This is the second Quality Engineering(QE) lecture.
Please see the document immediately. Click → QE2
p.1: In QE books, there are metaphor of target , arrows and pistol bullets, but I compare QE to a golf. I have never played golf. One of my colleagues who loves golf told me that QE does not relate to golf. And I've questioned QE consulting teacher who loves a golf that using QE will increase the score or not, but he have never answer clearly. Recreating the explanatory material is troublesome, so I will explain it as it is. I compare the method of examining the technical issues by the conventional method to golf practice. Usually (1)you shot at the hole, and then (2) you adjust deviation to small. In this case, it should be susceptible to disturbances and internal disturbances such as terrain, wind direction, physical condition and mental condition (referred to as noise factor in QE).
p.2: In the case of QE, you stabilize the swing and reduce the dispersion of distanace between the falling spot and a hole, and then you adjust the direction and flight distance to the target values (called tuning in QE).
p.3: Draws a diagram comparing the conventional method and QE.
p.4: In QE, a graph is drawn in which the horizontal axis plots the amount of energy and the vertical axis plots the characteristic values. This relation is called dynamic characteristics. In the case of golf, if the characteristic value on the vertical axis is the flight distance, the horizontal axis is the energy of flying. It can be expressed by the formula y = βθ at the lower right. Ideally, the flight distance y is proportional to the shoulder rotation angle θ. See the figure on the left. I show you the characteristics of professional and amateur. Professionals fly farther than amateurs and have less variability due to wind. Ideally, it should be better that β is large and dispersion is small.
p.5-7: I made a picture while remembering the lecture contents of the consultant teacher. I think that it is restored almost same graph. I show you the conventional two-stage design method. When there are three levels for parameters A to D, the target value of characteristic value y is a green horizontal line. The conventional method selects a level that matches the target value. We select A1 of 100 ° C for the temperature of parameter A. Similarly, the optimal condition selected is A1B2C1D3. (1) The distribution of characteristics obtained under these conditions is outside the upper and lower limits of the allowable width as shown in the lower left. (2)Next, the dispersion of the characteristic values is also reduced by managing the dispersion of the parameters A to D to be small. It is same as golf on p.1
p.6: These are two graph of dynamic characteristics. The left is the nonlinear response and the right is the linear response. The variation Δy of y for the same width of Δx is the distribution where ΔyA2 of the nonlinear response is the smallest. In the case of a straight line, the dispersion will be the same everywhere. Please keep this in mind.
p.7: The same characteristic graph as p.5 is shown for parameters A -D with levels 1-3. This time, we will select the level with the smallest variation, using theory of p.6. We will select the level of parameter B later. (1) The combination of A3C3D1D has the smallest dispersion. In this case, as shown in the middle below, the distribution is small, but it exceeds the upper limit. Therefore, the characteristic value is tuned to lower by selecting B3.
p.8: I suppose that you can see that the QE is superior to the conventional method. The conventional method selects A1B2C1D3 optimally, while QE selects A3B3C3D1 as optimal. The width of the blue rectangle is the permissible width of each parameter to make the characteristic value dispersion the same. Wider ones are easier to manage. How to select the level of QE may be a combination that can not be imagined with the experience so far. I think that skilled technicians worry about it and avoid using QE .