For a machine tool, the spindle is an essential element in the machining process; sporadic repairs or replacement of the spindle are a factor that directly hurts the machine tool user’s productivity, so being able to detect abnormal spindle conditions would be of enormous benefit to both the machine tool user and the machine tool manufacturer. In the present study, accordingly, the power consumption of the motor that turns the spindle was evaluated by using the error root mean square for the purpose of developing a system that could discriminate abnormal spindle conditions before they damaged the spindle. A previous study had indicated that in evaluating the S/N ratio not only the power consumption after the spindle rotation had stabilized but also the part of the power consumption waveform in which the power consumption ramps up abruptly during rotational acceleration should be considered, and that separate spindle specifications would require separate analyses, so these approaches were taken in the present study. As a result, by restricting analysis to spindles with just one specification, waveform discrimination within that specification became possible, but the need for data on spindles with abnormal conditions, which had been an outstanding issue, now became even greater.

View full abstract

In a series of remote lectures, the authors taught parameter design to fourteen graduate school students by having them perform experiments on paper helicopters at home. The students performed an L18 orthogonal array experiment aimed at increasing the flying time of a paper helicopter made public by NASA for use by educational instructors, and carried out parameter design using the S/N ratio of this time, which they treated as a biggest-is-best characteristic. All of the students improved on the flying time and S/N ratio of the NASA model, but encountered an increased noise factor contribution ratio in the analysis of variance. Having made the students aware of experimental error, as a further example we presented a dynamic-characteristic experiment in which the input was the initial altitude and the output was the flying time, showed that while the mean value achieved by the NASA model was small, the variation was also small, and pointed out that without a dynamic-characteristic experiment this feature would not be revealed. The last lecture addressed goals, issues, and difficulties in the students’ MS or PhD thesis work, relating these topics to the series of paper helicopter experiments, and had the students state factors lurking in the experimental equipment and procedures that had caused scatter in their data.

View full abstract

This is a report on the result of a functionality evaluation of the paper path in a copier, based on behavior simulation, in which we proposed a new functional definition concerning the transport of paper on a curved path. The definition states that the leading-edge locus length of the paper should match the paper feed length at the paper feed point. With this definition and with the same noise factor levels as in an actual copier, the new design was found to improve by 8.0 db over the previous design. Our definition is independent of the coordinate system and its dimensionality. It is thought to be applicable a wide range of functional evaluations in areas such as motion control and signal transport.

View full abstract