Paul, great advice on setting up a DOE, appreciate the best settings - it clearly shows the data and I agree it makes it much more obvious to see what the data is showing or if it's ill behaved. One note, the video stops at 11:08 but then continues as black for almost another hour.
Lance - If want more information on my DOE approach I've just released a new book... ' Design of experiments for 21st century engineers' you can get it from LULU.COM
Great, not really been involved in DOE but know it's value. Want to learn to input parameter data from injection moulding machines and be able to confidently explain results.
Paul, Nice video! Only one concern- you first set up the data in the spreadsheet with a certain sequence of experiments for the time, temperature and pressure. Once you entered the factors into Minitab and set up the experiment, Minitab prepared the experiments in a different order. Once you pasted the data from your spreadsheet, the data was not always the results matching the appropriate experiment.
Richard - thanks for posting a message it always helps with the channel. If want more information on my DOE approach I've just released a new book... ' Design of experiments for 21st century engineers' you can get it from LULU.COM
@@paulallen5321 this does not at all address the concern of richard and subedi. You make a mistake when copying the data from excel to minitab. The time/temp/pressure settings do not match between excel and minitab.
Hi Joost - Good to hear from you. You are correct of course. The aim of the video though was not show you how to copy and paste, it was to show you how minitab works. I reality you would type you experimental results straight into Minitab. I knew when I did it that the pattern was different, I just needed a data set to analyse and demo Minitab.....
question about the factors; do they need a correlation within the system? Like pressure and temp has got a relation, when the pressure goes up, also temp does. or is it also possible to put in factors like wait time in PLCs, or human factors? I want to know it because, I'm graduating right now, and I'm new to statistics from my study of mechanical engineering and I think I can use the DOE method but I need to know which factors I can use or not.
Tom - that is the point of DOE it will take into account interactions between variables like temp and pressure. I you can use any factor which you can set independantly. If 2 factors settings are linked together the DOE pattern becomes impossible...
Farzana - there isn't 5 responses, there are 5 repeated tests for each set up. this gives a sample size of 40. Sample size is very important to getting a successful experiment...
Is it possible to analyze when e.g one factor has 3 levels and the other two has 5 levels and more than one response? Any relevant video plz share the link.
Danish - There is a way to do this, but I'm not sure why you would do it? Drop me a line on paul.allen@allenp.co.uk and if i can assist further with this i will...
Mr. Paul ... Can you please make a tutorial video about how or in which cases we need to "Constraint" Factors. Preferable in manufacturing application. Thank you
Dear Paul, your videos are awesome have really helped me to use Minitab for DOE as a first time user of Minitab. The only nagging question I got is to do with "center points per block = 0" under the Designs option. It was my understanding that including 5 center points (for 5 replicates) for this tutoring will have been valid? Please, your thoughts on using center points on factorial designs in general will be great. Thanks you very much
Charles good to hear from you....firstly there is video on this subject ua-cam.com/video/fqpnW8zGEug/v-deo.html But I don't use centre points in the DOE I use them afterwards as the confirmation point. To prove the model is linear. But you can also do them as the first row a middle row and last row to prove the experimental discipline is good...
Hello Sir, I would like to choose some mixes for testing the model after analyzing the full factorial DOE. Is there any way in which the program choose the mixes?? Thanks for the tutorial :).
Rajaa - ixture DOE's are very specialised. Personally I like to use a slightly different approach. If want more information on my DOE approach I've just released a new book... ' Design of experiments for 21st century engineers' you can get it from LULU.COM
Great video Paul. Just wondering how you retrieved 5 values (Y1-Y5) for each of the 8 tests - Did you repeat the test 5 times at each setting measuring each time? And, would it have been the same if as setting it up in minitab for 5 replicates and recording it once?
Reuben - yes if you select 5 replicates in Minitab it will create column that is 5x the 8 run pattern. 40 rows long. In my software it is in a table so that you can see the 5 data points for each set up..
Zeliha - good to hear from you. Drop me al ine on paul.allen@allenp.co.uk and I'll see what i can do. Before you use mintab though you need to design your experiment. Watch this video first ua-cam.com/video/gsD8V2_eZ0A/v-deo.html
Dear Paul, you are doing great and your videos are really helpful, I need help and suggestions from you, I have 9 parameters ( Factor) and 3 levels. What Model will be suitable? Please let me know.
Stark good to hear from you, you have 2 options. Run the L18 which will only take 7 variables at 3 levels, so you'll have to fix 2 of the variables. Or you could only test at the 9 variables at 2 levels and use a Taguchi L12. Both of these options are screening DOE's. It's very rare to model 9 variables as it's too complex and expensive. Remember we're not looking for perfect knowledge, we're looking for knowledge that allows us to make the process land on targets that we need to please the customer...
Paul, great advice on setting up a DOE, appreciate the best settings - it clearly shows the data and I agree it makes it much more obvious to see what the data is showing or if it's ill behaved. One note, the video stops at 11:08 but then continues as black for almost another hour.
Lance - If want more information on my DOE approach I've just released a new book...
' Design of experiments for 21st century engineers' you can get it from LULU.COM
Great videos Paul
Hi John - Thanks for the support, I'll keep them coming as i think of the next subject...
Great, not really been involved in DOE but know it's value. Want to learn to input parameter data from injection moulding machines and be able to confidently explain results.
John - Here's my moulding DOE video
ua-cam.com/video/bIZ99T1xb7k/v-deo.html
@@paulallen5321 thanks, I will watch and learn
Paul,
Nice video! Only one concern- you first set up the data in the spreadsheet with a certain sequence of experiments for the time, temperature and pressure. Once you entered the factors into Minitab and set up the experiment, Minitab prepared the experiments in a different order. Once you pasted the data from your spreadsheet, the data was not always the results matching the appropriate experiment.
Actually this was exactly what I was thinking. You cannot just copy paste the data as Minitab automatically sets experiment in different order
Richard - thanks for posting a message it always helps with the channel. If want more information on my DOE approach I've just released a new book...
' Design of experiments for 21st century engineers' you can get it from LULU.COM
@@paulallen5321 this does not at all address the concern of richard and subedi. You make a mistake when copying the data from excel to minitab. The time/temp/pressure settings do not match between excel and minitab.
Hi Joost - Good to hear from you. You are correct of course. The aim of the video though was not show you how to copy and paste, it was to show you how minitab works. I reality you would type you experimental results straight into Minitab. I knew when I did it that the pattern was different, I just needed a data set to analyse and demo Minitab.....
question about the factors; do they need a correlation within the system? Like pressure and temp has got a relation, when the pressure goes up, also temp does.
or is it also possible to put in factors like wait time in PLCs, or human factors?
I want to know it because, I'm graduating right now, and I'm new to statistics from my study of mechanical engineering and I think I can use the DOE method but I need to know which factors I can use or not.
Tom - that is the point of DOE it will take into account interactions between variables like temp and pressure. I you can use any factor which you can set independantly. If 2 factors settings are linked together the DOE pattern becomes impossible...
Hello - I am wondering why do you have 5 responses? how did you determine that? kind of stuck there.
Farzana - there isn't 5 responses, there are 5 repeated tests for each set up. this gives a sample size of 40. Sample size is very important to getting a successful experiment...
Hi mr paul, may i know, this one way anova or 2 way anova factorial design?
Hi Faizal - Neither, I don't use Anova in my DOE analysis, I use multiple regression...
Is it possible to analyze when e.g one factor has 3 levels and the other two has 5 levels and more than one response? Any relevant video plz share the link.
Danish - There is a way to do this, but I'm not sure why you would do it? Drop me a line on paul.allen@allenp.co.uk and if i can assist further with this i will...
Mr. Paul ... Can you please make a tutorial video about how or in which cases we need to "Constraint" Factors. Preferable in manufacturing application. Thank you
Hi truth seeker - I have made that video for you, but i have laptop problem which will be sorted by next when I'll upload it...
@@paulallen5321... Take your time sir! Thank you.
Dear Paul, your videos are awesome have really helped me to use Minitab for DOE as a first time user of Minitab. The only nagging question I got is to do with "center points per block = 0" under the Designs option. It was my understanding that including 5 center points (for 5 replicates) for this tutoring will have been valid? Please, your thoughts on using center points on factorial designs in general will be great. Thanks you very much
Charles good to hear from you....firstly there is video on this subject
ua-cam.com/video/fqpnW8zGEug/v-deo.html
But I don't use centre points in the DOE I use them afterwards as the confirmation point. To prove the model is linear. But you can also do them as the first row a middle row and last row to prove the experimental discipline is good...
Hello Sir, I would like to choose some mixes for testing the model after analyzing the full factorial DOE. Is there any way in which the program choose the mixes?? Thanks for the tutorial :).
Rajaa - ixture DOE's are very specialised. Personally I like to use a slightly different approach. If want more information on my DOE approach I've just released a new book...
' Design of experiments for 21st century engineers' you can get it from LULU.COM
Great video Paul. Just wondering how you retrieved 5 values (Y1-Y5) for each of the 8 tests - Did you repeat the test 5 times at each setting measuring each time? And, would it have been the same if as setting it up in minitab for 5 replicates and recording it once?
Reuben - yes if you select 5 replicates in Minitab it will create column that is 5x the 8 run pattern. 40 rows long. In my software it is in a table so that you can see the 5 data points for each set up..
I have an assignment about DOE of minitab, can I get your help if it is possible?
Zeliha - good to hear from you. Drop me al ine on paul.allen@allenp.co.uk and I'll see what i can do. Before you use mintab though you need to design your experiment. Watch this video first ua-cam.com/video/gsD8V2_eZ0A/v-deo.html
@@paulallen5321 Many thanks for your reply, I will send you an e-mail soon.
I saw this video.
Dear Paul, you are doing great and your videos are really helpful, I need help and suggestions from you, I have 9 parameters ( Factor) and 3 levels. What Model will be suitable? Please let me know.
Stark good to hear from you, you have 2 options. Run the L18 which will only take 7 variables at 3 levels, so you'll have to fix 2 of the variables. Or you could only test at the 9 variables at 2 levels and use a Taguchi L12. Both of these options are screening DOE's. It's very rare to model 9 variables as it's too complex and expensive. Remember we're not looking for perfect knowledge, we're looking for knowledge that allows us to make the process land on targets that we need to please the customer...