We manufacture RF filters for cell phone base stations and use calibrated torque drivers for the mechanical assembly phase of our process. We calibrate our own electric torque tools. Occasionally, we will find a tool out of tolerance when we do a routine calibration check (usually every 3 months). Of course, that means product built with this tool from the last cal may be suspect. I would like to hear what practical methods or procedures others are using to control such a situation. Thanks

We had a torque transducer in the manufacturing area, and torque wrenches were required to be set against this daily, or at start of new job.

Change the calibration interval from 12 weeks to 10 weeks.

OK here goes my first post!

I am new here and relatively new to metrology but here goes.

The first thing I would do is figure out if this is a calibration issue or an equipment issue.

What type of gages are you using? If they are those 36 in/lbs click screwdrivers then good luck. Some equipment is manufactured to an accuracy spec of +/- 6% as per ISO 6789. I have found that some manufacturers are not accredited to ISO 17025 and that they do not have any measurement system analysis programs in place to determine what their uncertainty is. That could mean with uncertainty your gages aren't capable of the process you are trying to measure.

Considering the cost of recall I would investigate getting better equipment if the equipment is suspect.

If it is a calibration issue how do you set the calibration intervals? If this gage is drifting out of spec how long does it take? What is the manufacturers specs for drift over time? You may just have to reduce the calibration intervals. Also look at your process on line. Is the operator using the tool as a hammer? (don't laugh I had this happen before)

Hope this helps. Please post back with your results.

Well, this is my first post too.

I am wondering if Mr. Chip is looking for methods used to trace back and determine which tests may, or may not be required to be repeated? This is my question too, and what has brought me to this site.

Regardless of the calibration interval, there will be a time when an instrument is found to be out of tolerance. It can be very difficult to go back and figure out where it was used, and what was tested. A lot depends on the piece of equipment, the type of testing, and how well information is recorded and/or entered into a database, which is the delimma I am faced with now.

I may have to start a separate thread on this discussion alone, unless this is indeed what Mr. Chip is looking for. Then I will just follow this thread.

Thank you,
D_Wood

How about using something like a Stability Test (AIAG - MSA) to monitor the condition of the equipment on a regular basis and using a Control Chart to detect trends/ deterioration?

Regardless of the calibration interval, there will be a time when an instrument is found to be out of tolerance. It can be very difficult to go back and figure out where it was used, and what was tested. A lot depends on the piece of equipment, the type of testing, and how well information is recorded and/or entered into a database, which is the delimma I am faced with now.

At our facility we maintain strict documentation regarding the history of each lot of product. The calibrated equipment used in the manufacture and inspection of the lot is included in the documentation. When a calibration fails, we can go back to the documetation and identify which lots would be suspect.

This documentation also helps the users to be aware of the calibration status of the instruments they use since they are writing the information down on a regular basis.

I apologize that I have been on vacation doing volunteer work in Haiti for 17 days and just got back to the office. I usually tell Marc when I'll be gone, but I believe I forgot this time...

In our company we have a process called "Material Review Board". Any process "glitch that could have compromised product integrity (quality, meet specs, etc..) is reviewed by engineers and they make a call (and document it) as to any non-conforming situation did or did not affect product integrity.

In your case of out of tolerance torque wrenches, we would send a report of the out of tolerance conditions to the material review board and let them decide if this OOT did or did not have impact.

I would not recommend a blanket interval change. I would recommend (if you are not already doing so) making sure you have a target %IN-TOLERANCE (% confidence any given reading is in tolerance), and adjust calibration intervals in accordance with that.

If you are recording data readings on your torque wrenches, you could do some analysis on the data and apply some statistics to them to determine what kind of confidence you have in readings, and perhaps even be able to give a probability of at what point during the previous calibration interval (to some level of confidence) that a given torque wrench went OOT (rather than having to trace back to the beginning of the calibration interval.

Just a few very belated thoughts.