Cinnamond

This is an entry level problem, reads like it is straight out of a text book. You will learn much more if you look it up yourself. Give a man a fish…

Carolit, I found this short thread interesting, you seem to have the most information about the very question you are asking. If you already have a lot of experience in the subject then you would be a good person to ask the same question you asked. (My opinion is that design has outputs like: on time, on budget, problems were predicted, etc.…[Read more]

This is very basic, you should find a fairly simple explanation in your training manuals, or go have a discussion with your LSS mentor. You will be all the wiser…

Performance refers to the past. Capability refers to the future.

Weight is a Normally distributed variable. This may or may not be hidden in your data depending on how variable your product is (as you have already found out). The question to focus on is whether or not the resolution you have on your gauge is sufficient to keep your process in control (you didn’t mention your spec limits). If you cannot…[Read more]

These are basic probability questions that any first year student should be able to answer.   Oddly enough they read like they came from a text book, not a real life quality problem….  Good luck.

I would not try to find one metric since it would likely hide too much information.   Assuming you (and management) would use the metrics for CI, you should consider calculating several different metrics.   Certainly the overall scrap rate is great, throughput yields are nice and a metric that shows the scrap rate at each step individually woul…[Read more]

I would suggest that you actually have more than one distribution here.   There appears to be an event from data points 6 – 12.   These points may be from a different distribution in that something in the process had changed.   I understand that capability analysis is an exercise to show current state but in this case it may be fair to say that t…[Read more]

You do not have to use binary data.   If you have the before and after values, then do all your work on the difference.   This should be continuous and if done correctly will be a more powerful analysis.   Good luck

You will also need alpha level and the size of the difference you want to detect.   Minitab’s Power and sample size app works great for this.   Fill in all but one space and Minitab will tell you the missing value.

Just going back to the original posting and question…   One piece flow is definitely not always the answer.   It seems that this is actually somewhat of a misnomer (or maybe it’s a theoretical perfect state?) and causes confusion in some companies.   For example how could a metal stamper implement one piece flow?   The best approach to take wi…[Read more]

Use
xbar +/- N*1.96*(st. dev./root(n))/N

I wouldn’t want to disagree that the process needs to become stable and that Normally distributed data is desired.   Yet when we develop a new product or tooling or mold to make a product, what we have is what we have.   The purpose for calculating a Cp or Cpk is simply to find out what we might be “C”apable of in spite of how our data is curre…[Read more]