Linearity Testing of Inertial Test Equipment: Principles, Methods, and Standards

Linearity testing of inertial test equipment verifies that accelerometers, gyroscopes, and inertial measurement units produce outputs that remain directly proportional to their input stimuli across the full dynamic range. This technical article explains the underlying principles, sample types, testing items, calibration methods, and applicable standards—such as the ISO 16063 series and GB/T 20485 series—commonly applied by accredited third-party laboratories. Understanding these procedures helps manufacturers and end users ensure measurement accuracy, metrological traceability, and regulatory compliance for inertial sensing and excitation systems.

Technical Overview

Linearity testing evaluates the degree to which the output of inertial test equipment remains directly proportional to its physical input—linear acceleration or angular rate—over the entire specified measuring range. In practice, the calibration data of a device under test are compared against an idealized straight line, and the maximum deviation, normalized to the full-scale output or to the local reading, is reported as the non-linearity or linearity error. Because inertial navigation, guidance, and vibration monitoring systems rely on a constant scale factor, any uncorrected non-linearity translates directly into systematic measurement error that increases with input magnitude.

Linearity verification covers two complementary categories of equipment. The first category comprises inertial sensing devices, including piezoelectric, piezoresistive, capacitive, and MEMS accelerometers, as well as MEMS, fiber-optic, and ring-laser gyroscopes and integrated inertial measurement units (IMUs). The second category comprises the excitation and reference instruments themselves—electrodynamic vibration exciters, precision centrifuges, and rate tables—whose motion or generated acceleration field must be verified for linearity before the instruments can serve as calibration references. Both categories are addressed within a common metrological framework that maintains traceability to the International System of Units (SI).

Several definitions of the reference straight line are in common use, and the chosen definition must always accompany the reported result:

• Terminal-based (endpoint) linearity: the straight line connecting the output values at zero input and full-scale input.
• Least-squares linearity: the best-fit line obtained by linear regression over all calibration points, which minimizes the sum of squared residuals.
• Independent (minimax) linearity: the line that minimizes the maximum deviation, producing the smallest possible non-linearity figure.

The principal normative references include the ISO 16063 series for the calibration of vibration and shock transducers, ISO 5348 for accelerometer mounting, ISO 2041 for mechanical vibration and shock vocabulary, and the corresponding GB/T 20485 and GB/T 13823 series. Measurement uncertainty is evaluated in accordance with ISO/IEC Guide 98-3 (the GUM), and accredited third-party laboratories typically operate under an ISO/IEC 17025 quality management framework.

Testing Samples

Inertial Accelerometers

Servo (force-balance) and navigational-grade accelerometers, industrial piezoelectric and IEPE accelerometers, piezoresistive and capacitive accelerometers, and MEMS inertial sensors. Measuring ranges typically extend from a few g for navigation devices to tens of thousands of g for shock and ballistic applications, and each range must be exercised across representative amplitude points to characterize linearity.

Gyroscopes and Angular Rate Sensors

MEMS Coriolis vibratory gyroscopes, dynamically tuned gyroscopes, fiber-optic gyroscopes (FOG), ring-laser gyroscopes (RLG), and hemispherical resonator gyroscopes. Rate ranges span from fractions of a degree per hour for strategic-grade sensors to several thousand degrees per second for automotive and stabilization applications.

Inertial Measurement Units and Systems

Multi-axis IMUs, attitude and heading reference systems (AHRS), and complete inertial navigation systems, where the scale factor linearity of each accelerometer and gyro channel is verified individually and the residual cross-axis coupling is evaluated.

Excitation and Reference Equipment

Electrodynamic vibration exciters, precision centrifuges, precision rate tables and turntables, angular vibration exciters, and shock calibration systems. Linearity verification of this equipment confirms that the applied stimulus amplitude, velocity, or angular rate is proportional to the commanded value over the working range.

Inclinometers and Gravity-Reference Sensors

Tilt sensors and inclinometers whose angular output must remain linear with respect to the applied tilt angle, verified using precision indexing tables with angular resolutions in the arc-second range.

Testing Items

A typical linearity detection program for inertial test equipment includes the following measured characteristics:

• Scale factor (sensitivity) and scale factor non-linearity over the full input range
• Amplitude linearity at a fixed reference frequency across the dynamic range
• Frequency response flatness and linearity over the operating bandwidth
• Bias (zero offset), bias stability, and bias repeatability
• Hysteresis evaluated through ascending and descending input sequences
• Asymmetry between positive and negative input directions
• Cross-axis (transverse) sensitivity and inter-axis coupling
• Threshold, resolution, and dead band at small signal levels
• Residual non-linearity after least-squares or polynomial curve fitting
• Short-term repeatability and long-term reproducibility of the calibration

Results are reported as non-linearity in percent of full-scale output (% FS) or percent of reading, together with the expanded measurement uncertainty and the type of reference line used in the evaluation.

Testing Methods

Static Multi-Position (Tumble) Method

Earth's gravity serves as the reference input. The sensor is rotated in discrete angular steps—commonly 8, 12, 16, or 24 positions—on a precision indexing table or dividing head, covering the full ±1 g static acceleration range. A sine function is fitted to the output-versus-angle data to extract bias, scale factor, and the deviation of the measured points from the fitted response, yielding the static non-linearity. The method is the standard approach for navigational-grade accelerometers.

Precision Centrifuge Method

For accelerations beyond ±1 g, the device is mounted on a precision centrifuge that generates centripetal acceleration as a function of rotational speed squared and effective radius. The acceleration is increased in discrete steps across the range, and the output is regressed against the computed input to determine non-linearity, hysteresis, and scale factor asymmetry. Accurate radius measurement, static index alignment, speed stability, spindle runout control, and cross-axis acceleration compensation are critical to the uncertainty budget.

Vibration Comparison Calibration

Performed in accordance with ISO 16063-21, the device under test is mounted back-to-back with a reference standard accelerometer on an electrodynamic exciter. Sinusoidal excitation is applied at a fixed reference frequency (commonly 160 Hz) while the amplitude is stepped through the dynamic range, allowing amplitude linearity to be assessed point by point. Frequency sweeps at constant amplitude then characterize frequency response linearity. The reference transducer itself is periodically calibrated by primary interferometric methods per ISO 16063-11 to preserve traceability.

Primary Interferometric Calibration

For the highest accuracy requirements, laser interferometry is applied directly to the moving mounting surface. Fringe counting, minimum-point, and sine-approximation methods defined in ISO 16063-11 provide primary vibration calibration, while ISO 16063-13 covers primary shock calibration by interferometry. These techniques establish national reference standards and verify reference-grade transducers used in comparison chains.

Shock Calibration for High-g Linearity

Comparison shock calibration in accordance with ISO 16063-22, using shock exciters or Hopkinson bar systems, subjects the device to short-duration acceleration pulses ranging from hundreds to hundreds of thousands of g. Comparing the peak output with the reference measurement extends the linearity characterization into the shock regime relevant to ballistic, pyrotechnic, and impact-testing applications.

Rate Table Method for Gyroscopes

The gyro output is recorded at a series of stable constant angular rates in both positive and negative directions spanning the full measurement range. A least-squares regression of output versus applied rate defines the scale factor, and the maximum residual from the fitted line is reported as rate non-linearity. Reversal sequences also reveal hysteresis and scale factor asymmetry, while a zero-rate dwell before and after the sequence verifies bias repeatability.

Data Processing and Evaluation

All methods share a common analytical workflow:

• Least-squares fitting of the calibration data and residual deviation analysis
• Expression of non-linearity as percent of full scale or percent of reading
• Separation of hysteresis, repeatability, and pure non-linearity contributions
• Compilation of a complete uncertainty budget in accordance with the GUM
• Assessment of conformity against manufacturer specifications or class limits

Applications

Linearity detection services for inertial test equipment support a broad range of industries and use cases:

• Aerospace and aviation: flight-critical inertial navigation systems, flight data recorders, and structural vibration monitoring
• Automotive: airbag triggering sensors, electronic stability control, and crash-test instrumentation
• Defense and guidance: missile guidance IMUs, munition fuze sensors, and stabilization platforms
• Industrial machinery: condition monitoring, predictive maintenance accelerometers, and balancing systems
• Energy sector: downhole surveying, drilling vibration monitoring, and wind turbine structural monitoring
• Rail transportation: axle-box accelerometers, ride quality measurement, and derailment detection
• Robotics and unmanned systems: attitude estimation, flight stabilization, and odometry for UAVs
• Consumer electronics: qualification of MEMS sensors in mobile devices and wearables
• Metrology and calibration laboratories: reference-grade transducers and excitation equipment in national and accredited calibration chains

FAQ

What is linearity error and how is it expressed?

Linearity error (non-linearity) is the maximum deviation of the actual calibration curve from a specified reference straight line, expressed as a percentage of full-scale output or as a percentage of the reading at the deviation point. The definition of the reference line—endpoint, least-squares, or independent—must always be stated, because different definitions produce different numerical values for the same data set.

What acceleration range can be covered during linearity testing?

The static gravity method covers approximately ±1 g, precision centrifuges typically extend from a few g up to several hundred g, and shock calibration systems reach from hundreds of g to more than 100,000 g for ballistic-grade sensors. Vibration comparison calibration covers the continuous dynamic range, commonly from fractions of a g to several thousand g.

Which standards govern linearity testing of inertial equipment?

The ISO 16063 series defines vibration and shock transducer calibration methods, ISO 5348 specifies accelerometer mounting, ISO 2041 establishes terminology, and the GB/T 20485 and GB/T 13823 series provide corresponding national specifications. Laboratory competence is generally demonstrated under ISO/IEC 17025, and uncertainty evaluation follows ISO/IEC Guide 98-3.

How is linearity distinguished from hysteresis and repeatability?

Linearity describes systematic deviation from a straight line when the input is approached monotonically. Hysteresis is the difference in output for the same input approached from opposite directions, and repeatability is the scatter observed when the same input is applied repeatedly under identical conditions. A complete characterization separates and quantifies all three contributions.

How often should linearity verification be performed?

Reference transducers and excitation equipment are typically recalibrated at twelve-month intervals, or sooner following overload, mechanical shock, exposure to extreme temperature, or repair. Production inertial sensors are generally verified per manufacturing lot or per unit according to the applicable quality plan, with periodic surveillance to confirm long-term scale factor stability.

What measurement uncertainty can accredited laboratories achieve?

Primary interferometric vibration calibration can reach expanded uncertainties below 0.5 percent for sensitivity, comparison calibration typically achieves 1 to 3 percent, centrifuge-based static linearity testing commonly falls between 0.1 and 1 percent of applied acceleration depending on level, and the achievable uncertainty always depends on the sensor type, range, and environmental conditions, and is documented in the calibration certificate.