Using SPC charts for Verification of Job Setups

[chen]

"job setups shall be verified whenever a setup is performed(e.g.initial run of a job,material changeover,job change,significant time periods lapsed between runs,etc.)" - from QS9K 4.9.4 P37

"setup verification
a recommended method is to produce enough product to constitute a subgroup of the size used for spc.the parts are measured and the results are entered on the control chart.if these results fall within the central third of the control limit zone,the setup can be approved for production.if the results fall in the outer two thirds,a second subgroup should be manufactured,measured,and plotted.if this point falls in the same outer two thirds,the setup should be adjusted and this sequence repeated.if the point falls in the central third,the setup can be approved for production."
- from QS9K GLOSSARY P132

I want to use the method to verify the job setup of our press production.
but there have some questions:
1.control limit zone: USL-LSL or UCL-LCL ?
2.results: all datas of subgroup or Xbar ?
they used "results" and "point",so i do not know the meaning of these words.

pls tell me if you have any information about job setup verification.

have a nice day. Mar.,28/1999

 

[Batman]

Hi Chen!

We have used almost this exact definition for several years. Where appropriate, when a job is setup, the set up person takes a normal sample for the SPC chart and plots it. If the Xbar and sigma or range is in control, the set-up person initials that plotted point and the set-up is "verified." This only applies to set-ups that have run n the past, in control.

Yes, they are talking about the UCL and LCL. Quickly, your Xbar control limits are +/- 3 sigma from the central line. Imagine dividing the area between the UCL and LCL into thirds, three zones. The middle third is the +/- one sigma zone. If your first plotted group is within this middle zone, the process is allowed to go on. If the plotted point is in the next zone towards teh UCL or LCL, the requirement is to take another sample and plot that one. If that sample is in the second zone, an adjustment is required. If the first sample plotted is in the outer zone, that is near the UCL or LCL, an immediate adjustment must be made, and start again.

The range is normally assumed to be (or remain) in control from the last run. In general (note, in general), if the range chart is out of control, this is a machine or tooling problem, one that normal 'post-setup' adjustments cannot fix.

This is of course one "recommended" method, using an SPC chart. There are certainly other methods, and they may work for you.

 

[chen]

Hi batman:

Thanks for your information!

I also have some questions:
1)do you use Xbar and r chart for job setups
at the same time and verify the job setups
by using all the results from Xbar&R chart?
2)this is the first time we use this method
to verify the job setups.so i use the spec
limit(usl-lsl) zone as the control limit
zone.how do you think so? if you have any
good ideas for this case pls tell me.

thanks a lot/have a nice day. Mar.,31/99

 

[Batman]

Hi chen!
Sorry for the delay, I did not realize you had further questions.

If you are going to use the SPC chart for job set-up verification, take the last chart used and plot the values from the samples from the set-up. This plotted point should be "in control" according to the rules stated above. If it isn't, adjustments should be made towards moving "in control" and another plot made. The SPC chart will already have upper and lower control limits set. You do not use any sort of "specification" chart here.

Also, this is for product that has been run before. New product - first run stuff - needs data collected first before establishing control limits. "Set-up verification" does not apply here. New part data has to be recorded first, then control limits calculated.