Exempt from Calibrating our Tape Measures and Micrometers

[The Specialist]

This is a very valid point. However, this is step 2.

Step 1 is determining if you even care if the instrument is accurate or not.

Again, reasonable minds can differ on these issues. I take the more extreme approach. Too much money is wasted on equipment many times. If you don't need it and it doesn't matter, get rid of it. What's left matters, and should have some verification made in it.

You are right, of course.

I made an assumption, based on the OP, (as a specification of accuracy requirement is mentioned) that “step 1” has already been done!

 

[BradM]

I agree that you will not routinely ‘calibrate’ a steel rule. However, you will need to purchase one that has been tested against a reference standard and can be determined to be ‘accurate' from purchase.

The term ‘calibrated’ is a bit loose in this instance.

If is it needed to be verified. For example, if I am measuring boxes to send off to Fed Ex, who cares? However, if it's required to assure the 1/8 inch mark is truly 1/8 inch, then yes, a verification is needed.

And yes, the term calibrated is loose. Not sure if you have ventured into that jungle of a thread Jennifer mentioned on calibrating/verifying tape measures, but's a good thread talking about this very subject.

 

[BradM]

You are right, of course.

I made an assumption, based on the OP, (as a specification of accuracy requirement is mentioned) that “step 1” has already been done!

I think we're probably in more agreement here that appears. And knowing Jim, this is more what he is pushing at.

He's been here a long time, and we've seen many posts (and many realistic situations) where people get things calibrated simply because of the interpretation of the clause, and little to do with assessing the need/purpose/process, etc.

Certainly I may be wrong, but I think that's what he's getting at.

 

[Jim Wynne]

How do you suggest that you determine the accuracy of a measurement taken from a measurement device that has not been calibrated and the accuracy of the device is unknown?

Please read again: where necessary to ensure valid results...

I suggest that the accuracy is determined by the results, after reasonable deliberation by an experienced person who has good reason to believe, given a relatively large amount of product tolerance, that the method in question is adequate. The necessary bit is confirming the hypothesis by way of objective evidence.

When there are instances where a given measurement indicates that the product size at or approaching a tolerance limit, a calibrated device should be used to verify the findings. In that case, calibration is needed to ensure valid results and a rational decision may be made as to how to proceed.

The OP is talking about using a micrometer to measure something where the tolerance is ± 1/8". Not only that, but he's also talking about using linear scales in the same application, and those (a) can't be adjusted and (b) don't change over time, except changes that can be dealt with by throwing one out and getting a new one.

Let's also not forget about the utility of MSA in the grand scheme of things. GR&R can be used to understand how (in this case) measurement error contributes to process variation, and the results can further inform the need or lack thereof for periodic calibration.

 

[vanputten]

I agree with Brad M. Why would you want to get away with not knowing the measurement results are valid? And I ask this because you are stating that you perform measurements. Or you could study the variation in whatever you are measuring and conclude measurement and monitoring is not even needed.

7.6 states that measuring equipment shall be calibrated or verified... You could try a method of verifying the equipment instead of calibrating them.

At some point in the process I would think you would want to know if the supplied raw materials meet requirements.

 

[The Specialist]

Such demonstrates how tricky it can be to give good advice here. Notice how I went along, tippy tappy happy, and offered advice on gage blocks without knowing for sure that poprock7 needs them. I still don't know if they are needed. Did I mislead the OP? I don't know. I hope not.

To all:

And so it goes: in a place like this, advice is offered without specifc guarantee as to its accuracy. Our diversity is our strength, but that sometimes sends mixed signals. We can add, subtract and divide; we can correct ourselves and each other as a group, and we should. But let us do so gently please.

Well put!

I think (reading between the lines) that most of us actually agree on this issue.

I made the mistake of making certain assumptions, based on loose statements, about the OP and the advice the OP requests.

Certainly, I am new to this forum (and not been a member of any other) and it would be wise for me to take heed of some of the poster's remarks (yourself, Jim Wynne and BradM).

It is easier than I realised to construct your own interpretation of the OP.

Thanks for pointing this out Jennifer.

 

[The Specialist]

Please read again: where necessary to ensure valid results...

Jim,

At this point I think it is important that you understand that I agree with you on this!

I thought that my "I agree with Brad and Ron" statement encompassed my view that a review of the requirement (of calibration) was necessary, as this had been highlighted in their posts.

I gleaned (perhaps wrongly) from the OP that this requirement had indeed been established by way of reference to a specification. Again, it does appear that I was incorrect in my assumption.

I hope you will read through my other responses on this thread, as I feel that we ‘over-all’ agree with one-another.

Please accept my apology for ‘jumping in’ on this occasion. As I stated in my last post; I am new around here and have a lot to learn about how to give constructive and informative advice in posts.

I will endeavour to do this, as I am enjoying participating in this forum and have the best interests of all members at heart.

Many thanks.
Justin.

 

[Jim Wynne]

Jim,

At this point I think it is important that you understand that I agree with you on this!

I thought that my "I agree with Brad and Ron" statement encompassed my view that a review of the requirement (of calibration) was necessary, as this had been highlighted in their posts.

I gleaned (perhaps wrongly) from the OP that this requirement had indeed been established by way of reference to a specification. Again, it does appear that I was incorrect in my assumption.

I hope you will read through my other responses on this thread, as I feel that we ‘over-all’ agree with one-another.

Please accept my apology for ‘jumping in’ on this occasion. As I stated in my last post; I am new around here and have a lot to learn about how to give constructive and informative advice in posts.

I will endeavour to do this, as I am enjoying participating in this forum and have the best interests of all members at heart.

Many thanks.
Justin.

Please don't apologize for anything; you didn't do anything wrong. When there are divergent points of view among experienced people, it doesn't necessarily mean one person is right and the other one is wrong; sometimes ambiguity in the standard is the culprit, and in other cases there are just different approaches. Hardly a day goes by when I don't learn something here, and everyone who takes the time to register and post contributes to that.

 

[bobdoering]

Please read again: where necessary to ensure valid results...

And this is where the fun begins. You emphasize results, I may emphasize ensure valid. The audit question is always "How do you know?" If "it is intuitively obvious" works for you and your customer, then you may be OK (as in the FedEx box scenario or in a TCE quality system).

Let's also not forget about the utility of MSA in the grand scheme of things. GR&R can be used to understand how (in this case) measurement error contributes to process variation, and the results can further inform the need or lack thereof for periodic calibration.

The best way to state is it that GR&R and calibration (bias) can be used to understand how measurement error adds to process variation (along with other variations) to generate the observed variation or total variation. Because of gage, measurement, sampling and other variances, you never really observe the process variation from the collected data. The goal is to understand and either statistically eliminate (by making statistically insignificant) or properly accommodate for the variances.

 

[Jim Wynne]

And this is where the fun begins. You emphasize results, I may emphasize ensure valid. The audit question is always "How do you know?" If "it is intuitively obvious" works for you and your customer, then you may be OK (as in the FedEx box scenario or in a TCE quality system).

At this moment, all over the world, people are building things and achieving valid results using measuring devices that have never been calibrated, except perhaps a the factory where they were made. I emphasized results because where calibration is concerned, in the end it's the results that matter. If I can reasonably demonstrate that formal calibration control won't add value to anything in a given instance, that should be enough, and we shouldn't be saddled with the pedantic notion that if the results matter, calibration should be done. In some cases that's an illogical construction, and I do my best to avoid those.