Historical Standard Deviation or Select Parts for MSA

[Black arrow]

Hello,

I have always selected parts from the production when i have made my MSA but i understand from this forum and also from AIAG it recomends to use historical standard deviation instead. But i have to be clear how to approach this alternative. Should the standard deviation be calculated from a different measurementsystem (gauge)? Or same gauge that is in focus in my MSA?

 

[reynald]

I cant answer your question but have you looked into this post?:

http://elsmar.com/Forums/showthread.php?t=55103&highlight=msa+historical+standard+deviation

 

[Bev D]

if you have historical standard deviation then it is almost always better than using the standard deviation form the study. If you use the AIAG recommended number of parts (10) then you really have too few parts to make a decent estimation of the observed standard deviation. with small sample sizes this is compounded by a 'chunky' selection of parts to span the range of expected variation.

the historical standard deviation must be from the same gage used for the study. Remember that the observed SD includes the part variation and the measurement system variation, so if use a historical SD from a different gage you will have an incorrect observed SD for use in the study....