Quality engineering(QE) 20

This is the explanation material. →　QE20

p.1　Many people who are unfamiliar with quality engineering（QE) and those who are skeptical about QE say, "Is the number of experiments N = 1 all right for each row of the orthogonal array?". I answer such questions .

p.2: Please see column B of the orthogonal table on the left. You can see that B1, B2 and B3 are evaluated 6 times each. Now, rearrange between columns B and H while leaving column A as it is. Looking at column B, we still evaluate B1, B2 and B3 six times each.

p.3: In the orthogonal table, there are 2 × 37 = 4,374 combinations of control factors A to H, and it is possible to efficiently evaluate the effect with 18 combinations. In the L18 orthogonal array, although 18 experiments are labor-saving, the effect of the level of each control factor is confirmed 6 times each. Moreover, the other parameters are evaluated by randomly changing 6 types.

I have answered the question of orthogonal tables over the past three times, but could you understand a little about the usefulness of orthogonal tables?　First, you should try the orthogonal array experiment without suspicion.