It's easy enough to use excel to calculate mean effects and signal to
noise and plot the results. It takes more effort to use excel to
calculate residuals so most of the time I use Minitab. Not necessarily
because it's the best, but because it's what I have available. Minitab
has the capacity to analyze Taguchi arrays.
Whether an experiment was set up as a "Taguchi" design or not, I always
analyze them as plain old fractional factorials. Not to start a
religious debate, but I don't care for signal to noise ratio because I
don't have an intuitive feel for it. I find it much easier to make
judgements about the practical effect of a variable when I'm dealing
with units I care about (mm, newtons, degrees, etc). I like to look at
the main effects and two (sometimes 3) way interactions on both mean and
standard deviation. Seeing the effects of the factors on standard
deviation is easier for me to grasp than S/N. I just create an extra
column in my worksheet and calculate standard deviation for each
treatment combination. I still calculate S/N for my coworkers who like
to see it.
Any statistics package should be able to analyze a Taguchi design
whether it has that feature built in or not. Every modern package I've
used have functions to do logs and sums of squares which are all you
need to calculate S/N. Just create an extra column for the S/N
calculation and point the software at that column to calculate main
effects.
Scott Post
P.O.U. 1415-00M2-00P6
2151 E. Lincoln Road
Kokomo, Indiana 46904-9005
765-451-2983 (Phone)
765-451-0287 (FAX)
-----Original Message-----
From: TechNet [mailto:[log in to unmask]] On Behalf Of Ken Bloomquist
Sent: 01 June 2009 15:42
To: [log in to unmask]
Subject: Re: [TN] Taguchi
Hi Rex,
I too have used Taguchi in the past and found it quite useful. I had
some old DOS software once that didn't work very well. I'm curious what
software you use and how user friendly it is.
Thanks,
KennyB
-----Original Message-----
From: Rex Waygood [mailto:[log in to unmask]]
Sent: Monday, June 01, 2009 12:25 AM
To: [log in to unmask]
Subject: Re: [TN] Taguchi
I don't think I could give you a step by step approach to Taguchi but I
recommend it as an experimental method.
We were waving 50,000 assemblies a week when the customer introduced a
new design. There was an RF can surrounding the pcb which had to be 100%
soldered to the edge of the pcb. Initial work showed that the board was
too small for the can and we had 8 people per shift hand soldering the
bits the wave missed. The can was manufactured using a progressive tool,
megga money to change, the pcb was punch and crunch, expensive to
change. Attempts to improve the wave process failed so we called in two
wave consultants. One was kind enough to say it was impossible at the
start of the day and didn't charge, the second spent a day, did charge
and said it was impossible.
So we did a Taguchi and our first attempt got us down to 1.5 reworkers
per shift. Further honing of the Taguchi plus some other improvements
got us to 0.5 reworkers per shift.
In terms of understanding the wave process and being able to improve the
process, Taguchi is fantastic.
I recommend that you carefully tell your operators why you are doing the
Taguchi and how it works. We had a later Taguchi sabotaged by a wave
operator who didn't like running the machine on settings he knew would
produce rubbish, he doesn't work here now!
You have to gain the co-operation of your production department as to
get points on your graphs you will have to build product that is 'bad'.
You must carefully decide what your attributes are to be so opens,
shorts and pull through might be your choice. On our first one we were
only concerned with the % of the frame soldered to the board. The beauty
of Taguchi is that if you change your mind about the importance of the
attributes you can change the machine settings in line with a paper
exercise on the Taguchi.
Picking your independent variables for a wave is not easy and is best
done in a brain storm once your team realise the definition of
independent variable. I.e. preheater setting is not independent, topside
pcb temp, is as conveyor speed would change topside temp for a given
preheat setting.
We have some software which helps with the arithmetic and array choice.
I really do think Taguchi is a great tool.
Rex
Rex Waygood
Technical Manager
Hansatech EMS provides value manufacturing through engineering and
quality
Hansatech EMS Limited
Benson Road
Nuffield Industrial Estate
Poole
---------------------------------------------------
Technet Mail List provided as a service by IPC using LISTSERV 15.0 To
unsubscribe, send a message to [log in to unmask] with following text in
the BODY (NOT the subject field): SIGNOFF Technet To temporarily halt or
(re-start) delivery of Technet send e-mail to [log in to unmask]: SET
Technet NOMAIL or (MAIL) To receive ONE mailing per day of all the
posts: send e-mail to [log in to unmask]: SET Technet Digest Search the
archives of previous posts at: http://listserv.ipc.org/archives Please
visit IPC web site http://www.ipc.org/contentpage.asp?Pageid=4.3.16 for
additional information, or contact Keach Sasamori at [log in to unmask] or
847-615-7100 ext.2815
-----------------------------------------------------
************************************************************************
****************
Note: If the reader of this message is not the intended recipient, or
an employee or agent responsible for delivering this message to the
intended recipient, you are hereby notified that any dissemination,
distribution or copying of this communication is strictly prohibited. If
you have received this communication in error, please notify us
immediately by replying to the message and deleting it from your
computer. Thank you.
************************************************************************
****************
****************************************************************************************
Note: If the reader of this message is not the intended recipient, or an employee or agent responsible for delivering this message to the intended recipient, you are hereby notified that any dissemination, distribution or copying of this communication is strictly prohibited. If you have received this communication in error, please notify us immediately by replying to the message and deleting it from your computer. Thank you.
****************************************************************************************
---------------------------------------------------
Technet Mail List provided as a service by IPC using LISTSERV 15.0
To unsubscribe, send a message to [log in to unmask] with following text in
the BODY (NOT the subject field): SIGNOFF Technet
To temporarily halt or (re-start) delivery of Technet send e-mail to [log in to unmask]: SET Technet NOMAIL or (MAIL)
To receive ONE mailing per day of all the posts: send e-mail to [log in to unmask]: SET Technet Digest
Search the archives of previous posts at: http://listserv.ipc.org/archives
Please visit IPC web site http://www.ipc.org/contentpage.asp?Pageid=4.3.16 for additional information, or contact Keach Sasamori at [log in to unmask] or 847-615-7100 ext.2815
-----------------------------------------------------
|
