Crombes
7th January 2009, 03:37 AM
We are using XR charts to monitor our productionprocess
we begin to notice that the R values of the chart are getting bigger & bigger
( so a bigger difference between the 5 measured parts) i was wondering how this can be related to the production process ( The X values are still within spec)
So basically my question is : A big(ger) R value is an indication of what ?
can anyone help out ?
Thx
Miner
7th January 2009, 07:47 AM
First question: Has the Range chart triggered an out-of-control condition?
If the Range chart is definitely out-of-control, the short term variation of your process has increased. This may be due to a number of things. You did not identify the process, so it is impossible to speculate.
Some examples could be: wear on equipment, increased variation in a process input variable, etc.
pagnonig
7th January 2009, 09:47 AM
What kind of characteristic are you monitoring? How and when do you collect the data?
This may assist in giving an explaination of higher R values.
Craig H.
7th January 2009, 10:23 AM
How did you arrive at n=5?
David DeLong
7th January 2009, 10:25 AM
We are using XR charts to monitor our productionprocess
we begin to notice that the R values of the chart are getting bigger & bigger
( so a bigger difference between the 5 measured parts) i was wondering how this can be related to the production process ( The X values are still within spec)
So basically my question is : A big(ger) R value is an indication of what ?
To answer your question, the range correlates to the estimate of the standard deviation. The larger the range, the larger the spread or larger the estimate of the population standard deviation.
Now that I have said that, I don't believe that your SD' is increasing. It may have something to do with your sample size not being homogeneous. Taking 5 samples on a 12 station machine is an example but your averages should also jump all over.
Your measuring method should be confirmed. You really should take an R & R study and get something pretty low such as less than 10% of the tolerance width or process if you are automotive related. I have to admit that the R & R studies which were automotive driven are extremely helpful.
bobdoering
7th January 2009, 11:11 AM
It really does depend on your process and how the characteristic relates to it to determine the meaning of that signal - if a meaning can be determined. :cool:
Daniele Italy
7th January 2009, 12:28 PM
I agree with all the other guys statements: it depend to the process. In my experience I noticed that if the R increases too much it is related to some problems on the machine equipment mantainance....
Frictions of moving parts and or other not well maitained parts, or sometime also the strating of wearing of clutchs ....:notme: