Measurement System Analysis (PDF)

This is a practical guide for the essentials that you should know about measurement system analysis.

It can be used as a training material to learn the essentials about Measurement System Analysis (MSA) in accordance with the 4th edition of the AIAG manual.

Summary:
•  What is a measurement system and why should it be evaluated?
•  History of measurement system analysis
•  The main characteristics of a Measurement System
•  The different types of R&R studies
•  Crossed R&R study: Methods, process and interpretation of results
•  Attribute Agreement Analysis: Process and interpretation of results
•  Attribute Gage Study: Process and interpretation of results

Measurement is a main part of our decision making because it is how we get data that we need to monitor a process or confirm a product compliance.

If the variation caused by the measurement system is important, then the measurements will be false and will not represent the actual values of what we want to measure. This may cause a bad decision making.

If measurement is not precise or accurate, we could easily make bad decisions with all the consequences that this may have.

The goal of the measurement system analysis is to Evaluate the variation generated by the measurement system which must be small compared to the actual variation between parts and compared to the specifications of the Product.

As with any process, the measurement of a product characteristic or a process parameter can record a variation. This Observed variation in measured values is the combination of two variation:
• Actual variation between parts
• Variation caused by the measurement system
If the variation caused by the measurement system is important, then the measurements will be false and will not represent the actual values of what we want to measure. This may cause a bad decision making.

The main sources of variation of a measurement system are related to its components:
• Measuring instrument: Factors related to the Measuring instrument (status, resolution, ...)
• Methods: Factors relating to Measurement Methods (procedure, ...)
• Product: Factors relating to the Product of which we want to measure a Characteristic (state, definition of the Characteristic, ...)
• Personnel: Factors relating to the persons carrying out the measurement (training, …)
• Environment: Environmental factor (vibration, temperature; …)

The methods of analyzing a measurement system differ depending on whether we are dealing with continuous or attribute data.