Quality in Design - Medical Device Design: Innovation from Concept to Market - page 130

The next question is: What do we do with this data? There are two interpretations. The first
concerns minimizing manufacturing errors. If you want to make your quality control more
robust then you must examine the outliers and see how to best control them or remove them
from the design. Clearly, if you are able, then these should be tightly toleranced. All others
have a small effect but your level of control need not be so high, if at all.
The second way of looking at the data is if you are trying to maximize the effect; now the
data tells you which parameters are the most important and hence which you should vary to
improve your outcome.
As you can see, 2 k experimentation (or “design of experiments”) is an extremely powerful
tool. I have only presented the most basic of introductions; if you wish to understand more
then a good text is Montgomery (2001) . If you would like to play with this method conduct
a web search for “Taguchi Paper Helicopter” and you will find links to a great example of
“design of experiments.”
7.4 House of Quality
We have appreciated the importance of customer/end-user input into design. But how do we
know that we have really taken them into account, and how can we use this knowledge to
differentiate ourselves in the marketplace? One of the most valuable tools is the House of
Quality (or HoQ as I shall now term it). Figure 7.7 is a schematic of an HoQ structure.
The HoQ is split into zones. The first “room” is customer requirements; in this room the
individual requirements, as detailed by your customers and end-users are tabulated as
Cross-talk
Technical specs
Correlation
Targets
Figure 7.7
Typical House of Quality structure.