Gauge R&R explained: how to run an MSA study.
Before you trust a single measurement, you have to trust the measuring. Gauge R&R — the heart of Measurement Systems Analysis — tells you how much of your observed variation is real and how much is the gauge fooling you.
What Gauge R&R is
Gauge Repeatability and Reproducibility (Gauge R&R) quantifies the variation that comes from the measurement system itself — the gauge and the people using it — rather than from the parts. It is the most common MSA study and a required element of PPAP. If your gauge variation is large, your Cp/Cpk numbers, your pass/fail calls, and your FAI results are all built on sand.
Repeatability vs reproducibility
- Repeatability (equipment variation): the same operator measuring the same part with the same gauge gets different readings. This is the gauge's own scatter.
- Reproducibility (appraiser variation): different operators measuring the same part get different readings. This is the human/method scatter.
Gauge R&R combines both into one figure, expressed as a percentage of either the tolerance or the total process variation.
How to run a study
The standard crossed study is 10 parts × 3 operators × 2–3 trials:
- Pick 10 parts that span the normal range of variation — not 10 identical good ones.
- Have 3 operators each measure all 10 parts, in random order, without seeing prior results.
- Repeat for 2 or 3 trials per operator.
- Feed the readings into a Gauge R&R calculation to get %GRR and the number of distinct categories (ndc).
The free Gauge R&R calculator does the arithmetic and returns %GRR and ndc from your pasted data.
Acceptance criteria
| %GRR | Verdict | Action |
|---|---|---|
| < 10% | Acceptable | Use the gauge with confidence |
| 10% – 30% | Marginal | May be acceptable depending on cost/criticality |
| > 30% | Unacceptable | Fix before trusting any data |
Also check ndc ≥ 5 — the measurement system should resolve at least five distinct categories across the part range. Below 5, the gauge cannot reliably tell parts apart.
Fixing a failing gauge
Fix repeatability with a better-resolution or better-maintained gauge and proper fixturing; fix reproducibility with a clear, written method and operator training so everyone measures the same feature the same way.
FAQ
What is an acceptable Gauge R&R percentage?
Below 10% is acceptable, 10–30% is marginal (acceptable depending on cost and criticality), and above 30% is unacceptable.
What is the difference between repeatability and reproducibility?
Repeatability is variation from one operator and gauge repeating a measurement; reproducibility is variation between different operators measuring the same parts.
How many parts and operators do I need?
The standard study is 10 parts, 3 operators, and 2 to 3 trials each, with parts spanning the normal range of variation.
What is ndc?
Number of distinct categories — how many groups the measurement system can reliably distinguish. Aim for 5 or more.
Is Gauge R&R required for PPAP?
Yes. MSA, including Gauge R&R, is one of the PPAP elements and underpins the dimensional results you submit.