June 20, 2017 at 4:50 am #55750
i would like to get your advice please
i calculate ppk per machine for every week.
one week includes observations of 3 shifts – same employee in each shift in same week – but it’s different between each week.
so in one week i have internal variance from different shifts and maybe from different days.
The quantity of the observations of the one shift (/day) is different from other shift (/day) and when i calculate ppk, first – I refer to it as one subgroup
second – I do not include exceptional observations
what do you think about it?
Ben0June 20, 2017 at 7:31 am #201585
why not include “exceptional” observations0June 20, 2017 at 7:44 am #201586
I dont have a good answer
i guess it because most of the exceptional observations are before the proccess was stable
what do u think about 1 subgrop althouth i have 3 shifts in week?
Ben0June 20, 2017 at 7:54 am #201587
well, you should confirm the process is stable and in “process control” before doing a process capability but if you do a process capability study, you can check for stability and control after you’ve gathered the data. An example…one doesn’t do a smog study on a catalytic converter within the first few minutes of starting a car after a long rest because it’s known to not be stable and in control if operating correctly.
i’m skipping by the “make sure your measurement system is adequate” but it shouldn’t be assumed that your device is precise and accurate.0June 26, 2017 at 7:17 am #201599
why dont you share the results, will help to understand better.0June 26, 2017 at 9:12 am #201602
@Benjamin Here is what I would do. I would identify more exceptional observations and throw them out. I would continue to do this until the data was behaving the way I thought it should.
So you could do that which is of course the easy answer or you could leave all your data in (after you follow @Chris Seider advice and run an MSA) and then go see the process and figure out what is causing the variation.
You might even try a multi-vari chart (don’t worry the computer given the correct software will do that for you just like Ppk (only different – figure that out). Then you may actually be using a tool that can help you identify and fix a problem.
Process capability is a good tool to convey information to management. Beyond that it doesn’t have much value. Have you seen the movie from National Lampoon where Chevy Chase and the family are staring at the Grand Canyon (actually it could be @Mikecyger this week). That is the image I get in my head whenever any one puts up a Capability study0June 27, 2017 at 4:19 pm #201605
I visit iSixSigma everyday to see @Mike Carnell ‘s recent replies. I’ve learned a lot and laughed a lot more.0June 27, 2017 at 5:11 pm #201606
@Albert Sandoval You have made my night. I had to take my computer and show my wife, Consuelo. I have told her for a long time that I crack me up and I get one of those “seriously” looks. She never gets the stats jokes. She thought throwing out the data seemed like a good idea.
I hate to sound like an old timer but a few years ago we had the usual suspects (Darth, Stan, HBGB 2, and Stevo) and we averaged 40 posts per day. We were what Mike Cyger (owns iSixSigma) called his Jerry Springer Factor. That was fun. Bill Hathaway, Moresteam owner, refers to it as food fights.
Thank you. I appreciate the compliment about learning stuff. There are some awesome people on here giving out great advice. There is always something to learn.0
You must be logged in to reply to this topic.