Value of Measurement will be Promoted

15 December 1998

The UK Department of Trade and Industry's National Measurement System Directorate (NMSD) has announced the National Measurement Partnership program which, through a new body called the National Measurement Forum, aims to increase:-

• industry's uptake of traceable measurement
• the skill base in measurement
• access to measurement experts

The contract to run the program has been awarded to a consortium led by NPL and UKAS. NMSD and NPL have details of this program and other measurement improvement initiatives.

• NMSD
• NPL

Leap Second Announced

15 July 1998

• IERS

New DIS17025 Distributed for Comment

01 June 1998

In late May a drafting committee version of the latest "Draft International Standard" (dated 30 March 1998) was distributed to key representatives of (US) industry for comment. The official document has been sent to ISO/CASCO members for a five month commenting period during which national standardization bodies (ANSI, BSI, etc..) will co-ordinate their country's views. Agreed changes, which are expected only to be minor in nature, will be made and it then moves to the last stage in the process as a "Final Draft International Standard" (FDIS17025). ISO/CASCO members will then have a final one month balloting period for a simple yes/no vote.

So, all seems on-track for publication of the new ISO17025 standard in the first half of 1999.

More Millennium Misery for Metrologists ?

15 April 1998

It looks like the year 2000 could turn out a costly one for business. Apart from the infamous "millennium bug" that's currently resting dormant in some computer and measurement equipment, two more events are planned.

1. A 1980 European Directive (80/181/EEC) is scheduled to prohibit any mention of imperial (English) measurement units in sales or support material (advertisements, datasheets, catalogues, user or service manuals, etc.), or product packaging and labeling. Currently, although metric measures must be shown, the Directive allows a transitionary period when a dual marking system of "supplementary" units can be used but from January 1^st 2000 only metric measures must appear. However, international representations, such as those made at the EC Transatlantic Business Dialogue Briefing in Rome last November, may yet postpone it. According to a UK Department of Trade and Industry source, a Council meeting in Brussels last month (March'98) apparently indicated that member states offered no objections to rescheduling to 2010, although no decision on amendment was taken.
2. Is your lab still using a LORAN-C receiver as a frequency standard or a means to maintain traceability? If so, be aware that the US Department of Transportation plans to turn-off these transmissions in 2000. Most navigators have already adopted the superior GPS, as indeed have many calibration users.

Guide 25 Migrates to Standard

15 March 1998

The ISO's Conformity Assessment Committee (CASCO) has a Working Group (WG10) which is endeavoring to update ISO/IEC Guide 25. Following agreement in 1997 to use their "fast-track" process, the ISO has announced that the forthcoming revised version will become a standard, ISO17025. The draft writing group met in Holland last October and December to condense the considerable feedback resulting from the public consultation held during 1997.

The Working Group met in Switzerland in late-February'98 to consider these further changes -- see Jim Cigler's unofficial report. Informed sources believe that the public comment was so extensive and controversial that the committee won't be able to deliver an "FDIS17025" (final draft international standard) from this round of discussion and that further delay is likely before it goes back to the national CASCO members for voting. On the European scene, the EN45001 standard is being revised in parallel so that it is fundamentally identical. UKAS intends to scrap its own criteria (NAMAS M10) and adopt the new EN45001 when published.

ISO10012 also in Revision
For the past couple of years another group (ISO/TC176/SC3/WG1) has been revising this standard which gives advice on managing Quality Assurance for Measurement Equipment. Cited in ISO9000 standards, it's also stipulated in UKAS measurement accreditation criteria. A new ISO10012-1 addressing the Metrological Confirmation System was published in 1997 and in November a committee draft (ISO/CD10012-1.2) was released for the second part on Control of Measurement Processes.

ISO/IEC Guide 25 Revision Update : Unofficial Report from ISO/CASCO WG10

11 March 1998

The chief of the US accreditation scheme (NVLAP) at NIST, wrote the following summary report on progress made in drafting the new ISO17025 (currently ISO/IEC Guide 25:1990). Key topics discussed in Geneva were its compatibility with ISO9000, the clauses on traceability and, possibly the most contentious matter, guardbanding. These are not the official minutes; Mr.Cigler's summary represents his perception of significant points, as reported to the NCSL.

Date: March 11, 1998
Author: James Cigler, NIST
Summary of ISO/CASCO Working Group 10 Meeting in Geneva, Switzerland, February 26-27, 1998.
The WG 10 Chairman, Dr. Peter van de Leemput of RVA, convened the meeting on February 26. A moment of silence marked the passing of Dr. John Rodgers of UKAS who had served on Working Group 10 and on the drafting committee for the revision document.

Dr. van de Leemput reported that ISO/CASCO had decided to ballot the revision document as a standard, now known as ISO 17025. He stated that the working group's objective was to review comments received after circulation of the last previous draft, and to develop a revised draft to be sent to ISO/CASCO members for a five month comment period. Working Group 10 will meet again to incorporate all new comments and to prepare a final draft of the standard which will be submitted to ISO/CASCO for a two month ballot period.

After review of some comments on the introductory sections, including references and definitions, most of the first day was spent on section 4, Management Systems. Many comments related to incorporating ISO 9000 clauses to make the standard more compatible with the ISO 9000 standards were discussed. This work continued for the first hour of the second day, with the remaining time being devoted to Section 5, Technical Requirements.

The most notable discussion centered on Section 5.6, Measurement traceability. This section has been divided into consideration of traceability of calibration laboratories, and separate consideration of traceability of testing laboratories. The current text requires that all calibrations and measurements made by calibration laboratories be traceable to the SI units of measurement. Allowance is made for testing laboratories when traceability to SI units of measurement is not possible and/or is not relevant. Confidence in measurement results may be obtained in such cases by various possible methods, including participation in interlaboratory intercomparisons or proficiency testing, use of suitable certified reference materials, and replicate testing or calibrations using different methods.

I raised the issue that it is not always possible for calibration laboratories to achieve traceability to SI units, and that the same allowances afforded to testing laboratories should be made for calibration laboratories. As examples, I cited the areas of hardness, ratiometric measurements, and chemical calibration using standard reference materials. Opinions regarding this issue differed. The chairman will write to Terry Quinn of BIPM asking him to coordinate a position from the national laboratories in each country that might be used in the standard.

Another substantive issue arose in Section 5,10.4, Calibration certificates. The document being reviewed by the working group contained a statement that, when performing calibrations that ensure that parameters are within specified tolerances, "the measurement value(s) extended by the estimated uncertainty of measurement have to be within the appropriate specification limit". This technique is sometimes referred to as "100% guardbanding". With this approach, the customer of the calibration laboratory is guaranteed that the tolerances are not exceeded, however, there is an 18 % risk that the laboratory rejects results that are within tolerance. This can be very expensive when calibrating large numbers of devices and results in unnecessary adjustments or rework to achieve the restricted tolerances.

In the US, France, Australia, and some other countries represented at this meeting, 100% guardbanding is sometimes used, but it is also acceptable to use statistics associated with the test accuracy ratio between the test system and the device being tolerance-tested. Customers can choose to accept a stated probability that measurements are correct within tolerance limits. For a test accuracy ratio of 4:1, the probability of accepting measurements that are within tolerances is 99.2%. There was considerable disagreement among the working group members on this issue. The chairman achieved group consensus to remove references to 100% guardbanding in this section. While this will probably be acceptable to US interests, the difference in tolerance testing philosophy needs to be addressed as an international consideration. By remaining silent on the issue, each country will be left to its own interpretation of what is appropriate. This may resurface as an issue that could make international recognition agreements difficult to achieve.