TOLERANCE & INSPECTION
Plan and interpret a Gauge R&R study for CNC-part measurements without confusing it with calibration or measurement uncertainty.
Gauge R&R estimates how much observed variation comes from the measurement system under a defined study. It helps decide whether the system can distinguish meaningful part differences for a specific production decision. The conclusion belongs to the studied characteristic, instrument, fixture, method, environment, operators, and part range; it does not automatically transfer to another feature or program.
Repeatability concerns variation when the same measurement is repeated under specified repeat conditions. Reproducibility concerns changes between appraisers or other deliberately varied conditions. In automated CMM work, operator effects may move into loading, datum setup, probe selection, cleaning, and program control rather than disappear.
| Component | What varies | Typical sources |
|---|---|---|
| Repeatability | Repeated readings under defined conditions | Resolution, force, repositioning, instrument noise |
| Reproducibility | Appraiser or reproducibility condition | Alignment, fixturing, technique, appraiser–part interaction |
| Part-to-part | Representative production parts | Actual process variation in the study range |
Calibration links indication to a reference and identifies error under stated conditions. Measurement uncertainty quantifies reasonable doubt around a result. GR&R observes variation in a particular experimental design. A calibrated instrument may still perform poorly in use, while a low GR&R does not prove absence of bias or metrological traceability. Confirm whether the customer needs GR&R, bias, linearity, stability, uncertainty, or a combination.
Select parts spanning the process range, appraisers who use the system, repeated trials, and a randomized or blinded sequence. Include removal and replacement if routine measurement includes it. Control cleaning, temperature, fixture, work instruction, and force. Do not copy a sample size without considering risk, destructive measurement, cycle time, data type, and customer rules.
Reports may compare measurement-system variation with study variation, tolerance, or both. Those denominators answer different questions. Organizations may define acceptance bands, but the governing customer manual and characteristic risk must control the decision. Review plots, appraiser effects, interactions, resolution, and discrimination alongside the headline percentage. An artificially narrow part range can distort interpretation.
Burrs, oil, roughness, thin-part deflection, temperature, contact force, and fixture alignment can dominate a study. CMM results also depend on datum construction, point pattern, stylus, scanning conditions, algorithms, and program revision. GR&R reveals variation but does not diagnose every root cause; observation and focused trials are still required.
Destructive tests, attribute gages, automated systems, and single-appraiser methods need designs different from a conventional crossed study. State the characteristic, tolerance, method, fixture, sample structure, customer MSA rule, report format, and re-study trigger in the RFQ. VietTech can plan inspection within confirmed capability; GR&R scope and acceptance criteria must be agreed before quotation.
This article references the NIST/SEMATECH e-Handbook of Statistical Methods measurement-process chapter and AIAG Measurement Systems Analysis terminology. The applicable manual edition, statistical method, and acceptance rule are contractual matters. Results remain valid only for the studied characteristic, instrument, people, fixture, environment, program, and part range.
FAQ
No. Repeatability covers defined repeat conditions; reproducibility covers appraisers or other deliberately varied conditions.
It may. Calibration and GR&R answer different questions about error and use-condition variation.
No. Use the customer or applicable MSA rule and consider risk, plots, resolution, and study design.
Not automatically. Loading, fixturing, datum setup, probe strategy, and program control can still vary.
After significant changes to equipment, fixture, program, operators, environment, characteristic, or at the interval required by the quality plan.
TOLERANCE & INSPECTION