Dear all,

Could you please help me take over my metrology ignorance.


I found out that I really cannot explain what it means to register a base for calibration of a measuring device, when there is no standard for it (7.6 a). Please, help me by an example.

Another question is again probably to my poor English. What is really the difference between calibration and verification in 7.6 Russian translation unfortunately is not perfect. This particular point looks pretty misleading in Russian.


Thanks


Anton

Verification - To verify a piece of equipment is within tolerance between calibration intervals.

If you send a piece of equipment out for calibration to an accredited lab, say on a yearly basis, the checks you do on that equipment during that year is verification.

Your first point. Do you have a piece of equipment that needs to be calibrated but there is no procedure to do it against?

Anton,

I do not understand your first question. Do you have something that needs to be calibrated but there is no national or international standard? If so, I have faced this before and what I do is this: Create your own standard based on an agreed value for an agreed part (standard) between you and your customer. If you are measuring parameter "X" on a part that you are selling to your customer, you and your customer can pick a part, agree on the value of parameter "X" for that part, and use it to calibrate your measuring device. You will need to write a procedure as to how to calibrate, how often, etc.

I would also check with the OEM of the equipment for advice on calibration. They should have a procedure they use.

If not, Mike S. gave a solution I would follow.

Thanks,

As far as I understand registration in 7.6.a in ISO 9k2k does not imply going to a standardization body or an accredited lab to get calibratiuon basis registered. I just have to work the basis out, describe it and agree with customer (or other interested parties).

Anton

Anton,
You don't have to have the standard agreed by a customer. Make your own documented procedure for calibration.
Also be able to prove that there is no standard that applies.

Verification is often same/like samples tested by multiple labs. The results demonstrate/verify that the labs involved have consistency of method - or not, as is often the case.

Reading the new standard it looks to me like if you decide you need to do a calibration, and a national/international standard exists, you must use that national/international standard as a basis for the calibration. However, I can see cases where you (the vendor) and your customer (or a whole industry) may agree to use a different method/standard for calibration even though an international standard may exist -- perhaps because of cost, a history of doing it differently before the international standard came out, etc.

This would be one of those cases to me where you should do what is best for your company, not something different just to satisfy the standard. But, maybe you need the ISO cert. To avoid being in a pissing contest with an auditor over this issue how would you folks suggest this issue be handled? A letter of agreement with the customer saying that such-and-such is the agreed-upon way to calibrate machine "X" regardless of the availability of an international standard??? Other ideas???

Also, is the burden of proof on the company being audited to prove that a national or international standard does not exist for a given measurement? Sometimes it is hard to prove something does not exist.

Section 5.4.1 of ISO 17025 states "The laboratory shall use appropriate methods and procedures."

Section 5.4.2 of ISO 17025 states "Use appropriate calibration methods, preferably international, regional or national standards"

Further down it states "Laboratory methods or methods adopted may be used if approprite for the intended use and validated. Client informed of method selected."


Ya gotta love these guys. Nothing open to interpretation here:D


Also, is the burden of proof on the company being audited to prove that a national or international standard does not exist for a given measurement? Sometimes it is hard to prove something does not exist.

Just for S&G's, let the auditor prove otherwise. As long as the method you choose or write is appropriate for you and the client, the auditor should have no concern.


Now, ISO 9K2K might be another story.

Ken,

That was the jist of my question -- in reference to 9K2K. 17025 seems clearer to me than 9K2K. "Appropriate" and "preferably" are two good weasel words -- not to weasel out of doing a good job but to give you room to operate and use non-international calibration standards/methods when that works for you and your customers. IMO ISO should not prescribe what standards you use.

At this point, some clarification of terms is in order. ISO considers the word "method" to be very different from "procedure" or "standard". When ISO 17025 states "laboratory methods" versus "nationally or internationally recognized methods", what it means to say is how the data is derived.

For example:
I have to calibrate a micrometer, but I have no procedure. I write a procedure that uses gauge blocks for spindle linearity, and optical parallels with a monochromatic light for anvil flatness and parallelism. You use tolerances that were prescribed by the manufacturer of the micrometer on the specification sheet.

From this I can conclude:
You HAVE used nationally or internationally accepted methods, namely gauge blocks for length and optical parallels for flatness and parallelism. You have not invented a new METHOD, you just wrote your own procedure using standard methods. If you used a doppler laser to gauge spindle travel, your method would be non-standard, and you would have to validate it, get customer approval, etc.

I don't know how well this post will translate for our non-English friends, but this is how the English versions of ISO standards are meant to be interpreted.

Ryan