ISO 9001:2008 Sec 7.6 - Can we just log measurement devices as 'Pass/Fail'?

[BradM]

That is not a good analogy, but since you brought it up, sometimes "day time"/"night time" is enough.

Oh... It's a perfect example. That is what a true calibration program is, Jim. It's more than giving a thumbs up; it's managing the instruments used ( including remedial action with a failure). It would seem like a nightmare managing calibrations with 0's and1's.

And you know for a surity that four angels can fit on the head of a needle? Why not thousands? Why not one would cover many needles? This falls into the realm of the unknowable.

I agree Jim. Instead of encouraging the OP to develop a robust measurement system, and/ or encouraging him/her to engage the auditor more about the finding; we've invested our time arguing amongst ourselves about trivial matters. Church leaders arguing about angels, instead of feeding the poor.

I would agree that it is a good practice to write down the readings, after all, they had to be at least mentally noted in the process of determining pass/fail. But that isn't the question here. The question is really very simple. Is it a requirement. No matter how much you would like it to be, it simply is not.

I have no arguement that it is a good thing. It just isn't a requirement.

And I do think its a requirement. But we can agree to disagree.

 

[BradM]

Re: ISO 9001:2008 Sec 7.6 - Can we just log the devices as 'Pass/Fail'?

What more would you know if you did have the actual measurements? NOTHING. You know it failed. You know when you found it to be out of specification. You know the last time it was checked. So everything measured in the interval from the last time it passed until it failed becomes suspect. How is that changed because you recorded the actual measurements? It hasn't.

Ah... Now were getting somewhere.

The degree of error has everything to do with the investigation.

say my tolerance is +/-1.0 psi and the error is +1.2psi. While out of tolerance, it is borderline insignificant and I can limit my investigation ( with appropriate accuracy ratios).

Same thing with a standard. I can compare the magnitude of error to the tolerance of instruments calibrated. If I made no adjustments, and I have an acceptable ratio, I can conclude the error was insignificant.

And my decision tree continues based on the magnitude of error.

In my investigation, I may determine there was drift in the instrument and the interval should be shortened.

But if I have no data and I have to tell my customer there was a failure but I have no idea how big; my whole quality system is worthless to the customer.

Thee is much value in that data.

 

[Jen Kirley]

Re: ISO 9001:2008 Sec 7.6 - Can we just log the devices as 'Pass/Fail'?

I never mentioned the 10:1 rule. You have brought it up as a defense, and it just doesn't fit the scenerio.

Actually, the 10:1 rule isn't always appropriate. From earlier posts here from those that mainly deal with 17025, sometimes (and if I remember correctly mor often than not) it is 4:1.

The real point is that the 10:1 (or 4:1) rule has no bearing on this topic.

You have moved into requirement creep, imagining that there is more to it that what is in the written word.

How long are you going to try to keep this spinning?

My reference was meant to show one of many examples in which we enforce good practices that don't get included in the standard. It's less valid a point though, than the simple fact that a calibration check performed on a dimensional instrument results in a numerical figure that, if we want to be purists, is what 7.6 asks to be recorded; not, I maintain, the subsequent decision. If you didn't find it useful I would understand. I would like to know, though, what out OP thinks about this discussion.

 

[BradM]

Re: ISO 9001:2008 Sec 7.6 - Can we just log the devices as 'Pass/Fail'?

I would like to know, though, what out OP thinks about this discussion.

Me too. We've had a good discussion, but in the end, we want to provide them with answers they were seeking, on both sides.

 

[Big Jim]

I'm sure he got scared away after the first few posts. It was fully answered by the third post.

 

[Big Jim]

Re: ISO 9001:2008 Sec 7.6 - Can we just log the devices as 'Pass/Fail'?

Ah... Now were getting somewhere.

The degree of error has everything to do with the investigation.

say my tolerance is +/-1.0 psi and the error is +1.2psi. While out of tolerance, it is borderline insignificant and I can limit my investigation ( with appropriate accuracy ratios).

Same thing with a standard. I can compare the magnitude of error to the tolerance of instruments calibrated. If I made no adjustments, and I have an acceptable ratio, I can conclude the error was insignificant.

And my decision tree continues based on the magnitude of error.

In my investigation, I may determine there was drift in the instrument and the interval should be shortened.

But if I have no data and I have to tell my customer there was a failure but I have no idea how big; my whole quality system is worthless to the customer.

Thee is much value in that data.

If it is out of spec it is out of spec. It is not a matter of degree any more than being slightly pregnant.

 

[somashekar]

I'm sure he got scared away after the first few posts. It was fully answered by the third post.

I don't believe it is required. Show me the shall.

This is post # 3 or the third post as one can take it.
If QMS was about seeing the 'shall' then let alone calibration, the whole of the QMS can be let to the winds and one can hold few 'shall' documented procedures and make a kite of the ISO certificate ...

 

[AndyN]

Re: ISO 9001:2008 Sec 7.6 - Can we just log the devices as 'Pass/Fail'?

If it is out of spec it is out of spec. It is not a matter of degree any more than being slightly pregnant.

An overly simplistic analogy and incorrect, IMHO. You need to know how much out of specification the result is, for comparison to the product/process characteristic and any applicable tolerance being measured - that will help to dictate what course of action is taken. Indeed, when implementing a calibration system my clients have installed an activity where the out-of-tolerance condition is compared to the specification - since we may be using the 10:1 rule (or even higher) and have recorded the decision NOT to do anything, since the error was adjudged not to have been significant.

 

[Jen Kirley]

Re: ISO 9001:2008 Sec 7.6 - Can we just log the devices as 'Pass/Fail'?

There is no universally applicable 10:1 "rule" or 4:1 for that matter either. It might become a rule if I make it one, for my system, but otherwise it's only a guideline.

I agree, that must be why people sometimes ask us here if micrometers can be used to calibrate gauge pins.

You're being very selective in your definitions. How about, as a definition of "result," A consequence, effect or outcome of something? Is "Pass" not an outcome?
This last makes no sense, I'm afraid. There is nothing anywhere (except in the tortured interpretations of some in this thread) that says that numeric values must be recorded.

I'm being precise, which I have seen you be elsewhere but here you've chosen to apply a qualitative definition to a quantitative term. The result of a micrometer's calibration is mathematical; what one chooses to do with it is an outcome, consequence or effect.

If we wanted to be vague, which isn't a virtue in auditing, we could assign all sorts of definitions to a term. I see the same thing when people list # of audits, # of NCs and # of OFIs as audit results in management review. But I digress.

Like I said, if our OP wants to contest the writeup - if all we're after is simple compliance to the shalls - he can probably get it reversed.

Do you suppose this sort of thing is one reason why some people claim a company could make cement life preservers and get ISO 9001 registered?

 

[Jim Wynne]

Re: ISO 9001:2008 Sec 7.6 - Can we just log the devices as 'Pass/Fail'?

I agree, that must be why people sometimes ask us here if micrometers can be used to calibrate gauge pins.

Micrometers can be used to calibrate pin gages if the tolerance of the things being measured with the pins is such that the method makes sense. That's what's most important about determining tolerances for devices.

I'm being precise, which I have seen you be elsewhere but here you've chosen to apply a qualitative definition to a quantitative term.

No---in fact the opposite is true--you've chosen to use the quantitative without demonstrating that the qualitative term is inadequate. When you calibrate a device, the result is not expressed in numbers. You're confusing data with results. The result, even when data is recorded, is either pass or fail, or accept/reject. If you've just calibrated a device, and someone asks what the result was, are you going to answer with numbers? I don't think so. You'll answer that the thing was accepted or rejected.

The result of a micrometer's calibration is mathematical; what one chooses to do with it is an outcome, consequence or effect.

Again, the result is not the data. The data is what allows one to assign a result to the calibration. The result of calibration of a micrometer, or anything else, is accept or reject.

If we wanted to be vague, which isn't a virtue in auditing, we could assign all sorts of definitions to a term. I see the same thing when people list # of audits, # of NCs and # of OFIs as audit results in management review. But I digress.

Calibration isn't management review, but since you brought it up, if management wanted to know something about the calibration system, one of the things they would likely want to know is the number of devices accepted or rejected after calibration.

Like I said, if our OP wants to contest the writeup - if all we're after is simple compliance to the shalls - he can probably get it reversed.

It's not about "simple compliance to the shalls." It's about effectiveness, and effectiveness of a process can be judged only by its success in achieving its objectives.

Let's try a hypothetical case: There are two organizations and two calibration processes. Company #1 has voluminous measurement data to support its contention that its calibration process is effective. Company #2 has recorded only pass/fail status. Using your criteria, company #1's process is effective but company #2's isn't. But, company #1's data has been fudged in its entirety. Not only are the measurement numbers bogus, but no calibration has been performed at all. At company #2, calibration always proceeds in accordance with a documented procedure for each device, a procedure that is consistent with effective calibration. In both cases, there is no evidence that any NC ever occurred due to a device being out of calibration. How is an auditor supposed to determine that either process is not effective?

Do you suppose this sort of thing is one reason why some people claim a company could make cement life preservers and get ISO 9001 registered?

Whether or not cement life preservers are acceptable or not depends on the specifications, no?