Seeking Help: Design of Experiments - DOE

[carvind]

Design of Experiments

How do I apply the concepts of designof experiments for cycle time reduction whether it is new product development or an existing manufacturing process. I would appreciate any inout on the topic. Thanks

 

[Jon O]

carvind,

Your DOE results are a starting point. Now that you have a "optimal" window you can run your process in at a lower cycle time, you need to test your results. Your DOE has showed that you can produce product within specification at your new settings however it has not proven whether you can produce capable product at the new settings. You need to go ahead and conduct a capability assessment on the new settings. If your capability study shows that the new process settings are capable, and you are confident in the new settings, present this to your team and move forward with the required quality system changes.

Regards,

Jon

 

[carvind]

Design of Experiments - Help

Please help me in creating a plan for implementing a DOE for the following situation.

The position of the leads on an electronic component is important to get a satisfactory mount of the component to an electronic printed circuit board. There is concern that with manufacturing an electronic tester of the component function is bending the component leads. To monitor physical changes from tester handling, the leads from a sample of components are noted before and after the tester.

(a) Create a plan for implementing a DOE that assesses what should be done differently to reduce the amount of bending on each component. Consider the selection of measurement response, selection of factors, and results of validation.
(b) Create a plan that could be used in the future for a similar machine setup.

As I am not familiar with the process, I would appreciate if someone can define the key process input and response variables for this situation.

Thanks

 

[Marc Richardson]

DOE

Let's start at the beginning. First of all, have you proved that the mountings are being bent during testing? I would conduct a T-Test on the before and after measurements to be sure that that is where the damage is ocurring. If there is a statistically significant difference, the T-Test will quantify the amount of change and then perform a correlation study to determine if the testing was making a preexisting condition worse or not. After that, we can identify the variables that effect the bending of the mounts.
Marc Richardson
Sr. Q.A. Eng.

 

[carvind]

Hi Marc,

Thanks for your response. Could you please suggest the possible input and output variables assuming that whatever you have suggested has been verified.