SPC Short Run Advice of a Stamping Process

Plegreid

Registered
Looking to dive into SPC of a stamping process;

The end result is a welded Aluminum part.

We stamp one half in the morning conduct a tool swap (changeover) and stamp the opposite half in the afternoon.

Reading through the various books about SPC our process is considered a short run. Since collecting the data would not be complete to after the run is finished.

Based on the General guidelines I would need to complete 12 subgroups of 5, collecting every half hour.

The Key Path variable I want to study is the profile of the flanges, the profile has a direct affect on the welding operation when the halves are welded together.

Creating a program and accurately measuring this profile would take our only CMM offline to other product.

Does anyone have ideas on completing this study? Or is there a better option that i missed?
 

Steve Prevette

Deming Disciple
Leader
Super Moderator
Re: SPC Short Run Advice

You say you one half in the morning, swap tools and stamp the other half. What do you then do "tomorrow"? The key is looking at what is continual or at least repetetive in the process and to stay out of the Short Run SPC trap.
 
D

ddouglas

Instead of doing SPC try looking into other means of looking at your process, maybe pre-control charting??
 

bobdoering

Stop X-bar/R Madness!!
Trusted Information Resource
I recommend following the proper steps of determining what you need to track. See this blog. Traditional Shewhart charts monitor the process looking for unusual or "special cause" conditions. You need to sort them from the common cause variation. What special causes would you expect to see? Tool damage? Excessive tool wear? How would it manifest itself? Dimensional changes? Process variable changes? Do set-ups cause variation? Do you put all data for each die on a chart to track tool wear (which is generally a long process in typical stamping.) If you end up tracking tool wear, then you need to understand X hi/lo-R charting and not attempt to use Shewhart charts. They are useless for tool wear.

The LAST thing you want to do is attempt to rubber stamp any type of charting methodology on a process, short term or long term (even though historically this has been the favored approach.) Determine what tool will allow you to make decisions that are meaningful to the process.
 
Top Bottom