msa-gauge-rr

Measurement System Analysis (MSA) / Gauge R&R

When to use

Use this skill when:

• Qualifying a measurement system for a new part or process (PPAP requirement)
• Interpreting Gauge R&R results — is this gauge acceptable?
• Auditing a supplier's MSA study for correctness and adequacy
• Selecting the right study type for a given measurement situation
• Investigating a quality problem where measurement error may be a factor
• Responding to a customer request for MSA data on a specific characteristic

Prerequisites

• The characteristic to be measured (product or process)
• The gauge or measurement system to be studied
• Production parts spanning the expected process variation (10 parts minimum)
• At least 2 trained appraisers who normally perform the measurement
• The specification (tolerance) for the characteristic

Workflow

Step 1 — Select the MSA study type

| Study type | When to use |

|-----------|-------------|

| Gauge R&R (crossed) | Variable data, 2–3 appraisers, each measures all parts (most common) |

| Gauge R&R (nested) | Variable data, parts are destroyed during measurement (e.g., tensile test) |

| Attribute MSA | Pass/fail, go/no-go, visual inspection — data is not a number |

| Bias study | Accuracy of a single gauge vs. a reference standard |

| Linearity study | Whether gauge accuracy is consistent across its measurement range |

| Stability study | Whether gauge accuracy drifts over time |

For PPAP: Crossed Gauge R&R is required for all variable gauges on special characteristics.

---

Step 2 — Crossed Gauge R&R study procedure

Setup:

• Minimum 10 parts selected to represent the full process variation range (not cherry-picked from centre of tolerance)
• Minimum 2 appraisers (3 preferred)
• Minimum 2 trials per appraiser per part (3 preferred)
• Parts randomly numbered and labelled — appraisers must not see each other's measurements
• Measurement conditions must match normal production conditions

Standard study design:

• 10 parts × 3 appraisers × 2 trials = 60 measurements
• 10 parts × 2 appraisers × 3 trials = 60 measurements

Execution:

1. Appraiser A measures all 10 parts in random order — record results
2. Appraiser B measures all 10 parts in random order — record results
3. Appraiser C measures all 10 parts in random order — record results
4. Repeat the cycle for trial 2 (and trial 3 if applicable)
5. Appraisers must not see their own previous results or other appraisers' results during the study

Do NOT:

• Allow appraisers to adjust the gauge between trials
• Use parts that are all near the nominal value (no spread)
• Record only one trial (single measurement per appraiser per part is insufficient)

---

Step 3 — Interpret the results

%GRR (Gauge R&R as % of Total Variation or % of Tolerance)

| %GRR | Interpretation | Decision |

|------|---------------|---------|

| < 10% | Excellent | ✅ Gauge accepted |

| 10% – 30% | Marginal | ⚠️ May be acceptable based on application — requires engineering review and customer approval |

| > 30% | Unacceptable | ❌ Gauge not suitable — investigate and improve before use in production |

Two calculation methods:

• % of Study Variation (% of TV): GRR / Total Variation × 100 — preferred when process is in control
• % of Tolerance: GRR / Tolerance × 100 — use when process capability is unknown or for attribute gauging

For PPAP, %GRR < 30% (tolerance method) is the typical customer acceptance criterion. <10% is the target.

Number of Distinct Categories (ndc)

ndc = 1.41 × (Part Variation / GRR)

| ndc | Interpretation |

|-----|---------------|

| ≥ 5 | ✅ Gauge can distinguish adequate number of categories |

| 3 – 4 | ⚠️ Gauge can be used for go/no-go decisions only |

| 1 – 2 | ❌ Gauge cannot distinguish parts — unacceptable |

ndc ≥ 5 is required for measurement systems used on special characteristics.

Repeatability vs. Reproducibility

| Component | Description | Common cause |

|-----------|-------------|-------------|

| EV (Equipment Variation / Repeatability) | Variation when same appraiser measures same part multiple times | Gauge imprecision, worn parts, environment |

| AV (Appraiser Variation / Reproducibility) | Variation between different appraisers measuring the same part | Training inconsistency, measurement technique, gauge setup |

If EV > AV: investigate gauge (calibration, maintenance, resolution)

If AV > EV: investigate training, measurement procedure, gauge fixture/setup

When %GRR > 30% — improving the measurement system

A result >30% means the gauge is not suitable for production use. Do not proceed to PPAP — investigate and retest. Common root causes and actions:

| Root cause (high EV) | Action |

|---------------------|--------|

| Gauge resolution too coarse | Replace with a gauge of finer resolution (rule: resolution ≤ 10% of tolerance) |

| Gauge worn or damaged | Inspect, recalibrate, or replace the gauge |

| Environmental interference (vibration, temperature) | Move measurement to a stable environment; add fixture if needed |

| Inconsistent part fixturing | Design a repeatable fixture or measurement aid |

| Root cause (high AV) | Action |

|---------------------|--------|

| Measurement technique varies by appraiser | Develop a standardised measurement instruction (WI with photos/video) |

| Gauge difficult to read or position | Redesign fixture; add a datum locator; use a self-positioning gauge |

| Training gap | Retrain all appraisers using the standardised measurement WI; repeat the study |

After implementing improvements: re-run the full study. Do not accept a %GRR > 30% result with a customer waiver unless the characteristic is non-critical and the customer explicitly agrees in writing.

---

Step 4 — Attribute MSA (pass/fail gauges)

For go/no-go gauges, visual inspection, and any pass/fail decision:

Expanded attribute study (recommended for PPAP):

• 50 parts spanning the full range (include parts near the accept/reject boundary)
• 3 appraisers × 3 trials per part = 450 measurements
• Calculate % agreement (within appraiser and between appraisers)
• Calculate kappa statistic (Cohen's kappa ≥ 0.9 is acceptable)

Short method (minimum acceptable):

• 20 parts × 2 appraisers × 2 trials
• 90% agreement within and between appraisers required
• Include borderline parts near the specification limit

---

Step 5 — Audit an existing MSA study

When reviewing a supplier's or internal MSA, check:

• [ ] Parts selected represent the full process variation range (not all near nominal)
• [ ] Minimum 10 parts, 2 appraisers, 2 trials confirmed
• [ ] Appraisers did not see each other's or their own previous results
• [ ] %GRR result stated clearly (tolerance method or study variation method — which one?)
• [ ] ndc ≥ 5 for all special characteristics
• [ ] EV vs. AV breakdown analysed and interpreted
• [ ] %GRR < 30% (conditional: <10% preferred)
• [ ] For attribute gauges: kappa statistic calculated
• [ ] Study performed in production conditions (not in metrology lab if production measurement is on the line)
• [ ] Gauge calibration certificate current at time of study

---

Validation criteria

An MSA study is acceptable for PPAP when:

• %GRR < 30% (tolerance method) — <10% preferred
• ndc ≥ 5
• Study conducted with production appraisers using production gauges in production conditions
• Parts selected from the full process variation range
• Results documented with part numbers, appraiser IDs, gauge ID, and calibration reference

Common mistakes

• Parts selected near the nominal value only — artificially inflates ndc and understates GRR
• Appraiser A re-measures all parts immediately — not the same as separate trials (must separate in time)
• Using metrology lab gauges for an MSA of a production line gauge — study must use the same gauge in the same conditions
• %GRR > 30% submitted in PPAP without explanation or customer waiver
• Gauge calibration expired at time of MSA study — study is invalid
• ndc < 5 on a special characteristic — automatic PPAP finding in IATF audits
• Attribute MSA with only 2 appraisers and no borderline parts — study has no discriminating power

Output Format

At the start of each use, ask the user:

> "How would you like to receive the output?

> A — Structured Markdown (formatted tables and sections, ready to copy)

> B — Plain tables (simplified structure for Excel or Word)

> C — Narrative report (flowing text for a formal document or email)

>

> Default: A."

Adapt all output sections to the chosen format. If the platform or session context already defines a format preference, skip this question.

Changelog

| Version | Date | Author | Change |

|---------|------|--------|--------|

| 1.0 | 2026-06-06 | @RBraga01 | Initial release |

| 1.1 | 2026-06-06 | @migmcc | Added improvement guidance for %GRR > 30% — root cause table for high EV and high AV with corrective actions |