Minitab and Control Charting Time to Complete each Repair Over Time

I have a need ot control chart repair data time to complete each repair over time. Even Minitab's instructions of I-MR charting suggests you should not change the UCL and LCL. How do you continue to add data to Minitab without changing the parameters. I thought it does this automatically and is why we use the SW< but what if i want to keep the UCL and the LCL a constant until there is a dramatic change in the process, or some other statistically relevent reason to change the UCL and LCL. Does anyone know how to do this in Excel?

If I have 12 benchs to monitor this way who are doing 20 repairs/day. Ho wlong do I need to collect data to baseline the UCL and LCL. Its hard ot fathom 25 points of data. I am thinking more like 1-2 weeks approx 1200 data points/week. How does one determine how long to colect data for before alllowing the UCL and LCL to become constant as Minitab v16 help tutorials suggest??

Is there any info on how ot predict the future goals for these repair techs using Control Charts? I have seen some data on this, but I am a bit numb on how much is out there (including here) on correct methodolody. Can I add an expected 10% improvement (repairs/day) to these predictions in Minitab assuming we change the process to make it more efficient.

I don't know how minitab does this, but JMP allows you to heard enter the LCL and UCL. On teh rare occassions when I use JMP for a control chart, I first enter the data that represents my stable process, have JMP calculate the limits, then I go in and hard enter those limits into the space allowed.

I'm sure a Minitab expert will be along soon to tell you how this is done in Minitab.

Hi, Hami812 and Bev D -

This article entitled "Optimize the Performance of Your Control Charts" on the Minitab site includes directions for setting control limits. There are two ways to do it, depending on what you're trying to accomplish. Here are pertinent excerpts from the article:


Since I don't yet have five posts here, I can't include the URL for the article directly, but a quick google search on "Optimize the Performance of Your Control Charts" will take you right to it.

StatsQuo said:

.....Since I don't yet have five posts here, I can't include the URL for the article directly, but a quick google search on "Optimize the Performance of Your Control Charts" will take you right to it.

Click to expand...

OK, here's the link:

Optimize the Performance of Your Control Charts

Hey I found it. Thanks so much. The more I delve into Control Charting and SPC in general the more I see how this is a completely seperate branch of Stats that could be a subset of Stats in itself

JMP???? I am not familiar

Bev do you find that being allowed to enter limits directly can potentially allow users to enter the specifications instead of control limits. I am always wary of that.

If you want the equivalent use in Minitab 16, use the assistant control charts, and the I-MR in there either allows you to estimate the limits from the data or put fixed limits in there.

But make sure they are control limits, not specs.

Allattar said:

....do you find that being allowed to enter limits directly can potentially allow users to enter the specifications instead of control limits. I am always wary of that.

Click to expand...

Not really...Mainly because we use JMP as an offline tool for engineering analysis. Our on-line control charts have governance and control limits aren't allowed to be spec limits. Although we do have spec limit charts and goal charts we are very clear which are control limits and which are specifications.

We still struggle some sectors that apply spec limits or 'action limits' - drives me crazy that 30 years after Deming's comeback we STILL don't get this fundamental concept...

Drives me crazy too, that and something multiplied by 1.5