Love your videos - so information dense yet clearly explained. Brushing up my SPC knowledge both for work and for making my own RUclips video about it, but I already know I'll just be pointing people over to your videos for how to set up the graphs. Great stuff! There is one thing that I have some doubt about, hope you can help clarify it -> your control lines are based on the average value of the used datapoints (and even the number of probes in one sample, since that effects SE); wouldn't you normally decide on all control values beforehand and not change them unless you have a good reason to do so (like a real change in your proces setup)? In short, I would have expected the control limits to be fixed numbers instead of changing when new data is entered - what's you take on that?
Glad you found them useful! You bring up a good point/question! Here’s my take... First off, How would you determine limits before you have any data? Control limits are always calculated based on the data. They can however be locked after a period of time - say to include only the first 30 samples. As far as locking limits...If, when, and how you you lock limits is really up to the user so I didn’t talk about it. There are plenty of cases where I use a control chart without ever needing or wanting to lock control limits. It typically requires some process knowledge, sufficient time/data, and confidence of a stable process before you would look to lock limits. Locking limits could probably warrant a whole video in itself. There are also some good ways to do this within these spreadsheets dynamically.
@@TheEngineeringToolboxChannel I would imagine that changing control lines based on a different subgroup sample size is actually a good thing (although the effect should be minor when you're working with a group size big enough to use the S-chart). Of course we need a good set of data before setting limits in the first place. Just never really thought about locking or updating control limits based on the data you get when testing 'work as usual'. Indeed a nice topic for a next video :)
Hi, I have a Q... when I do the close and load instead off adding the columns it creates a new tab with only two columns instead of what u get on the video. What am I doing wrong???
Hi Glenda, thanks for the question! I am not sure why that is happening. It is normal for it to create a new table in a new tab, but it should have four columns. It would help a lot if i could see your workbook to determine what exactly is going on. Can you send it to theengineeringtoolbox@gmail.com? Id be happy to take a look and help!
People often confuse sample size (n) with number of samples. E.g. over a one month period 100 lots of 50 parts were produced. From each lot I sampled and inspect 5 parts (sample size n=5) Number of samples = 100 (each lot) Sample size (n)=5(# parts inspected from each lot) Total samples = sample size * # of samples = 500
Can you give me context on what your scenario is? What are you sampling? How often are you sampling? And how long does it take to gather the 200+ samples?
Sorry to give answer question with question, but typically if you are running into a situation where you have greater than 25 size, there is are opportunities to improve sampling methods.
You lost me at the 5 minute mark using a table that was not introduced in the video beforehand. I had no context. I would be nice to know what to do without use of the table lookups
Thanks for watching! What'd you think? I'd love to hear your thoughts
That was an excellent video and I learned a lot. How did you learn so much about excel?
Love your videos - so information dense yet clearly explained. Brushing up my SPC knowledge both for work and for making my own RUclips video about it, but I already know I'll just be pointing people over to your videos for how to set up the graphs. Great stuff!
There is one thing that I have some doubt about, hope you can help clarify it -> your control lines are based on the average value of the used datapoints (and even the number of probes in one sample, since that effects SE); wouldn't you normally decide on all control values beforehand and not change them unless you have a good reason to do so (like a real change in your proces setup)?
In short, I would have expected the control limits to be fixed numbers instead of changing when new data is entered - what's you take on that?
Glad you found them useful!
You bring up a good point/question! Here’s my take...
First off, How would you determine limits before you have any data? Control limits are always calculated based on the data. They can however be locked after a period of time - say to include only the first 30 samples.
As far as locking limits...If, when, and how you you lock limits is really up to the user so I didn’t talk about it. There are plenty of cases where I use a control chart without ever needing or wanting to lock control limits. It typically requires some process knowledge, sufficient time/data, and confidence of a stable process before you would look to lock limits.
Locking limits could probably warrant a whole video in itself. There are also some good ways to do this within these spreadsheets dynamically.
@@TheEngineeringToolboxChannel I would imagine that changing control lines based on a different subgroup sample size is actually a good thing (although the effect should be minor when you're working with a group size big enough to use the S-chart). Of course we need a good set of data before setting limits in the first place.
Just never really thought about locking or updating control limits based on the data you get when testing 'work as usual'. Indeed a nice topic for a next video :)
Nice work! Hope you dont mind, can you share the excel file for the constants table? Thanks.
Google “control chart constants”
Where can I find your table of constants for calculating the Sigma value?
There are many online, you can search “control charting constants”
Hi, I have a Q... when I do the close and load instead off adding the columns it creates a new tab with only two columns instead of what u get on the video. What am I doing wrong???
Hi Glenda, thanks for the question! I am not sure why that is happening. It is normal for it to create a new table in a new tab, but it should have four columns. It would help a lot if i could see your workbook to determine what exactly is going on. Can you send it to theengineeringtoolbox@gmail.com? Id be happy to take a look and help!
@@TheEngineeringToolboxChannel Thank u very very much, I will do so right now.
What if the sample size is bigger than 25?
Hi Fred, I would ask…why the sample size is that large?
People often confuse sample size (n) with number of samples. E.g. over a one month period 100 lots of 50 parts were produced. From each lot I sampled and inspect 5 parts (sample size n=5)
Number of samples = 100 (each lot)
Sample size (n)=5(# parts inspected from each lot)
Total samples = sample size * # of samples = 500
@@TheEngineeringToolboxChannel is there a way to calculate n larger than 25? My sample size is large thus i need to get the c4 for 200+ (n)
Can you give me context on what your scenario is? What are you sampling? How often are you sampling? And how long does it take to gather the 200+ samples?
Sorry to give answer question with question, but typically if you are running into a situation where you have greater than 25 size, there is are opportunities to improve sampling methods.
You lost me at the 5 minute mark using a table that was not introduced in the video beforehand. I had no context. I would be nice to know what to do without use of the table lookups