We manufacture gears that are manufactured on a CNC lathe that produces 48 pieces from 1 bar. Therefore, in my study I will have 3 different bars within my 100 pieces. I have been told not to make any adjustments for the new bars and I have heard to make adjustments since this is a known cause. If I do not make any adjustments you can tell on the x-bar chart where the new bar starts. What is the correct method for this study?

A capability study is done to determine the variation in a normally run process. You should make no adjustments OTHER THAN adjustments which are normal and intrinsic in the process. If your process includes re-setting for a new bar then in my opinion it should be done during your study. The point made about not making any adjustments is meant to keep people from "tweaking" the process during the run. It isn't meant for restricting your normal process. If for example you were applying paint to something, you would certainly continue to refill the paint reservoir. To run the paint dry avoiding an adjustment would make no sense.

If there is any doubt, ask your customer.

Dave

We manufacture gears that are manufactured on a CNC lathe that produces 48 pieces from 1 bar. Therefore, in my study I will have 3 different bars within my 100 pieces. I have been told not to make any adjustments for the new bars and I have heard to make adjustments since this is a known cause. If I do not make any adjustments you can tell on the x-bar chart where the new bar starts. What is the correct method for this study?
To add to Dave's good advice, it's a good idea to make a note of adjustments made and material changes along with the time you make them so that someone looking at the charts (or other records) later on knows what happened.

Thanks for the advice.

Thanks for the advice.
And..... Welcome to the posting side of the Cove :bigwave:

I think you could study the bar's capability first , if the capability is ok , then , you mustn't to change the control line , otherwise, you should improve the bars manufacture process. Just for reference.
p.s : I came from china and my english isn't very good , so , if some mistake occured in my sentence, please tell me how to revised . Thank you very much.

I am making an assumption that you manufacture few hundred pieces per day and you mentioned that one batch of raw material makes 48 jobs. This means that your shop appears to be consuming multiple bars on same day.

In such cases, calculations of capability from one bar only out of 48 pieces is likely to give more uniform products and therefore tighter control limits which also means false security that your process is better than what reality is.

A better approach can be is that you decide what to call as short term. If you decide that one day's production is considered as short term and if you are making 480 jobs per day ( 10 bars), you may take 4 samples from each bar to capture maximum sources of common cause variation like small material variation, tool wear, start up or warmed up machine etc to get a true picture of the process.

We manufacture gears that are manufactured on a CNC lathe that produces 48 pieces from 1 bar. Therefore, in my study I will have 3 different bars within my 100 pieces. I have been told not to make any adjustments for the new bars and I have heard to make adjustments since this is a known cause. If I do not make any adjustments you can tell on the x-bar chart where the new bar starts. What is the correct method for this study?

This is the wrong way to do a capability study for precision machining. To do it correctly see:

The Statistical process control for precision machining (http://elsmar.com/Forums/blog.php?b=79)