Evaluate measurements
Start with repeated observations, build a Type B uncertainty budget, and define specification limits. The result records each engine step instead of hiding the path behind a single number.
Workspace · local only
Measurement & Tolerance Workspace
Evaluate repeated measurements, build an uncertainty result, and make a conformity decision—or switch to a dimensional tolerance stack. Both modes run locally and reuse existing verified engines.
Design tolerance workflow
Tolerance stack
Compare RSS and worst-case limits for an independent one-dimensional dimension chain.
Stack model
Scenario A
Independent 1-D stack. All dimensions use the scenario unit; no implicit unit conversion.
| Dim | Nom | +Tol | −Tol | Dir |
|---|
Add dimension
Direction + adds to the stack; − subtracts. Statistical basis stays Magnitude (drawing limits).
Stack result
Scenario A
RSS is a magnitude estimate for independent contributors, not a guaranteed assembly limit.
Stack range
Worst-case envelope with the RSS magnitude centered on nominal.
Worst-case limits — RSS magnitude —
Magnitude basis · no probability coverage implied.
Analysis methods
Reuse the same model at the analysis depth the decision requires.
Statistics-backed workflow
Measurement → uncertainty → decision
Enter repeated observations and one or more Type B contributions. The workflow calculates Type A statistics, combines standard uncertainty by RSS, expands it by k, then applies a simple or guarded conformity rule.
Conformity decision
EXAMPLE DATA · NOT A DECISION RECORD
INCOMPLETE
Complete the inputs and run the workflow. No measurement data leaves this browser.
Measurements→Statistics→Uncertainty→Decision
Decision position
Guarded by USpecification limits Guarded acceptance zone Evaluated mean
Measurement model
Guarded mode uses the calculated expanded uncertainty U = k·uc as the guard width. It is a binary ILAC G8-style rule, not a JCGM 106 risk model.
Result details
Statistics and standard-uncertainty composition for the evaluated mean.
Uncertainty contribution
Share of combined variance, u²/uc².
Method & evidence
Workflow v1.2.0 · 3 versioned contracts
Inspect versioned Evidence: Type A · Type B · Decision rule
How this workspace works
One local scenario holds a measurement decision and a saved 1-D stack.
Edit the stack
Use the independent dimension-stack path when the task is a design tolerance problem. From Stack-up, Worst-case, or Monte Carlo, click Add to workspace. Nothing leaves this browser.
Keep the model boundary visible
The measurement path assumes independent Type A and Type B components. Correlated GUM propagation stays on Uncertainty propagate; distributions and yield stay on Monte Carlo.