Is it possible to calculate a single sided distribution for a process spread where the lower end (LSL) is 0?

I have a customer who wants to see variable data stats - in particular a "Cpk" for his spec of 0 to 1.5 with zero being the desired (but unattainable goal).

Customer wants to see our process capability at being able to come in under his threshold tolerance of 1.5

Any ideas?
Michael

Please scroll down to the bottom of the page and you will see previous discussions of this subject. Review each to find your results.
Hope this helps

As Ralphsulser Said, check this site, is full of info, try to dig, and you will be surprised.

To be short, yes you can (i don't like it but..), in the most of the cases as the related treads said, if LSL really doesn't exist (unreachable) is considered as one sided spec (only Upper) system.

The problem for such cases could be non normality (non gaussian distribution) that the mean and standard deviation take in account. Can you post some of your data?

Hi Micheal,

For clarification and better understanding..... what feature are you measuring, is it something which cannot physically go below zero, for example, ovality of a turned part?