The dispenser jets 100 times for each run, and the weight of a single jetted dot is measured each time. The measured results are shown in
Figure 5. Subsequently, the average weight and standard deviation of the results for each run are calculated and summarized in
Table 6. The average weight of a single jetted dot ranges widely from 0.337 g (0.001 standard deviation, 5th run) to 0.477 g (0.002 standard deviation, 12th run) with respect to the operating parameter set. The measured results are analyzed by using a commercial statistical software package, Minitab (trial version 17 Minitab Inc., State College, PA, USA), and the level of significance is set to 0.1 (10%).
The main effects and the interaction effects plots for the average weight of a single jetted dot are depicted in
Figure 6 and
Figure 7, respectively. For the main effects plot, the parameters with a steep slope display a significant effect and the parameters with a gentle slope display a small effect on the average weight of a single jetted dot. As shown in
Figure 6, the needle stroke (A) and open time (B) are evidently associated with a high effect on the average weight of a single jetted dot, and the rising time (C) and cooling pressure (D) exhibit a low effect on the average weight of a single jetted dot. The average weight of a single jetted dot decreases with decreases in the needle stroke and open time. With respect to the interaction effects plot, the two parameters exhibit a high interaction effect when the two lines cross each other, whereas the parameters do not display any interaction effect when the two lines are parallel. The six interaction effects for the four parameters, AB, AC, AD, BC, BD, and CD, are plotted in
Figure 7. As shown in
Figure 7, the parameters exhibit small interaction effects on the average weight of a single jetted dot. The Pareto chart clearly explains the effectiveness of each parameter and is shown in
Figure 8. It is noted here that any parameter with a bar that extends beyond the vertical red line (significant reference line) is potentially important. With the Minitab Software, the reference line for statistical significance is determined on the basis of the significance level and the Lenth’s method [
21]. As shown in
Figure 8, open time and needle stroke are the main parameters associated with a significant effect, and the rising time is also a main parameter with a low effect. The interactions stroke–open time, stroke–cooling pressure, and open time–cooling pressure are meaningful parameters, although their effect is very low for the average weight of a single jetted dot.
The effects and estimated coefficients for the average weight of a single jetted dot are listed in
Table 7. The
p-value for each term is tested considering the null hypothesis in which the coefficient is equal to zero (no effect). A low
p-value (less than 0.1 in this work) indicates that the null hypothesis can be rejected. In other words, a predictor that has a low
p-value is likely to be a meaningful addition to the regression model because the changes in the predictor’s value are related to the changes of the response variable. Conversely, a larger
p-value (higher than 0.1 in this work) suggests that the changes in the predictor are not associated with the changes of the response [
21]. Then, any parameter in the table with a
p-value lower than 0.1 is an important parameter that affects the average weight, because the level of significance is set to 0.1.
Table 7 shows the effective parameters with a shaded background, and the results are exactly in accordance with those of the effects plots shown in
Figure 6,
Figure 7 and
Figure 8. The column ‘Effect’ denotes the vertical distance between two points in
Figure 6 and
Figure 7. The column ‘Coefficient’ is the same as half of the column ‘Effect’, and the values are the coefficients in the regression equation. The column ‘
t’ is obtained by dividing the ‘Coefficient’ with the ‘SE Coefficient’. The values of column ‘
t’ are shown in the Pareto chart in
Figure 8. A high magnitude of ‘
t’ indicates that the parameter displays a high effect. Thus, it is concluded that the most effective method to minimize the weight of a single jetted dot involves reducing the needle stroke and the open time, and this significantly affects the average weight with a high
t-value. In
Table 6, a run order of 1, 2, 5, and 6, with low needle stroke and low open time displays a lower average value when compared with other run orders. The average values for the run orders 9, 10, 13, and 14 slightly increased because of the higher needle stroke when compared to the run orders 1, 2, 5, and 6. However, the average values for the run orders 3, 4, 7, and 8 sharply increased because of the higher open time when compared to the run orders 1, 2, 5, and 6. The effects and estimated coefficients are reinterpreted with only the meaningful parameters, and the result is depicted in
Table 8. The coefficients in the third column in
Table 8 are used to express the regression model to estimate the average weight of a single dot as follows: