Does TS16949 require MSA on all measurement systems in the Control Plan?



We currently carry out MSA on all types of gauge requiring calibration, some laboratory analysis methods used to control significant characteristics within our process and visual inspection of appearance attributes.

However, we do not carry out MSA on every measurement system referred to in the Control Plan.

Our 3rd party auditor has written an NC because we had not carried out a statistical analysis for some destructive tests carried out on painted parts – Specifically an adhesion test (cross cut method), a test for curing (solvent wipe) and a pencil hardness test.

It would be difficult for us to produce study samples having a range of performance levels in order to carry out the studies – We would have to deliberately adjust the process parameters outside of specification or skip parts of the production process to achieve this.

The auditors understanding of the standard is that statistical studies are required for all measurement systems in the Control Plan and he therefore wrote an NC.

I would appreciate any views on this – Is the auditor correct, or is there any scope in TS16949 not to carry out MSA on a control plan test/measurement method?

What about tests (such as corrosion test) carried out by ISO 17025 accredited laboratories?


Quite Involved in Discussions
Section 7.6.1 of the TS standard specifies to do MSA for the test that is defined in the Control Plan. We also have the same problem with the auditor because we do not do the MSA using the ANOVA gage R&R. We make fluids for OEM's and according to the PPAP blue book section F.9 "MSA studies are not required where standardized tests (e.g. ASTM methods) are used. However it is still important for the organization to understand the measurement component of variation in the test methods used." We define the type of MSA that we do in our lab scope. The auditor only knows the ANOVA R&R but since we document what we do (e.g. verifying the ASTM R&R) and when we fail we do corrective actions, it seems to be OK.


Thank you for your response.
I see that section F.9 of the PPAP manual refers to suppliers of bulk materials.
There are some characteristics of our process which has some similarity to those involved in the production of bulk materials and this may be of help to us.
For example, measurements made on chemical treatment baths used to process supplied parts may be similar to those used to those made on bulk raw materials and the level of variation higher than would be found in mechanical measurements.
I would be interested in any other thoughts on this subject.


Quite Involved in Discussions
Thank you for sharing your MSA. Our MSA is defined in the Lab Scope as : benchmark the performance of their laboratories (against around 60+ other plants) in carrying out standard Quality Control tests on fluid finished product samples. Here's how it works.Four samples of finished products are despatched each year to laboratories.
Analysis are carried out by laboratories using the Standard method ASTM, ISO, etc...;Data from all laboratories are analysed. Laboratories are informed of the results. Assessment is based on z-score concept (ISO/REMCO N263) where z = (x-X)/STD = (x-X)/(R/1.96(SqRt 2));x = Reported result
X = Robust mean
STD = Standard deviation of the test method
R = Reproducibility of the test method
I the laboratory failed to meet the requirements of the z-score, they have to determine the root cause and implement corrective action by having 3 operators perform the analysis until they get within the specified z score. Everything is documented. If the laboratory gets a 100% acceptable z scores for the tests run, they do nothing.

Once again we have to explain what we do in accordance to the F.9. My advice to you is to document everything you do for MSA. If you do not follow the ASTM method then you need to do the linearity, bias, and ANOVA R&R and ndc >= 5.
Hope it helps


Trusted Information Resource
Short answer is yes. But it is one of those areas which defies the real world as MSA may be unrealistic or over kill in certain situations, as you point out. But there is a conflict between the standard and the "Blue books." 2.2.8 of the PPAP book indicates that msa may not be applicable for bulk materials. You can try to work from that end. Otherwise, you'll just have to come up with something workable. Good luck.


Quite Involved in Discussions
Thank you for giving me more confidence to deal with MSA which the auditor likes to ask. What we do is applicable to our process as define in the Lab scope.
Top Bottom