[in Japanese]

View full abstract

This report presents a method to generate the DIMOS (Digital Model of Solid) directly from digitized data of a master model. At present, it is still necessary to use a digitizing process in the current die manufacturing system. Therefore the method to generate the DIMOS can apply both CAD data and digitized data. In this proceeding, a method to detecte shape characters from the digitized data is referred. And the special weight factors are introduced in calculating the location points from a set of discrete data.

View full abstract

A versatile vacuum-chuck which can be adapted to a variety of parts shapes is needed to be developed to realize a true flexible assembly system. For this sake, a vacuum-chuck with single- exhaust multisuction configuration has been developed. A prototype was designed and tested. The experimental result showed a good performance of the new vacuum-chuck.

View full abstract

Designing shapes of products is one of the most primary activity in design process. In this paper, a new methodology of representing free form shape features with the aim of making the system capable of holding and manipulating their features after synthesizing them to support designer in their early design stage is proposed. The key idea in this study is Shape Feature Generating Rules and specifying the rules which generate feature shapes. Developmental biology is applied to devise a computational model of the representation called Cell Model using Pitt approach of classifier system. Finally, a computer program is developed to evaluate the methodology by combining two existing shapes to examine if the features of shape are preserved in combined shapes. In addition further examination is conducted to specify which rules are actually hold the feature of shape. It is demonstrated that the methodology can show variety of combined shapes with original, often exaggerated, features and the rules are specified which hold the part of feature of shapes.

View full abstract

This paper describes a nobel method of force/torque sensorless control for a single link flexible arm. The flexible arm includes multiple oscillation modes and its dynamical behavior is generally described as multiple mass spring system. In such a system, it is necessary to suppress the vibrations induced by the multiple oscillation modes. To address the above issue, first, this paper shows that the system stability increases by a feedback of reaction torque. In the proposed approach, the reaction torque is identified by using disturbance observer. Furthermore it is possible to estimate the position where the flexible link arm contacts with target environment. These features make it easy to realize the force/torque sensorless control. Second, the analytical comparison between the sensor-based approach and sensorless approach is described. These analyses show that sensorless approach is effective for a realization of stable force control in the flexible link arm system. The validity of the proposed method is confirmed in several experimental results.

View full abstract

Hydrostatic journal gas bearing with a large pocket in the bearing surface and with fluid damper has high stiffness without causing pneumatic hammer instability. But the settling time cannot be lower than two or three times of natural period time owing to the large pocket volume even if the damping coefficient of fluid damper can be set high. In this paper, the settling time of hybrid restrictor type hydrostatic journal gas bearings is investigated. The stiffness of hybrid type is about 60% less than that of pocket type, and the pocket volume of this type, where pocket depth is about two or three times as deep as bearing clearance, is about 2-5% of the pocket volume of the pocket type. The following conclusion is obtained : the settling time of hybrid type can be set to one natural period time (critical damping) by the use of the appropriate damping coefficients of fluid damper.

View full abstract

This paper describes the intelligent visual evaluation system for mirror finished surfaces of heat treated steel using a neural network. The automated evaluation system for mirror finished surfaces by an image processing has almost not been reported in the past, because of the complexity of image processing. If the features of image processing data are recognized by a vision system and a neural network, the degree of surface roughness and/or kinds of finishing methods will be clarified in relation to the image data of surfaces, and also the surface integrity of mirror finished surfaces will be clarified by the correlation analyses between the features of image processing data of surfaces. The objectives of the paper are to analyze the surface image of mirror finished surfaces using texture analyses, and to construct the neural network system which is able to evaluate the surface roughness and kinds of finishing methods. It is found that the features of image data of mirror finished surfaces can be decided by texture analyses. As the results, it is possible to construct the neural network system which is composed of the features of surface image as input data and the degree of surface roughness and/or kinds of finishing methods as output data.

View full abstract

In order to establish a delta operator technique in fixed-point arithmetic, the first report proposed a modified delta operator newly and verified its availability for the first-order system. The present report tries to extend the application to the nth-order systems and to examine how to adjust the n design parameters T₁-T_n. Let the sensitivity smaller, T_j must be set as small as possible. However, their procedure makes underflow occur easily at the integral operation, namely a delta operation. To overcome this defect, a new error reduction method named Variable Modulation Method (V.M.M.) is proposed. The principle is to make the product of the integral being zero and one in turn at sampling period whenever any steady underflow occurs. Having applied this V.M.M. to the feedback control systems attached with phase-lag compensation, which could never be controlled accurately by the conventional shift and delta operator, the availability of V.M.M. was made clear in a limited word length and constraint of design parameter. The simulated results show that the V.M.M can get accurate controllability in such cases when the systems cannot be controlled only by the modified delta operator.

View full abstract

A new simple method to detect the scanning position of a laser beam from a miniature two-dimensional optical scanner is proposed in order to develop a scanning type miniature two-dimensional image sensor which is as small as possible. In the method, the scanning position is directly detected from the wave of the driving voltage of the piezo-actuator by using the following relation : the phase delay between the driving voltage and the reflection angle of the mirror is 90°, since the two-dimensional vibrator, on which the mirror is mounted, is driven at the resonance frequencies. Additional mechanical components to detect are not hence required;and then the size of the miniature two-dimensional image sensor is 25mm×26mm×15mrn, kept as small as that of the main component, namely the two-dimensional optical scanner. The accuracy of detection experimentally obtained was 0.1-0.2mm, and the resolution of detection was 0.2mm.

View full abstract

Recently, high-precision positioning is increasingly required in precision manufacturing, measuring instrument and so on. The purpose of this study is to develop a new type mechanism which is capable of fine positioning with five DOF, high rigidity and heavy load capability. It consists of three plates and six piezoelectric actuators. Each actuator has a flexible end which is deformed and inclined to perform the motions of X, Y, Z, α and β. with controlled piezo expansions. Since this mechanism has neither guide way nor lubrication, it is operated without friction, wear and backlash. The new principle allows for extremely simple design, operating conditions and high-precision positioning. The maximum displacements were ±2μm in X and Y directions, ±5μm in Z direction and ±20 grad in a and β directions. The positioning errors and motion errors were within ±0.01 pm in translation and ±0.06 μad in rotation which are corresponding to the performance of measurement system. The positioning resolutions for the translation and rotation were less than 10nm and 0.1μad respectively. For external loading of up to 400N, its performances were remained. The settling time was 25ms for each step response of 1μm in X direction and 5 grad in a direction.

View full abstract

In grinding processes, the surface profile of a grinding wheel is an important factor to determine the surface roughness of the work after grinding. There are many studies which deal with measuring the surface profile of the grinding wheel using various methods. However, the height of cutting edges of the grains hasnot been clarified absolutely because these methods used in them were two dimensional. Consequently, it has been impossible to estimate the vary of height of cutting edges by wear. The aim of this study is to measure the surface topography of the grinding wheel in three dimensions. Additionally, the relation between the surface topography of a grinding wheel and the surface roughness of the work after grinding is clarified in order to improve the grinding conditions. Specially, this paper describes the new measuring system developed using the laser stylus unit on the machine in order to measure the surface topography of a grinding wheel. By this system, it is possible to measure the surface of the grinding wheel in circumference. It is shown that these measurements are very precise in 0.8μm.

View full abstract

Structure of polycrystalline silicon films was investigated by X-ray diffraction, reflection electron diffraction and transmission electron microscopy. Furthermore, electric-al and optical properties of the films were studied with the measurements of IR absorption, optical absorption, electron spin resonance and electrical conductivity. The films were prepared by rf planar magnetron sputtering, in which process atomic hydrogen radicals were utilized in order to make the polycrystalline silicon films on the substrate at low temperature. It was shown from X-ray diffraction, reflection electron diffraction and transmission electron microscopy that most of crystal silicon grains grew with a [100] texture, and that the grain sizes were 10 - 20nm. The results obtained from IR absorption and optical absorption measurements were as follows; hydrogen content decreased with the increase of hydrogen partial pressure (P_H2) in the range of 2 ≤ P_H2 ≤ 7mTorr, which was caused by the formation of Si Si bonds, and in the range of 7 ≤ PH₂ ≤ 12mTorr, however, hydrogen content increased gradually with P_H2, which suggested that dangling bonds in the silicon films were more terminated in the higher hydrogen partial pressure. These facts were also explained by the electrical conductivity measurement, and it was found that the polycrystalline silicon film prepared under the highest hydrogen partial pressure had the lowest hopping site density in the films observed in this work.

View full abstract

A new grinding wheel has been developed, in which very fine SiC whiskers are normally aligned to the grinding surface. Using this grinding wheel, a smooth surface of hardened die steel could be produced with the roughness of 16 nmR_max (1.5 nmRa). However, one problem was observed in the grinding process : the grinding wheel exhibits a high tendency to loading, and the grinding efficiency thus decreases. In this study, a method to improve the grinding efficiency was suggested. A small amount of very fine abrasive grains is added to the grinding fluid. The experimental results showed that alumina of 0.6μm grain size is most suitable and a concentration of 0.25wt% is sufficient to prevent loading. The grinding efficiency and grinding ratio were remarkably improved by this method.

View full abstract

Using lapping tape coated with SiC abrasive grain on tetoron taffeta backing, polishing fluids are developed for titanium alloy (Ti-6Al-4V). The experiments for the development are applied to the principle of super-finishing and used some emulsions varied these components. By measuring finished surface roughness and its time-dependence, the difference in polishing characteristics is ascertained. As the result, it is found that excellent mirror-like surface is obtained by using emulsion containing surfactants with fatty acid, especially with it having sulfide.

View full abstract

Intensity distribution of heat source in grinding process is a very important factor in analysis of heat conduction or thermal stresses in the workpiece. Its function is usually postulated to be a triangle, for example, regardless of grinding condition. In this paper, a method of estimation of intensity distribution of heat source in surface grinding is proposed. It is assumed that intensity distribution of heat source during grinding process depends on stock removal by abrasive grains, which can be estimated using surface generation theory. Influence of some grinding parameters on intensity distribution of heat source is investigated. The present method is applied to estimate theoretically structural changes of affected layer in grinding process. A visual simulation system for structural changes of carbon steel has been already constructed by authors. In this system, initial structure of the workpiece is input to an EWS through a microscope, a TV camera and a video board, which is binarized with the image processing technique and used for numerical computations of carbon reconstruction and structural changes. The simulated results of distribution of temperature, microstructure and carbon concentration are visualized on the display. Experiment of surface grinding was carried out and its results were in good agreement with simulation. Structural changes process can be simulated to a certain degree for any given hypoeutectoid or hypereutectoid steel under surface grinding.

View full abstract

This paper deals with a new prototyping method called a dot-matrix method by electrical discharge machining (EDM). In this method, shapes are machined on a workpiece by scanning bundled wire electrodes. The machining process by the dot-matrix method is similar to the printing process with a dot-matrix impact printer. A prototype of a machining unit has six miniaturized electrode feeding devices. It measures 55×55×80mm. Wire electrodes of 0.3mm in diameter are arranged with the pitches of 0.72mm. This machining unit can control the length of each electrode independently. Planes and slopes are machined on a silicon wafer with a planetary motion. By compensating the wear of the electrode during EDM, various shapes with the accuracy of micrometers order can be obtained. Since it is unnecessary to produce a tool electrode before machining shapes, this method is available to reduce the total machining time.

View full abstract

A surface durability of case-carburized and hardened steel rollers is examined by a highly loaded two rollers contact fatigue testing machine. The surfaces of test rollers are finished to a surface roughness of about 3μmR_max by a cylindrical grinding machine. Thereafter, in order to improve the ground surface, the processes of shot peening, barrelling and hydraulic horning are employed. These processes can be applied to surfaces with complex configuration, such as gears. Each surface roughness of rollers finished by above machining processes is measured before/after running test, and residual stress in the vicinity of the surface of test pieces is also detected by an X-ray diffraction method. It is confirmed that the surface durability of rollers whose surfaces are subjected to the barrelling process after the shot-peening process is appreciably improved.

View full abstract

Deep levels in semi-insulating (SI) GaAs single crystals have been studied by the photo-induced current transient spectroscopy (PICTS) and photo-hall effect transient spectroscopy (PHETS) A method of measuring defect level density in SI-GaAs using PHETS has been developed and apphed for the measurement of the EL6 density Properties of the defect levels EL6 and EL2 were investigated by companng their density distnbution in the crystals and their annealing behavior The density distnbution of EL2 in the wafer showed the similar pattern to the dislocation density The density distribution of EL6, however, showed the inversely related pattern to that of EL2 The density of both EL2 and EL6 levels increased similarly by the annealing at 600 to 1000°C and diminished by the annealing at 1100°C These expenmental results were well explained by the tentatively assumed defect models of EL2 (As_GaV_Ga) and EL6 (As_GaAs₁)

View full abstract

Instantaneous force with static deflection feedback model is applied to predict cutting force and dimensional surface error generation in peripheral milling with irregular tooth pitch end mills. It is demonstrated that the irregular tooth pitch end mill reduces the dimensional surface errors according to the chip regeneration mechanism. Unlike in case of uniform tooth pitch cutters, chip loads on each flute are different in milling with irregular tooth pitch cutters. The excess surface error generated by the flutes which have larger chip loads is removed by the following flutes which have smaller chip load due to the irregular tooth pitch. In order to predict precise cutting force and dimensional surface error for the irregular tooth pitch end mill, the chip regeneration mechanism or the static tool deflection has to be considered to derive the accurate chip loads. Predicted cutting force and dimensional sur face error show good agreement with experimental results. Up to 20% reduction in surface errors is achieved by using the irregular tooth pitch end mills tested.

View full abstract