GIFSQ Tools · Module 15 — Metrology, Validation & Laboratory Accreditation · Tool 296 of 500
Judge a measurement system by AIAG MSA: repeatability from the average range, reproducibility and part variation combined into %GRR and the number of distinct categories.
Try sample values — fills the form with the worked example below.
–
Repeatability SD
–
%GRR
–
Distinct categories (ndc)
–
Total SD
Equation: Repeatability SD = R̄ ÷ d₂*. GRR SD = √(repeatability² + reproducibility²). Total SD = √(GRR² + part²). %GRR = GRR ÷ total × 100. ndc = floor(1.41 × part SD ÷ GRR SD).
Method: Measurement system analysis per the AIAG Measurement Systems Analysis Reference Manual (MSA, 4th edition): the average-and-range method. %GRR under 10% is acceptable, 10–30% may be acceptable, over 30% needs improvement; ndc of 5 or more is required.
Assumptions: the repeats are taken under repeatability conditions; d₂* matches the number of repeats (1.128 for 2, 1.693 for 3, 2.059 for 4); part variation is estimated from a representative set of parts.
Limitations: the average-and-range method is an approximation — the ANOVA method separates operator-by-part interaction, which this tool does not; %GRR depends on the parts chosen, so use real production parts.
Rounding: repeatability and total SD to 4 decimal places; %GRR to 2 decimal places; ndc is a whole number; intermediates unrounded.
This is a calculation aid, not a laboratory report — confirm against your accredited method before citing a value.
Method: Measurement system analysis per the AIAG Measurement Systems Analysis Reference Manual (MSA, 4th edition): the average-and-range method. %GRR under 10% is acceptable, 10–30% may be acceptable, over 30% needs improvement; ndc of 5 or more is required.
Assumptions: the repeats are taken under repeatability conditions; d₂* matches the number of repeats (1.128 for 2, 1.693 for 3, 2.059 for 4); part variation is estimated from a representative set of parts.
Limitations: the average-and-range method is an approximation — the ANOVA method separates operator-by-part interaction, which this tool does not; %GRR depends on the parts chosen, so use real production parts.
Rounding: repeatability and total SD to 4 decimal places; %GRR to 2 decimal places; ndc is a whole number; intermediates unrounded.
This is a calculation aid, not a laboratory report — confirm against your accredited method before citing a value.
Worked example: R̄ = 0.05 mm, d₂* = 1.128, reproducibility SD 0.02 mm, part SD 0.15 mm: repeatability SD = 0.0443 mm, %GRR = 30.84%, ndc = 4, total SD = 0.1577 mm — needs improvement.
Enter values above to see results.
Accept: %GRR under 10% and ndc 5 or more. Conditional: 10–30%. Over 30% or ndc under 5: improve the gauge or the method. Tool 296 of 500 — GIFSQ
