A simple method to obtain the bulk generation time and surface generation velocity in the completely depleted state is presented. The method is based on measuring both the gate current and high frequency capacitance at the time, t=0, soon after the gate voltage is switched from a depletion to an inversion state on an MOS capacitor. The gate current Igo and high frequency capacitance Chfo at t=0 is measured for various step gate voltages, and the plots of Igo/(1Chfo/Cox) against (Cf/Chfo1) is obtained as the straight line, from which the two generation constants are calculated. In respect of some advantages compared to already existing methods, it is pointed out that our method does not require to know the impurity density and depletion layer width of semiconductor for the determination of the two generation constants.

A method is presented to synthesize many-valued majority elements which can correct errors and simultaneously realize logic functions.

Differential detection of π/2-TFSK, which turns to MSK and π/2 shift PSK in the limitting cases, is analyzed in order to find conditions under which the error performance is not degraded severly and to make clear the relationship between error characteristics of MSK and PSK differential detection in the multipath interfering channel. A two-wave model of two Rayleigh waves is adopted as a frequency selective channel. The conditions under which the π/2-TFSK scheme is able to be implemented and detected with the differential detection are investigated. It is indicated that these can be realized by CPFSK. It is shown that existence of a flat part in the excess phase transition exhibits an important role in order not to be severly corrupted by the multipath interference.

The minimum current of a two-junction SQUID gate has been studied theoretically. An analytical expression for the current-voltage characteristics of the SQUID gate has been obtained, which includes the minimum current of the SQUID gate. Studies have been made of the dependences of the minimum current on parameters of the SQUID gate such as an interlinked magnetic flux, a loop inductance and a resistance. It is shown that the minimum current of the SQUID gate depends strongly on the interlinked magnetic flux, while it becomes the same as that of a single junction in the absence of the magnetic flux. It is also shown that the obtained analytical results agree well with those of computer simulation.

The principle of measuring with random vibration which is well known in industrial vibration testing is applied to measuring the mechanical properties of bio-materials. In this method it is possible to measure in short time (about 10 seconds) and measuring values are not disturbed by noise and the region of the body vibrated is not limited.

Exact equipotentials and lines of electric flux between two thin conducting plates in an anisotropic medium are theoretically presented by using an affine transformation and a conformal mapping technique.

The harmonic distortion and the intermodulation distortion of a laser diode are calculated to be severely influenced by the coherent interaction with reflected waves and are given to be about 2030 dB larger than intrinsic values of the laser diode without reflected waves.

Mobile radio propagation tests in the UHF band were carried out in urban areas. Signals transmitted from radio base stations were received by dipole antennas installed on an automobile and a small cart. Median values of field strength on streets and sidewalks were then measured. In addition, floor height gain characteristics of field strength were measured for a building in the downtown area.

Brent, Kung and Thompson have presented suitable VLSI models, and discussed area-time and/or area complexities of various computations such as discrete Fourier transform and multiplication. Although the VLSI models by Brent-Kung and Thompson are suitable for analyzing VLSI circuits theoretically, their models are not yet sufficiently practical from the viewpoint of the current VLSI technology. In this paper, effects of the practical assumptions such as the boundary layout assumption and a restricted location assumption on bounds of the area complexity are discussed, and some technique for obtaining the area lower bounds of combinational circuits is presented on a VLSI model with the boundary layout assumption. To obtain the area lower bounds, a relationship between locations of I/O ports on the boundary of a combinational circuit and it's area is derived. By using the result, we show that a combination circuit to compute the addition or the multiplication requires area proportional to a square of the input bit size, if some I/O port locations are specified. Also we show that combinational circuits to compute the multiplication, the division and the sorting require area proportional to a square of the input bit size, independently of the I/O port locations. These lower bounds are best possible for the multiplication and the division, and are optimal within a logarithmic factor for the sorting.

This letter presents the precise stack-controlling LALR(1) parser that can perform finite state recognition locally without associated stack manipulations. The presented parser can be applicable to the subclass of the LALR(1) grammars.

A self-complementary basis of a finite field corresponds to the orthonormal basis of a vector metric space. Seroussi and Lempel showed that a finite field GF (qn) has a self-complementary basis over GF (q) if and only if either q is even or both q and n are odd. In this paper, firstly we show that by using the complementary basis of a polynomial basis we can write a self-complementary basis explicitly. Since a polynomial basis and a normal basis are the most popular bases in finite fields, in this paper we consider whether a polynomial basis or a normal basis can be self-complementary. Secondly we show that any polynomial basis can not be self-complementary. Thirdly we tabulate the numbers of all the different self-complementary normal bases of GF (qn) over GF (q) for various q and n. From this table we present a conjecture concerning the existence of nonexistence of self-complementary normal bases.

A regulation model of the platelet count was proposed. All the characteristics of the components of the model in the normal state were determined from clinical and experimental results using numerical simulation procedure. Platelet counts of a subject with cyclical thrombocytopenia, measured over a period of 244 days, revealed that the maximum platelet count was 890,000/µl and the minimum count was less than 1000/µl. In addition, nine platelet survival curves were measured within the same period. Analyzing the 9 platelet survival curves, one of the factors causing cyclical thrombocytopenia appeared to be the periodic elevation of the platelet destruction rate. From simulations done under condition that the destruction rate varied cyclically, it was found that the variation rate of peak value of platelet counts was affected by four major parameters of the model. The average daily destruction rate during the period of measurement of survival curves was estimated in such a way that the square error between the estimated and observed curves was made small by adjusting one of the four major parameters. The second factor was found to be a slow variation of the maximum production rate of magakaryocytes (precursor cells of platelets).

A polarization interferometer capable of measuring film thicknesses ranging from 10 to 3000 is described. A suitable system for measuring thickness of films in resist patterns is designed. The film thickness is measured in terms of the azimuth angle rotation of a plane polarized light. Experimental details are given, along with the results of thickness measurement of Alluminium films. Possible sources of errors and various parameters affecting the precision of measurement are discussed. The measurement error is estimated to be less than 0.1% . The precision of measurement is found to be 30 under normal laboratory conditions and can be improved to 1 by protecting the system from environmental disturbances. With some design modifications, the system could be used for the measurement of very narrow profile of films such as one we meet in the case of the resist pattern. Calculations show that resist films with widths of profile as small as 6 µm can be measured by this method.

Computer image processing is becoming a helpful tool even in industrial inspections. A computerized method for welding visual inspection is proposed in this paper. This method is based on computer image processing of X-ray photograph of welding, in which the appearance information of weldments such as shape of weld bead really exists. Structural patterns are extracted at first and seven computer measures for inspection are calculated using those patterns. Software system for visual inspection is constructed based on these seven measures. It was experimentally made clear that this system can provide a performance of more than 0.85 correlation to human visual inspection. As a result, the visual inspection by computer using X-ray photograph became a promising tool to realize objectivity and quantitativity of welding inspection. Additionally, the consistency of the system, the possibility to reduce computing costs, and so on are discussed to improve the proposed method.

A new stapler-type fiber cutting tool for obtaining good fiber end surfaces is described. T he size of the tool is about 2 cm wide, 3 cm high, and 10 cm long. The weight is only 90 g. In principle, the fiber is scored and than fractured by applying the tensile stress and bending stress simultaneously. The optimum conditions for cutting the fiber are investigated: The blade pressure is from 15 to 32 g, tensile stress from 6 to 13 kg/mm2, and bending stress around 10 kg/mm2. The average breaking angle of 0.6is easily obtained. It is confirmed that the tool is fully endurable for use in field.

By psychological measurements, data of two kinds of subjective judgements--perceptual and cognitive judgements--for Kanji and Kanji-like patterns were obtained, and it was shown that two stages of information processing--perception and cognition--in human pattern recognition has different functions respectively.

The finite-element method is adapted to handle scattering from periodic structures. The formalism developed by us treats both TE and TM waves in a unified form. The validity of the method is confirmed by comparing numerical results with the earlier theoretical results.

Transverse electromagnetic transmission cells are analyzed with high accuracy using a mixed discretization based on both of constant and linear elements in a boundary element method, in which we can take account into discontinuous values of a normal derivative of the potential at corners in the cells.

Properties of the flux-flow type Josephson oscillator coupled to a quasiparticle detector through a planar stripline cavity are investigated in the millimeter wave region. The present coupling scheme is essential for magnetic field isolation between the oscillator and the Josephson rf-devices, as well as filtering of spurious harmonics.

Some properties π on graphs are characterized in terms of a set S(π) of forbidden graphs, that is, a graph G satisfies property π if and only if G has no subgraph homeomorphic (subcontraction isomorphic) to graph in S(π). Among such properties, πplanarity" is the most famous, where S(π){K3, 3, K5}. This paper clarifies the relations among such properties and presents necessary and sufficient conditions for a property to be characterizable in terms of k-connected (k1, 2, 3) forbidden graphs.