The ideas of SN Ratio , Two Step Analysis , and Classification of factors are made to realize Robust Design . The outside arrangement of orthogonal array is used to realize these ideas.
The figure, "Data analysis of two unevenness", is the basic style to estimate unevenness. In this case, same sets of conditions are repeated.
The repeat the set of condition is necessary to evaluate unevenness. For example, we cannot divide the effects of "the unevenness by the repeat of A1" and "difference between B1 and B2". (General style of Design of Experiments often adopts the latter.)
The outside arrangement is used to evaluate Error Factor .
The figure, "Outside Arrangement of Error Factor", changes the level of the error factor. The process of data analysis is same to "Data analysis of two unevenness".
Repeat of the experiment is not done because of the idea that "Unevenness is strengthened by the error factor."
The other use is to evaluate signal factor . The figure, "Outside Arrangement of Signal Factor", changes the level of the signal factor.
For example, V1 is correlation coefficient calculated by (D1, Y11), (D2, Y21), (D3, Y31).
More than 3 levels are necessary to evaluate the signal factor.
The figure, "Outside Arrangement of Error Factor and Signal Factor", is to evaluate both factors. When you study the dynamic characteristics by some books, you may meet this table as the first step to learn.
SN ratio of dynamic characteristics in the page of " Characteristics and SN Ratio " is to evaluate only the signal factor. But many books show the SN ratio for the mixed case of the error factor and the signal factor.