Introduction
MSA is an acronym for Measurement Systems Analysis. It is a method for estimating the variability in a measurement system using specific tools. In MSA analysis, a special study (such as a mini-experiment) is performed to estimate the variability.
This article is an introduction to Measurement Systems Analysis, discussing the concept of MSA, its history, major terms and the standards used in the industry.
I invite you to read it!
History of MSA
The history of MSA is younger than the SPC or FMEA history. Interestingly, I have encountered an opinion that the origins of MSA go back to ancient times because measurement units were already being used then, but to me, this is a rather stretched theory. Using such logic, I could claim that SPC or FMEA are ancient since the ancients used sampling for quality control (a'ka SPC) and certainly figured out how to build things to work reliably (a'ka FMEA). So let's drop the theories about the ancient origins of MSA.
And now, to the specifics:
- 1970+ In 1977, the BIPM organization initiated work on establishing the principles for determining measurement uncertainty. Due to the topic's international scope and technical challenges, this work continued throughout the 1980s.
- 1990+ The AIAG organization published the first MSA manual in 1990 for the American automotive industry. This document was updated in 1995 to meet the requirements of the QS-9000 management system. Simultaneously, the ISO organization released the "Guide to the Expression of Uncertainty in Measurement" (GUM) in 1995, which marked the culmination of the work that began in 1977.
- 2000+ The MSA method in the automotive industry continued to develop, with AIAG publishing the third edition of MSA in 2002. The VDA organization published the VDA 5 manual, "Measurement and Inspection Processes," which became the standard for the German automotive industry. ISO published "ISO/IEC Guide 98," which is consistent with the GUM.
- 2010+ Further development of the MSA method. Updates to AIAG MSA (2010) and VDA 5 (2010). The ISO organization released the ISO 22514-7 standard in 2012.
- 2020+ Update to ISO 22514-7 (2021) and VDA 5 manual (2021).
Measurement Systems Analysis to some extent, uses tools developed by Walter A. Shewhart in the 1920s. During that time, the SPC control charts were created and are now part of Statistical Process Control (SPC).
What is the MSA used for?
When taking any measurement, it is important to remember that the result obtained is never ideal, we can assume that it is always subject to some measurement error. This error is related to the variation caused by the measurement system.
The purpose of MSA analysis was interestingly described by G. Larsen (2003): "The purpose of measurement system analysis (MSA) is to separate the variation among devices being measured from the error in the measurement system.""[2]
In conclusion:
MSA analysis is able to separate the so-called
Measurement system in MSA
The measurement system consists of the various elements associated with the measurement being performed, which are also a source of variability. Typical factors are:
- Instrument. Caliper, multimeter, micrometer, manometer, gauge, etc.
- Operator. The person performing the measurement.
- Method. The adopted procedure followed when performing a measurement.
- Environment. Temperature, humidity, vibration, lighting, etc.
- Part. The parts measured (their measured characteristics).
- Standard. Known reference value, standard, visual evaluation criteria, etc.
- Assumptions. Physical constants, rules for rounding results, etc.
- Other. For example, software used for analysis, etc.
In the MSA context, we can describe the measurement system using two popular models: the S.W.I.P.E model and the P.I.S.M.O.E.A model.
If we make significant changes to the measurement system, e.g., changing the measuring instrument from a caliper to a micrometer, we should formally redo the MSA analysis.
Variation in MSA
Variation in the measurement system can be divided into that which affects the bias (accuracy) of the measurement results and that which affects the spread (precision) of the measurements.
The MSA method defines how to estimate:
- Bias, Linearity, and Stability of the measurement instrument.
- Repeatability. The impact of the measurement instrument on the measurement of a given characteristic.
- Reproducibility. The impact of operators on the measurement of the characteristic under study.
The study of the repeatability and reproducibility of an instrument is referred to as Gauge Repeatability and Reproducibility, abbreviated as GR&R or GRR. Sometimes, an even shorter acronym R&R, is used.
MSA study
An MSA study is a kind of "mini-experiment" to estimate the variation in a measurement system. Depending on the characteristic being measured, different types of studies are used.
MSA for Variable Data
Variable data refers to characteristics whose values can take a continuous range of measurements. Examples of such characteristics include length, angle, temperature, resistance, pressure, etc. The MSA methods applied are:
- Bias, Linearity, and Stability of the measurement instrument.
- Type I Study: Measurement Equipment Evaluation. Estimation of Cg and Cgk indices.
- Type II Study: GR&R Study with Operator Influence.
- Type III Study: GR&R Study without Operator Influence.
Type II and Type III studies use different calculation methods. Common formulas are the Average and Range Method (ARM) and the Analysis of Variance (ANOVA). It is important to note that the ANOVA calculation method is recommended and may be required by customers.
MSA for Attribute Data
Attribute data refers to characteristics with discrete, non-continuous values. Examples include the number of scratches, number of holes, number of defective products, product color, etc. Examples of MSA methods for attribute data include:
- Kappa Method (required by AIAG MSA).
- Simplified Method.
- Simplified Method with Reference.
- Signal Detection Method.
- Analytical Method.
MSA Acceptance Criteria
The results of an MSA study have specific limits that determine whether a measurement system is acceptable or not. If customers do not have other requirements, the typical approach is to follow the guidelines established in the selected manual or standard.
For example, according to the most commonly used AIAG MSA approach, the following limits are set for GRR[3]:
- GRR ≤ 10%: The measurement system is acceptable.
- 10% < GRR ≤ 30%: The measurement system is conditionally acceptable.
- GRR > 30%: The measurement system is unacceptable.
Additionally, AIAG MSA specifies a second index called NDC (Number of Distinct Categories). If the measurement system is to be used in SPC, the NDC should be 5 or greater[3].
If customers require the use of other standards or have their own criteria, their expectations should be followed.
MSA Standards
- AIAG MSA "Measurement Systems Analysis".
- VDA 5 "Capability of Measurement Processes".
- SAE AS13003 "Measurement Systems Analysis Requirements for the Aero Engine Supply Chain".
- RM13003 Measurement Systems Analysis. A manual published by AESQ, supporting the SAE AS13100 standard.
- ISO 22514-7. Statistical methods in process management - Capability and performance. Part 7: Capability of measurement processes.
Advantages of MSA
- Improvement of the measurement system: MSA allows for the estimation of variation within the measurement system and helps in its improvement. Knowing what needs to be improved enables corrective actions.
- Better product quality: Reduced measurement variability leads to better quality control and production processes. Fewer losses translate to greater savings.
- Meeting requirements: The method is very useful for meeting the requirements of various management systems. For example, the use of MSA is a requirement in the IATF-16949 standard. All measurements included in the control plan should undergo an MSA study.
Disadvantages of MSA
- Time-consuming: MSA analysis requires time for preparation, execution, and analysis of study results.
- Requires training: Applying the MSA method requires prior training of personnel.
- Difficulty in a variable work environment: MSA should be conducted every time there are significant changes in the measurement system. If the changes are introduced very frequently, repeating the MSA becomes very challenging.
Summary
Improving the quality of anything first requires that it can be measured. After all, we need to know exactly what needs to be improved. Once actions have been taken, we should also check whether there has truly been an improvement, or if it is only our perception. This is why measurement is a crucial aspect of quality improvement.
However, when taking measurements, it is important to remember that the measurement system itself is a source of errors, or in other words, a source of variation in measurement results. If this variation is too high, our analysis of the results may be incorrect.
In such cases, MSA comes to the rescue by verifying whether our measurement system is capable of providing reasonably reliable results. Through MSA analysis, we can determine whether the measurement system itself needs improvement, or if we should focus instead on improving the production process.
In conclusion, MSA analysis shows how much we might be wrong when making measurements. It is essential to know how much we can trust the measurement results, especially when they pertain to something significant.
Footnotes
- Joint Committee for Guides in Metrology (JCGM). Evaluation of Measurement Data—Guide to the Expression of Uncertainty in Measurement, International Bureau of Weights and Measures (BIPM), Sèvres, France (2008), URL, https://www.bipm.org/documents/20126/2071204/JCGM_100_2008_E.pdf/cb0ef43f-baa5-11cf-3f85-4dcd86f77bd6. BIPM, IEC, IFCC, ILAC, ISO, IUPAC, IUPAP and OIML, JCGM 100:2008, GUM 1995 with minor corrections.
- G. Larsen, "Measurement System Analysis in a Production Environment with Multiple Test Parameters," Quality Engineering, vol. 16, pp. 297-306, 2003, doi: 10.1081/QEN-120024019.
- AIAG (Automotive Industry Action Group), Measurement Systems Analysis (MSA) Reference Manual, 4th ed., AIAG, Southfield, MI, USA, 2010.