I am frequently frustrated with my peers/management signing for parts that fail a hard gage check or overlay based on a statement that "its not that much out, it shouldn't affect the function...".

Rather than take his or her word for it, can I insist such acceptable deviations be documented or perhaps not gaged at all (this feature might instead be left to visual assessment or something less rigid than a gage) if the characteristic is indeed non-functional?

One particular example is that of overlay check. I have suggested that if we are to allow the profile to deviate from the tolerance band (in a certain area), should we not remove that portion of the profile from the overlay plot? Or perhaps we open the tolerance band an amount in that area to reflect the lesser importance of the form in this area.

I would like to hear arguments around the topic of "grey area" - where Quality must draw on something other than prints and standards to accept a part that deviates from requirements. Perhaps I am being too particular in my inspection, but if we cant define exactly how much more is too much, perhaps we shouldnt be making such exceptions(the call can change from day to day).

How I wish I could rely on my gages, prints, and visual standards without question.

Isn't this just a slightly different version of a question you asked before: When is a nonconformity clearly scrap? (http://elsmar.com/Forums/showthread.php?p=203431#poststop)?

I would like to hear arguments around the topic of "grey area" - where Quality must draw on something other than prints and standards to accept a part that deviates from requirements. Perhaps I am being too particular in my inspection, but if we cant define exactly how much more is too much, perhaps we shouldnt be making such exceptions(the call can change from day to day).Your organization's regulatory compliance policy states:
Quality System Management
Product design, manufacturing, packaging, labelling, distribution, installation and servicing will comply with Quality System Standards and regulations. These include standards such as ISO 13485, Medical Devices – Quality Management Systems – Requirements for Regulatory Purposes, FDA CFR Part 820 (QSR), Pharmaceutical Regulations, e.g., CGMP. All subsequent clauses of this policy, Clauses 3 to 12, will be in compliance with all standards and regulations that apply to these clauses.Both 13485 and FDA QSR require a formal disposition of non-conforming products. The scenario, as described, seem to violate the formal tenets of a NC MRB process. Isn't this just a slightly different version of a question you asked before?Yep.

Thank you for your responses, and your patience. I apologize for the repost. I thought it went dead in the other thread (perhaps because I replied to my own thread); I hadnt noticed a response for a few days.

My approach will be to consult Development to confirm any and all features that are Critical.
I will then consult my group to ask that our gages (e.g., overlay) be adjusted to measure only those features that are Critical and propose that we use other means (perhaps visual comparison to a physical model), to assure our Incidental features are "good enough".
In this way, the shop floor will not be confused when using overlay for example, as the plotted profile will measure only the Critical section of the part profile and nothing more - these profiles come from the CAD model and no attempt at this time is made to restrict them to the critical portion.