Scientists Asked to Leave a Deposit

Here's an amusing story with a tenuous measurement link.....

According to the Guardian newspaper, staff members at the UK Laboratory of the Government Chemist (LGC) were sent an e-mail last week asking male staff to donate semen by visiting the men's toilets "on the first floor nearest to Block 1 between 10am and 5pm".

Apparently, the samples were wanted by the Crime Scene lab to make practice forensic exhibits and for training in DNA extraction techniques. The message continued "If you are able to provide a sample in one of the tubes and place this in the ice box, this would be much appreciated."

An LGC spokeswoman later confirmed the e-mail to be genuine but wouldn't be drawn on whether any donations had been made.

Visitors to the country's national metrology institute, the National Physical Laboratory, may be aware that NPL shares its site in Teddington, south-west London with the LGC.

• Guardian Unlimited

Meaning of "Alignment" Debated

ISO's Conformity Assessment Committee (CASCO) has established a Working Group (WG25) with the aim of aligning 17025:1999 with 9001:2000. Nineteen of its thirty-eight members came together for the first meeting in Geneva, Switzerland at the end of January to discuss how to carry out the mandate. Peter Unger of A2LA provided the following report.....

Several documents were presented beforehand, in particular one from NATA Australia that provided an analysis of existing clauses of 17025 against 9001. Several alternatives on how to proceed were discussed. There was general agreement to not change too much, but be able to say that labs accredited in accordance with 17025 also complied with ISO 9001. Defining what alignment meant was not agreed. It appeared to be agreed that alignment is a matter of degree but what would be an acceptable degree of alignment was not determined. Many proposed recommending to CASCO to slow down the process until the transition to both standards was farther along. Many also suggested that alignment may not be desirable, but it was conceded that such a position would be contravening the CASCO mandate.

There was strong support to survey the laboratory community on what they believe should be done but the convenor, Peter van de Leemput, said that ISO could not sponsor one and that such a survey would probably result in a split view and we do not have criteria to decide what to do on the basis of the outcome. There was a strong push just to revise clause 1-6 of 17025 (the claim of compliance with ISO 9001:1994) and to revise the cross-reference annex to the 9001:2000 version from the 1994 version and declare the alignment is already sufficient. The convenor suggested that this was not going to be acceptable so no vote was taken on the proposal. However, he summarized the points upon which we did seem to agree:

1. 17025 is not a sector-specific quality management system (QMS) standard and therefore how far the ISO sector-specific QMS directive applies in this case was debatable;
2. 17025 should be a stand-alone document to ensure that there would be no need to purchase ISO 9001:2000 or any other references;
3. The full text of ISO 9001 would not be adopted, but there would probably be some tailoring of parts of the text to make it more understandable for labs;
4. We should not be eager to change the words; keep changes to a minimum as much as possible (no "major surgery");
   [Although it was my expressed opinion that major surgery may be needed to fully align 17025 QMS requirements with 9001 but that we should follow the proposal from the American National Standards Institute (ANSI) and the proposed structural principles from CASCO Working Group 23 on common elements of conformity assessment bodies]
5. No changes should be made to section 5 of 17025 (i.e., no changes to the technical requirements will be proposed).

After further discussion on the way forward, it was agreed that a small
task group (comprising Peter van de Leemput, Netherlands; Pierre de Ruvo, Switzerland; Christian Ranson, France; Alberto Pederneschi, Italy; Merih Malmqvist, Sweden; Ned Gravel, Canada) would:

1. Analyze the lack of correspondence and degree of equivalence between 17025 and 9001:2000;
2. Attempt to justify no change with provision of objective evidence.

The work of the task group is to be completed by 5 June ready for the next meeting, scheduled for 5 July 2002. Peter Unger would welcome your comments about the way forward.

• A2LA

Spreadsheet Users Warned

Britain's National Physical Laboratory is leading a research program on software used in metrology. One of the topics is validation of commercial-off-the-shelf (COTS) packages used extensively by labs, such as the Microsoft® Excel™ spreadsheet application. To minimise costs, metrologists and other engineers often utilise the capabilities of their business software suite rather than buying specialised statistical packages, for example.

The tests made on Excel™ have revealed that several functions, such as for standard deviation and regression, are based on defective algorithms or formulae and can produce highly inaccurate results. These are the very functions that feature heavily in metrologists' measurement analyses!

If this news comes as a shock and you want to quickly see an example of one of these problematic calculations, select several adjacent spreadsheet cells and type the same number in each. In another cell, enter the formula =STDEV(cellrange). The result should, of course, be zero but depending on the chosen number a variety of non-zero results are returned. For example, 83.7 in 5 cells produces a reported standard deviation of 1.35E-06. It's not only when the numbers are the same.

Luckily, for this function, there is a fairly simple work-around which is to subtract each cell's value from the mean of the group and perform the standard deviation function on those differences instead. The result will now be correctly calculated and is mathmatically equivalent.

NPL strongly advises users of Excel™ to download reliable algorithms as add-in functions for use in place of the defective intrinsic functions.

• Software Support for Metrology Program
• Summary of Excel Validation

®™Microsoft Corporation

Green Light to Europe's GPS

After protracted debate, the European Union finally gave the go ahead to development of its own global satellite-based navigation system. Galileo, not to be confused with NASA's Jupiter mission of the same name, is a 3 billion euro (about US$3B) project that will compete with, and complement, the aging US GPS and Russian GLONASS solutions.

In late March the EU released half a billion euro which, together with a similar contribution from the European Space Agency, will enable the project's developmental stage. This is expected to establish the test infrastructure and 4 satellites in orbit for validation by 2005. If all goes well, the fully operational constellation of 30 space vehicles is targeted for 2008.

Like GPS, Galileo will not only provide for the navigation business but will benefit the precise timing market as well.

• Galileo European Satellite Navigation System

Certification Offered for Calibration Personnel

Prompted by an increasing focus in quality and competency standards such as ISO9000:2000 and ISO/IEC17025, qualification of laboratory technical personnel is becoming a hot topic in several countries worldwide. Formally recognized training and certification schemes for calibration technicians and metrology professionals alike have already been implemented in Britain and Mexico and most recently in the US by the American Society for Quality (ASQ) under their Certified Calibration Technician (CCT) program.

ASQ CCT is designed to be a peer recognized certification program that provides authentication on the level of knowledge and experience associated with the duties and responsibilities of a mid-level calibration technician based on the consensus expectations of hundreds of calibration / metrology professionals that contributed to the development. The focus of the CCT program on the mid-level calibration technician is deemed by the program committee to be the most advantageous for industry. Full details of the program's scope, requirements, etc. is expected to be available in October.

• ASQ CCT

Steps Towards 17025 Revision (9001 Alignment)

ISO/IEC 17025, General requirements for the competence of calibration and testing laboratories, was published in 1999 taking into account the quality management requirements specified in ISO 9001 and 9002:1994. Subsequently, ISO 9001:2000 was published. In order to align ISO/IEC 17025:1999 with ISO 9001:2000, ISO/CASCO Working Group 25 is revising ISO/IEC 17025 and met last week in Geneva, Switzerland.

The first working draft, that had resulted from the last WG meeting in July, was reviewed with considerable discussion and suggestions for changes. Several key issues were agreed including how the alignment should be described in the Scope section of the standard. This is that alignment is with the "principles" of 9001 rather than its "requirements". There was also concensus that 17025 is not a sector-specific quality management system (QMS) standard despite incorporation of the 9001 principles, nor is it intended for certification purposes.

The next step will be production of a second working draft next month that will be circulated to WG25 and Technical Committee 176 (responsible for ISO9001) for comments due in March. This document will not be made public.

Depending on the significance of the feedback, the working group may meet again in April. Assuming all goes well, including agreement by CASCO to proceed with a shortened process (and after translation is completed), a draft amendment to 17025 would be circulated to ISO CASCO membership for the required five-month review and voting period. This would be equivalent to the process for a draft international standard (DIS) and is expected to take all of 2003. An amended 17025 will then not be available until sometime in 2004. Of course, there would be an agreed transition period for implementation of any new requirements contained in the amendment. Consequently, the full review and any possible major revision of 17025 would not be initiated until early 2005 in accordance with ISO five-year review rules.