Hi Paul, I love your simple to the point explanations! 🙂 One point I'd like your perspective about is a "too high" CPk . We should always keep in mind the long term relevance of our process limits and whether they are reflecting the quality we require from it
Nice job. Thank you sir, many of us are grateful for your straight forward explanation and I personally enjoy the way you present this without pushing statistical software. I have watched too many videos presented by SS experts and I learn so very little because they relied entirely on software to get their point across. I prefer to learn the methodology first then use software to make the analysis quicker Once again, from many of us, thank you. Have a nice day
Another great video Paul. You are certainly a valued member of the ILSSI community and we all look forward to seeing you again at our next conference !
Every time I watch one of your videos I look forward to the "and my name is PAUL ALLEN." This needs making into a meme! Seriously though, great videos!
Staminski - Thanks for the positive comment, don't tell anyone but I stole that intro from another youtuber! if you like my video content take a look at my book 'Drink tea and read the paper' you can get it from LULU.COM...and keep watching!
I am from india. I am searching from 3 year for this answer but today i found this answer.i am very thankful for you and very appreciate from you . Thank you for this video Please make more videos on six sigma .
@@paulallen5321 i see your video you clarify all doubt and all videos are fullfill from knowledge from other people. My suggestion is for you can start online course.
@@paulallen5321 i watch 200+ videos from other channels but when i see your videos you clear my all doubt . I am a engineer in a manufacturing line i say about your channel very perfect from other i share your ourse to my all friends for grow and better knowledge...
Dear Paul, great way to explain, I want to make note of your description, so my request is pl leave the board for some time by standing in side not in front to make note or to take screen shot.
Hello Paul, I must say that you do have clarity of thoughts and it comes out very well in your videos. Much appreciated!! I have one question, is there any rule or guidance on how one should do sampling (number of samples) for a given process based on CPK values. For example, if we have a process that is giving CPK value of 1.33, do "x" number of samples randomly.
Hi Paul, thanks a lot for your clear explanation, during my work in production side, 1.33 cpk is preferred, questions like "how to determine this value and how to control yield rate" stay in my head. However, this book link seems expired, will you update it?
Thank you again ... Now I can tell, I understood Six Sigma! Never without you,Sir. I have this question: what factories usually do to keep there SPC diagrams (Sigma Shift) on the look out! How often they take sample and how many each time? ... Is it a daily/weekly/monthly/by shift!!! ... am I talking about "Process Control Phase"! as an engineer I'll need to take care of my machine production continuously.
Truth Seeker - The sampling rouitne is defined by how often you think a variable moves. If you think the process moves in 15 mins then you should sample every 5 mins. But if the process is unlikely to move after 4 hours, then checking every 2 hours should be ok...It's a practical decision not a rule...
thank you for this great video! we have some supplier processes where we require ppk of 1.02 and around there. Are you saying that the process would have to shift 2+ sigma before we "notice" a shift, unless we are sampling 100% ?
Diana - I'd have to look specifically at your numbers but potentially yes. What you're trying to see is that the distribution has shifted. Your process could have moved down, but by luck of sampling you see a high value and miss the shift this hour, next hour you get a lower result, but maybe not outside control limits, still no indication of the shift and the third point falls in an unusual place and you spot the problem. But the problem could have existed for 3 hours....but if the shift is dramatic, then the signal will appear in your data more often and therefore you'd see it earlier...by the 100% inspection is a poor way to gaurantee quality....
Lokesh - There are a few video's explaining the capability statistics. But maybe the 3 you mention might fit in the video as a good group to understand..
Thank you very much for the great explanation, I would like to ask please, what if I have limited amount of typical measurements (less than 10) - but in addition I have corner parts measurements (+/-3sigma). How can I calculate the Cp/Cpk?
Thanks for these videos! Ordering the book for sure! I have a weight control problem (bread manufacturing) where the mean is significantly off the customer specification target. But, I also have a very wide distribution that even if the mean was centered I would be creating a lot of defects. The specification range is 20 grams so target is 335 lsl 325 usl 345. I have a standard deviation of 11.86. Which is terrible. Is the Cpk number i am getting in Minitab( -0.50) skewed by the target mean thus giving me the extremely low Cpk? Second, I assume I should try and center first and then reduce variation. Any input is appreciated. I am a one man CI department.
hi Jim good to hear from you...sounds like you have a challenge on your hands...yes the poor cpk is being effected by the results being off target.....job one would be 30 consequtive bread rolls, measure the weight and send over the results. paul.allen@allenp.co.uk
I have checked out all the videos on internet explaining process capability and performance indices, but none pointed out about data distribution. What if this data does not fit for normal, non normal distributions, and any other transformations. Mistakenly, many quality practitioners continue using cpk as the most important index in process capability, however, ppk is more critical.
Renzo - it's a good question. It depnds on what the lack of normality is. I beleive there is some explanation in this video th-cam.com/video/5i1ya7ofHQ4/w-d-xo.html
dear sir ,i well understand the reason why the, cpk must be large ,however my only doubt is when it shift why it must shift 1.5 sigma.pls teach about 1.5 sigma(from where come the 1.5
Hi sir. Great explanation in you videos. I would appreciate if you could help me clarify. In my assignment, there are two data set of one company. The first has less samples but more observations for sample and the second data set has increased samples but lower observstions. Cp of the latter is 1.48 while the first set has 1.09 Cp. Does this means more sample with less observation is always better and why we dont just use go with that instead of getting less samples with a lot of observation?
Mukesh - that depends on your process and the variables that change and how fast they change. For example if you process is high volume and you use a new batch of raw material 4 x per day. Each material change will cause a change. If that change in the material only happens once per month, then the movement is slower...
I watched 20 videos and still couldn't understand it. Very good explanation on your tutorial 👌
Hi Paul,
I love your simple to the point explanations! 🙂
One point I'd like your perspective about is a "too high" CPk . We should always keep in mind the long term relevance of our process limits and whether they are reflecting the quality we require from it
I'll make a video and send you the link....
A brilliantly succinct and clear explanation- thank you
let's see paul allen's cpk
Very concise and easy to understand explanation.
Nice job. Thank you sir, many of us are grateful for your straight forward explanation and I personally enjoy the way you present this without pushing statistical software. I have watched too many videos presented by SS experts and I learn so very little because they relied entirely on software to get their point across.
I prefer to learn the methodology first then use software to make the analysis quicker
Once again, from many of us, thank you.
Have a nice day
Kwok - thanks for the comments. I might make a video or two with the theory followed by the software method to combine the two....
Another great video Paul. You are certainly a valued member of the ILSSI community and we all look forward to seeing you again at our next conference !
Every time I watch one of your videos I look forward to the "and my name is PAUL ALLEN." This needs making into a meme! Seriously though, great videos!
Staminski - Thanks for the positive comment, don't tell anyone but I stole that intro from another youtuber!
if you like my video content take a look at my book 'Drink tea and read the paper' you can get it from LULU.COM...and keep watching!
Super cool explanantion. So great
very good explanation
It's a nice video. A simple and elegant way of explaining.
Glad you liked it!
Thanks for your simple explanation
Glad it was helpful!
it is brilliant how you explained it that simply
Fatma - most things are very simple it's when academics explain things that it get complicated...
I am from india. I am searching from 3 year for this answer but today i found this answer.i am very thankful for you and very appreciate from you .
Thank you for this video
Please make more videos on six sigma .
Hi Mohit - I think you'll find over 200 video's on six sigma on my channel, please enjoy...
@@paulallen5321 thank you for
@@paulallen5321 i see your video you clarify all doubt and all videos are fullfill from knowledge from other people. My suggestion is for you can start online course.
Hi Mohit - good idea! I'm just about to do that....
@@paulallen5321 i watch 200+ videos from other channels but when i see your videos you clear my all doubt
. I am a engineer in a manufacturing line i say about your channel very perfect from other i share your ourse to my all friends for grow and better knowledge...
Your videos are really helpful, thanks for doing these!
No problem Robin...
Now I just understand 1.33cpk and 1.67 and 2.0 cpk. Very good sir. Keep uploading video. I just subscribe in your channel.
Thanks and welcome
Excellent explanation sir...tq u
Simple but brilliant. TNX
Cheers Milan
Dear Paul, great way to explain, I want to make note of your description, so my request is pl leave the board for some time by standing in side not in front to make note or to take screen shot.
Rajesh - I will certain try to do that in future...
Superb explaination sir, thank you
Rahul - thanks for your positive comments, hope you find them all useful...
Hello Paul, I must say that you do have clarity of thoughts and it comes out very well in your videos. Much appreciated!! I have one question, is there any rule or guidance on how one should do sampling (number of samples) for a given process based on CPK values. For example, if we have a process that is giving CPK value of 1.33, do "x" number of samples randomly.
Sumitpal - good to hear from you. No there is no rule. Use your own sense of cost versus benefit for the product your you're producing...
so good explanation, thanks
Floyd thanks for your positive comment, if you like this you'll enjoy my book ;drink tea and read the paper you can get if from LULU.COM
Hi Paul, thanks a lot for your clear explanation, during my work in production side, 1.33 cpk is preferred, questions like "how to determine this value and how to control yield rate" stay in my head. However, this book link seems expired, will you update it?
Thanks you
No problem
Thank you again ... Now I can tell, I understood Six Sigma! Never without you,Sir. I have this question: what factories usually do to keep there SPC diagrams (Sigma Shift) on the look out! How often they take sample and how many each time? ... Is it a daily/weekly/monthly/by shift!!! ... am I talking about "Process Control Phase"! as an engineer I'll need to take care of my machine production continuously.
Truth Seeker - The sampling rouitne is defined by how often you think a variable moves. If you think the process moves in 15 mins then you should sample every 5 mins. But if the process is unlikely to move after 4 hours, then checking every 2 hours should be ok...It's a practical decision not a rule...
@@paulallen5321 Thank you for your time
thank you for this great video! we have some supplier processes where we require ppk of 1.02 and around there. Are you saying that the process would have to shift 2+ sigma before we "notice" a shift, unless we are sampling 100% ?
Diana - I'd have to look specifically at your numbers but potentially yes. What you're trying to see is that the distribution has shifted. Your process could have moved down, but by luck of sampling you see a high value and miss the shift this hour, next hour you get a lower result, but maybe not outside control limits, still no indication of the shift and the third point falls in an unusual place and you spot the problem. But the problem could have existed for 3 hours....but if the shift is dramatic, then the signal will appear in your data more often and therefore you'd see it earlier...by the 100% inspection is a poor way to gaurantee quality....
Very well explained sir. Can you please make a video on differences between Cp, CpK, Pp, PpK and Cm, CmK.
Lokesh - There are a few video's explaining the capability statistics. But maybe the 3 you mention might fit in the video as a good group to understand..
I think this might already be covered in this video..
th-cam.com/video/zwwSDPduj_g/w-d-xo.html
Thank you very much for the great explanation,
I would like to ask please, what if I have limited amount of typical measurements (less than 10) - but in addition I have corner parts measurements (+/-3sigma). How can I calculate the Cp/Cpk?
Hi Ofir - you'll have to explain to me what corner parts are??
Thanks for these videos! Ordering the book for sure! I have a weight control problem (bread manufacturing) where the mean is significantly off the customer specification target. But, I also have a very wide distribution that even if the mean was centered I would be creating a lot of defects. The specification range is 20 grams so target is 335 lsl 325 usl 345. I have a standard deviation of 11.86. Which is terrible. Is the Cpk number i am getting in Minitab( -0.50) skewed by the target mean thus giving me the extremely low Cpk? Second, I assume I should try and center first and then reduce variation. Any input is appreciated. I am a one man CI department.
hi Jim good to hear from you...sounds like you have a challenge on your hands...yes the poor cpk is being effected by the results being off target.....job one would be 30 consequtive bread rolls, measure the weight and send over the results. paul.allen@allenp.co.uk
Hi, thank's for your videos..
If the Cpk is more than 2 like 10 what does it tells us about the process ?
It menas that you are using 10 % of given tolerance meaning tighter process control
Paul Allen is back up and running!!
I have checked out all the videos on internet explaining process capability and performance indices, but none pointed out about data distribution. What if this data does not fit for normal, non normal distributions, and any other transformations. Mistakenly, many quality practitioners continue using cpk as the most important index in process capability, however, ppk is more critical.
Renzo - it's a good question. It depnds on what the lack of normality is. I beleive there is some explanation in this video
th-cam.com/video/5i1ya7ofHQ4/w-d-xo.html
Renzo - let me ask you what do you think the difference is between Cpk and Ppk?
dear sir ,i well understand the reason why the, cpk must be large ,however my only doubt is when it shift why it must shift 1.5 sigma.pls teach about 1.5 sigma(from where come the 1.5
Thanks for the comment - that video has already been made
th-cam.com/video/yodjfhhKNE0/w-d-xo.html
Hi sir. Great explanation in you videos. I would appreciate if you could help me clarify. In my assignment, there are two data set of one company.
The first has less samples but more observations for sample and the second data set has increased samples but lower observstions. Cp of the latter is 1.48 while the first set has 1.09 Cp.
Does this means more sample with less observation is always better and why we dont just use go with that instead of getting less samples with a lot of observation?
Trung were you averaging the the observations?
@@paulallen5321 yes the mean of observations are given sir. The given table contains 3 collumns: Sample number, mean of sample and range of sample.
That is why the Cpk is different, the bigger the group in each sample the better the cpk will be...
Sir how long does it takes for 1.5s shift?
Mukesh - that depends on your process and the variables that change and how fast they change. For example if you process is high volume and you use a new batch of raw material 4 x per day. Each material change will cause a change. If that change in the material only happens once per month, then the movement is slower...