Gage Family Calibration - Setting Calibration Frequencies by Gage Family

[Matt Swartwood]

We are one of five satellite facilities within a corporation who have traditionally outsourced calibration. Due to the exponential growth of the corporation in the last couple of years, outsourced calibration has become no longer feasible. Corporate Management has assigned our plant to be the primary source of calibration for all of the facilities. We have begun making changes procedurally to address the majority of calibration being performed "in-house" and are progressing fairly well.

Our facility uses large families of working gages on a daily basis (i.e.: a 780 piece pin gage set), which may be included in our in-house calibration schedule. Because this would become a monumental and labor-intensive task for us to take on, I began digging deeper in my research. I found a clause in the ISO/DIS 17025 standard which states:

“The laboratory shall have a sampling plan and procedures for sampling when it carries out sampling of substances, matrices, materials or products for subsequent testing or calibration. The sampling plan as well as the sampling procedure shall be available at the location where sampling is undertaken.

Note 1: Sampling is a defined procedure whereby a part of substance, matrix, material, or product is taken to provide for testing or calibration a representative sample of the whole. Sampling can also be required by the appropriate specification for which the substance, matrix, material, or product is to be tested or calibrated.

Note 2: A sampling plan should describe the allocation, withdrawal, and preparation of a sample or samples from a substance, matrix, material, or product to yield the required information.”

Will this clause allow us to set each family of gages on a sampling plan based upon criteria such as frequency of use (providing they are not used as standards) to help alleviate some of the workload? Will a procedural reference to this clause be deemed satisfactory by a QS-9000 auditor? If I have misinterpreted the standard and there are no other avenues we could explore, we may opt to continue the use of an outside source for gauge families--considering we will be responsible for the calibration of several thousand gages.

 

[Jerry Eldred]

I am not a QS9000 expert, so PLEASE, understand I am only replying from a Metrology point of view. My company is QS9000 registered, and I am trained as an ISO lead auditor, but I don't consider myself an expert in the finer points of QS9000.

I will make an observation, though. My inclination is that in order maintain the pin gages under an SPC sampling plan, you will most certainly need enough data to consider your sample significant. In order to do that, a consideration must be that you are using many pin gages in identical application with essentially identical measurands and tolerances.

For example, if on a production line you have 100 0.010" pin gages used by 100 different operators using the same kind of pin gage for the same kind of test with the same product specification limits, I could see the potential to apply a sampling plan effectively. However, if there are a variety of applications with a mix-match smattering of numerous different pin gage sizes, and numerous different product specifications, I would be skeptical. There will be differences in wear and tear, and differences in how much the same amount of wear and tear matters if there are a variety of product tolerances. So the sampling plan may more than adequate in one application, and insufficient in another.

Another alternative method that intelligently addresses this would be a statistically based calibration interval adjustment program. This way they would all be adequately calibrated (i.e.: at appropriate intervals), and like the sampling plan methodology, keeps you from expending too much resources on the calibrations. In a way, sampling plan and statistically based calibration interval adjustment are parallel methodologies. In the sampling plan, you statistically derive a confidence level in the accuracy of your pin gauges (drawback being that you don't account for a variety of uses, and there are potentially escapes due to some units with potentially higher wear and tear being missed). In statistically based interval adjustment, every instrument is checked periodically. The interval between checks is based on the pin gauge's historical performance (or the performance of the set).

I reiterate, this is not meant as a QS9000 response, only a metrology based observation. Hope this is of some help.

------------------

 

[CarolX]

Matt,

A method I used for pin gages was to mark the set "calibrate before each use" and a procedure that required using a micrometer to verify the pin size before using the pin. The micrometer accuracy was to .0001, and part of the regular calibration system, so we had the necessary tracebility back to NIST.

Don't know if this may help, but good luck!

 

[MarkR]

Matt,
Here's another idea. Seal each pin so it is obvious which have been used (wax, wrap in tissue, etc.). Then calibrate only those that have been used since the last calibration. This probably requires an update to your calibration procedure.

Good luck.

 

[e.s.deo]

The question as appears is how to club gages as families!
Whether such clubbing is allowed? To what extent?
You may club similar gages performing similar jobs as
A family.This clubbing is acceptable unless disapproved
By your customer specific requirements.
However the references indicated from 17025 are for sampling
Of products and not gages for calib or msa.
However the family concept is only applicable for MSA & not
For calib.

Thanks

 

[Matt Swartwood]

Thank you for detailed replies. I have been benchmarking other companies who calibrate their own equipment and have decided to dip them in wax and calibrate only those which have the wax removed. Because our calibration schedule is based on useage, I will be using the signout log for a year or so to determine an adequate calibration scedule.

Thanks again.