I’m curious as to how others document reactions to Out of Control situations

more specifically, do you react to 7 point Runs, if so how do you handle a reaction to a process that cannot be adjusted there but rather must be adjusted upstream at an earlier process, especially if the upstream process seems to be running in control and does not merit an adjustment.


Because of some of the functionality of our new SPC software, we get a real-time calculation of the current process’s Cpk so we can monitor when it begins to drop below our target of 1.67


Because of this we have implemented a procedure that our operators document the current Cpk and are allowed to not make an adjustment to a run (as long as its still with in the Control Limits) if our Cpk is above 1.67

Just looking for others opinions on this approach, and how they react to similar situations.

Why would you want to make an adjustment to a run, if its within the control limits and you have a CPK 1.67.

What type of machine (or industry)

"more specifically, do you react to 7 point Runs, if so how do you handle a reaction to a process that cannot be adjusted there but rather must be adjusted upstream at an earlier process, especially if the upstream process seems to be running in control and does not merit an adjustment. "

I would think if the upstream process is running "in control" and has a CPK of 1.67 then its within specification. If the engineering was done correctly and the control limits were set correctl, then the next process should be within tolerance.

Check for the attachement and Wheelers article, I added on another tread.

http://elsmar.com/Forums/showthread.php?t=6572&highlight=darius

There is no ONE SIZE FITS ALL. Use discretion, not all rules apply to any case.

But think if you need to reduce variation on the upstream process (you may need), one thing is that it's in control, the other that can be improved to reduce variation on the downstream process.

:bigwave:

I have little help here but questions also.
We are a metal stamper how do you run capability charts on a part feature that is affected by metal thickness varriation? we have parts that are formed into a "U" and the customer asks us to provide capability info on the OD of the legs. It changes based on stock thickness so it kills long term capability studies. Any wisdom on this?
Mark

Use a run chart or a historgram.

Process capability is how well the process is currently behaving in regards to output specifications.

Also start using subgroups or if you have to operator precontrol

If you know the relationship between thickness and OD you can translate the measured value to a calculated to a "standard" thickness. So you measure to a thickness and you chart the value at the "official reporting standard".
:bonk:

If you know the relationship between thickness and OD you can translate the measured value to a calculated to a "standard" thickness. So you measure to a thickness and you chart the value at the "official reporting standard".
:bonk:

I agree with your saying. But adding additional inspection points for each piece would not be feasable for the stamping enviorment.

Sampling , subgroups...better way

I agree with your saying. But adding additional inspection points for each piece would not be feasable for the stamping enviorment.

Sampling , subgroups...better way
Well the other thing to consider is that with these parts material chemistry is as important as thicknes. I am not really sure how I got my job being that I have no prior metal forming training but I am picking it up. The YTE and N along with the carbon levels in the steel all affect the part form too. So I guess I am wondering about keeping a process controlled that has so many variables from one heat of steel to the next. On one of these parts we need Boron in order to keep it from getting compression cracks. It all plays a factor and I am not sure how a run chart or histogram with subgroups is going to take out or limit my variation from one master coil to the next. depending on where steel is available from we may be high for a day or a month then swing down low sweet chariot for the same undetermined time. It kills my charts. My cpk stays where it needs to because we have put limits on the steel thickness to reduce our process variation but we still have many points out of the control limits all the time.

So if thats the case then use the system Darius wrote. If you have to test each picece you can take the time and measure it.

Nice job Darius