Why Calibrate ? Or "Calibration? How does that help me?"

British scientist Lord Kelvin (William Thomson 1824-1907) is quoted from his lecture to the Institution of Civil Engineers, 3 May 1883...
"I often say that when you can measure what you are speaking about and express it in numbers you know something about it; but when you cannot express it in numbers your knowledge is a meager and unsatisfactory kind; it may be the beginning of knowledge but you have scarcely, in your thoughts, advanced to the stage of science, whatever the matter may be."

This famous remark emphasizes the importance that measurement has in science, industry and commerce. We all use and depend upon it every day in even the most mundane aspects of life -- from setting your wristwatch against the radio or telephone time signal, to filling the car fuel-tank or checking the weather forecast. For success, all depend upon proper calibration and traceability to national standards.

As components age and equipment undergoes changes in temperature or sustains mechanical stress, critical performance gradually degrades. This is called drift. When this happens your test results become unreliable and both design and production quality suffer. Whilst drift cannot be eliminated, it can be detected and contained through the process of calibration.

Calibration is simply the comparison of instrument performance to a standard of known accuracy. It may simply involve this determination of deviation from nominal or include correction (adjustment) to minimize the errors. Properly calibrated equipment provides confidence that your products/services meet their specifications. Calibration:

* increases production yields,
* optimizes resources,
* assures consistency and
* ensures measurements (and perhaps products) are compatible with those made elsewhere.

By making sure that your measurements are based on international standards, you promote customer acceptance of your products around the world. But if you're still looking to justify that the cost of calibration does add value, check-out some of the calibration horror stories that have been reported.

Paperless Calibration

Bulging filing cabinets or over-full hanging files are a common office scene. But as far as calibration records are concerned, is the "paperless office" taboo ?

What do You Keep in Your Drawers ?

Most quality managers keep calibration results and certificates in their drawers! From discussions that have taken place with many people in industry whose responsibility includes the control of test instruments, a filing cabinet full of paper calibration ‘evidence’ is an integral part of the quality system -- without which audits would fail and the business would crumble. But when pressed for a rationale for such belief, three main reasons to maintain paper records emerge:

• They believe that auditors would not accept any alternative
• They believe that ISO9000 or accreditation agencies demand it
• It is historical; they have always done it and it is a comfort factor.

Alternative Feared

During these discussions a potential alternative option based around electronic records being retained by the calibration supplier, to be provided electronically on demand, was met with a mixed reaction.

On one hand, the positive aspects of fewer papers to handle, file, retain, refresh, retrieve, etc. were enthusiastically supported. However, the conflicting dilemma that such a change might have an impact on audit success tempered that initial enthusiasm. Equipment managers' fundamental belief is that both ISO/IEC Guide 25 or EN45001 (ISO17025) and ISO9000 audit bodies would not recognize or be comfortable with such a ‘virtual’ record system. This fear alone would deter them from seriously considering any such change.

This collective feedback formed the basis of a discussion between Hewlett-Packard in Britain and senior officials from the United Kingdom Accreditation Service, the agency responsible for both accrediting calibration/test labs and overseeing quality management system registrars. The goal was to establish, for the record, whether UKAS would endorse a paperless system. The outcome of this meeting is summarized in a letter to Agilent Technologies from Brian Thomas, Technical Director of UKAS, in which he summarizes that the responsibility of the user of calibration services (the customer)

"....is to be able to demonstrate to the assessor that it can, and does when needed, obtain evidence of calibration and that it has an effective records system enabling tracking back of full calibration data and certification for the defined period."

This doesn’t mean that records are necessarily kept locally by the equipment-user in paper form but that they could, indeed, be retained by the supplier of the service and provided when needed at any time in the future. In most cases, the only data a company needs in real-time relates to parameters found to be outside the instrument’s specification when initially tested (on-receipt status) so that a potential product-recall process may be invoked. But even this doesn’t need to be provided on paper -- it could be made available to the customer via the Internet (e.g. e-mail or a secure web server) or through a variety of other electronic means (fax, floppy disk, etc.).

Control is Crucial not Mechanism

Whichever medium is most appropriate, it is the evidence of control that is imperative, not the evidence of paperwork. In Brian’s words:

"In principle, your customers would be able to contract you to retain their calibration records; this arrangement would then become part of their system for retention of records. UKAS assessment of such a customer would address whether this system provided access that was easy, quick and reliable and controlled from the point of view of security, confidentiality and accuracy. Assuming this to be so in practice then the system would be acceptable to UKAS."

This alternative solution is, therefore, one which UKAS would support provided that the customer and the supplier met some key requirements. Those requirements are concisely detailed by Brian as:

"The documentation of such records and certification is acceptable in any form of medium, hard copy, electronic, etc. provided that it is legible, dated, readily identifiable, retrievable, secure and maintained in facilities that provide a suitable environment to minimize deterioration or damage and to prevent loss."

Dispelling Reluctance

So, the voice of industry is clear. It would like to take advantage of contemporary technology by contracting-out its data and certificate storage requirements and, provided that their suppliers could satisfy their needs (echoed by the needs of UKAS above), they are willing to forego historical practices by trusting virtual documentation. But the most significant reason that they are reluctant to take this step is fear of audit failure.

Agilent Technologies believe that a major step forward would be made if quality system and accreditation consultants and assessors could advise their clients that, far from impeding audit success, such a move could enhance it -- whilst at the same time saving space, time and ultimately money for both the equipment owner and calibration provider.

Calibration, Verification or Conformance?

When discussing calibration requirements with a potential supplier it's obviously important to understand what's being offered. Other articles in this section should help you to establish your requirements and distinguish the differences between available services. But one of the variations has, sometimes, even confused calibration laboratories and quality auditor. It's a matter of the difference between calibration, verification and even conformance.

Similar to the often confused specification terms accuracy and precision, a myth became "established wisdom" that calibration and verification are differentiated on the basis of quality or integrity.

* Specification terminology

Popular opinion being that verification is a quick-check of performance perhaps made without any real traceability, whereas calibration provides genuine assurance that the product really meets its specification. In fact, the US national standard ANSI/NCSL-Z540 defines "verification" as being calibration and evaluation of conformity against a specification. This definition originated with the now obsolete ISO/IEC Guide 25 but neither its replacement (ISO/IEC 17025) or the International Vocabulary of Measurement (VIM) currently have it or any alternative. The only relevant international standard that includes terminology covering the process of both calibrating and evaluating a measuring instrument's performance against established criteria is ISO10012 which uses the rather cumbersome term "metrological confirmation".

Calibration is simply the process of comparing the unknown with a reference standard and reporting the results. For example:
Applied= 1.30V, Indicated= 1.26V (or Error= -0.04V)
Calibration may include adjustment to correct any deviation from the value of the standard.

Verification, as it relates to calibration, is the comparison of the results against a specification, usually the manufacturer's published performance figures for the product. (e.g. Error= -0.04V, Spec= �0.03V, "FAIL"). Some cal labs include a spec status statement on their Certificate of Calibration. (i.e. the item did/did not comply with a particular spec).

Where no judgment is made about compliance, or correction has not been made to minimize error, it has been suggested that Certificate of Measurement would be a more descriptive title to aid recognition of the service actually performed. Some suppliers also use Certificate of Verification where no measurements are involved in the performance testing (such as for certain datacomm/protocol analyzers), rather than Certificate of Functional Test as this latter term is often perceived as simply being brief, informal checks as might be performed following a repair (often termed "operational verification").

Verification can also relate to a similar evaluation process carried out by the equipment user/owner where the calibration data are compared to allowances made in the user's uncertainty budget (e.g. for drift/stability between cals) or other criteria such as a regulation or standard peculiar to the user's own test application.

Verification is not intermediate self-checking between calibrations. Such checks are better termed confidence checks, which may also be part of a Statistical Process Control regime. The results of confidence checks may be used to redefine when a "proper" calibration is required or may prompt modification of the item's working spec as assigned by the user.

But what about conformance, especially regarding the meaning of a Certificate of Conformance ? Typically available when an instrument is purchased, it is now generally recognized that such a document has little value as an assurance of product performance. Of course, the manufacturer expects that the product conforms to its spec but, in this sense, the document simply affirms that the customer's purchase order/contract requirement has been duly fulfilled.