MSA Studies on each Equipment, each Measurement System or Family in Control Plan?

[bobdoering]

Re: MSA studies: on each Equipment, each Measurement System or Family in Control Plan

I don't know how long the tools have been in use. I don't know their understood variation over time.

I agree with you points, except that these items refer to calibration and stability - not gage R&R. A valid gage r&r has no time function - but calibration and stability definitely do!

 

[qusys]

Re: MSA studies: on each Equipment, each Measurement System or Family in Control Plan

My organization just performed Gage R&R studies on 2 identical measuring tools. Both are under the "same" maintenance and calibration process. Both measure the exact same product. Both are used in the same way, in the same production environment.

One tool failed and the other passed.

Does this prove anything? No.

But consider if we had performed a Gage R&R on the tool that passed and applied the results to the "family" of 2 tools? What could the impact have been?

Or conversely, what if we had performed the Gage R&R on only the failed tool and applied the results to the family?

I am not sure if approaching all tools listed on a Control Plan or in a measurement system, as a family, is always the most effective approach.

Who would know? How would they know?

ISO TS does not mention statistical studies on each equipment on CP, so this could not be a requirement. Different thing if the requirement should come from the Customer , for example inside MSA studies in PPAP documentation. What about??

 

[bobdoering]

Re: MSA studies: on each Equipment, each Measurement System or Family in Control Plan

ISO TS does not mention statistical studies on each equipment on CP, so this could not be a requirement. Different thing if the requirement should come from the Customer , for example inside MSA studies in PPAP documentation. What about??

Yes, some customers have specific requirements. Ford appears to looking for Gage R&R for each device on the control plan - as well as capability and distribution of the characteristic. Others ask for Gage R&R annually - even though it is not a time function - calibration and stability are the time functions for gages. So...you have to keep your eyes open in TS - especially for customer specifics.

 

[vanputten]

Re: MSA studies: on each Equipment, each Measurement System or Family in Control Plan

Bobdoering: Please notice that the original posting referred to MSA Studies. The original posting was not about only Gage R&R. Therefore, I included the time parameter in my thoughts because of stability studies.

Also, I am going to guess (Note - a total guess with no empirical data) that 20 year old identical tools, even if under the "same" PM and calibration controls, will exhibit more variation between the tools than brand new tools. I am going to guess that PM processes and calibration process may not be effective enough to control for all tool variation over time. This can be debated to infinity. I am simply providing this thought for thought.

Maybe we need an operational definition from the original poster as to what they are really asking. Are they asking about doing only gage r&r assessments for tools or are they asking about studies of their measurement system?

 

[vanputten]

Re: MSA studies: on each Equipment, each Measurement System or Family in Control Plan

Qusys:

7.6.1 refers to "measurement systems referenced in the control plan."

What is a "measurement system?" The answer to this question may answer your original questions.

 

[Jim Wynne]

Re: MSA studies: on each Equipment, each Measurement System or Family in Control Plan

Bobdoering: Please notice that the original posting referred to MSA Studies. The original posting was not about only Gage R&R. Therefore, I included the time parameter in my thoughts because of stability studies.

MSA and calibration are separate categories. It's assumed that MSA is being performed on properly calibrated devices; thus the stability aspect is assumed.

Also, I am going to guess (Note - a total guess with no empirical data) that 20 year old identical tools, even if under the "same" PM and calibration controls, will exhibit more variation between the tools than brand new tools. I am going to guess that PM processes and calibration process may not be effective enough to control for all tool variation over time. This can be debated to infinity. I am simply providing this thought for thought.

What you're referring to is a problem with the calibration process, not the MSA process. If calibration intervals are set such that devices are found to be out of calibration at the end of the calibration period, someone needs to investigate. Note that calibration requirements apply to all measurement devices "when necessary to ensure valid results." The results of MSA are indicative of whether or not a give device (or type of device) is (when properly calibrated) appropriate to the task at hand. If the assumption is that measurement systems (device + operator + measured feature) in a given family might perform differently from one another in production, there's something wrong with the gage (addressed by calibration) or the operator (addressed by MSA).

Maybe we need an operational definition from the original poster as to what they are really asking. Are they asking about doing only gage r&r assessments for tools or are they asking about studies of their measurement system?

A GR&R is a study of a measurement system, although it's not the only type of study.

 

[bobdoering]

Re: MSA studies: on each Equipment, each Measurement System or Family in Control Plan

MSA and calibration are separate categories. It's assumed that MSA is being performed on properly calibrated devices; thus the stability aspect is assumed.

I have always interpreted (based on the AIAG book) MSA as the overall analysis of the measurement system, with the subcategories of: Location (or Calibration) Variation -Accuracy, Bias, Stability, Linearity; Width Variation - precision, gage R&R, Capability, Performance, Sensitivity, Consistency and Uniformity; and System Variation Capability, Performance and Uncertainty (AIAG MSA 4th ed page 5-8).

It is somewhat academic - as long as it is done properly and understood. But, one this is clear; MSA is not Gage R&R, and I hear that misconception a lot. Gage R&R is a subset of MSA.

 

[qusys]

Re: MSA studies: on each Equipment, each Measurement System or Family in Control Plan

Bobdoering: Please notice that the original posting referred to MSA Studies. The original posting was not about only Gage R&R. Therefore, I included the time parameter in my thoughts because of stability studies.

Also, I am going to guess (Note - a total guess with no empirical data) that 20 year old identical tools, even if under the "same" PM and calibration controls, will exhibit more variation between the tools than brand new tools. I am going to guess that PM processes and calibration process may not be effective enough to control for all tool variation over time. This can be debated to infinity. I am simply providing this thought for thought.

Maybe we need an operational definition from the original poster as to what they are really asking. Are they asking about doing only gage r&r assessments for tools or are they asking about studies of their measurement system?

First thanks for response to you, Jim and BoB.
My original question is related to how to accomplish ISO TS for clause 7.6.1.
In the control plan there are reported several equipment measurement released , they belong to same technology and subjected to the same calibration and maintenance plan.
On this measurement equipment production material is measured by operator and equipment are completely automated. The operator charges the materials on one of them , the robot selects some sampled materials ( according to a sample plan) and measures some caractheristics according to a program .
The question is : to make G R&R for all of them or I can select some samples of them? For example I have 15 equipment of a certain technology, instead of doing 15 G R&R I could do 3 according to some criteria, for another technology I have 20 equiment , instead of doing 20 G R&R , I could do 5 ( according to some criteria that we could establish).
What about?
Which could be the criteria?

 

[bobdoering]

Re: MSA studies: on each Equipment, each Measurement System or Family in Control Plan

First thanks for response to you, Jim and Bob.
My original question is related to how to accomplish ISO TS for clause 7.6.1.
In the control plan there are reported several equipment measurement released , they belong to same technology and subjected to the same calibration and maintenance plan.
On this measurement equipment production material is measured by operator and equipment are completely automated. The operator charges the materials on one of them , the robot selects some sampled materials ( according to a sample plan) and measures some characteristics according to a program .
The question is : to make G R&R for all of them or I can select some samples of them? For example I have 15 equipment of a certain technology, instead of doing 15 G R&R I could do 3 according to some criteria, for another technology I have 20 equipment , instead of doing 20 G R&R , I could do 5 ( according to some criteria that we could establish).
What about?
Which could be the criteria?

Here is one of the best answers - ask your SQE. If your customer only wants GR&R for critical characteristics, fine. If they want Gage R and stability for automated gages, fine. If they want every gage Gage R&R reported - better to know before you submit PPAP, rather than have it rejected and spend a couple frantic weeks trying to make up for it.

The criteria for performing gage R&R is NOT the number of characteristics or machines you have, but the adequacy of the knowledge you have of the machines to assure they are appropriate for the job. You may have agreement with your SQE that the machines and their function are so similar that sampling is acceptable, but otherwise there is nothing that would support that - especially if the only supporting argument is time and effort. The other side of that argument is if you do not do the Gage R&R, how do you KNOW the machine is adequate? Guessing that they are kind of similar? Not good enough.

 

[qusys]

Re: MSA studies: on each Equipment, each Measurement System or Family in Control Plan

Here is one of the best answers - ask your SQE. If your customer only wants GR&R for critical characteristics, fine. If they want Gage R and stability for automated gages, fine. If they want every gage Gage R&R reported - better to know before you submit PPAP, rather than have it rejected and spend a couple frantic weeks trying to make up for it.

The criteria for performing gage R&R is NOT the number of characteristics or machines you have, but the adequacy of the knowledge you have of the machines to assure they are appropriate for the job. You may have agreement with your SQE that the machines and their function are so similar that sampling is acceptable, but otherwise there is nothing that would support that - especially if the only supporting argument is time and effort. The other side of that argument is if you do not do the Gage R&R, how do you KNOW the machine is adequate? Guessing that they are kind of similar? Not good enough.

Good point!
Are matching studies among several equipment acceptable for this?
To return to your post, sometimes it is a question of time and effort. Sometimes special characteristics changes and your previous job of G R&R fades away or there are brand new special characteristics that are communicated in a second time!
To also consider that PTR is between 10% and 30% , this is communicated to the customer but imporvment programs are very difficult to implement, other you should change completely the equipment given that they are on leading edge.