Vol. 8, No. 11

November 2006

PQ Systems

Racing to quality with GAGEpack

Quality Quiz: With a video!

MSA with Jackie Graham

Data in everyday life

Bytes and pieces

FYI: Current releases


Send Quality eLine
to a friend!

Just type in your friend's email below:


Sign up
If you received this newsletter from a friend and want your own subscription to Quality eLine, click below.

Subscribe to Quality eLine




MSA with Jackie Graham
The problem may be in the system.

This article is the first in a series in which Jackie Graham, Ph.D., will explore measurement issues and measurement systems analysis. Graham is director of PQ Systems Australia, Pty.

A critical aspect of managing processes is the ability to measure well. One needs to be confident in the integrity of the data collected so that appropriate decisions are made about acceptable product and process changes. Have you considered that product declared 'out-of-specification' may be the result of a defective measurement system rather than a poor process?

Costs associated with poor measurement systems include: defective product shipped to a customer; unnecessary process changes (over control); acceptance of poor product or rejection of good product, leading to unnecessary adjustments. The consequent negative effect on profits is obvious and avoidable. If you have ever re-sampled and re-tested because initial results were poor, or adjusted a production process based on bad results, only to find that the results got worse, you have likely been 'touched' by a poor measurement system. These symptoms are real indicators that measurement is the issue rather than the process.

When assessing any product, particularly when applying the principles of statistical process control, we assume that the data being collected and analyzed is accurate and reliable. We rely on the measurement system to provide good data. This assumption for many people is based on the fact that they are collecting data in a logical manner using approved procedures with an experienced tester and have an efficient calibration program for their measurement devices. But is this enough? Frequently, the answer is "no!"

Let's take some time to think about what constitutes a measurement system. It is important to look beyond the measuring instrument/equipment and consider the environment and personnel as well.

Have you wondered what impact different testers can have on results? Over time, it is common for people to develop bad habits when using measuring devices and test equipment. It is part of human nature to look for short cuts. A good example is driving a car. When we learn to drive, to avoid failing the driving test, we are taught good techniques and learn to avoid bad driving habits. Experienced drivers, however, often slip into sloppy habits, driving one-handed, for example. In a similar fashion, it is easy for testers to develop poor measuring habits. The result is that the accuracy of the data is adversely affected. From tester to tester there is no guarantee that everyone will achieve the same result. In measurement systems analysis terms, the ability of different testers to achieve the same result is known as reproducibility. It follows that when poor measurement techniques are followed, it is likely that individual testers will find it difficult to get the same results consistently. The ability of one tester to achieve consistent results is known as repeatability. A good reproducibility and repeatability (R&R) statistic is an essential precursor to the successful application of statistical process control and, more importantly, the ability to consistently produce within specification. Frequently, having good calibration programs for measurement equipment is seen as sufficient, but there is no provision for calibrating personnel or environmental factors. It's like checking a car's brakes and assuming it will always stop safely without considering the driver's ability.

Our understanding of the principles of statistical process control tells us that every process contains a component of inherent or expected variation in output. It is important to understand that variation in output also contains a measurement system component. This can be thought of as an equation:.

Process variation + Measurement system variation = Output measurement

Problems arising from a poor measurement system can be avoided by conducting a thorough assessment and implementing appropriate strategies to remedy any errors found and to ensure on-going accuracy. Such an assessment includes an examination of:

Measuring instruments
Environmental issues

The methods employed, known as R&R analysis or measurement systems analysis (MSA), follow a thorough, statistically robust methodology. They accurately define the ability of the various components of measurement systems. Many measurement systems, when assessed using these techniques, are found to be inadequate and are adversely affecting output.

We will be talking more about measurement systems analysis in future additions of Quality eLine, including how to set up a measurement study and analyze the resulting data.

We recommend GAGEpack for all your measurement assessment needs.

Copyright 2006 PQ Systems.
Please direct questions or problems regarding this web site to the Webmaster.