A hexapod can pass a basic stroke test and still produce incorrect motion at the cockpit, payload interface, antenna fixture, or vehicle cab. A few thousandths of an inch of actuator offset, an incorrect joint coordinate, or encoder scale error can become a noticeable pose error once six axes work together. Knowing how to calibrate hexapod actuators therefore means validating the complete motion system, not simply zeroing six linear actuators.
For professional simulation equipment, calibration should establish a traceable relationship between the commanded pose and the actual pose under defined operating conditions. The work combines actuator-level measurements, mechanical geometry verification, controller configuration, and independent platform validation. The correct sequence protects equipment, reduces rework, and produces data that can support acceptance testing, maintenance records, and application-specific compliance requirements.
Define What the Calibration Must Prove
Before moving the platform, define the required coordinate frame, payload condition, motion envelope, and allowable error. A motion base supporting a light visual system will not necessarily use the same acceptance criteria as a high-payload flight simulator or a precision antenna test fixture. Requirements may address position, orientation, repeatability, velocity, acceleration, latency, or a combination of those parameters.
Separate actuator calibration from platform calibration. Actuator calibration confirms that each servo axis reports and achieves the correct extension. Platform calibration confirms that the inverse kinematics model converts a requested surge, sway, heave, roll, pitch, and yaw pose into the correct six actuator lengths. Both are necessary. Correct individual leg positions cannot compensate for an incorrect base-joint or platform-joint geometry model.
Establish a controlled test configuration. Record the payload mass, center of gravity, tooling, cable routing, hydraulic or pneumatic utility restraints if present, ambient conditions where relevant, and controller software version. A calibration completed with no payload can be useful for commissioning, but it may not represent the deflection, friction, and drive loading seen in operational service.
Start With Mechanical and Safety Readiness
Calibration data is only as credible as the mechanism being measured. Inspect the actuator mounting points, universal or spherical joints, fasteners, mechanical stops, cable carriers, brake function, lubrication condition, and structural interfaces before enabling motion. Check for joint binding and verify that the platform can reach its intended neutral position without interference.
Confirm that safety circuits, overtravel limits, emergency stops, servo enable logic, and following-error limits are functional. Start with conservative velocity, acceleration, and torque limits. High-speed motion is not needed to establish home references or identify geometry, and it can hide a problem until the platform reaches an unfavorable attitude.
The calibration process also requires a clear distinction between a software limit and a physical limit. Software travel limits should preserve a margin before mechanical stops under all commanded poses. That margin must account for actuator following error, structural compliance, deceleration distance, and the effect of combined multi-axis motion.
Establish a Repeatable Home Reference
Each actuator needs a repeatable datum. Depending on the design, this may be a hard-stop reference, a home switch, an absolute encoder reference, a linear scale index, or a precision external measurement at a defined mechanical condition. The selected method must be repeatable enough for the required platform accuracy.
Home every actuator independently and record the encoder count or feedback value at the reference. Then repeat the homing cycle several times, approaching the reference from the same direction used in normal operation. Compare the resulting values. A home reference that varies from cycle to cycle may point to switch hysteresis, backlash, compliance, coupler movement, servo tuning issues, or inconsistent contact at a hard stop.
For systems with absolute position feedback, do not assume that absolute feedback removes the need for verification. Absolute encoders retain position through power loss, but their reported position must still align with the mechanical coordinate system. Confirm the zero offset against a physical datum and document it in the controller configuration.
Calibrate Actuator Stroke and Feedback Scale
With the actuator referenced, verify the relationship between reported position and actual extension. Use a measurement method appropriate to the tolerance: a calibrated linear scale, laser tracker, laser interferometer, displacement transducer, or other traceable metrology equipment. Measure at multiple points across the usable stroke rather than checking only the ends.
A practical test includes positions near the low and high software limits plus several intermediate points. Command each point from both directions. The difference between approach directions reveals backlash, lost motion, compliance, or friction effects that a one-direction test can miss.
Calculate the feedback scale and offset from the measured data. For servo-driven ball screw actuators, the controller may derive position from motor encoder counts, screw lead, and gear ratio. Any incorrect value in that chain produces proportional stroke error. Systems using a direct linear encoder require separate verification of the encoder scale, alignment, and mounting stability.
Do not use a simple offset correction to hide a nonlinear error. If error changes across the stroke, investigate screw lead variation, encoder resolution or interpolation, mechanical wear, thermal effects, drive configuration, or measurement setup. The appropriate correction may be a scale adjustment, a compensation table, or mechanical service. The decision depends on the source and magnitude of the error, as well as the platform’s required fidelity.
Verify Hexapod Geometry and Kinematic Parameters
The most consequential part of calibrating a six-degree-of-freedom platform is identifying its actual geometry. The controller model needs the location of all six lower joints in the base coordinate frame, all six upper joints in the moving-platform frame, the neutral platform pose, actuator length conventions, and axis sign conventions.
Measure joint centers, not approximate bracket edges. On a spherical joint, the relevant point is the center of rotation. On a universal joint or clevis, the effective center depends on the mechanical design and the kinematic model used by the controller. A small error in a joint center can create coupled position and attitude errors throughout the workspace.
Use consistent coordinate definitions. The base frame must have a documented origin and orientation, while the moving frame must be tied to a physical datum on the payload interface. Confirm whether the controller uses right-hand coordinate conventions and whether roll, pitch, and yaw are applied in the expected rotation order. Errors in sign or rotation order often appear as motion in the wrong direction, but subtler errors can create cross-axis coupling only at larger attitudes.
Enter measured geometry into the inverse kinematics model, then command the neutral pose. All six actuators should settle at their expected nominal lengths without residual force caused by a conflicting model. If the platform must be mechanically forced into neutral, revisit the geometry, actuator offsets, or definition of neutral.
Validate the Actual Platform Pose
A controller reporting the requested pose does not prove that the platform achieved it. Measure the moving platform independently with a laser tracker, photogrammetry system, precision inclinometer, coordinate measurement system, or application-appropriate fixtures and indicators. The instrument must be accurate enough that its uncertainty is small relative to the platform tolerance.
Test neutral first, then single-axis translations and rotations. After that, test combined poses that represent the intended application, such as heave plus pitch, surge plus roll, or a sustained high-angle condition. Combined commands are essential because kinematic and structural errors may not be visible during isolated axis tests.
For each pose, evaluate static error, repeatability, hysteresis, and settling behavior. Then return to neutral and measure whether the platform comes back to the same location and attitude. A platform that reaches target positions but does not return consistently may have friction, joint loading, structural compliance, control-loop issues, or an unstable reference method.
Dynamic validation should follow static calibration. Run representative trajectories at controlled speeds and accelerations while monitoring following error, servo current, vibration, and controller latency. The acceptable result depends on the simulator objective. Flight training applications may prioritize cue fidelity and low latency, while antenna test systems may prioritize settle time and absolute pointing accuracy.
Tune Only After the Position Model Is Correct
Servo tuning cannot correct bad geometry or an inaccurate home offset. First establish valid feedback scaling, actuator references, and kinematic parameters. Then tune the motion control loops to achieve stable tracking with the installed payload.
Review position following error during ramps, reversals, and representative motion profiles. Excessive gain can create vibration or oscillation at the platform, particularly where structural modes, cable forces, or payload flexibility are present. Insufficient gain can create lag and position error. Feedforward, filtering, notch filters, and acceleration limits should be selected from measured system behavior rather than applied as generic settings.
Document the final controller parameters, geometry file, feedback offsets, test instruments, payload configuration, and acceptance results. This baseline is valuable when a drive is replaced, an actuator is refurbished, a payload changes, or an operator reports degraded motion quality.
Treat Calibration as a Lifecycle Control
Hexapod calibration is not a one-time factory event. Recheck it after actuator replacement, joint service, encoder work, structural modification, controller updates, transport, major payload changes, or an impact event. For high-utilization simulators, periodic verification can identify drift before it affects training fidelity or test validity.
Servos & Simulation approaches calibration as part of the complete motion-system lifecycle: mechanical condition, servo performance, controller behavior, and application-level motion accuracy must agree. The most useful calibration result is not merely six corrected actuator values. It is documented evidence that the platform produces the intended physical motion, safely and repeatably, where the customer’s equipment actually operates.
