Die Switchover Control Chart
- December 3, 2017 at 5:47 am #55887
Hi everyone, cuurrently I am constructing individual and moving range Control chart for time consumed during exchange of die in Extruder line. As current state is highly unstreamlined and unpredictable there happens to be wide Control limits for individuals due to high range average. Consequently all points drop within control limits. How How to evaluate that kind of process? Poor but stable?December 4, 2017 at 7:04 pm #202033
It appears that your control chart is confirming something you already knew. A basic principle is to use DMAIC. Control charts are most useful during C. In short, find out why the range is so wide and do something to narrow it.December 5, 2017 at 6:12 am #202034
@Straydog Hope you were trying to generate blog traffic. There are some very passionate folks that LOVE using SPC as an investigation tool in the Measure/Analyze phases.
I have used it in the M,A,C phases myself. I find the use in measure phase QUITE useful when the folks have no ideas about variation causes or sources in their process.
Thanks for the banter as always.December 6, 2017 at 9:34 pm #202040
@cseider You might convince me that SPC is useful in the A phase, after you have reasonably reliable data. Before that a control chart with questionable data can lead you down the garden path.December 7, 2017 at 8:27 am #202043
For this sort of thing, I like to just use the average time. The system isn’t IN control yet, so a control chart won’t be able to tell you when it goes OUT of control. In that case, some good old descriptive stats might get the attention of your audience.
* The Range is 45 minutes with some changes taking 10 minutes, and others taking 55 minutes.
* The Mean time is 30 minutes.
* The Standard Deviation is 18 minutes.
* The Future State goal is 9 minutes.
Descriptive Statistics are pretty easy for most folks to understand.December 7, 2017 at 9:05 am #202044
I don’t understand the statement “SPC won’t tell you when the system is out of control”.
I suspect I know your intent but SPC is a tried and true technique to know when things are statistically showing evidence of being out of control (with an alpha error of course).
Yes, your statement that descriptive statistics are great but don’t forget that time series plots and histograms are super also to help understand.
You must be logged in to reply to this topic.