sigma value
Six Sigma – iSixSigma › Forums › Old Forums › General › sigma value
 This topic has 8 replies, 7 voices, and was last updated 13 years, 12 months ago by Anup Kumar.

AuthorPosts

October 11, 2008 at 5:09 am #51109
Hello,
If a process has a cpk value of 0,66 (2/3) and a cp of 1 (3/3) implying that the process is not centered, is it correct to say that the sigma level is 3 ? or is it 2 ?
Thanks
Vincent
0October 11, 2008 at 12:10 pm #176651how about saying your cpk is 0.66?
0October 12, 2008 at 1:17 am #176658There are various rules of thumb for equating Cp & Cpk to Sigma scores, such as 1.33 Cp = 4 Sigma(short therm). If you can calculate Cpk, then you can calculate the Z score (assuming normal distribtuion). Cpk is the minimum of [(USL – X(bar))/(3*StDev)] and (X(bar)LSL)/(3*StDev).
If you have all the details to fill those formulas out, you can calculate the % yield using the Z distribution and then convert that to a Zscore. If you have MINITAB go Calc>Probability Distributions>Normal Distribution. Put your average and standard deviation in, do a cumulative probabilty, and put your LSL in for the constant. Write down the result. Do the same for the USL. Subtract the two (this is your % Yield). Then go back to Calc>Probability Distributions>Normal with the mean of 0 and standard deviation of 1. Put your % yield (calculated before) in the constant field. This is the Z score (sigma level).0October 12, 2008 at 5:50 am #176659Nik,
Thanks very much for this detailed answer.
If I go back to my example (cpk=0,66) , the Z score is 2 and therefore the sigma level is 2. Is that right ?
Thanks
Vincent
0October 12, 2008 at 1:48 pm #176661Yes.
0October 13, 2008 at 1:53 am #176674
Dr Peter ThompsonParticipant@DrPeterThompson Include @DrPeterThompson in your post and this person will
be notified via email.“Sigma levels” are totally meaningless.
DO NOT USE THEM.
3.4 dpmo is also totally meaningless. There are plenty of papers published on how M Harry derived this rubbish.
0October 13, 2008 at 3:57 pm #176683Dr. Peter,So what metrics do you use?
0October 13, 2008 at 7:26 pm #176687
Dr. PhilParticipant@Dr.Phil Include @Dr.Phil in your post and this person will
be notified via email.Jeff,
Go do a little homework.
Cp, Cpk, Pp, Ppk, Cpm all predated the “sigma” metric. They all do as good or better in describing the behavior of a process.0October 14, 2008 at 5:41 am #176703
Anup KumarParticipant@AnupKumar Include @AnupKumar in your post and this person will
be notified via email.Yes. This is right. Basically if the value is above 1 it is considered to be a good process. IF the value is 2, then it is a six sigma process.
0 
AuthorPosts
The forum ‘General’ is closed to new topics and replies.