sensors. This article provides a comprehensive technical overview of hysteresis error determination for smart sensors, covering failure mechanisms, applicable international standards, sample categories, measured parameters, calibration procedures, and calculation methods. It serves as a practical reference for sensor manufacturers, quality engineers, and design teams seeking accredited third-party laboratory verification of sensor performance.
Technical Overview
Hysteresis error refers to the phenomenon in which a smart sensor produces different output values at the same input magnitude depending on the direction from which the input is approached. When the input is swept from zero to full scale (upscale cycle) and then returned from full scale to zero (downscale cycle), the two output curves do not coincide. The maximum deviation between the upscale and downscale outputs at any given input point, expressed as a percentage of the full-scale output span, is defined as the hysteresis error of the sensor.
For smart sensors, hysteresis originates from multiple physical mechanisms: elastic aftereffects and micro-yielding in diaphragm or spring elements, internal material friction in strain gauges and bonded adhesives, mechanical play and friction in linkages, domain wall motion in ferromagnetic and magnetoelastic materials, moisture absorption and desorption lag in capacitive humidity elements, and thermal mass effects in temperature probes. Although embedded digital electronics can linearize gain and offset, hysteresis is a history-dependent, non-deterministic error that cannot be fully corrected by static software compensation, which makes direct measurement indispensable.
The determination of hysteresis follows the general principle established in IEC 61298-2, where the sensor is subjected to a full-range loading and unloading cycle under reference conditions, and the output deviation between corresponding ascending and descending test points is evaluated. Mathematically, the hysteresis error is calculated as: Hysteresis Error (percent of full-scale output span) equals the maximum absolute difference between the upscale output and the downscale output at the same input value, divided by the full-scale output span, multiplied by 100 percent.
Key standards and references commonly applied to hysteresis error determination include:
- IEC 61298-2, covering general methods and procedures for evaluating the performance of process measurement and control devices under reference conditions, including hysteresis and dead band.
- IEC 60770 series, specifying performance evaluation methods for transmitters used in industrial process control systems.
- IEC 60751, which defines tolerance classes and evaluation procedures for industrial platinum resistance thermometers.
- GB/T 17614 series and GB/T 15478, Chinese national standards for industrial-process transmitters and pressure transmitters, harmonized with the corresponding IEC methodologies.
- ASTM E74 and ASTM E4, covering the calibration and verification of force-measuring instruments and force verification of testing machines.
- JIS B 7505 series, JIS C 1602, and JIS C 1604, Japanese industrial standards for pressure gauges, thermocouples, and resistance thermometers.
- IEC 60068-2 series and ASTM E104, relevant to temperature cycling, damp heat cycling, and controlled relative humidity generation used in environmental hysteresis testing.
- ISO/IEC 17025, specifying the general competence requirements for accredited third-party laboratories performing these calibrations.
Testing Samples
Pressure Sensors and Pressure Transmitters
Gauge, absolute, and differential pressure smart sensors based on piezoresistive silicon, capacitive ceramic, sputtered thin-film, or strain-gauge metal diaphragm technologies. These are the most frequently tested sample category because diaphragm elastic aftereffects make hysteresis a dominant error source, particularly at high overpressure ratings.
Temperature Sensors
Platinum resistance thermometers (Pt100, Pt1000), thermocouples, thermistors, and digital temperature probes with integrated transmitters. Thermal hysteresis is evaluated by comparing indicated temperature during heating and cooling sweeps at identical reference bath or furnace set points.
Force Sensors and Load Cells
Strain-gauge load cells, magnetoelastic force sensors, and smart weighing modules. Force hysteresis is determined by applying ascending and descending force steps using reference force standards in accordance with ASTM E74 practices.
Displacement and Position Sensors
Linear variable differential transformers (LVDT), capacitive and inductive displacement probes, magnetostrictive linear position sensors, Hall-effect position sensors, and rotary encoders with integrated signal conditioning.
Humidity Sensors
Capacitive and resistive humidity elements with digital output. Humidity hysteresis is a characteristic specification for these devices and is measured by comparing readings during adsorption and desorption sweeps in a humidity chamber.
Other Smart Sensor Types
Torque sensors, flow meters, level transmitters, inclinometers, accelerometers, and gas sensors, where direction- or history-dependent output deviation must be quantified as part of overall accuracy verification.
Testing Items
Depending on the product specification and applicable standard, a hysteresis determination program typically includes the following measured items:
- Hysteresis error, expressed as a percentage of full-scale output span or of the rated input range.
- Dead band, defined as the range through which the input may be varied in both directions without producing a detectable change in output.
- Repeatability error, evaluated from multiple upscale and downscale cycles at identical test points to separate random scatter from true directional deviation.
- Nonlinearity (independent, terminal-based, or least-squares end-point linearity) measured on the same calibration data set.
- Combined accuracy or total error band, integrating nonlinearity, hysteresis, and repeatability contributions.
- Temperature hysteresis, evaluated as the indication difference between heating and cooling sequences at identical set points.
- Humidity hysteresis, evaluated between adsorption and desorption sweeps for humidity-sensitive devices.
- Zero and span drift before and after cycling, confirming that hysteresis testing does not reveal permanent set or signal instability.
Testing Methods
Static Upscale-Downscale Calibration Method
The reference method follows IEC 61298-2. The sensor is first conditioned by executing several preliminary full-range cycles to seat mechanical elements and stabilize the electronics. A defined number of test points, typically five to eleven evenly distributed values including zero and full scale, is then applied. Starting from zero, the input is increased stepwise to full scale, with a defined stabilization dwell at each point, and outputs are recorded. The input is then decreased stepwise through the same points back to zero. The sequence is repeated for at least three full cycles, and cycle data may be averaged point by point before calculating the maximum upscale-downscale deviation.
Pressure Hysteresis Determination
Pressure generation is performed with deadweight testers, gas or hydraulic pressure controllers, or pressure comparators traceable to national pressure standards. Venting, leak checks, and adiabatic settling are controlled so that the measured deviation reflects sensor hysteresis rather than media instability. Evaluation is conducted in accordance with IEC 60770, GB/T 17614, and GB/T 15478 as applicable to the transmitter class under test.
Temperature Hysteresis Determination
Temperature probes are swept through ascending and descending set points in calibrated stirred liquid baths, dry-block calibrators, or climate chambers with reference thermometers per IEC 60751, JIS C 1602, and JIS C 1604. Sufficient soak time at each set point ensures that both the device under test and the reference standard reach thermal equilibrium, so that residual deviation is attributable to sensor hysteresis.
Force and Load Hysteresis Determination
Ascending and descending force increments are applied using reference force standard machines or calibrated proving rings in accordance with ASTM E74. Load application rates, seating loads, and return-to-zero checks are controlled so that elastic aftereffects in the loading frame do not contaminate the sensor hysteresis result.
Humidity Hysteresis Determination
Humidity elements are cycled between low and high relative humidity set points in a climate chamber, with saturated salt verification per ASTM E104 where applicable. Outputs are compared during the adsorption sweep and the desorption sweep at identical humidity points, with sufficient equilibration dwell at each step.
Data Processing and Reporting
For each test point, the upscale and downscale outputs are compared, and the maximum absolute difference across all points is divided by the full-scale output span to obtain the hysteresis error. Test reports issued by accredited third-party laboratories typically include the calibration curves for both directions, a tabulated point-by-point deviation analysis, the reference standards used with their traceability and uncertainty, environmental conditions during testing, and a conformity statement against the applicable standard or manufacturer specification.
Applications
Hysteresis error determination is applied throughout the sensor product lifecycle, from research and development to production release and periodic recalibration:
- Industrial process control, verifying pressure, level, flow, and temperature transmitters before installation in refineries, chemical plants, power stations, and water treatment facilities.
- Automotive engineering, qualifying pressure sensors for brake, fuel, and engine management systems as well as position sensors for throttle, transmission, and chassis applications.
- Aerospace and defense, where flight-critical sensors require documented hysteresis limits within strict total error budgets.
- Weighing and metrology, confirming force and load cell conformity for commercial weighing instruments and batching systems.
- Medical and laboratory instrumentation, ensuring that pressure and flow sensors in ventilators, infusion devices, and analyzers meet directional accuracy requirements.
- Building automation and IoT, validating humidity, temperature, and air quality sensors used in smart buildings and environmental monitoring networks.
- Quality assurance and product certification, where manufacturers submit samples to accredited third-party laboratories for independent verification against IEC, GB/T, ASTM, EN, UL, or JIS requirements.
FAQ
What is hysteresis error in a smart sensor?
It is the maximum difference between two output readings taken at the same input value, once when the input is approached from below and once from above. It is normally reported as a percentage of full-scale output span.
How is hysteresis error calculated?
The sensor is cycled upscale and downscale across its full range. At each test point, the absolute difference between the averaged upscale and downscale outputs is computed. The largest difference is divided by the full-scale output span and multiplied by 100 to obtain the hysteresis error in percent.
What is the difference between hysteresis, dead band, and repeatability?
Hysteresis is the directional deviation between ascending and descending outputs at the same input. Dead band is the input range over which no output change is observable when the direction reverses. Repeatability is the random scatter of repeated readings taken in the same direction at the same input. All three are evaluated within the framework of IEC 61298-2 because they represent distinct components of total sensor error.
Can digital compensation in smart sensors eliminate hysteresis?
No. Embedded electronics can correct deterministic errors such as offset, gain, and nonlinearity, but hysteresis depends on the prior loading history and internal material behavior. Digital compensation can only partially reduce it, so physical measurement remains the only valid basis for specification and conformity assessment.
What is a typical acceptable hysteresis value?
Acceptance limits depend on accuracy class and application. General industrial pressure transmitters commonly specify hysteresis at or below 0.1 to 0.3 percent of span, high-precision devices may require less than 0.05 percent, and humidity elements are frequently specified between 1 and 3 percent relative humidity. The applicable product standard or manufacturer datasheet defines the governing limit.
Which test conditions are used for hysteresis determination?
Reference-condition testing is performed in a controlled laboratory environment, typically at a stable ambient temperature and humidity as specified in IEC 61298-2, with traceable reference standards, defined stabilization dwells, and documented measurement uncertainty as required by ISO/IEC 17025 for accredited third-party laboratories.
How often should hysteresis verification be performed?
New designs are tested during type approval and design qualification, production units are sampled during routine quality control, and in-service sensors are verified during periodic recalibration intervals, with increased frequency in high-cycle or high-fatigue applications where mechanical hysteresis can grow over service life.