I am looking for some recommendations for acceptance criteria for QS-9000 linearity studies. The MSA manual does not define acceptance criteria it just says that regression line should have a good linear fit(R squared) and good linearity or low slope. The auditor says I must have values or targets stated for these characteristics. I can't find any guidelines. Any ideas?

As I understand it, you take the value of the instrument's range deviation and look at it as a % of the spec range. Was this brought up with respect to your uncertainty budget?

Dear Bharder,

You hv to consider linearity as a proportion of the total r&r
As well as total tolerance.
We should not consider this in isolation!
If further details are provided we can better discuss this!

Thanks
e.s.deo