GB Standard Testing for Machine Tool Positioning Accuracy: GB/T 17421.2 Methods, Parameters and Comp

Machine tool positioning accuracy testing to GB standards is the core procedure for verifying how precisely the numerically controlled axes of a CNC machine can reach commanded positions, based on GB/T 17421.2 (aligned with ISO 230-2). Measured primarily with laser interferometry, the test quantifies positioning accuracy, repeatability and reversal value for acceptance inspection, periodic verification and error compensation. This guide explains the applicable GB standards, test items, measurement procedures, data evaluation and typical applications for manufacturers and accredited third-party laboratories.

Technical Overview

Positioning accuracy describes the ability of a numerically controlled (NC) axis to approach a commanded target position within a defined tolerance, and it directly determines the dimensional accuracy of machined workpieces. In the Chinese national (GB) standard system, this performance is evaluated according to GB/T 17421.2, "Test code for machine tools — Part 2: Determination of accuracy and repeatability of positioning of numerically controlled axes", which adopts ISO 230-2 with identical technical content. The test yields statistical parameters — positional accuracy, repeatability of positioning, mean positional deviation and reversal value — that are compared with the permissible limits stated in the machine's product standard or technical specification.

The main standards relevant to positioning accuracy verification include:

• GB/T 17421.2 — determination of accuracy and repeatability of positioning of numerically controlled axes (aligned with ISO 230-2)
• GB/T 17421.1 — geometric accuracy tests of machine tools under no-load or quasi-static conditions (aligned with ISO 230-1)
• GB/T 17421.7 — geometric accuracy of axes of rotation, applied to rotary tables and swivel heads (aligned with ISO 230-7)
• GB/T 16462 series — test conditions and acceptance requirements for numerically controlled turning machines
• GB/T 20957 series — test conditions for machining centres

Testing is normally performed by machine manufacturers during factory acceptance, by accredited third-party laboratories for independent verification, or by end users for periodic calibration and maintenance. Because thermal effects dominate modern machines, the standards require the machine and the measuring instruments to be thermally stabilized and the ambient conditions to be recorded and, where necessary, compensated.

Testing Samples

The object of the test is a complete machine tool or a specific motion axis (linear or rotary). Typical test objects include:

CNC Machining Centres

Vertical, horizontal and gantry machining centres, including five-axis machines with simultaneous rotary axes. Each linear axis (X, Y, Z) and each indexing rotary axis is tested over its full travel.

CNC Lathes and Turning Centres

Numerically controlled lathes and turning–milling centres, with tests on the X and Z axes and, where applicable, on the positioning performance of turrets and driven tool stations.

Grinding, EDM and Special-Purpose Machines

Surface and cylindrical grinders, wire-cut and die-sinking EDM machines, CNC boring machines and dedicated equipment whose process capability depends directly on axis positioning performance.

Machine Subassemblies and Retrofitted Equipment

Ball-screw-driven slides, linear-motor stages, rotary index tables, and machines after CNC retrofit, ball screw replacement or rebuild, which are verified before being returned to production.

Testing Items

GB/T 17421.2 defines a complete set of statistical parameters derived from the measured positional deviations at each target position, approached from both directions:

• Unidirectional positional accuracy (A↑ and A↓) — the extreme spread of positional deviation when approached from one direction only
• Bidirectional positional accuracy (A) — the overall accuracy considering both approach directions
• Unidirectional repeatability of positioning (R↑ and R↓) — the scatter band of deviations at each position for a single approach direction
• Bidirectional repeatability of positioning (R) — the scatter band including both directions, incorporating the reversal effect
• Mean bidirectional positional deviation (M) — the mean of the mean unidirectional deviations over all target positions
• Reversal value (B) at each position and mean reversal value (B̄) — the difference between mean positional deviations approached from opposite directions, commonly related to lost motion or backlash
• Systematic positional deviation (E) — the range between the largest and smallest mean positional deviations along the axis

These items are reported together with graphs of positional deviation versus target position, which reveal characteristic error patterns such as ball screw pitch error, compensation residuals, slope caused by thermal expansion or scale misalignment, and periodic single-pitch errors.

Testing Methods

Measuring Equipment

The reference method is laser interferometry: a frequency-stabilized helium–neon laser with beam steering optics, a retroreflector or plane mirror mounted on the moving component, and environmental compensation sensors for air temperature, pressure and humidity, plus material temperature sensors attached to the machine structure. The laser wavelength must be traceable to national metrology standards, and the test uncertainty ratio relative to the tolerance should be adequate, typically at least 3:1. Alternative equipment includes calibrated step gauges or line scales used with displacement indicators, and grid encoders for dynamic error separation.

Test Conditions and Preparation

The machine is installed and levelled according to the manufacturer's instructions, thermally soaked in the test environment, and exercised through warm-up cycles before measurement. Ambient temperature is held as stable as practicable around the 20 °C reference temperature; deviations, gradients and material temperatures are recorded so that compensation can be applied. Non-essential auxiliary equipment is switched off, and the state of the CNC's built-in pitch-error and backlash compensation (active or disabled) is agreed and documented, since the result applies only to the configuration tested.

Measurement Procedure

• Select at least five target positions unevenly distributed over the useful travel of the axis, including points close to both ends of travel; the spacing must not coincide with the ball screw pitch period.
• Position the moving component at a starting point outside the measurement span, then approach each target in sequence at the agreed feed rate (usually rapid traverse) with a brief settling dwell.
• Repeat the cycle in the reverse direction; perform at least five cycles in each direction (n ≥ 5) to establish the mean and standard deviation at every target for each direction.
• Record the positional deviation at every approach together with axis position, direction, feed rate and environmental data.

Data Evaluation

For each target position, the mean positional deviation and the experimental standard deviation are computed for each approach direction. Positioning accuracy is derived from the extreme values of the statistical estimator defined by the standard — the mean deviation combined with a coverage factor of 2, corresponding to approximately 95 % confidence, in the editions aligned with ISO 230-2:2014, whereas earlier editions used a 3σ band. Repeatability at a position equals four times the standard deviation (bidirectional repeatability additionally includes the reversal value), and the reversal value is the difference of the mean deviations of the two directions. The reported parameters A, R, B, E and M are compared with the tolerance limits specified by the applicable GB product standard, the declared accuracy class of the machine, or the contractual specification.

Compensation Verification

The test is frequently repeated before and after CNC error compensation. Comparing the two datasets verifies the effectiveness of pitch error compensation and backlash compensation and documents the achievable accuracy for acceptance. Any retest after adjustment is performed under the same conditions to ensure comparability of results.

Applications

• Factory and site acceptance of new CNC machine tools against GB product standards and purchase specifications
• Periodic verification and calibration of installed machines supporting quality management systems and process control
• Verification after maintenance, rebuild, ball screw or guideway replacement, and CNC retrofit
• Development and validation of CNC error compensation tables, including pitch error and thermal compensation
• Independent test reports issued by accredited third-party laboratories for tenders, disputes and export compliance
• Benchmarking of machine tool performance for procurement decisions and capability studies

FAQ

What is the relationship between GB/T 17421.2 and ISO 230-2?

GB/T 17421.2 is the Chinese national standard that adopts ISO 230-2 with identical technical content. Test procedures, symbols (A, R, B, E, M) and evaluation rules are therefore fully compatible with international practice, allowing results to be recognized in both domestic and international contexts.

How many target positions and measurement cycles are required?

The standard requires at least five target positions per axis, spread unevenly over the travel and including the extreme positions, and at least five measurement cycles in each direction. Longer axes use more target positions so that periodic screw errors are captured reliably.

What does a large reversal value indicate?

The reversal value reflects lost motion between opposite approaches — commonly backlash in the ball screw nut, gearing or bearings, loss of preload, or elastic deflection. A reversal value beyond specification usually calls for mechanical adjustment before compensation is applied.

How is the permissible positioning accuracy determined?

GB/T 17421.2 defines how to measure and evaluate, not the tolerance itself. The limits come from the machine's product standard (for example the GB/T 16462 or GB/T 20957 series), the manufacturer's declared accuracy class, or the acceptance terms agreed between supplier and buyer.

Can the test be performed on-site?

Yes. Portable laser interferometer systems allow testing in the user's workshop, provided the environment is thermally stable and conditions are documented. On-site testing by accredited third-party laboratories supports independent acceptance and periodic verification.

Why does temperature matter so much?

Steel expands by roughly 11 μm per metre per kelvin, so a 1 K error in material temperature compensation can shift results on long travels by more than the entire tolerance. This is why environmental sensing, thermal soak and, where required, temperature correction are mandatory elements of a valid GB/T 17421.2 measurement.