Calispec MSA Engine v2.0 · Screens derived from process-flow diagrams

Data Inputs & Method Selector

This prototype converts the GRR method selector flow and the Bias/Linearity acceptance flow into operator-facing screens. It captures study context, validates data sufficiency, recommends the correct method, and prepares an audit-ready payload.

4Study families: Variable GRR, Attribute GRR, Stability, Bias/Linearity

9Selector checks: eligibility, design, sample, policy, CI, final decision

3Execution states: AUTO_RUN, REVIEW_REQUIRED, BLOCKED

1Normalized payload for compute APIs and audit trail

Study header

Audit identity

Study ID

Tenant / Plant

Control plan ref.

Gauge ID

Gauge type

Part / CTQ

Study objective

Study family selector

Routes to process flow

Variable GRRContinuous measurements; crossed, nested, unbalanced GLM, REML and range fallback. Attribute GRRNominal or ordinal classification; agreement, Cohen/Conger, Fleiss/Light, weighted kappa. StabilityX-MR chart for master-reference observations with Western Electric and Nelson rules. Bias & LinearityAccuracy at a reference point and across the operating range. Dedicated AIAG/IATF screen.

Core data inputs

Selector-critical fields

Design

Execution mode

Rigor

Parts

Operators / appraisers

Repeats

Missing cells %

Interaction suspected?

Scale type

Rater panel

Stability observations

Gauge resolution

Tolerance

Validation gate

Policy chain preview

Selector flow alignment

Variable studies check destructive/nested → missing cells → sample size → interaction. Attribute studies check scale → CI requirement or rater panel. Stability checks observation count. These gates mirror the uploaded method selector tree.

Measurement data entry scaffold

Example rows, import-ready

Part	Operator	Repeat 1	Repeat 2	Repeat 3	Reference / category	Notes

Recommendation

AUTO_RUN

99SCORE

Crossed ANOVA GRR

Primary method for balanced crossed design

Candidate methods ranked

Composite score = design fit + rigor + sample + policy

Process route

Screen-to-flow map

Audit actions

What the user must do next

Bias study inputs

Accuracy at one reference point

Reference available?

Reference value

Readings count

Mean reading X̄

Std. dev. s

Tolerance

Linearity study inputs

Accuracy across range

Reference levels

Total readings

Slope b

Process variation / range

Tolerance

Slope p-value

Bias/Linearity decision route

Acceptance flow

Required evidence

IATF-ready attachments

Evidence	Mandatory	Purpose
Calibration certificate	Yes	Traceable reference and gauge calibration proof.
Raw reading sheet	Yes	Individual readings for bias and all reference levels.
Regression chart	Yes	Bias vs reference line, slope, R² and p-value.
QE approval	Conditional	Required for MARGINAL, REVIEW_REQUIRED, or range restriction.

Method selector process flow

Original uploaded diagram

Screen linkage

The Study Setup and Data Inputs screens feed the same gates shown here: destructive/nested, missing cells, sample size, interaction, attribute scale, rater panel and stability observations.

Bias & Linearity process flow

Original uploaded diagram

Screen linkage

The Bias & Linearity screen feeds the same gates shown here: reference availability, minimum readings, bias %, p-value, reference levels, regression slope and linearity %.

Normalized result payload

API handoff

Screen coverage checklist

Built from process flows

Flow node	Screen field / section	Status