Xbar / S – chart sample size.
Six Sigma – iSixSigma › Forums › Old Forums › General › Xbar / S – chart sample size.
 This topic has 4 replies, 5 voices, and was last updated 16 years, 2 months ago by Andhale.

AuthorPosts

April 10, 2002 at 8:14 pm #29223
Devesh DahaleParticipant@DeveshDahale Include @DeveshDahale in your post and this person will
be notified via email.I have a situation where end of shift data is summarized for three machines automatically and printed in the form of a report. Typically each machine runs about 800 to 1000 parts every shift. The mean and standard deviation of the parameter of interest of the parts run on each of the three machines during the shift is obtainable from the report. Mean and standard deviation is available for each machine. So, essentially I am trying to apply a control chart for a sample size of over 800 pieces (data poiints).
What value of “B4” should be used for calculating the control limits for the “S – chart”?
What value of “A3” should be used for calculating the control limits for the Xbar chart?Devesh Dahale0April 10, 2002 at 10:34 pm #74206Devesh,
Why would you be interested in so many values? I’m assuming, since you can calculate a meaningful mean and standard deviation that you have contiuous data. So, if that’s the case why would you make the rational subgroup so high? Values of constants that I’ve seen don’t typically go higher than a n=100. If you do a COV study and show minimal variation within shift, why not simply pick a random sample of 10 or 20 measures from the shift to represent the output. Typically this is the range of subgroups where we shift to the XbarS chart.
Regards,
Erik0April 11, 2002 at 1:12 am #74210
Ashok SarkarParticipant@AshokSarkar Include @AshokSarkar in your post and this person will
be notified via email.If all the data is available in the system why we need Xbar S cahart. If it is possible to get the data then we may try different control chart like XMR, CUSUM, EWMA depending on what shift we are interested to monitor.
Otherwise we may sample the data for charting purpose!0April 11, 2002 at 3:15 am #74212
Marc RichardsonParticipant@MarcRichardson Include @MarcRichardson in your post and this person will
be notified via email.My question follows those that preceeded me. You have population data. You know what every part measures. Why do you need to use statistics? Statistics are used to infer population characteristics when all one has is sample data.
Marc Richardson
Sr. Quality Assurance Engineer0November 3, 2005 at 4:20 pm #129290Is this Devesh Dahale of Pune, India…. If yes, hello.. this is Deepak Varma — mail me at [email protected] OR [email protected]
Phone : +919840730391
rgds
deepak
0 
AuthorPosts
The forum ‘General’ is closed to new topics and replies.