On 10/1/99 4:06:31 PM, Richard Helle wrote:
>What I would like to know is how to calculate Pp and on this feature. The formula for this is: USL-LSL/6s. For true position that would be i.e. .010-0/6*(whatever standard deviation is on the population of parts). When mmc becomes involved, the USL becomes variable. Sometimes its .025, sometimes its .0252.
The method I use now is to calculate the percentage of tolerance used and then my
formula becomes 100-0/6*standard deviation of
the percentage. Is this correct? If not, why and what is the correct method.>

Richard-

No it is not.

First, "True Position" data is not normally distributed. Normality is an underlying assumption of process capability and performance (Cp, CpK, Pp, PpK)calculations.

Second, the bonus tolerance available for the position of a measured feature is dependent upon that feature's departure from mmc IN THAT PARTICULAR PART. It (bonus tolerance) is good for individual part compliance, NOT FOR PRODUCTION CAPABILITY.

That being said, if the size deviation of the feature is consistent through production one can temporarily add the bonus tolerance to the positional tolerance to qualify the parts. But if the size wanders through its full range you never know how much bonus a particular part offers.

Mike Matusky
Dimensional Management
mfmatusky@aol.com