GaugeConnection
¿Es usted consultor de calidad? Perfil público, referidos de clientes y acceso gratuito a todas las herramientas.
Más información

Qué is MSA?

Tabla de contenidos

MSA is a discipline, not a single calculation

Análisis del sistema de medición (MSA) is the practice of quantifying how much of the variación you observe in your datos comes from the medición proceso itself rather than from the piezas you are medir. It is not one estudio or one number - it is a set of preguntas you ask about a sistema de medición, each respuestaed by a different estudio.

Everything in MSA rests on one identity. The variación you observe is the sum of the true pieza-to-pieza variación and the variación your sistema de medición adds.

σ2observed = σ2part + σ2measurement
Variances add. Standard deviations do not.

The consequence is direct: if medición variación is large relative to pieza variación, you will reject good piezas and accept bad ones, and every proceso improvement decision you make from that datos will be built on noise. You cannot improve a proceso you cannot measure.

The five propiedades of a sistema de medición

MSA evaluates five distinct propiedades. Each one respuestas a different pregunta, and each is detected by a different estudio. A R&R del sistema de medición estudio addresses only the last two.

Property Qué it means Study that detects it
Bias A systematic offset between your mediciones and a known referencia value. The gauge reads consistently high or consistently low. Bias estudio against a calibrado patrón
Linearity Whether the sesgo changes across the medición rango. A gauge may be accurate at mid-rango and badly off at the extremes. Linearity estudio across multiple referencia points
Stability Whether the sistema de medición drifts over time. Today's readings should agree with last month's. Control chart on a master pieza over time
Repeatability Scatter when the same operador measures the same pieza repeatedly. Also called equipment variación (EV). R&R del sistema de medición
Reproducibility Differences between operadores medir the same piezas. Also called evaluador variación (AV). R&R del sistema de medición

This is the most common and most costly misunderstanding in MSA: running a R&R del sistema de medición and calling it "our MSA". A R&R del sistema de medición says nothing about sesgo, linealidad, or estabilidad.

Accuracy versus precisión

The five propiedades split cleanly into two groups. Bias, linealidad, and estabilidad are about exactitud - where your mediciones sit relative to truth. Repeatability and reproducibilidad are about precisión - how tightly your mediciones cluster. These are independent, which produces four possible situations.

Accurate and precise

Readings cluster tightly around the valor verdadero. This is the goal.

Accurate but imprecise

Readings scatter widely but average out near truth. A R&R del sistema de medición will catch this.

Precise but inaccurate

Readings cluster tightly but in the wrong place. Precisely wrong. A R&R del sistema de medición will pass this gauge.

Neither accurate nor precise

Readings scatter widely and sit off-target. The sistema de medición is unusable.

The key insight
A gauge can pass a R&R del sistema de medición with excellent numbers and still be badly sesgoed. R and R only sees the precisión axis. If you have never run a sesgo check against a calibrado patrón, you do not yet know whether your mediciones are correct - only that they are consistent.

Cuál GaugeConnection tool covers which propiedad

Here is an honest map of what GaugeConnection currently addresses, and where you will need another método.

Property Herramienta
Bias No dedicated tool Assess today from as-found / as-left readings on your calibración certificates.
Linearity No dedicated tool Assess today by comparing calibración resultadoados at several points across the rango.
Stability Stability Tracker
Repeatability R&R del sistema de medición
Reproducibility R&R del sistema de medición
Resolution / discrimination Gauge Capability

Dedicated sesgo, linealidad, and atributo concordancia estudios are not yet available in GaugeConnection. Dónde a tool does not exist, the note above describes how to assess that propiedad with datos you already have.

Variable datos versus atributo datos

Everything described so far assumes variable datos - continuous mediciones such as a diameter in millimetres or a torque in newton-metres. Variance components only make sense when the medición is a number on a continuous scale.

Attribute datos is different. Go / no-go gauges, visual inspection, and pass / fail decisions produce categories, not mediciones, and require a completely different método called atributo concordancia análisis. It measures how often evaluadores agree with themselves, with each other, and with a known patrón. If most of your inspection is visual or go / no-go, a R&R del sistema de medición is the wrong estudio.

Cuándo to run an MSA

MSA is triggered by events, not by the calendar alone. Run or refresh the relevant estudio when:

  • You introduce a new gauge or medición método.
  • A new operador begins taking mediciones.
  • You start medir a new pieza family or a materially different tolerancia.
  • A gauge has been repaired, adjusted, or failed a calibración.
  • You are preparing a PPAP submission or an IATF 16949 audit.
  • Your annual review cycle comes due, or proceso datos starts behaving in ways you cannot explain.

Dónde MSA is defined

  • AIAG MSA, 4th edition -- The referencia manual most automotive customers expect you to follow. Defines the estudio métodos and the acceptance criteria.
  • IATF 16949, clause 7.1.5.1.1 -- Requires statistical estudios to analyse the variación in the resultadoados of each type of inspection, medición, and test equipment identified in the control plan.
  • ISO 22514-7 -- The ISO treatment of medición proceso capability, used more widely outside the automotive sector.

Rejoining the server...

Rejoin failed... trying again in seconds.

Failed to rejoin.
Please retry or reload the page.

The session has been paused by the server.

Failed to resume the session.
Please retry or reload the page.