Calibration Frequency - Control of Monitoring and Measuring Devices - Clause 7.6

[3lutz]

Hello to everyone, I'm implementing the ISO 9001:2000 but I do not know how to approach the clausule 7.6 Control of Monitoring and Measuring devices, I do understand the importance of this point, but how do I determine the frequency of calibration for measuring devices? or how do I elaborate a plan to meet the requirements in this point? Is there a standard that rules it?

The company which I work for produces gears and some mechanical parts for trucks, so we work with verniers, gages as well as other devices.


Thanks in advance and I love the forum, I think you have done a great job!

Karla

[fuzzy]

Welcome to the Cove. No hard and fast rules exist in the calibration frequency that I am aware of, but others may provide more solid information. For me, I would base my frequency around usage: gages with very heavy use might require calibration every six months to start, less frequently used gages might go a year, infrequently used devices 18-24 months. Start with a framework something like that, and adjust based upon results. Certain devices, in certain areas may not last the full cycle without "drifting" out of calibration. This may lead you to shorten the frequency, while great results may allow you to stretch the interval. These reactions (especially when out of standard "as found") may be more important than the chosen frequency. Remember the other requirements such as NIST traceability where possible, and a risk assessment when devices are found out of cal. Good luck!!!

[juliov]

Fuzzy is correct, I was about to suggest the same guidelines to you. Basically review the usage, look out for the the environment where the devices are used, then determine frequency, make a schedule, enter info.

[Grizz1345]

Hello and welcome,

I would suggest that if you have no history for your tools not to go longer than 12 months between calibrations. After extablishing some history then you can adjust your cycle times accordingly. I have found that encasing gage in a coating of plastice dip has saved my company a goodly amount of money and time. If the seal is not broken when it comes time to re-cal then I just roll it over due to lack of use (which is proven by the unbroken seal). Good luck and if you would like more info feel free to e-mail me.

[qcman]

We do all of our measuring devices and gages (bought and built) annually - unless of course we know it was mishandle.

[Hershal]

3lutz,

First, contact CENAM and see if they have guidelines. CENAM is the National Measurement Institute for Mexico. Also, your traceability for calibration is through CENAM.

If they do not have guidance, then check with your accredited calibration provider. If they have any personnel who are PROCERH certified then they should be able to help you determine the best calibration frequency, based on usage.

Hope this helps.

Hershal

[Louis Reimer]

I will add to the excellent advise, It is recommended that you identify the accuracy of your measuring devices and then identify the accuracy that is required of the devices that are being measured.

If you find that the accuracy of the measuring device is 001 and the accuracy of the devices are 010, 10:1 ratio you have some justifications for annual interval based on limited risk. If you find that the the ratio becomes less than 4:1, internal checks or more frequent calibration schedules should be considered since the risk of false positive reading increases as the ratios get smaller. Also using the manufacturer recommended interval until history is establish has been used by many other companies as well.

[potdar]

Among other good things my friends have mentioned, do not forget to define your acceptance limits for error in each MMD. The standards define very tight limits for acceptance and if you dont define, the calibrators are very happy to declare your instruments as "not fit for use" by referring to the standards.

Your acceptance limits should be based on what you do with the instrument rather than what the instrument is supposed to be capable of doing.

A very coarse example : One of my clients has afoundry and a machining unit in the same premises. Verniers "not fit for use" in the machine shop are marked with red and sent to the foundry where they are perfectly "fit for use". The acceptance error limits for verniers and "verniers marked in red" are different.