Hi everybody,
I have lot's of testing equipments and many of them were calibrated by external labs. And same of them were calibrated by internal.
But I need acceptance criteria for equitmens, what will the maximum tolerance for each caliber, meter, micrometer, thermometer, etc etc. Of course it depend on where the equipment using. But I have to have a systematic approach

Hello there!

I have lot's of testing equipment

I guess I'm wondering what a lot is to you. Are we talking about 50, 500, or 5000?

Of all these instruments, are many of them the same, and are most of them different?

many of them were calibrated by external labs. And same of them were calibrated by internal.
But I need acceptance criteria for equitmens, what will the maximum tolerance for each caliber, meter, micrometer, thermometer, etc etc. Of course it depend on where the equipment using. But I have to have a systematic approach

I am wondering how you have been having these instruments calibrated without having tolerances set in place. Did you have tolerances for the internal calibrations, or just adjust them when you found error?

I would think your external calibration certificates should have the information on it you are looking for. Did they not put tolerances, etc. on the certificates?

But I have to have a systematic approach

I like that statement-developing a systematic approach. Here are some general thoughts; don't know if they help any or not.

1. Have you started an excel spreadsheet/database etc. of the instruments that you have? If you don't, you need to start one ASAP. If you talk to your external vendor, they should have information for you on this.

2. Determine the capability of the equipment that you have. Most all the equipment you listed has mfg. specifications for them. That should be first.

3. Next, I would recommend determining what your requirements are. What do you need the equipment to do? Depending on how many instruments are in your system, you may be able to establish classifications for the instruments. Say, for thermometers, you develop four classifications of required accuracy.

4. Analyze the capability of your equipment with your need. Make sure your equipment can do what you need it to do. Too, if your process requires +/-5 PSIG and you're using a +/-.05 PSIG measuring device, you may can use that device somewhere else, or lower the required tolerance.

5. Be sure you are performing the internal calibrations appropriately. This is having the proper environment, the suitable standards, appropriate accuracy ratios, generating appropriate documentation, proper recall intervals, etc. I'm sure you have all that in place, but figured I would bring it up. If you are confident on this one, disregard.

These are some general ideas. There are some metrologists on this site that I respect greatly. Be sure and search the other threads on tolerance, setting up a lab, performing internal calibrations, and the like.

Feel free to repost with more questions.

Thanks for your fast reply,
Of course there are some tolerances on the calibration certificate that comes from the external lab. But for example if the deviation is 0,06 mm of height gauge for 600 mm, is it ok or not ok? if ok, how can it ok?

So is the external vendor just reporting deviation observed? Are they not reporting in tolerance/ out of tolerance to a specification?

Basically, if it is OK/not OK must be your call. It will depend on your requirement, and how it impacts your process. In the absence of nothing, this deviation should be compared to the mfg. specification. If you have process specifications, then compare it to that.

For example, you stated a .06mm deviation. If I am measuring a part of some kind and I am reporting to 1.0 mm, then .06mm is not that impacting. If I am concerned with +/.5mm, then the error is still not significant, but more significant than at 1 mm. Reporting at .1mm, then the .06 becomes something I need to look at.

Hopefully, that helps some.

Yes, external lab just reporting deviation.
And now, I 'm using the technique that you said. But auditor said we need a systematic approach. Because the old technique depends on me, a man.

How about we calculate the uncertainty at 95% for the random error only

t(0.025 , n-1) X σ difference / √n

Which n is the number of the sample test.:cool:

Yes, external lab just reporting deviation.
And now, I 'm using the technique that you said. But auditor said we need a systematic approach. Because the old technique depends on me, a man.

OK, now I am really interested, especially that an auditor is involved.

Exactly which technique are you referring to, determining what tolerance you need? Yew's helpful post is along this path.

Or keeping a list of instruments with due dates? Or the system of classifying what is in/out?

What industry requirement are you having to satisfy?

Are you confident with the lab you are using? What do they state as their credentials? I'm intrigued they are not listing tolerances on their certificates.

Thanks for the follow-ups.

Hi Again,
We had audited and I have minor NC about the acceptance criteria of measurement equipments. ( In fact, this can be a major, but not, thanks god :) And I found the solution. There are some standards for acceptance criteria of each equipment. Sorry for that, but I cant remember the standard number now. And I will write an instruction for that to classify the equipments for the tolerance.

10X rule for choosing the gage to use, so if a part has a ± .01 tolerance use a gage that can show .001 resolution, and be accurate to .001.

The gage manufactures provide the ± of the gage.

http://mitutoyo.com/TerminalMerchandisingGroup.aspx?group=1388

Notice on a caliper ±.001 is the accuracy, so as long as the gage calibrates within those set numbers it is to be considered acceptable.