IMU calibration service establishes and documents the sensor errors that determine navigation accuracy, typically through factory six-position static tests and rate-table dynamic runs, with bias and scale-factor models fitted across temperature. A calibration certificate should state the method, reference equipment, test points, residuals, and validity period. Recalibration intervals are set from observed drift, environmental exposure, and application criticality rather than a single universal schedule.
An inertial measurement unit combines accelerometers and gyroscopes whose outputs carry deterministic and random errors. Calibration identifies the deterministic portion so the navigation solution can compensate for it. The principal terms are bias (output with zero input), scale factor (output error proportional to input), and misalignment (sensitive-axis departure from the defined case axes). Nonlinearity, cross-axis sensitivity, and quantization effects are characterized where the application demands it.
For an INS, the IMU is only part of the chain. Calibration also covers the transformation between the IMU frame and the vehicle or case frame, and the alignment procedure used to initialize attitude. A complete service therefore separates sensor-level calibration from system-level alignment checks, because a well-calibrated IMU installed with an uncorrected mounting rotation will still produce heading error.
Factory calibration is performed on precision rate tables, multi-axis positioners, and thermal chambers under controlled laboratory conditions. It can resolve small errors, characterize temperature dependence across the operating range, and produce a full coefficient set. It requires the unit to be removed from service and shipped, which introduces downtime and handling risk.
Field calibration uses the host system, known reference orientations, and sometimes the vehicle's own motion to estimate a reduced error set. It is valuable for verifying that a unit has not shifted, for correcting installation misalignment, and for updating bias estimates before a mission. Field methods generally cannot separate all error terms as cleanly as laboratory equipment, so they complement rather than replace factory calibration. Many programs alternate: factory calibration at overhaul, field checks between.
The six-position static test places the IMU in six orthogonal orientations, typically with each sensitive axis pointing up and down. In each position the accelerometer senses gravity plus its own bias, and the differences between opposing positions separate bias from scale factor and misalignment. The method is simple, repeatable, and well suited to accelerometers and to coarse gyro bias checks using Earth rate.
Rate-table testing rotates the unit at precisely known angular rates about each axis, at multiple rates and in both directions. This separates gyro bias, scale factor, and nonlinearity, and reveals asymmetry between positive and negative rates. Temperature is swept during either method, or the unit is soaked at fixed temperatures and tested at each, so that bias and scale factor can be modeled as functions of temperature rather than treated as constants.
Inertial sensors drift with temperature, and the relationship is often nonlinear and differs between units. A temperature model is built by recording sensor output at a series of stabilized temperatures with the unit held in known orientations or rates. The resulting data are fitted with polynomial or piecewise functions, and the residuals are reported so the buyer knows how much error remains after compensation.
Good practice includes hysteresis checks, because a sensor's output at a given temperature may depend on whether it was approached from above or below. Models are stored in the unit or in the host software, and their coefficients become part of the configuration baseline. If the model is regenerated, the new coefficients should be versioned and traceable to the calibration record.
A useful certificate identifies the unit by serial number and configuration, the procedure and revision used, the reference instruments with their calibration status, and the environmental conditions during test. It should list the test points, the measured values, the fitted coefficients, and the residuals or uncertainties at those points. The date of calibration and the recommended next-due date belong on the document.
Traceability matters: the certificate should name the standards against which the reference equipment was calibrated and the laboratory that performed it. Where a specification is claimed, the certificate should state the acceptance limits and whether each was met. A certificate that reports only a pass or fail without data is difficult to audit and offers little basis for predicting future drift.
There is no universal interval. Recalibration frequency is set from the sensor technology, the stability observed in prior calibrations, the thermal and vibration environment, and the consequence of navigation error in the application. A common starting point is annual calibration for precision units in benign conditions, shortened for high-vibration or wide-temperature service, and extended where repeated calibrations show little change.
The defensible approach is to track calibration history and adjust the interval from evidence. If successive calibrations show bias shifts well inside the error budget, the interval can be lengthened; if drift approaches the budget, it should be shortened. Field checks between factory calibrations provide early warning. The preserved Applied Analytics inertial navigation record is referenced in this archive as a historical example of inertial system documentation practice, not as a current certification.
IMU calibration characterizes the accelerometers and gyroscopes themselves: bias, scale factor, misalignment, and temperature dependence. INS calibration adds the system-level relationship between the IMU frame and the case or vehicle frame, plus alignment and initialization checks. A complete service addresses both, because sensor accuracy alone does not guarantee navigation accuracy.
Yes, within limits. Field methods use known orientations and host-system references to estimate bias and installation misalignment and to verify that a unit has not shifted. They generally cannot separate all error terms as precisely as laboratory rate tables and thermal chambers, so field calibration is usually a check between factory calibrations rather than a full replacement.
Look for unit identification, procedure revision, reference equipment and its traceability, test conditions, test points, measured values, fitted coefficients, residuals or uncertainties, acceptance limits, and the calibration and next-due dates. A certificate reporting only pass or fail without supporting data is hard to audit and gives little basis for predicting drift.
This page is an archival specification note for informational purposes only. The former plant is not quoting from this domain. Nothing here is a qualified procedure, an inspection report, or a design calculation; confirm current requirements with the issuing standards body and your material supplier.