Gauge Capability: Analyzing the Performance of Measuring Instruments

A gauge capability study evaluates a single measuring instrument measuring a single reference part under stable environmental and operational conditions. The study isolates the instrument’s inherent measurement spread from operator-to-operator and part-to-part variation. This isolation distinguishes gauge capability from the broader measurement system analysis (MSA), which includes all sources of variation within a measurement system.

Gauge capability applies to all measuring instruments that produce quantitative measurement results, including tactile probes, laser triangulation sensors, confocal sensors, and infrared cameras measuring surface temperature distributions. The study requires a minimum of 25 repeated measurements on a single reference part, performed under identical conditions by a single operator.

Gauge capability and process capability measure different properties of a production system. Process capability indices Cp and Cpk quantify how well a manufacturing process keeps part dimensions within tolerance. Gauge capability indices Cg and Cgk quantify how well a measuring instrument captures those dimensions. A process capability study has no validity if the measuring instrument used to collect the data has insufficient gauge capability.

3 international standards and industry documents define gauge capability methodology:

• VDA Volume 5 (Measuring Process Suitability, 2nd edition) — the German automotive industry framework
• ISO 22514-7 (Statistical methods in process management — Capability and performance — Part 7: Capability of measurement processes) — the ISO framework
• AIAG MSA Manual (Measurement System Analysis, 4th edition) — the North American automotive supply chain methodology

Cg is the ratio of a defined fraction of the tolerance band to 6 times the standard deviation of the repeated measurements. The factor 0.2 defines the study range as 20% of the total tolerance band — the threshold fraction the instrument spread must fit within. $T$ is the bilateral tolerance of the characteristic; $\sigma$ is the standard deviation of the repeated gauge measurements.

Cg=0,2⋅T6⋅σCg​=6⋅σ0,2⋅T​

Cgk extends the Cg calculation by incorporating the systematic offset between the measured mean and the reference value. xˉxˉ is the mean of the repeated measurements; xrefxref​ is the nominal reference value of the reference part. Cgk decreases when the instrument measures systematically above or below the reference value.

Cgk=0,1⋅T−∣xˉ−xref∣3⋅σCgk​=3⋅σ0,1⋅T−∣xˉ−xref​∣​

The threshold value for both Cg and Cgk is 1.33 according to VDA Volume 5. An instrument with Cg ≥ 1.33 and Cgk ≥ 1.33 is classified as gauge-capable and released for production measurement. 3 result conditions determine the required action:

%GR&R (Gauge Repeatability and Reproducibility, expressed as a percentage of total study variation) is a complementary metric calculated within a full Measurement System Analysis (MSA) study. Unlike Cg and Cgk, %GR&R captures both within-operator repeatability and between-operator reproducibility. A %GR&R below 10% is classified as acceptable. The full methodology for %GR&R calculation is detailed in the Measurement System Analysis (MSA) article.

The reference part used in a gauge capability study is a physical part or artifact whose nominal value for the measured characteristic is known with traceability to national measurement standards. 4 requirements define a valid reference part:

1. The nominal value lies within the middle third of the tolerance band of the characteristic.
2. The part is stable over time and does not deform, oxidize, or change thermally during the study.
3. The surface properties of the reference part match the surface properties of the production parts — critical for optical measurement systems.
4. The part is documented with a calibration certificate that provides metrological traceability.

Environmental conditions during the study must match the conditions under which the instrument operates in production. Temperature, vibration, and humidity levels are recorded and documented as part of the study protocol.

A Type-1 gauge capability study consists of 25 or 50 repeated measurements performed by a single operator on a single reference part without repositioning the part between measurements. The operator measures the reference part, returns the instrument to a defined starting position, and repeats the measurement cycle. The study isolates the instrument’s inherent measurement repeatability from repositioning and operator effects.

For optical 3D sensors and infrared cameras from AT Sensors, the measurement cycle includes sensor triggering, data acquisition, and extraction of the measurement result from the point cloud or thermal image. All 3 steps must be identical for each of the 25 measurement cycles to ensure valid repeatability data.

The evaluation calculates the standard deviation $\sigma$ from the 25 repeated measurements, then computes Cg and Cgk. The documentation records 7 data elements:

The standard deviation $\sigma$ used to calculate Cg corresponds to the repeatability component of the expanded measurement uncertainty. An instrument with a large measurement uncertainty relative to the tolerance will produce a low Cg value and fail the capability threshold. Reducing measurement uncertainty — through improved sensor hardware, thermal stabilization, or vibration isolation — directly increases the Cg value.

ISO/IEC Guide 98-3 (GUM — Guide to the Expression of Uncertainty in Measurement) and ISO 22514-7 define the formal relationship between gauge capability indices and expanded measurement uncertainty for production measurement processes. VDA Volume 5 provides the automotive-sector adaptation of this relationship.

A measuring instrument’s resolution limits the minimum detectable change in the measurand. An instrument whose resolution is coarser than $\sigma /4$ of the study cannot produce a valid gauge capability result, because the discretization of the measurement signal introduces artificial repeatability errors. For 3D sensors, the point cloud resolution and the subpixel interpolation algorithm of the sensor determine the effective resolution available for the capability study.

Linearity errors introduce a systematic deviation that varies across the measurement range. A sensor with poor linearity produces different Cgk values at different positions within its measurement range. Sensor resolution and linearity are each evaluated in dedicated articles within the Metrology node.

Metrological traceability connects the measurement result of a gauge capability study to national or international measurement standards through an unbroken chain of calibrations, each contributing a documented measurement uncertainty. VDA Volume 5 requires the reference part used in the Type-1 study to be calibrated by an accredited laboratory in accordance with ISO/IEC 17025. Traceability ensures that Cg and Cgk values obtained at one measurement station are comparable with values obtained at a different station, a supplier, or a customer facility.

Automated 100% inspection measures every produced part in the production line. In these systems, the measuring instrument performs thousands of measurement cycles per shift without operator intervention. Gauge capability in 100% inspection systems must account for 3 additional influencing factors:

• Thermal drift of the sensor during extended operation
• Vibration transmitted from the production environment to the sensor mounting
• The effect of part-to-part surface variation on optical measurement systems

AT Sensors 3D sensors use laser triangulation and structured light principles to measure geometric characteristics such as height, step height, flatness, and profile dimensions. Gauge capability studies for these sensors use a calibrated step artifact with a certified step height within the measurement range as the reference part. The Cg and Cgk values are calculated for the step height measurement and must meet the 1.33 threshold before the sensor is released for the geometric characteristic it measures in production.

AT Sensors infrared cameras measure thermal characteristics such as surface temperature, temperature distribution, and thermal gradients. Gauge capability studies for infrared cameras use blackbody radiators as reference parts, whose emitted radiance temperature is calibrated with traceability to national temperature standards. The 1.33 threshold applies equally to temperature measurement capability as to geometric measurement capability.

Quality management systems compliant with IATF 16949 and ISO 9001 require documented evidence of measuring instrument qualification. The gauge capability study report — including raw data, indices, and the release decision — forms part of the measurement system documentation required for customer approval in the automotive supply chain (PPAP — Production Part Approval Process). The capability study document is retained as a quality record for the lifetime of the production program.