What is your industry standard for the total number of pieces needed for Cap Studies and First Article Inspection parts?
We supply a range of customers and it varies. I am trying to set a standard of whats 'typical' for a business model.
Class '8' Truck - most require 30 piece capability and 1 piece FAI
ATV / Sports - require 100 piece capability and 5 piece FAI.
What is the statistical 'break point' for cost vs benefit? (on non-safety critical parts).
The idea behind a capability report (and what the automotive industry generally requires) is to do--where appropriate--standard control charts during a "significant production run," where "significant" depends on individual circumstances. The primary reason for this is to be able to establish at least a semblance of statistical stability before making any capability calculations. The chronological component is important in this regard.
There is a basic idea in statistics that in a normally-distributed population, the difference between the sample standard deviation and population standard deviation will approach being negligible when the sample size reachs 30. What this means in a general sense is that for many (certainly not all) populations, returns diminish with samples sizes >30. It's a dangerous generality though, with one reason being that an assumption of normality can be lead to unexpected outcomes.
Som how this misunderstood 30-piece sample idea became a sort of meme, with a lot of people believing that it makes perfect sense to calculate Cpk for a small sample that has been chronologically mixed. It doesn't, but the idea has taken on a life of its own, so it becomes one of the many time-wasters we have to put up with because of customers who don't understand what they're asking for.