Page 236 - Design for Six Sigma a Roadmap for Product Development

DFSS Transfer Function and Scorecards 207

■ The scorecard is driven by the DFSS algorithm, and concepts such
as coupling and loss function will be stressed.
■ Noise factors and some times external signal factors can’t be speci-
fied by the design team even when knowledge about them is readily
available.
■ We are using the physical structure for illustration purposes only, and
the discussion here applies equally for process structure scorecards.
Both structures set of scorecards are needed in the DFSS algorithm.

Notice that we numbered the entries with the scorecard for ease of
reference in Fig. 6.14 (where DPMO is defects per million opportunities).
After documenting the hierarchical level and scorecard scope in terms
of the design mapping, the team needs to populate the scorecard with
the following entries (where the listed numbers correspond to column
numbers in Fig. 6.14):

0. Axiom 1 measures, which include type of design mapping in the
structure addressed by the scorecard, reangularity (calculated)
estimate using Eq. (6.4) and semangularity (calculated) estimate
using Eq. (6.5). Both measures accurately estimate the degree of
axiom 1 satisfaction, in particular, when nonlinear sensitivities
exist. Entries of these two equations are documented in the sensi-
tivity matrix in column 19 in Fig. 6.14.
1. List of all FRs within the scorecard scope as represented by the
transfer function and design mapping.
2. Units used per the FRs; measurable and continuous FRs are
expected.
3. The distribution of each FR is documented. Usually, “normal” is a
safe assumption per the central-limit theorem (CLT) when the
number of (P K L) is greater than 5.
4. The transfer function equation per each FR is entered in this col-
umn. The team can make use of column 20 (Fig. 6.14) and the sen-
sitivity matrix in column 19.
5. FR type according to robust design is documented to indicate the
direction of optimization. Robust design optimization requires the use
of one of four classifications of responses.These quality characteristics
are classified by Dr.Taguchi as “the smaller, the better” (e.g., minimize
vibration, reduce friction), “the larger, the better” (e.g., increase
strength), “the [most] nominal, the best” (where keeping the product
on a single performance objective is the main concern), and “dynamic”
(where energy-related functional performance over a prescribed
dynamic range of usage is the perspective).