WoluTools

Measurement system analysis · crossed ANOVA

Gage R&R Calculator: %GRR, NDC and the ANOVA table from your measurements

Type or paste parts × operators × trials. The result below updates as you type. The calculator runs in your browser: nothing is uploaded and no daily job is used.

  • 2–30 parts
  • 2–6 operators
  • 2–5 trials
  • Paste from Excel

Gage R&R calculator

1 · Your measurements

Loaded with a fictional example (readings in mm). Overwrite it or clear the grid.

Tolerance is optional: the spec width (upper minus lower limit), used only for %GRR of tolerance. Review and error lines apply above 10% and 30% by default, as in the file workflow.

PartOperator AOperator BOperator C
T1T2T1T2T1T2
1
2
3
4
5
6
7
8
9
10

Needs review: %GRR 16.44% · NDC 8 · 13.81% of tolerance See the tables

Enter or the arrow keys move between cells. Pasting a block copied from Excel into any cell fills the cells to the right and below.

Paste a whole table from Excel or a CSV

Grid layout: one row per part with operator A trials first, then operator B, and so on; a label column and a header row are skipped. List layout: one row per measurement, with or without a header row. Tabs, semicolons, commas or spaces work as separators.

2 · Result

Computed from the grid. No values leave this page.

Needs reviewGage R&R is 16.4% of the study variation, between the 10% review line and the 30% error line.
%GRR (study variation)
16.44%
Distinct categories (NDC)
8
%GRR of tolerance
13.81%

10 parts × 3 operators × 2 trials · 60 measurements · grand mean 2.5028

Variance components

SourceVariance%Contrib.6×SD%Study var%Tolerance
Total Gage R&R0.00013252.70%0.0690516.44%13.81%
Repeatability0.000022720.46%0.028606.81%5.72%
Reproducibility0.00010972.24%0.0628514.96%12.57%
Operator0.00010972.24%0.0628514.96%12.57%
Operator × Part00.00%00.00%0.00%
Part-to-part0.00476897.30%0.414398.64%82.86%
Total variation0.004900100.00%0.4200100.00%84.00%

Two-way ANOVA with interaction

SourceDFSSMSFp
Part90.25760.028622416< 0.001
Operator20.0044130.002207186.3< 0.001
Operator × Part180.00021320.000011850.52150.925
Repeatability300.00068150.00002272
Total590.2629

Method: two-way crossed ANOVA with the operator × part interaction kept in the model; negative variance estimates are set to zero; NDC = 1.41 × part SD ÷ gage SD, rounded down. Same formulas as the file report (engine gage-rr-v1).

For the quality record

Full report from a CSV or XLSX file

The file workflow runs the same analysis on the server and returns a PDF, an evidence XLSX with every intermediate value, plots, CSV, JSON and a hash manifest. It names missing cells, duplicate rows and unbalanced designs instead of guessing. Download as CSV above writes your grid in the layout the upload expects.

Free
3 jobs a day, files up to 10 MB, no account needed. Failed jobs are never counted.
Pro
Up to 200 jobs a day and larger files · €12.99/month or €89.99/year
Formats
CSV, XLSX · 1 file · 50 MB service limit · 100,000 rows · 256 columns
File handling
Draft 2 hours · sources deleted after processing · encrypted result 24 hours
You receive
PDF, evidence XLSX, plots, CSV, JSON and manifest

File workflow

From one file to evidence you can review

  1. Add the file.One CSV or XLSX. Operator, part, trial and measurement columns are detected from their headers, or you name them.
  2. Confirm the job.File name, settings, tolerance and thresholds stay visible before the job starts. Nothing is counted until you confirm.
  3. Process in isolation.A one-job networkless sandbox receives only this job and cannot access accounts, queues, encryption keys or other customer storage.
  4. Download the evidence.PDF, evidence XLSX, plots, CSV, JSON and manifest. Every output hash is covered by the manifest.

What this product does not claim

The report exposes method, assumptions, intermediate values and limits; it is not a calibration certificate.

Clear answers

Gage R&R calculator FAQ

How is %GRR calculated here?

With a crossed two-way ANOVA that keeps the operator × part interaction in the model. Negative variance estimates are set to zero. %GRR is the gage standard deviation divided by the total standard deviation. The file report uses the same formulas.

How is the number of distinct categories worked out?

NDC = 1.41 × part-to-part standard deviation ÷ gage standard deviation, rounded down to a whole number.

Are the values I type into the calculator uploaded?

No. The calculator runs in your browser and does not use one of the daily jobs. Only the file workflow sends data to the server.

How must the measurement file be laid out?

One CSV or XLSX file with a row per measurement carrying the part, the operator, the trial number and the measured value. Download as CSV in the calculator writes exactly this layout.

What happens if the study is unbalanced?

The design is reported exactly as it was supplied rather than quietly corrected. You can see that a trial or an operator is missing before you read the percentage.

Where does the pass or fail line come from?

From the thresholds you set: 10% for review and 30% for error by default, in the calculator and in the file workflow. They are your team policy, not a published standard. The report names the value that was applied.

Can the report serve as a calibration record?

No. It documents the method, assumptions and intermediate values of one study. It is not a calibration certificate and does not stand in for a metrology review.

Available now

Know whether the measurement system is hiding the real variation.

Check a study in the calculator now. For the report, your file runs here with 3 free jobs a day and no account, or up to 200 a day with Pro.