Repeatability
Variation when the same appraiser measures the same part repeatedly using the same method and equipment under the same conditions.
Example: One inspector measures the same shaft three times with the same micrometer.
SPI
Special Processes Institute
sp-i.org
Measurement System Analysis Report
Measurement System Analysis & Gauge R&R Project App
Understand the complete Measurement System Analysis framework, learn how Gauge Repeatability and Reproducibility fits within MSA, explore practical examples, calculate a crossed variable Gauge R&R study, and build a controlled measurement-system project with portable JSON and a professional report.
Evaluate measurement confidence Working Gauge R&R calculator Validate before SPC or capability
5
Core studies: bias, linearity, stability, variable R&R and attribute agreement
EV + AV
Gauge R&R combines repeatability and reproducibility
0%
Current project completeness
JSON / PDF
Portable project data and report output
How MSA and Gauge R&R Relate
Gauge R&R is one study within the wider MSA discipline.
Measurement System Analysis (MSA) evaluates whether measurement data is sufficiently accurate, consistent and discriminating for its intended decision. Gauge R&R focuses on the portion of variation caused by repeatability of the equipment and reproducibility between appraisers or operating conditions.
Reproducibility
Variation caused by different appraisers, fixtures, stations or other defined measurement conditions.
Example: Three inspectors measure the same set of shafts using the approved method.
Part-to-part variation
The real variation between the selected parts. A useful study needs parts covering the expected process range.
Example: Ten shafts are selected across low, middle and high process output.
Variation Model
| Component | Meaning | Desired condition | Typical action when poor |
|---|---|---|---|
| Total observed variation | Combined part variation and measurement-system variation. | Dominated by real part-to-part differences. | Separate measurement error before interpreting process performance. |
| Equipment variation (EV) | Repeatability of the device, method, fixture and short-term conditions. | Small relative to process or tolerance. | Improve resolution, fixturing, method, environment, maintenance or equipment. |
| Appraiser variation (AV) | Reproducibility between appraisers or defined operating conditions. | Operators reach equivalent results. | Clarify method, datum, force, alignment, training and interpretation. |
| Gauge R&R | Combined EV and AV. | Appropriate for risk and intended use. | Reduce measurement-system sources before accepting data. |
| Part variation (PV) | Actual differences among study parts. | Clearly distinguishable from measurement noise. | Select a more representative study range if PV is too narrow. |
Important: A low percentage of study variation can still be misleading if the study parts do not represent the process range, the gauge resolution is poor, the measurement method is changed during the study, or measurements are not randomised and blinded. MSA is a controlled experiment, not simply repeated inspection.
MSA Study Types
Common Interpretation Guidance
| Indicator | Common interpretation | Important caution |
|---|---|---|
| < 10% Gauge R&R | Generally considered acceptable. | Still review ndc, bias, resolution, risk and study validity. |
| 10% to 30% | May be acceptable depending on application, cost, risk and improvement feasibility. | Document the technical justification and intended use. |
| > 30% | Generally unacceptable. | Investigate EV and AV separately before choosing corrective action. |
| ndc ≥ 5 | Often used as a minimum indication that the system can distinguish useful categories. | Higher discrimination may be needed for SPC, tight tolerance or key characteristics. |
| Resolution | A practical rule is that display resolution should be small relative to process or tolerance width. | Display resolution alone does not prove capability. |
Practical Examples
Machined shaft diameter
Study: Variable crossed Gauge R&R using 10 parts, 3 appraisers and 3 trials.
Finding: High EV but low AV.
Likely causes: poor fixture, inconsistent measuring force, worn contacts or insufficient gauge resolution.
Coating thickness
Study: Bias, stability and Gauge R&R.
Finding: Acceptable R&R but drift over several weeks.
Likely causes: probe wear, reference foil deterioration, substrate effects or calibration-control weakness.
Visual paint-defect inspection
Study: Attribute agreement analysis using known reference samples.
Finding: Inspectors agree with themselves but disagree with the reference standard.
Likely causes: unclear defect criteria, poor lighting, inconsistent viewing distance or inadequate boundary samples.
Heat-treatment temperature system
Study: Bias, stability and system-level comparison.
Finding: Recorder and independent reference differ across the operating range.
Likely causes: sensor location, extension-wire error, channel correction, calibration bias or thermal gradients.
Torque verification
Study: Crossed Gauge R&R with representative fastener joints.
Finding: Strong appraiser-by-part interaction.
Likely causes: different reaction technique, alignment, seating method or interpretation of breakaway torque.
Digital record validation
Study: Attribute agreement and software verification.
Finding: Data-entry rules are repeatable but the master-data reference is wrong.
Likely causes: incorrect specification logic, obsolete revision or uncontrolled lookup tables.
Crossed Variable Gauge R&R Calculator
Average-and-range estimate for 10 parts, 3 appraisers and 2 or 3 trials.
Used to calculate percent of tolerance as 6σGRR ÷ tolerance × 100.
Enter representative repeated measurements and select Calculate Gauge R&R.
MSA Project Builder
Plan phase0% complete
MSA Project Report
Generated from the project builder and Gauge R&R calculator.
Complete the project builder and select Generate Report.