Update: To read Craig Gygi’s responses, visit these two posts:
- Original Questions posed to Craig Gygi
- Craig Gygi answers Question on Process Capability
- Craig Gygi answers Question on Accelerated Life Testing
On May 25, I announced a contest whereby the readers of shmula could enter and win a Six Sigma for Dummies Workbook. In the proceeding posts, I’ll be posting Craig Gygi’s responses to those questions. Today are Craig’s comments on methods of achieving failure quicker on product or component testing and on developing supplier quality with your suppliers.
Some background first on the contest and to view the winners, please go here and Craig’s previous answers can be found here.
Below were the contest details:
Submit your Six Sigma questions and enter a chance to win Six Sigma Workbook For Dummies, an industry bestseller since its publication in 2005. We are fortunate to have Craig Gygi, the lead author of Six Sigma for Dummies and the companion Six Sigma Workbook For Dummies to answer reader’s questions for this raffle.
- Submit your question for Craig in the comments of this post.
- Comments will be turned off on June 8, 2007
- We will enter all the commentators names on my handy-dandy randomizer, using a hard-to-crack random seed, and 5 randomly selected contestants will win a Six Sigma Workbook For Dummies book.
- On June 11, 12, 13, 14, and 15, I will publish Craig’s answers to some or all of the reader’s questions.
- That’s it.
Accelerated Life Testing (ALT)
Gary Petersen said,
June 9, 2007 @ 4:30 am
I realize it is after June 8, but I’m finally getting around to asking other questions.
We have a test process that takes upwards of five days to run on a part. It is a wear test that just takes lots of cycles to complete. Performing a measurement systems analysis on that process seems highly impractical because of the time involved. Is there anything I can do to get a statistically valid feel for how capable my measurement system is without running many samples to failure?
We currently only have the ability to measure when the coating is worn through and we can detect the base metal. We cannot measure degrees of wear, say to know that we are half-way through the coating. Can we, for example, replace the test media with something incredibly more aggressive to get to failure quicker and make some
assumptions or comparisons with the actual process?
The idea you describe is often called accelerated life testing (ALT). But whether you’re doing the full or accelerated test, you need to determine if the test regimen validly simulates real field performance. For destructive tests like your wear test, you should plan on doing paired tests, where you take a group of equivalent-as-possible parts and divide them into two subsets—one to be exposed to real field conditions and the other to be sent through the wear test regimen (either full or accelerated). Then paired statistical tests can be performed between the subsets to confirm whether the wear test validly reflects real field conditions or not.
Gary Petersen said,
June 9, 2007 @ 4:34 am
Many of our suppliers are from across the oceans, so they are a long and expensive plane ride away. Most have not had Six Sigma training. What would you recommend for options to help them get started with either an entire Six Sigma program or with key elements to help them better understand their manufacturing processes? How can we help from so far away when we cannot send someone to each place for weeks to do the training?
What would you suggest, understanding that you don’t know our processes or issues, as the key tools for a supplier to start with? Is there a resource that could work with them remotely to help? Does something like that have any shot at being useful?
You need to give your distant suppliers incentive to enter into a Six Sigma effort on their own. This motivation can come in the form of granting preferential status to your Six Sigma-qualified suppliers, or it can originate from educating them on the financial benefits they’ll realize themselves. A critical point to know is that most suppliers don’t need to begin with a full/traditional Six Sigma deployment. But critical initial skills include topics like: collecting measurements, quantifying variation, and calculating capability. Subsequent topics will naturally follow from this start. Six Sigma is mature enough that resources and helps can be found throughout the world—even in remote areas of Fiji! I suggest the best way to train someone remotely is through writing—books, articles, field guides, emails, etc. Writing is extremely powerful; it is a thought transmission device that spans space and time. Through reading the writings of Deming, for example, your suppliers can access the thoughts, ideas, and methods of one of the giants of improvement and quality for themselves. Books like Six Sigma For Dummies and Six Sigma Workbook For Dummies give newcomers to Six Sigma accessible explanations of all the Six Sigma tools and methods and provide practice problems to develop competency and mastery.