New ISO 6789 standard brings uncertainties for torque wrench laboratories
Those of you who have read my previous blogs and especially my March 2016 update will know that there are lots of changes to ISO 6789 coming at the end of the year. In our ISO Working Group, the combined experts from several international manufacturers and calibration accreditation bodies have created a comprehensive standard for the traceable calibration of both setting and indicating torque tools.
ISO6789 part 2 is completely new and we have put thousands of hours of work into it. We know that it will confuse some people however, and this blog is aimed to reduce the confusion. If it doesn’t please contact me and ask.
Torque tools come in all shapes and sizes. Their mechanisms and performance do vary and just because most manufacturers claim that they are accurate to +/-4% it does not mean that they are all the same. Those of you who are experienced laboratory managers or technicians will understand the principle of applying an uncertainty to the observed readings. This allows us to say that the observed value has a 95% (in most cases) probability of being within W% of the reference value. The way that a laboratory estimates the components of uncertainty will have a significant effect on the value of the expanded relative uncertainty W. The purpose of ISO6789 part 2 is to ensure greater consistency between laboratories. This will mean that customers are purchasing a more consistent calibration and national accreditation bodies can more easily compare the performance of laboratories under their accreditation scheme.
One challenge that laboratories will have is to charge the correct price for the work that goes into a calibration. Following the requirements of the standard for evaluating the uncertainty budget takes about two hours and that means the laboratory should charge accordingly. If a laboratory is offering a calibration taking less than that, you should ask how they are doing it.
The exception comes when a laboratory has already calibrated at least ten similar examples of that specific torque tool. Similar could mean grouping new tools which behave differently to used but well maintained tools and which are different again to heavily used or badly maintained tools. ISO6789 part 2 then allows the laboratory to use statistical techniques to assign an uncertainty for that specific model of torque tool. The laboratory does still need to periodically evaluate the value by performing the full process but they do not need to do it for every calibration of that model. In our experience as a laboratory, we see some brands and models frequently and some are quite rare. Those rare tools are going to take longer to calibrate if we have not seen at least ten of them in recent months.
If you are a user, you will get a better but more expensive calibration. If you are a laboratory, you will need to look at your resources before the new standard comes out at the end of the year.