Sigma level of a process
Six Sigma – iSixSigma › Forums › Old Forums › General › Sigma level of a process
 This topic has 5 replies, 4 voices, and was last updated 20 years, 8 months ago by Keshri.

AuthorPosts

January 10, 2002 at 6:34 pm #28503
Ovidiu ContrasParticipant@OvidiuContras Include @OvidiuContras in your post and this person will
be notified via email.Looking at the Minitab output for Capability Analysis , we have Cpk and PPM “Overall” as direct Sigma level indicators .
My question is : If the data is normal and if the Sigma level given by Cpk*3 is different from the one resulted from the PPM , is this an issue? Which ( Cpk or PPM ) is the one that should be taken into account in this situation ?0January 10, 2002 at 7:25 pm #71158Ovidiu,
I hate to use this reply, but it depends. What are you looking to do with the data? If youre looking at developing a shortterm perspective of the process then you would use the Cpk calculation (and the associated PPM for ‘within’). Look at what Minitab give you for hints. Cpk statistics are created based on within which should trigger a shortterm estimate of the variability of the process. Remember from control charting, the basic questions revolved around what factors and noise are captured within and between subgroups and these were used to diagnose where the greatest source of variation lay in the process. The PPM statistics for within will give you the shortterm estimate of the process. The overall calculations for PPM are looking at the longterm picture of the process. The variability calculated for this process will not only factor the within component, but also the between component. Traditional interpretation of Cpk, or the ensuing Z score obtained, is the shortterm since standard deviation will be estimated by Rbar/d2. Assumptions typically include a state of control for confidence in this estimate. Bottom line, you’re looking at different ways of calculating variability (shortterm vs. longterm) when you look at the statistics that you’ve mentioned.
Regards,
Erik0January 10, 2002 at 7:42 pm #71159
Ovidiu ContrasParticipant@OvidiuContras Include @OvidiuContras in your post and this person will
be notified via email.Thanks for the reply Eric ,
now let’s look at the actual output : PPM total for “within” is 294,000 which gives a sigma level of about 2 ; Cpk is 0.2 , which gives another sigma level of 0.6 …..should I use Cpk ?0January 10, 2002 at 8:49 pm #71161Which parameter are you looking at? The ‘oberved’ or the ‘expected’? If you take the PPM from the ‘expected’ this should sync up with 3*Cpk.
Regards,
Erik0January 11, 2002 at 10:28 pm #71187I think that your data is not syncing because you did not factor in the 1.5 sigma shift. If you are using PPM and the standard normal tables to obtain your Process Sigma (Z), they do not correspond directly to Z (Cpk *3) from a normal PDF. Add 1.5 to your calculated Z then see if the PPM defectives match.
0January 12, 2002 at 4:09 am #71191Q – If the data is normal and if the Sigma level given by Cpk*3 is different from the one resulted from the PPM
If Cpk1 = (USL – Mean)/ sigma and Cpk2 = (Mean LSL)/ sigma
Cpk is the lower value of Cpk1 and Cpk2
PPM is calculated adding the two DPMO calculated as indicated by Cpk1 and Cpk2
So the Sigma level calculated on the basis of PPM and Cpk*3 will be diffrent. and Cpk*3 will be higher than the sigma value calculated on the basis of PPM.
0 
AuthorPosts
The forum ‘General’ is closed to new topics and replies.