Recommendations on calibration interval (frequency) ISO/IEC 17025:2005 para 5.10.4.4

[joecalibratio]

I.A.W. ISO/IEC 17025:2005 section 5.10.4.4 "A calibration certificate (or label) shall not contain any recommendation on the calibration interval except where this has been agreed with the customer. This requirement may be superceded by legal regulations."

Does anyone know the reason behind this paragraph? I am under the understanding that the manufacturers recommendations are typically utilized for calibration intervals. This paragraph has some ambiguity with "agreed with the customer" and "superceded by legal regulations".

Does this mean we leave the Expiration date blank? I do not believe the majority of our customers would be receptive to this.

Joel

[Jerry Eldred]

Although I'm not an expert on this, I will give it a shot. The "Certificate of Calibration" is essentially a statement of as found/as released quantitative information about how the instrument is performing at a given time. Establishing the interval is a separate issue.

For example, I am currently working on statistical evaluation of about 86,000 calibrations in my company's calibration recall database. This is based on NCSLI Recommended Practice method 3 (if I recall corrrectly - sorry, it's late on Friday afternoon). The interval is a statistical probability to a given confidence of how long an instrument may be expected to remain within tolerance. It may be lengthened when prior as received calibration history allows it, or may be shortened when prior history dictates that way.

The given users accuracy requirements, frequency of use, environmental factors, the "lemon factor" (if you have one bad one in an otherwise "good" population of instruments, it has a "lemon factor") and other factors can impact what the interval should be set to.

This is distinct from the Certificate of Calibration. Particularly if you send to the OEM or Third Party lab, since they don't have your internal data, it isn't fair for them to assign an interval. This may potentially create an undue increased risk of out-of-tolerance performance, and possible conflicts.

For (fictitious) example, suppose I have 100 Fluke 77 multimeters (no good or bad implications about Fluke 77; it's just a very common instrument). I've (fictitiously) been calibrating them for the last 10 years with an initial assigned interval of 12 months. After 10 years of history, I find that 5 of the meters have a higher out-of-tolerance history than the others. I research and find that these are used by a crew outdoors. They are frequently dropped, exposed to higher voltages, lay in the truck in the heat, and so forth. I've had to adjust that one little potentiometer in these five a few times, where I haven't ever adjusted any other in my entire population of Fluke 77's.

My recall system and calibration records state the assigned interval (which I statistically derived based on history - still, a fictitious example). However, I sent two of the bad meters back to Fluke, where they calibrated them twice. They sent back a Certificate of Calibration with them, but did not provide a due date, since it isn't appropriate for them to include that in the certificate. When I received them back from Fluke, even though I increased the intervals on all the remaining population to 30 months, these five are at 12 months, and 2 of them are at 6 months.

Hopefully this will shed some amount of light on the question. Have a good weekend. I'm going home.

[Marc]

Thanks for the detailed reply. Excellent post. Excellent example. Thanks, Jerry!

[Hershal]

Originally Posted by joecalibratio

I.A.W. ISO/IEC 17025:2005 section 5.10.4.4 "A calibration certificate (or label) shall not contain any recommendation on the calibration interval except where this has been agreed with the customer. This requirement may be superceded by legal regulations."

Does anyone know the reason behind this paragraph? I am under the understanding that the manufacturers recommendations are typically utilized for calibration intervals. This paragraph has some ambiguity with "agreed with the customer" and "superceded by legal regulations".

Does this mean we leave the Expiration date blank? I do not believe the majority of our customers would be receptive to this.

Joel

Actually, it does mean leave the due date blank, UNLESS there is an agreed-upon calibration cycle.

Why the clause? Users may simply use an annual cycle or may want to change the cycle based on some criteria or lack of criteria. So, this clause is for the protection of both the customer AND the cal lab. The cal lab is not suggesting or requiring some arbitrary date to generate business, and the customer is free to select their own date.

Now, most customers will ask their cal house for a recommendation. That's OK and desired.....I suggest it gets documented.

Hope this helps.

Hershal

[Hershal]

One last post.....Jerry, RP-1.

[Wayne]

The customer is the only one who can accurately specify the calibration cycle. The calibration lab is not in any way qualified to suggest the calibration cycle.

My experience is with hard gages. I frequently advise my customers that time is not the best unit of measure to use when determining calibration increment. When using hard gages, they have a tendency to wear or gaul with use. The operative variable here is ‘with use.’ With use brings its own variables to the table.

How many uses?
A hard gage that is used once a year, needs to have a greater interval than a gage used a thousand times a day.

Against what material is it used?
Some materials are more abrasive than others. Very soft materials being manufactured with dirty cutting fluid can pick-up worn particles of carbide suspended in the cutting fluid. These particles then become imbedded in the product and act as a lap against the gage causing rapid gage wear. Some materials are more gummy. They leave material welded to the gage surface thus causing the gage to grow.

How close to the mean is the product being checked?
If the product being measured is made close to the maximum material condition, the GO gage will have much more gage to product contact. With more contact comes more wear. If made closer to the minimum material condition the NOGO gage will be used more and thus experience more wear.

What type and amount of lubrication is used?
Cutting fluid is not lubricant. Lubricant does not effect the measurement properties of the gage. Keep the gage well lubricated and you will extend the life of the gage. Conversely, use no lubricant and, well, you know.

What are the work station conditions?
Kept clean. Gage not laying in a pile of chips. Machined product chips removed from the work piece before applying the gage. Chips between the gage and the work piece cause gage wear. Cutting tools kept sharp. Burrs left on the work piece get between the work piece and the gage cause wear.

Who is using the gage?
Popeye on spinach or Olive Oil? If more muscle is used the more abrasion can occur. More abrasion means wear.

Hard gages should be checked when first received from the supplier. If not used they should be checked again at one year to check for gage growth. After that the gage can be considered stable for long term storage. Calibration cycle should start at first use after calibration. Calibration cycle should be based on number of uses the gage sees. Using the number of uses provides a more accurate measure than time. Consideration should be given to the above variables when determining the specified calibration cycle. This is a good use for statistical evaluation.

In the final analysis, time is the almost the worst measure to use to determining calibration cycle for hard gages, but time is usually what is recommended by calibration labs.