What often eludes us, but is worth keeping in mind, is that measurement itself is a process and can be treated as such. Calibration and traceability are parts of that process, and all have the same fundamental limitation as all measurements: Uncertainty Rules!
Learning to Get It Right
One measurement is never enough. Every measurement has an accompanying error, and that immediately takes you into the world of statistics. To achieve measurement quality, you have to find the answer to such questions as: How many measurements are enough? How can you determine and express the magnitude of measurement errors? And how do you know you are measuring as accurately as you should?
For details on the methods and statistics involved in establishing and maintaining measurement quality, the MSA-3 manual—published by the Automobile Industry Action Group (AIAG) and available online for between $9 and $27—provides an excellent introduction to relevant methodologies and a rigorous treatment of measurement QA concepts and their part in statistical process control (SPC).
The MSA-3 manual is loaded with math, making it a formidable pill to swallow. For additional support, AIAG offers classes that thoroughly examine measurement-quality concepts and their application. These resources will help you understand the SPC picture, which is so fundamental to AIAG's QA program, QS-9000, and others such as Six Sigma.
In addition, there are lots of free resources on the Web that can help you cultivate an understanding of measurement uncertainty. One developed for college students can be found the on the Purdue University Web site. There is also a tutorial on the subject on Frostburg State University's Web site, so you can learn at your own pace from simple, online exercises. It is well worth a look. As is the Web site by Eurachem, which is devoted to the subject and features several advanced downloadable, detailed tutorials specifically designed for analytical chemists. However, the terminology and concepts apply to nearly all measurements.
Achieving Long-Term Success
An appreciation of measurement uncertainty is but a starting point, but it is essential if you are to effectively establish and run a competent measurement program in industrial automation. You can get along without it for a short while, and many do. But only organizations that come to grips with measurement uncertainty are successful in the long term.