Measure

MSA Type 1

Measurement System Analysis Type 1 — Repeatability (Cg) and Bias (Cgk)

Overview

The MSA Type 1 study checks a measurement system for bias (position) and repeatability (spread) — on one single reference part, with one operator, in a short time frame. It is the first step of any measurement-system analysis: if it fails, no deeper study (Type 2 / Gage R&R) is worthwhile.

Reference part: A part or master with a known, traceable reference value. Without this value, position cannot be judged — the bias part of the study would be blind.

Repeat measurements: Typically 25–50 measurements of the same reference with the same system under conditions as constant as possible. Only then does pure repeat variation become visible.

Bias: Systematic deviation: mean of repeat measurements − reference value. A bias ≠ 0 indicates a calibration offset or systematic shift.

Cg: Measure of repeatability relative to tolerance. Typical formula: Cg = 0.2·T / (6·s), where T is the tolerance width and s the standard deviation of repeats. Accepted when Cg ≥ 1.33.

Cgk: Extends Cg to include bias. Cgk ≥ 1.33 is accepted. High Cg, low Cgk → the system is precise but shifted (calibration issue).

Type 1 does not replace Gage R&R, it is the pre-test. Failing Type 1 means running a Gage R&R afterwards wastes resources.

Approach

Provide a reference part with a traceable value — ideally near mid-spec.
Use the system under production conditions, not lab conditions.
Take 25–50 repeat measurements on one day without re-setup.
Compute mean, standard deviation, and bias.
Compute Cg and Cgk and compare to the acceptance limit (usually 1.33).
Document the result; if it fails, investigate causes before running Gage R&R.

Pitfalls

No traceable reference value: A reference part without a known value disables bias assessment — the study then measures only repeatability. Without a reference, the result is incomplete.

Optimistic lab conditions: Testing a gage in air-conditioned calm produces nice Cgk — irreproducible in real production. The test must mirror actual use.

Too few measurements: With 10 repeats, s and bias are uncertain, Cg/Cgk unstable. 25–50 repeats are standard; below that it becomes unreliable.

Measurements "optimized": Freely dropping outliers produces nice numbers and poor reality. Record all measurements; remove outliers only with a documented cause.

Type 1 passed = system OK: Type 1 only checks bias and repeatability on one part by one operator. Variation from operators or part-to-part range shows up only in Type 2 / Gage R&R.

Acceptance threshold too lax: Cg/Cgk < 1.33 is not internationally accepted. Lowering the bar internally makes the study worthless the moment a customer audit appears.

Examples

This module ships with the following example datasets — load any of them in the app with a single click.

MSA Type 1: capable gage — Project template

25 repeated measurements of a 10.0 mm reference part (LSL 9.9 / USL 10.1, T = 0.2 mm). Very small spread, negligible bias → Cg ≈ 3.8 and Cgk ≈ 3.8 well above the 1.33 threshold. Textbook capable-gage case.

MSA Type 1: marginal gage — Project template

25 repeated measurements with moderate spread and slight positive bias. Cg ≈ 1.5 just above the 1.33 threshold, Cgk ≈ 1.0 just below — bias pulls capability down. Classic case for the role of Cgk alongside Cg.

MSA Type 1: incapable gage — Project template

25 repeated measurements with large spread and noticeable bias. Cg ≈ 0.6 and Cgk ≈ 0.3 — gage clearly fails the capability threshold. Textbook case for a measurement system that must be rejected.

Available in the following cycles

DMAIC: Measure
DMADV: Measure
8D: D2 — Describe the Problem