In order to solve quality problems at customer's hand, we have to improve a product design. If there are quality variations at the final production process, we can find the causes of the variations to improve it by using Taguchi approach in Mahalanobis-Taguchi system. First select unit space of product having same output. Then use the Taguchi method explained in Chapter 6.

It is not easy to use a micro-drill less than 1mm in diameter to drill a straight deep hole. When the material to be drilled is difficult to cut, as are the materials used in aircraft engine parts, the hurdles are particularly high and no established technology exists. A micro-drilling experiment was therefore performed to find optimum conditions for drilling difficult-to-cut materials. In the initial experiment, which used an L18 array, evaluation of dynamic characteristics was hindered by frequent tool breakage, so the evaluation was carried out by scoring four criteria on a smaller-is-better basis and calculating S/N ratios. In a follow-up L9 experiment centered on the optimum conditions that were obtained, no tool breakage occurred and the rpm flutter noted in the L18 experiment was eliminated. In the end, stable, breakage-free optimum machining conditions were found for a wide range of applications. In a subsequent life evaluation experiment, under the optimum conditions, drills 0.3mm in diameter lasted seven times as long as previously, each cutting 15 holes or more through plates 6mm thick.

A new method of evaluating optical elements is proposed for use in optimization of the optical element fabrication process by quality engineering, In the past,such evaluations have usually been carried out by evaluating quality characteristics such as shape, but this has been difficult because of the high precision required in the optical elements. The new evaluation method evaluates the transferability of a light spot obtained by focusing the exit light of the optical element with a lens. The proposed idea is to compare the shapes of the focused light spots of the mold and the molded product, the ideal condition being no change in the shape of the spot. The signal factor was the shape of the focused light spot from the mold. To introduce noise factors, comparisons were made in both the central region of best transferability and peripheral regions of poor transferability, taking profiles in both the x- and y-directions. From the experimental results, it was possible to tell what factors were important for raising the S/N ratio (for reducing variability). The authors are unaware of previous examples of this type of functional evaluation.

Real estate values are influenced by a complex set of factors, but when real estate is appraised, it is important to estimate the value accurately. In the appraisal of buildings and the sites on which they stand, in practice, the cost approach and the income capitalization approach are often taken; the sales comparison approach using actual sale prices is employed sparingly. Since it is possible to collect extensive sales data for secondhand housing, the TS method in the MT system was applied to the estimation of secondhand housing values. Niiza, a city near Tokyo, was selected for this study; thirty-three properties of average value and twenty-nine items were taken as the unit space, Fifteen (signal) properties were analyzed by the TS method, and accurate results were obtained. These results were compared with the results of the older MTS method and standard regression analysis. The MTS method gave inaccurate results. The accuracy of regression analysis was acceptable, but the number of usable factors was severely restricted, so in consideration of degree of influence on value, which is an important concern in actual appraisal, the TS method,which can deal with many factors, has the advantage.

The die-casting process has recently been receiving renewed attention and considerable use as a way to save weight and fabricate unitary combinations of parts. The natural trend is toward increasingly complex product shapes and higher quality requirements. Although die casters attempt to cope with this situation through various types of preliminary analysis, at present, the conditions for the actual die-casting process are set according to the experience of the casting engineer. In this study, die-casting of a test piece was evaluated by Taguchi methods. A single set of conditions, arrived at by raising the S/N ratio, was found to optimize both transferability and density, but repeatability remains a problem for further research.

The Atractylodes lancea rhizome used as an ingredient of Kampo medicines is a natural medicine derived fom plants in the aster family (Compositae) found in China. Regional differences in quality are observed. To produce Kampo medicines with consistent quality, a componential analysis of the Atractylodes lancea rhizome material was carried out by two types of chromatography, and the MT system was applied to the included quantities often components having pharmacological effects. A unit space was constructed, referenced to specimens from regions described in classical literature as producing superior quality. The distances between specimens from eighteen production sites in China were determined by the inverse matrix method. Components were selected for each region and a classification was made from differences in the contribution ratio of the signal-to-noise ratio (S/N) and sensitivity (S) of the ten components. As a result, it was found that the regions could be roughly classified into four north-south-east-west patterns centered on the production sites used in the unit space. The results obtained by the MT system also enabled the results of the componential analysis to be interpreted more precisely.

Electronic circuit simulation was used in this study to optimize a low-pass filter circuit by parameter design using a standardized S/N ratio and a nominal-is-best characteristic S/N ratio. Optimal methods of analysis were found for each method. When the standardized S/N analysis was compared with the nominal-is-best characteristic S/N analysis, the highest repeatability characteristic was obtained from the standardized S/N analysis. The standardized S/N analysis also best expressed the characteristics of the experimental circuit.

Quality engineering methods were applied to analyze power consumption as measured by a PZ 4000 power analyzer during milling of RC 55 hardened steel to improve the milling accuracy and find the optimum machining conditions. The results led to high speed milling of this difficult-to-machine material. Application of the quality engineering methods proved useful in providing an understanding of the optimum machining conditiens, which were, for RC 55: no tool runout, down-cut, distance x=20mm from workpiece to air nozzle, cutting speed V=200m/min, depth of cut t=0.9mm, feed rate f=0.25mm/tooth, pick feed fp=1.0mm, HIT tool type. Machining characteristics were improved, and the optimum machining conditions turned out to be similar to empirical precision machining conditions.