I

MEASURE FOR MEASURE

Why Buy Accredited?
by Philip Stein

recently taught a refresher class for the ASQ certified calibration technician exam. One of the students asked whether buying accredited calibrations was worth the money, and I thought you might be interested in the answer as well.

Some of you simply may have no choice. Many calibration customers, most notably those suppliers registered to the automotive industry’s QS-9000 standard, are required to use accredited labs. Sometimes, accreditation may be required by law or regulation. Either way, case closed. Also, if your lab is calibrating reference standards for a lower tier lab that is accredited, your lab will most likely need to be accredited. (Another approach is allowed, but it’s so hard to do it’s rarely followed.)

If neither of these requirements applies to you, using an accredited lab can still offer many important technical advantages, including:

• Traceability.

• A quality system.

• Tracking of defects.

• Staff training.

• Measurement assurance.

• Contract review.

• Service to the client.

• Measurement uncertainty.

Traceability. Every calibration, regardless of why it’s needed or who performs it, needs to be traceable. This means the measurements made when the calibration is done can be related to national standards by an unbroken chain, and each step in that chain includes a report of the uncertainty of the measurement. Without traceability, there’s no confidence the results have much meaning.

Quality system. An accredited lab has a quality system similar or identical to one registered to ISO 9000. As a quality professional, you will be familiar with the lab’s document control, corrective action and controlled processes. Although much of the benefit of a quality system goes to the laboratory, it’s nice to know the lab will have solid records of the work it did for you.

Tracking of defects. When a problem is found in an accredited lab, the ramifications are investigated and communicated. Your own quality system demands that any problems that are discovered be tracked backward and their consequences analyzed.

This requirement extends the responsibility to your supplier labs. When an accredited lab finds indication of possible trouble, it must analyze all work done during the uncertain time and let you know if there might have been an issue. Also, if the lab tries to calibrate your equipment and finds it out of tolerance or broken, it must let you know so you can start your own analysis.

Staff training. Accreditation guarantees calibration technicians are adequately trained and supervised. During an accreditation assessment, many technicians are interviewed. This interview digs much deeper than an ISO 9000 or Baldrige interview, where employees are quizzed about their knowledge of the company’s goals and products and the quality system and their responsibilities under the system.

An accreditation assessment asks about technical issues. Does the technician understand the scientific principles underlying the work he or she is performing? Does he or she understand the meaning of measurement influences and how they combine to form an uncertainty statement about his or her work? Knowing the technicians have these skills will increase your confidence in their work.

Measurement assurance. An accredited lab is required to carry out a program to increase the confidence in its measurement results. This is usually called measurement assurance. There is a wide range of options as to how this can be done, and a lab will often choose more than one option.

Foremost among these choices is proficiency testing. The laboratory will exchange items to be calibrated with other similar labs in an interlaboratory comparison. Statistically significant deviations from a reference “right” answer require root cause analysis and corrective action.

Other choices for measurement assurance have been previously discussed in this column, the most important of which is the use of control charts and check standards to ensure the measurement process is stable and in control during the time your equipment is being worked on.

Contract review. Any work undertaken by an accredited lab must be discussed with the client, and the lab and the client must come to a mutual understanding regarding the technical work that should be done. Although it’s not common, labs are moving toward including decision rules in this discussion.

The client—that’s you—needs to tell the lab what to do if a calibration result is close to a tolerance or specification limit—especially when measurement uncertainty could cause that result to go over the limit line. The lab really can’t make that decision for you, although that’s often what happens today. (For more about decision rules, see ANSI/ASME B89.7.3.1-2001-2002 Guidelines for Decision Rules: Consider-ing Measurement Uncertainty in Deter-mining Confirm to Specifications.)

Service to the client. This is a requirement of ISO/IEC 17025. Clients are to have reasonable access to observe their work being done. Access to several other activities, such as customer surveys, is recommended but not required. Nonconformances to this clause are rarely cited because most calibration labs are cooperative.

Measurement uncertainty. The reported results will include a statement of measurement uncertainty (or conformance to a tolerance or specification) calculated according to international standards.

While no calibration or laboratory result can be guaranteed perfect, using accredited labs and purchasing accredited services from them increases the probability the work they did and the results provided to you are correct and supportable. It may cost a little more, but it may be well worth it.

 

---

PHILIP STEIN is a metrology and quality consultant in private practice in Pennington, NJ. He holds a master’s degree in measurement science from George Washington University in Washington, DC, and is an ASQ Fellow.-

94 I JULY 2004 I www.asq.org

Using an accredited lab may help you rest assured your results are correct.

comment

If you would like to comment
on this article, please post your remarks on the Quality Progress Discussion Board at www.asq.org, or e-mail them to editor@asq.org.

Please

QUALITYPROGRESS I JULY 2004 I 95