InsightFab
Knowledge Base/Gage R&R: How Truthful Are Your Measurement Tools?
MSA8 min read

Gage R&R: How Truthful Are Your Measurement Tools?

A new engineer measures a part multiple times with slightly different results, while a colleague also gets a different reading. With a specification of ±0.05mm, this might seem acceptable, but is your calculated Cpk of 0.95 truly a process capability issue or primarily due to measurement system error?

Situation

A new engineer measured a part with calipers: 10.023mm. Five minutes later, he measured it again: 10.031mm. Suspecting he made a mistake, he asked a quality control colleague to measure it: 10.018mm.

The three readings differ by a maximum of 0.013mm. Your specification is ±0.05mm, which might seem acceptable.

However, if your Cpk calculation shows a process capability of only 0.95... is it truly a process problem, or does the measurement tool's inherent error account for a large portion of it?

Plain Language Explanation

Gage R&R breaks down the "total observed variation" into three parts:

  • True Part Variation (Part-to-Part): You want this to be large.
  • Repeatability: Same person measures the same part, different results each time → Instrument issue.
  • Reproducibility: Different people measure the same part, different results → Operator issue.

%GRR = (Measurement System Standard Deviation ÷ Total Standard Deviation) × 100%

If %GRR = 38%, it means that 38% of the "observed variation" is noise introduced by the measurement tool, not actual part variation.

Practical Assessment

Standard Design for Conducting Gage R&R:

  • 10 parts (should represent the process variation range, not solely good parts within spec)
  • 2–3 operators
  • Each person measures each part 2–3 times
  • Randomization: Do not let operators know previous measurement results

Main CauseDiagnosisCountermeasure
High RepeatabilityInstrument instabilityReplace with higher precision instrument, calibrate, anti-vibration
High ReproducibilityInconsistent operator techniqueStandardize SOP, fix measurement position, use fixtures

How InsightFab Does It

To run Gage R&R in JMP and Minitab, you first need to design tables, input the part/operator/replicate structure, and then interpret the ANOVA output. InsightFab allows you to upload a standard CSV format and directly provides traffic light diagnostics and improvement recommendations.

Golden Quote

"Performing Gage R&R before process improvement is like performing delicate surgery while wearing foggy glasses—you think you see clearly, but you don't."

Want to try it yourself?

Every tool mentioned in this article is available on InsightFab — just upload a CSV to analyze.

Go to Tools