Hello forum

Sorry for the poorly descriptive title, I don't have a copy of the standard with me today but I'll try to ask as clearly as possible...

We are in the process of creating in-house calibration procedures. We really like the idea of taking calibration measurements on equipment as found. However, ISO 9001:2000 7.6 states that we shall assess the effects of any measurement equipment found to be out of tolerance.

We do not track what measurement tools are used on what products, so how is this accomplished?

Hello forum

Sorry for the poorly descriptive title, I don't have a copy of the standard with me today but I'll try to ask as clearly as possible...

We are in the process of creating in-house calibration procedures. We really like the idea of taking calibration measurements on equipment as found. However, ISO 9001:2000 7.6 states that we shall assess the effects of any measurement equipment found to be out of tolerance.

We do not track what measurement tools are used on what products, so how is this accomplished?To accomplish this you need to know where the equipment is used. If you don't identify what equipment was used you can't go back and make the assessment. Identify in your calibration records what department is using the equipment and then on the inspection records identify the equipment used to perform the inspection or test.

We do know where the equipment is used at a department level, sometimes even operator/machine.

This is difficult. We were shooting some ideas around about it yesterday, but only came up with being able to re-verify product that is currently in process. We don't track what was made with what.

So let's say we're in a company that can track all the products verified with certain instrument, and the instrument comes back with a minor error after a year calibration interval. Does that company have to assess the effects upon every single product verified with that instrument piece by piece or can they say the error was small enough to forgive? Can they assess the effect of the error upon, say, machined product, create a record of this and satisfy the requirement?

I should note, the concern is with calipers, micrometers, depth gages, drop indicators, test indicators and the like. Tolerances are generally larger than .004 total. Anyone verifying tolerances smaller than .002 total knows just what they are doing.

IMO, it is a question of assessment of risk, how much risk your organization is willing to accept, and what the scope of your containment effort will be should you find a piece of equipment to be out of tolerance.

If you do not record where equipment is being used, you could have a very large (costly) recall effort on your hands should you discover an OOT piece of M&TE. Worst case, everything you have produced since the last positive verification could be subject to recall.

Thanks BadgerMan

I will review the requirement this morning and post a strategy this evening to see if we've got the right idea

--Erik

Hello again

Well, after looking up the requirement in Hoyle, I can say that this is definitely a difficult requirement to meet. This is going to be a major PITA actually.

In large organizations that own many pieces of equipment that are constantly being used in a variety of situations, meeting the requirements can be very difficult.

While I'm certain we will have an adequate response to this requirement, the response will not be elegant or efficient in any respect. It could be accomplished through corrective action but it will take a personal investigation to create this assessment and record. It will not be absolute or complete either.

This is a major disappointment to me. Let me present the situation and hopefully someone will have a more enlightened outlook.

It took a great deal of encouragement to get this company to document calibration procedures. We have made some major changes to the calibration regimen. We've made some major improvements, and I feel like those improvements will now be a liability in light of ISO 9001:2000; 7.6.

First, calibration procedures are being drafted that are rigorous and demanding. Micrometer acceptance was based on +/- .0003 across the board. This resulted in micrometers not being adjusted, because they were not out of tolerance. Now micrometer acceptance is based on GGG-C-105C which specifies +/- .0001 for 0-1" and +/- .0002 for 3-4" mics!

Second, calibration intervals have been extended from an average of 3 months to more like 9.

We will have more equipment found out of tolerance with the new, rigorous procedures. After a 9 or 12 month interval, there will be dozens of tons of product in question.

How are some people out there handling this? I have read that many professional calibration managers expect that at least 95% of tools calibrated will have remained in tolerance, and that they adjust their calibration intervals to achieve this number.

Does everyone else have computerized data collection and an absolute recall system or am I missing the mark?

Hello again

Well, after looking up the requirement in Hoyle, I can say that this is definitely a difficult requirement to meet. This is going to be a major PITA actually.

We will have more equipment found out of tolerance with the new, rigorous procedures. After a 9 or 12 month interval, there will be dozens of tons of product in question.

How are some people out there handling this? I have read that many professional calibration managers expect that at least 95% of tools calibrated will have remained in tolerance, and that they adjust their calibration intervals to achieve this number.

Does everyone else have computerized data collection and an absolute recall system or am I missing the mark?

It only says "assess" That's all it says.

Hopefully most mikes are doing in process checks. Where you need to worry is the gage that does the final acceptance. If that turns up wrong, the customer may have bad product.

So the first thing to find out, was there another check downstream? Was it OK? Document this brief investigation. You have met the requirement to assess. Hopefully 80% of the problems go away.

So, now you need to look at gages for final. Is a 1 year cal cycle reasonable? Or should they be mastered daily at the other extreme? It is a very unhappy day when the final gage proves to be out of cal. Almost always better to avoid this.

Another idea - you have been in business a long time. Any customer problems due to out of spec gages? If not, all that is gained here will be cost. Look at it again - ISO should not make us do things that make no business sense. If the old looser tolerance made customers happy - why adopt a tighter one? Fight it - you have the data!

Another trap depends on the risk tolerance (or risk ignorance) of the person doing the assessing. Some people are willing to accept anything. I've heard some "amazing" arguments as to why things are really OK if only I could take off my Quality hat and put on my Manager hat.

Good luck.

This might be partially :topic: but the need for recall is, in my experience, exceedingly rare. In fact, in all of my experience I can think of only a few instances where calibration was in issue in customer acceptance of a questionable product, and those involve measurement of gold plating thickness (measured in millionths of inches) by x-ray fluorescence and the discussions were over whether there was .0000024" or .0000025" inches of gold present.

When it comes to shop floor measurement using calipers, micrometers and indicators, there is almost always some form of redundancy at work that catches errors before they leave the building. If the spec says x +/- .0005, and the operator is reading something near the limit, someone else will look and make sure that the measurements are accurate, and that there is agreement between devices in order to eliminate gage error as a possible cause.

The question of positive recall involves making sure that there is some reasonable level of lot control, and that when anyone makes a critical recorded measurement, there's also traceability to the device used. Not a difficult thing to do, because the calibration system already requires that devices be uniquely identifiable; it's a matter of recording the device number on the checksheet or inspection report.

Also, it should be remembered that a positive recall system doesn't require recall when an out-of-calibration device is discovered. It requires that there's a system in place for review of the circumstances in order to determine whether recall (or customer notification) is necessary or not. Just because a gage is out of calibration doesn't necessarily mean that measurements made with it weren't accurate enough in a given application.

GGG-C-105-C is difficult to attain with some micrometers. I'm currently calibrating my 0-4" set of Brown & Sharpe economy mics and some of them eat the whole tolerance. They're practically brand new (economy Mitutoyo micrometer sets blow these out of the water).

However, 7.6 of ISO 9001 requires effects assessment of instruments found out of tolerance. My hope is that it will be easier to assess the effects when the micrometers are finely tuned and are only out by .0002", rather than .0004". More instruments will be found out of tolerance, but the effects will be reduced.

From a logistical standpoint, recording the instrument used for inspections would really add work and not much capability. Once it is recorded on the inspection sheet, that sheet goes into a filing cabinet. Might be OK if it was an employee owned micrometer, because you could probably track down all the jobs that employee worked on. The trail goes cold on a thread gage, borrowed tools, etc.

It seems that the only solution for positive recall is electronic data collection, is that correct?

Once the calibration system has gotten itself under control, the next big push is going to be for databased specifications and electronic data collection. 4 QA's for 100 people... we gotta keep up!