Page 125 - Design for Six Sigma a Roadmap for Product Development

100 Chapter Three

x s
x ns

m 1 y 1
y
m 2 2
m 3 Design Entity y 3
• •
• •
• •
• •
z ns
z s
Figure 3.19 The P-diagram.

symptomatic behavior as perceived by the customer. Let the array x be
split into significant and nonsignificant factors denoted as {x s ,0} and
nonsignificant factors {0, x ns }, respectively, that is, x {x s ,0} {0,
x ns }. Also, let the array z be split into significant and nonsignificant
factors denoted as {z s ,0} and nonsignificant factors {0, z ns }, respec-
tively, that is, z {z s ,0} {0, z ns }. Of course, significance and non-
significance are subject to physics by derivation or empirically from
experimentation.
Usually the nonsignificant factors are numerous while the signifi-
cant factors are few, assuming their existence. There are four possibil-
ities of a DFSS project from the standpoint of design-versus-noise
classifications in the context of this section. They are listed in
Table 3.2. The effects of nonsignificant factors, whether design para-
meters or noise factors, are usually weak and sparse in a manner that
bears creditability to the Pareto principle. As such, their existence does
not add to the complexity of the problem and/or its solution.
Only when the significant x s array does exist is there a potential for
the DFSS method to produce a Six Sigma capability in the concerned
FR. The conceptual change in the third classification is to be conducted
following the DFSS strategy proposed in this book.

TABLE 3.2 Possibilities of a Six Sigma Problem
x z z s exists z s does not exist
x s exists Six Sigma and DFSS have Trivial problem—may be
potentials solved by a DOE only
x s does not exist Need conceptual change—DFSS No problem—such design
has potential while Six Sigma entity may not exist
has no potential