Robotic arm repeatability positioning accuracy testing is the internationally recognized method for verifying that an industrial manipulator can return to a taught pose within a defined tolerance, cycle after cycle. This test plan outlines the applicable standards, sample preparation, test items, measurement procedures, and evaluation methods based on ISO 9283 and GB/T 12642, providing engineers and quality managers with a complete framework for validating robot performance in production, research, and third-party laboratory environments.
Technical Overview
Repeatability positioning accuracy defines how closely a robotic arm returns to the same taught position and orientation when approaching the same target multiple times under identical conditions. It is distinct from absolute positioning accuracy, which measures the deviation between the commanded pose and the actually attained mean pose. In practice, most industrial robots are taught by demonstration, so repeatability is the decisive performance specification for applications such as pick-and-place, precision assembly, welding, dispensing, and machine tending.
A rigorous test plan ensures that measurement conditions are controlled, test poses are representative of the working envelope, the number of observation cycles is statistically sufficient, and the evaluation method is traceable to an international standard. The primary reference is ISO 9283, which specifies performance criteria and related test methods for manipulating industrial robots, harmonized in China as GB/T 12642. Related references include GB/T 12642/ISO 9283 companion specifications for test path characteristics and the safety requirements of ISO 10218-1, which define the conditions under which performance verification is performed. The test plan described below can be executed by manufacturers, end-user quality departments, or accredited third-party laboratories.
Testing Samples
The test plan applies to all major categories of industrial and collaborative manipulators. Sample preparation and configuration must be documented before testing begins.
Six-Axis Articulated Robots
General-purpose industrial robots with a vertically articulated kinematic chain. These are the most common test subjects and typically exhibit repeatability values in the range of 0.02 mm to 0.08 mm depending on payload class and reach.
SCARA Robots
Selective Compliance Assembly Robot Arm units used for high-speed planar pick-and-place and assembly. Their rigid Z-axis and compliant planar mechanics typically yield repeatability of 0.01 mm to 0.02 mm.
Delta and Parallel-Link Robots
High-speed parallel kinematic machines used in food, pharma, and electronics packaging, commonly tested at very high cycle rates with lightweight payloads.
Collaborative Robots
Lightweight robots designed for human-robot collaboration, verified under both powered and force-limited operating modes, with repeatability typically between 0.02 mm and 0.1 mm.
Cartesian and Gantry Robots
Linear-axis systems used for large-envelope handling, dispensing, and CNC tending, tested along each axis individually and at three-dimensional target points.
Sample Preparation Requirements
- The robot must be mounted on the manufacturer-specified foundation with verified rigidity and leveling.
- A representative payload, normally the rated payload or the actual production end effector plus part mass, must be securely attached to the flange.
- Cable dress packs, tool changers, and peripheral components must be installed in the production configuration.
- The controller must be set to the intended production parameters, including payload data, speed overrides, and motion profile settings.
- No external mechanical stops, fixtures, or compliance devices may assist the positioning during measurement.
Testing Items
The test plan evaluates the following performance characteristics, all defined in ISO 9283 and GB/T 12642:
- Unidirectional pose repeatability (RP): the dispersion of attained positions when the robot approaches a target repeatedly from the same direction, expressed as the mean radial distance from the centroid plus three times the standard deviation of the radial distances.
- Orientation repeatability (RPa, RPb, RPc): the dispersion of attained orientation about the three rotational axes of the tool coordinate frame, evaluated as three times the standard deviation of angular deviations.
- Multidirectional pose repeatability (RPo): repeatability evaluated when approaching the same target from different directions, exposing backlash and servo hysteresis effects.
- Pose accuracy (AP, APa, APb, APc): the deviation between the commanded pose and the mean of the attained poses, measured as a supplementary item to fully characterize static positioning behavior.
- Pose stabilization time and overshoot: the time required for the end effector to settle within a defined zone and the amplitude of oscillation around the target.
- Drift of pose repeatability (dRP): the change in mean pose over extended cycling, used to assess thermal and mechanical stabilization.
- Path repeatability (RT): an optional extension for continuous-path applications such as welding, dispensing, and cutting, evaluated on linear and circular reference paths.
Testing Methods
Measurement Equipment
- Laser tracker with a spherically mounted retroreflector (SMR) attached at or near the tool center point, offering volumetric accuracy on the order of micrometers plus a few micrometers per meter of range; this is the preferred instrument for ISO 9283 verification.
- Telescoping ballbar systems for circular path repeatability and backlash screening.
- Coordinate measuring machines (CMM) for small-envelope robots and calibration artifacts.
- Optical motion capture systems with infrared markers for dynamic trajectory analysis.
- Precision dial indicators and gauge blocks for simplified on-site verification when traceable screening is sufficient.
- Calibrated environmental monitoring instruments for temperature, humidity, and vibration recording throughout the test.
Test Setup and Pose Definition
A reference test cube is defined within the working envelope of the robot, with edges aligned to the base coordinate system and dimensions as large as the workspace permits. Following the ISO 9283 standard test cycle, five measurement points are defined: point P1 at the center of the cube and points P2 through P5 located at the corners of a square centered on P1 in the diagonal plane of the cube, offset by half the cube edge length along the plane diagonals. A start position S is defined outside the cube, and the robot executes the programmed cycle S to P1, P2, P3, P4, P5 and back to S, pausing at each target long enough for the measurement system to capture a settled static reading. Tests are performed on the primary diagonal plane, with additional planes added for extended verification.
Test Conditions
- Test loads: rated payload as the primary condition; reduced load or no-load conditions added when the manufacturer specification requires them.
- Test speeds: 100 percent, 50 percent, and 10 percent of the rated velocity between poses.
- Environmental control: ambient temperature maintained at 20 degrees Celsius plus or minus 2 degrees Celsius, or within the manufacturer-specified range, with temperature gradient and humidity logged continuously.
- Warm-up: a minimum of 10 complete cycles executed before data acquisition to stabilize lubricant films, drivetrain preload, and controller thermal state.
- Vibration isolation: no measurable external vibration sources from presses, compressors, or transport equipment during measurement windows.
Measurement Procedure
- Record robot identification, controller and software versions, payload, tool data, and environmental baseline conditions.
- Install the measuring instrument, establish the reference coordinate frame, and qualify the setup with a gauge verification measurement.
- Teach the five target poses and the start position, and verify programmed approach directions and pause times.
- Complete the warm-up cycles, then execute 30 measurement cycles per point per condition as specified by ISO 9283, recording the full three-dimensional position and orientation at each settled target.
- Repeat the sequence for each load, speed, and plane combination included in the test matrix.
- Export raw data with timestamps and instrument uncertainty statements for evaluation.
Data Evaluation
For each target point and condition, the arithmetic mean of all attained positions is calculated as the centroid. The radial distance of every individual attained position from the centroid is computed, and pose repeatability is evaluated as the sum of the mean radial distance plus three times the standard deviation of the radial distances, which corresponds to the ISO 9283 formula RP equals mean radius plus 3 sigma. Orientation repeatability for each axis is evaluated as three times the standard deviation of the corresponding angular deviation about the mean orientation. Results are compared against the manufacturer datasheet specification and the acceptance limit agreed in the test plan, typically requiring the measured value to be equal to or better than the specified value under rated load and rated speed conditions.
Test Report
The final report documents the robot configuration, serial numbers, test load, speed settings, environmental records, instrument calibration certificates, raw measurement data, per-point statistical results, expanded measurement uncertainty, and a clear conformity statement against the acceptance criteria. Reports issued by accredited third-party laboratories additionally state the accreditation scope and traceability chain to national measurement standards.
Applications
- Acceptance testing of new robots at commissioning, verifying that delivered units meet contractual repeatability specifications before production release.
- Periodic verification in predictive maintenance programs, typically annually or semi-annually, to detect gradual degradation from wear, backlash, and calibration drift.
- Post-event verification after collisions, overloads, gearbox or encoder replacement, controller swaps, or robot relocation.
- End-of-line quality control by robot manufacturers and system integrators prior to shipment.
- Research and development benchmarking of prototypes, alternative drivetrain configurations, and calibration algorithms.
- Process qualification in automotive body shops, electronics assembly, semiconductor handling, aerospace drilling and fastening, medical device manufacturing, and precision dispensing, where documented positioning capability is a customer or regulatory requirement.
- Independent conformity assessment by accredited third-party laboratories for export certification, tender qualification, and dispute resolution.
FAQ
What is the difference between positioning accuracy and repeatability? Accuracy describes how close the mean attained pose is to the commanded pose, while repeatability describes the dispersion of attained poses around their own mean when returning to the same taught point. A robot can be highly repeatable yet inaccurate, which is acceptable for taught-point applications.
How many measurement cycles are required? ISO 9283 and GB/T 12642 specify 30 measurement cycles per target point per test condition, following at least 10 warm-up cycles, to provide statistically stable dispersion estimates.
What repeatability values are typical? Six-axis industrial robots commonly specify 0.02 mm to 0.06 mm, SCARA robots 0.01 mm to 0.02 mm, and collaborative robots 0.02 mm to 0.1 mm. Always compare measured results to the specific datasheet value at the declared load and speed.
Which standards govern the test? ISO 9283 is the international reference and GB/T 12642 is the corresponding Chinese national standard; both define the test poses, cycle counts, evaluation formulas, and reporting requirements used in this test plan.
Can testing be performed on-site at our facility? Yes. Portable laser trackers allow full ISO 9283 verification at the installation site, provided that ambient temperature, airflow, and vibration are controlled and documented during measurement.
What factors most influence the results? Payload and its offset from the flange, commanded speed, thermal state of the drivetrain, mounting foundation rigidity, backlash, and measurement uncertainty all contribute, which is why the test plan fixes these parameters explicitly.
How often should repeatability be retested? Best practice is at commissioning, after any collision or major repair, after relocation or recalibration, and at planned intervals of six to twelve months as part of a predictive maintenance program.
How is pass or fail decided? The measured RP value for each target, corrected for measurement uncertainty, must be equal to or smaller than the acceptance limit defined by the manufacturer specification or the agreed contractual value, evaluated under rated load and rated speed.
