Sidebar
ISO 17025 In Texas - ISO 17025 Laboratory Certification

ISO 17025 In Texas - ISO 17025 Laboratory Certification

Posted by Barry Ward ,Jun 27th 2016

Why 17025? Do you need to do an ISO 17025 calibration? First, let me describe 17025 as I see it. This standard basically says that a shop does things consistently. There are supposed to be documents that cover everything we do, and we are supposed to follow those documents. We have been accredited for about 6 months and I have already updated many of our documents. These are real, live, used documents. They discuss how we deal with our customers and their equipment. They discuss how we will calibrate a piece of equipment – what are our procedures? What will we actually test. For some things there might be industry standards and these are used; for others we have our standard procedure that is a compilation of the industry standards; for still others the manufacturer may have a recommended procedure that they make available.

anton-paar-inhouse-calibration.jpg

Now, having said that 17025 helps assure consistence, it doesn’t necessarily assure good consistency. A shop, or an individual, could follow the procedures and still be bad. With 17025 this will show up in the uncertainties. This is the term that most calibration shops use while the normal person might talk about accuracy or quality. In the calibration world uncertainties are the deviations from the ideal value. Some factors that influence this are: the resolution of the device being calibrated (how many digits are on the display?), the resolution of the calibrator, temperature changes, and the consistency of the operator. This last one, the consistency of the operator, seems difficult, but is actually fairly easy to determine. Have the person do a measurement 10 times, starting from scratch each time, and log the 10 values. If the device they are calibrating is sufficiently accurate then you will see some variation in the 10 values. Do some simple math on this and you can determine how inaccurate, or technically what is the uncertainty, of this person doing this check. Do the same for all of the people that might do this type of calibration check, and you can see if the process is consistent (the same uncertainty for each person) or if the person is inconsistent (different uncertainties for different people. ISO 17025 requires this type of testing to determine uncertainties.

ISO 17025 companies are audit before they are accredited and annually thereafter. During this audit all of the documents are subject to being audited. If something needs a record to show proof that it was done, then that record will be examined. In the case of a calibration procedure, a dummy calibration might be done. This allows the auditor to see if the calibration is done per the procedure and if the correct records are made and kept. We are also required to do proficiency testing. For this process, we will get an instrument within our scope of accreditation, and do a calibration on it. We report our results to the proficiency testing company. They will check our results against those of other company’s results. If we get the same results (within an acceptable error), then we are considered acceptable to the test company.

Then second part of having the procedures is to audit them and see if they are being followed or if they need to be changed. A procedure that works for a one-man company may not work when there are a dozen calibration technicians working for a company. If the calibration equipment changes the procedure may need to change to reflect the new equipment and its specifications and capability. Again, the documentation is a living, changing document that is being updated as needed. This is the part that makes an ISO 17025 company into a better company, not just a consistent company.

So, do you need an ISO 17025 accredited company? Maybe not, but you will get a consistent quality job. Even if you don’t need the accredited certificate, using that company will ensure that you get the calibration work you want.

photo credit: Anton Paar