| | | | | WHAT IS THE BEST ANGLE FOR THE SQUEEGEE BLADE TO BE ANGLED AT? WE ARE USING 45 DEGREE ON OUR POLYURETHANE BLADES, AND I BELIVE WE HAVE THE OPTION OF EITHER 45 OR 60 DEGREES ON OUR METAL BLADES. WHICH OF THE TWO IS BETTER, OR SHOULD WE BE USING ANOTHER ANGLE? | | | | Hi guys, | | | | Sounds sick but I miss this stuff. I saw the explanation and feel as though you have all been cheated. I'm not trying to whack you Ben but I bet you had Shainen into your facility to put on a little seminar. A word to the wise, you cannot learn DOE in one day. I shouldn't say that. Maybe there's someone out there who can but I know that I've taken 2 semesters worth of course material at the graduate and undergrad levels and I'm still learning more about this tool every time I use it. The purpose of DOE is to avoid the full factorial experiment (they're too costly). Imagine trying to run an experiment with every combination of 16 parameters. DoE was designed to pair down the number of factorial combinations in an experimental set, then through an Analysis of Variance, you are pointed in "the right direction." DoE can help you to improve your process average or your process variability or both depending on how it's used. You would perform an average analysis to improve process average and a signal to noise analysis to improve process variability. Guess what that means, Yup. Cp and CpK has yet another purpose. Find your process capability, Cp, and also how your process is really running, CpK, and you will know whether or not you need to improve your process average or your process variability. There are several "gurus" out there who have different methods of going about DoE. Shina, Shainen, Phadke, and Taguchi are just a few. Personally, I'm fond of the Taguchi method but some say it's not statistically strong enough. I disagree for several reasons, sound performance and great results being the two most influential. My advice to you would be to pick up a book on this stuff. Once you learn how to use and manipulate the experimental matrixes you have a very powerful engineering tool in your belt. Just when you thought you were safe, it always comes back to "sadistics." | | | | Oh yeah, Ben. If you want to break out of the "no learning" rut. Look a little harder into DoE. You'll be able to learn a lot more even if your company's process is stagnant and unchanging. | | | | Best Regards, | | | | Justin Medernach | | | So Justin, if I'm looking in the SMTNet or some other on-line book store, what book would you suggest for someone wanting to take a practicioner's approach to learning Taguchi? Dave F | | Dave, | | I looked on the SMTnet and found nothing on Taguchi DoE. I am not familiar with books on Taguchi method as I learned it in class. I have a number for a gentleman that can help you out. His name is Prof. Sam Shina of the University of Lowell. He sits on the NEPCON board of advisors and I was lucky enough to have him for a few courses in college. We've remained friends over the years and I'm sure he'll help you. His number is 978 934 2590 at the university. Give him a buzz. He'll probably try to sell you on a consulting gig. It's a great seminar but i'm not one for plugging people. | | regards, | | Justin Medernach | Dave, | Check out Amazon.com and type in "taguchi" under the subject search. You'll find all the info you need. | Justin Justin: Thanks. Chrys also recommended Prof Shima to me. I found one of those "quickie" discussions on DoE at the Reliability Analysis Center (link below) that recommends several Taguchi books. I have the one written by Ross. It's a dust collector. I just don't have the time to wade through all of his "stuff" to try to figure-out what he's talking about. Dave F
reply »