## 04 When the top metrology fails
This time I am going to talk about my job: metrology. No, not meteorology, but the "scientific study of measurement", which is what I do every day for a major car company. To understand most of the information in this issue, you need to understand some basics. All measurements are related to standards, a specific instrument that is used as a reference. I have references from the site where I work. They are directly linked to their references for a region or sometimes a country. And each nation participating in ILAC (ILAC International Laboratory Accreditation Cooperation), a global network of accreditation bodies that promotes confidence in laboratory test results, has its own reference or standard for all physical quantities such as a length, a volume, a pressure, a voltage, a current, a time, etc. ...
You can find information there :
=>
https://en.wikipedia.org/wiki/Metrology Wikipedia on Metrology
So I rely on the best labs to give me confidence in what I'm measuring. And for years I had confidence in these "top labs". I did what is called an interlaboratory comparison. It is used to assess the reliability of your lab's results. An interlaboratory comparison involves carrying out tests on similar samples and comparing the results. I do it with all the members of the laboratory, for example to validate a new member of the team. I'm the 'referee', but I also have to compare myself and my measurements with other referees. So the best labs are useful for that. The problem is that some of their results were strange, heterogeneous or even random without explanation. Errare Humanum Est... But in laboratories like my employer's, people who know their job perfectly are becoming rare. Retirement or simply a change of job are the reasons. It's a matter of experience to understand the results of a measurement. I know myself what is behind some results, if a small mistake has been made, if someone has copied the w
rong figures, etc...
We recently changed our reference laboratory for a material. We had strange results in France with a kind of oscillation around zero in opposite directions every year. We had to deal with the manufacturer and their labs in the UK and the US, which are linked to Japan. Of course, the UK is better for us because it's not so far away from France, despite Brexit. With the French labs, it was difficult to compare the results of similar materials. It was always one standard deviation (see 1) of 0,8%...but with the UK lab it was more like 0.15% and the variability with the last years is not that different now without oscillations. The method of measurement seems to be similar between the labs but the results are so different. I have done some measurements myself on some of this material to see if it is repeatable (see 2). It was, and the influence of the user is not that great. I know that the French laboratory that did this in the national laboratory consisted of 2 people. The head retired and the new head was the
former "apprentice"... But I don't even know the results for the last 10 years. It's a very specific measurement and very few laboratories can do it. Confidence is the secret and you are always confident when you see an ILAC label.
I also had a problem with one of my standards and I had to compare myself to my own different standard to prove it. It could have been a big problem with all the work I do and all the equipment I have to check. As usual, there was no response from the national laboratory and therefore they are not ISO 17025 compliant. I have also seen many other laboratories in France with similar problems with their measurements. It's now difficult to find a single technical auditor who complies with ISO 17025. I know that other European countries have similar problems. It's frightening for my profession and for the whole industry to see such a loss of skills. Industry and metrology are not attractive to those entering the job market. I'm still one of the youngest people in my profession, although I've been doing it for 25 years. I don't know if my little laboratory will be here in 2 years. At the moment, there are only 2 companies in France capable of carrying out metrology, with a concentration of this activity. For the n
ational laboratories, it's less people and less money every year, as we can see in other national research laboratories such as ....MeteOrology, Geography... I work with China, Morocco, Brazil, India and I know that many people are new to these jobs and sometimes they tell big lies out of ignorance of the subject because there is also competition between countries. Metrology should not be a matter of competition, but an alliance of the best. Many people have forgotten that.
=>
https://en.wikipedia.org/wiki/Standard_deviation 1- Wikipedia on Standard Deviation
=>
https://en.wikipedia.org/wiki/Repeatability 2- Wikipedia on Repeatability
2DÉ›
=> mailto:
[email protected] Comments by mail
or by a reply on your blog
