Repeatability positioning accuracy is one of the most decisive performance indicators of a CNC machining center, directly governing dimensional consistency, scrap rates, and long-term process capability. Determined in accordance with ISO 230-2 and the equivalent Chinese national standard GB/T 17421.2, this test quantifies how closely a machine axis returns to the same commanded position over repeated cycles. This article explains the technical basis, test objects, measured parameters, laser interferometry procedures, data evaluation methods, and practical applications of repeatability positioning accuracy determination for machining centers.
Technical Overview
Repeatability positioning accuracy defines the closeness of agreement between successive positioning movements of a numerically controlled axis when the same commanded target position is approached repeatedly under identical conditions, including the same approach direction, feed rate, and load. In a machining center, this characteristic reflects the combined influence of mechanical backlash, friction behavior in guideways and ball screws, servo tracking performance, encoder resolution, and short-term thermal drift.
Repeatability must be clearly distinguished from positioning accuracy. Positioning accuracy describes the systematic deviation of the achieved mean position from the commanded position across the axis travel, whereas repeatability describes the random dispersion of results around the mean at each target position. A machine may be highly repeatable yet inaccurate, a condition that can normally be corrected through CNC pitch-error compensation. Poor repeatability, by contrast, indicates random mechanical or servo-related defects that cannot be removed by software compensation and therefore represents the practical floor of achievable workpiece accuracy.
The determination is internationally standardized. ISO 230-2 specifies the test code for determining the accuracy and repeatability of positioning of numerically controlled machine tool axes, and it is adopted in China as GB/T 17421.2. Comparable procedures are given in ASME B5.54 and JIS B 6194. Modern precision vertical machining centers typically achieve bidirectional repeatability in the range of 1 to 5 micrometers, while general-purpose machines commonly fall between 5 and 10 micrometers.
Testing Samples
Because the measurand is a kinematic property of the complete machine, the test object is the machining center itself, evaluated axis by axis. Typical categories of equipment submitted for repeatability positioning accuracy determination include the following.
Vertical Machining Centers
The X, Y, and Z linear axes are tested individually using a laser interferometer positioned along each axis of motion. Vertical spindle machines, including high-speed drilling and tapping centers, represent the largest share of routine test volume. The Z-axis test is performed with the spindle head carriage under normal counterbalance conditions.
Horizontal Machining Centers
In addition to the three linear axes, horizontal machines frequently incorporate a rotary B-axis table whose indexing repeatability, often specified in the sub-arc-second range, is verified with an autocollimator and precision polygon or with laser angular optics. Pallet-changing mechanisms may also be checked for repositioning consistency.
Gantry-Type and Large Travel Machines
Long-travel gantry machining centers require special attention to optics support structures, air-path stability, and thermal gradients along the axis. Where travel exceeds the interferometer range, measurements may be executed in overlapping segments with a calibrated step-over technique to maintain traceability across the full stroke.
Five-Axis and Multi-Axis Machining Centers
For machines with swivel heads and rotary tables, linear axes are assessed per ISO 230-2, while rotary axes are characterized using angular measurement instrumentation. Complementary circular-interpolation tests with a telescoping ballbar, described in ISO 10791-6, are frequently added to expose servo mismatch and interpolation errors between coordinated axes.
Testing Items
A complete determination per ISO 230-2 and GB/T 17421.2 yields the following quantified characteristics for each tested axis.
- Unidirectional positional deviations at each target position for each approach direction, together with the mean positional deviation at every position.
- Experimental standard deviation of the positional deviations at each target position and direction.
- Unidirectional repeatability of positioning at each position, evaluated as four times the standard deviation.
- Overall unidirectional and bidirectional repeatability of positioning for the full axis.
- Reversal value at each position, defined as the difference between the mean positional deviations obtained from opposite approach directions, representing effective lost motion or backlash.
- Range of mean bidirectional positional deviations across the travel.
- Axis positioning accuracy, derived from the envelope of mean positional deviation plus or minus two standard deviations at every target position.
- Optional extended tests such as thermal drift behavior under defined duty cycles.
The final test report documents the measurement setup, the status of built-in compensation functions, environmental data, instrument identification and calibration traceability, the raw deviation data, the calculated statistics, characteristic deviation curves, and conformity statements against the specified tolerances.
Testing Methods
Laser Interferometry
The primary measurement technique is displacement interferometry using a frequency-stabilized helium-neon laser with a vacuum wavelength near 632.8 nanometers and a wavelength accuracy at the sub-part-per-million level. A linear interferometer optic and a retroreflector are mounted, one on the machine structure and one on the moving carriage, converting axis displacement into optical path change with sub-micrometer resolution. Air temperature, pressure, and humidity sensors feed the wavelength compensation algorithm based on the Edlén equation, while part temperature sensors apply thermal expansion correction. The laser source and optics are calibrated with traceability to national metrology institutes under an ISO/IEC 17025 compliant calibration system.
Test Procedure
The standardized measurement sequence is performed as follows.
- Thermally stabilize the machine, typically by running a warm-up program until key components reach operating equilibrium.
- Mount and align the laser beam coaxially with the axis of motion to minimize cosine error, and position the optics to reduce Abbe offset effects and dead-path error.
- Select a minimum of five target positions distributed unevenly over the travel, including positions near the stroke limits, deliberately avoiding multiples of the ballscrew pitch to prevent masking periodic screw errors.
- Command the axis to approach each target position at the specified feed rate without overshoot, repeating the cycle a minimum of five times from each direction.
- Record the positional deviation at each stop automatically, then repeat the entire procedure for every axis of the machine.
Data Processing and Evaluation
For each target position, the mean and the experimental standard deviation of the recorded deviations are calculated for each approach direction. In the current edition of ISO 230-2, repeatability at a position is evaluated as four times the standard deviation, corresponding to a plus-or-minus two-sigma band with approximately 95 percent coverage, whereas earlier editions applied a three-sigma band. The reversal value, the range of mean deviations, and the axis positioning accuracy are then derived, plotted as characteristic deviation curves, and compared with the manufacturer tolerance or contractual specification. Conformity decisions account for measurement uncertainty in accordance with the decision rules of ISO 230-1.
Complementary Methods
Where interferometry of linear axes is impractical, high-accuracy incremental linear scales or precision dial comparators referenced to gage blocks may be used for simplified verification, accepting larger measurement uncertainty. Rotary indexing is assessed with autocollimators and calibrated optical polygons. Ballbar circular tests per ISO 10791-6 and standardized machining test pieces, such as the NAS 979 cutting test, provide complementary verification of the combined static and dynamic behavior of the machine under real cutting conditions.
Environmental and Machine Conditions
Valid results require a controlled environment referenced to 20 degrees Celsius, with ambient variation typically held within plus or minus 2 degrees and restricted rates of change. Vibratory sources are isolated, the machine is level and clamped as in production, and the status of backlash and pitch-error compensation is explicitly recorded, because results with compensation enabled and disabled are not interchangeable. Measurement instruments are soaked to thermal equilibrium before data acquisition begins.
Applications
Repeatability positioning accuracy determination serves a wide range of engineering and commercial purposes.
- New machine acceptance testing and manufacturer outgoing inspection before shipment.
- Periodic verification and condition trending within preventive and predictive maintenance programs.
- Requalification of machines after overhaul, retrofit, ball-screw or guideway replacement, and relocation between facilities.
- Generation and validation of CNC compensation tables for pitch error, backlash, and sag.
- Machine capability studies supporting production of aerospace structures, automotive powertrain components, precision die and mold tooling, medical implants, and semiconductor equipment parts.
- Independent verification by accredited third-party laboratories for trade settlement, import and export inspection, and technical dispute resolution.
- Objective evidence for audits under quality and measurement management frameworks such as ISO 9001, IATF 16949, and ISO 10012.
FAQ
What is the difference between positioning accuracy and repeatability positioning accuracy?
Positioning accuracy is a systematic error, describing how far the mean achieved position deviates from the commanded position over the axis travel. Repeatability is a random error, describing the scatter of results when returning to the same position repeatedly. Systematic error can usually be compensated in the CNC, while random error sets the ultimate limit on achievable part consistency.
Why is repeatability expressed as four times the standard deviation?
The current edition of ISO 230-2 evaluates repeatability over a plus-or-minus two-sigma interval, giving a coverage probability of approximately 95 percent for the normally distributed positioning scatter. Earlier editions of the standard applied a three-sigma interval, so results from different standard editions must be compared with care.
Which standards govern the determination?
The governing international test code is ISO 230-2, adopted in China as GB/T 17421.2. Equivalent or closely related procedures are specified in ASME B5.54 in North America and JIS B 6194 in Japan, while ISO 230-1 provides the underlying test setup and decision rules.
How many target positions and measurement cycles are required?
The standards specify a minimum of five target positions distributed over the axis travel and a minimum of five approach cycles in each direction at each position. For high-precision machines, contractual disputes, or tight tolerances, laboratories commonly increase both the number of positions and the number of cycles to strengthen the statistical basis.
Can poor repeatability be corrected through CNC compensation?
No. Compensation tables correct only systematic, repeatable deviations such as pitch error and constant backlash. Random scatter arising from worn guideways, bearing damage, inconsistent friction, or servo tuning problems must be addressed mechanically or through control optimization, which is exactly why the repeatability test is such a powerful diagnostic indicator.
How often should machining centers be tested?
Best practice calls for full determination at machine acceptance, followed by periodic verification at intervals of six to twelve months depending on utilization and tolerance criticality, and immediately after any event that could affect axis kinematics, such as collision, rebuild, relocation, or major component replacement.