早乙女 英夫

<p>  The dynamic magnetic loss in ferrites is obtained by subtracting the hysteresis loss, which is independent of the excitation frequency, from the iron loss. In the high frequency excitation region, the dynamic magnetic loss is the dominant component of the iron loss in ferrites. The iron loss in ferrite is temperature-dependent and this dependence has been described in product catalogs, where the hysteresis and dynamic losses are not separated. The catalog data are measured using sinusoidal wave voltage excitation, whereas ferrite cores are commonly used under rectangular wave voltage excitation in DC-DC converters. In this paper, the experimentally obtained temperature characteristics of the hysteresis and dynamic magnetic losses for rectangular wave voltage excitation are shown separately, and it is found that the two are different. This suggests that the physical mechanisms involved are different as well. Thus, it is important to consider the temperature characteristics of the dynamic magnetic loss to produce low-loss ferrites.</p>

The B-H loop of ferrites consists of two areas: one is the DC hysteresis loop and the other corresponds to the dynamic magnetic loss. The former is temperature dependent whereas the latter is temperature independent. The difference in the temperature dependence of these two areas suggests that the physical mechanism for the dynamic magnetic loss is different than that of the DC hysteresis loss. The eddy current loss in ferrite grains is a candidate for the dynamic magnetic loss. The conductivity of the ferrite grain was estimated from the experimental results of the dynamic magnetic loss. Based on the results, it was found that the conductivity is too large for iron oxide. This fact leads to some important suggestions.

Ferrite power loss consists of four parts: the hysteresis, eddy current, equivalent dielectric, and dynamic magnetic losses. The eddy current and dielectric losses are negligible in ferrites compared to the other two losses. It is well known that the hysteresis loss is constant with respect to the exciting frequency but is dependent on temperature. It was previously reported by one of the authors that the dynamic magnetic loss is dominant in the high frequency excitation region. This paper shows that the dynamic magnetic loss is independent of temperature. We experimentally verified this with Mn-Zn and Ni-Zn ferrite cores that were magnetized by rectangular waveform voltages.

A novel method for the contactless sheet resistance measurement is presented. The sheet resistance of a silicon wafer has been conventionally measured using the four-point probe technique. However, the high mechanical pressure of the probe causes pitting and structural damage in semiconductor wafers. Therefore, to prevent such damage, a contactless measurement technique was devised that employs electromagnetic induction. The conventional contactless method is implemented using high frequency sinusoidal magnetic excitation with a Robinson marginal oscillator as an excitation current supply. However, the oscillator gain drifts due to an increase in temperature during continuous scanning of the wafer, which causes an error in the resistance measurement. The proposed method does not require an analogue electronic circuit; therefore, more precise measurement is expected. This novel method utilizes an impulse voltage to generate an eddy current in the sheet under test. However, no high voltage source is required, because the impulse voltage is generated by cutting off the DC current flowing through the winding of a ferrite core facing the sheet.

We have designed a hybrid dc-dc converter that includes a four-legged transformer and alternately operates as forward and flyback converters. To miniaturize dc-dc converters, it is necessary to reduce the ripple current that flows through output smoothing capacitors. A transformer applied to the hybrid dc-dc converter has a particular air-gap length that reduces the ripple current of the output smoothing capacitors during forward and flyback operations. We applied the air-gap length conditions to four-legged transformers as well as EI core transformers and experimentally obtained the output current ripple characteristics.

A transformer possesses several functions, such as insulation, voltage transformation, impedance transformation, and so forth. Similarly, an inductor or a reactor is used for current ripple reduction, resonance, energy storage, and so on. Those magnetic power devices are separately placed and connected to circuitries so far. On the other hand, cost reduction is strongly required in designing power supplies. Therefore, power circuit engineers always consider the ways to reduce the number of circuitries.<br>In this article, we introduce recently developed technologies that reduce the number of circuitries in DC-DC converters by synthesizing the functions of a transformer and an inductor.

We applied a four-legged transformer that integrates a transformer and an output choke function on a hybrid-type dc-dc converter that alternately operates forward and flyback conversions and successfully reduced the footprint of the magnetic devices used in the converters. To further miniaturize hybrid-type dc-dc converters, it is necessary to reduce the output smoothing capacitors' footprint. In turn, reducing the size of output smoothing capacitors requires the ripple current that flows though them to be reduces. As a result of analyzing the magnetic circuit of four-legged transformers, we found that a transformer has an air-gap length that reduces the ripple current of the output smoothing capacitors to zero during forward operations. In addition, we found that the initial value of the ripple current during flyback operations matched the last value under this condition of air-gap length. In this paper, we discuss the design conditions for a four-legged transformer and the experiments we conducted to verify the effects of the design.

Mn-Zn ferrite cores are used for DC-DC converters under the excitation conditions of frequencies ranging from tens of kilohertz to several megahertz. The development of a mathematical model enabling the power loss of Mn-Zn ferrites to be estimated in high-frequency regions, should contribute to improving the design of the converters. The spatial network method (SNM) has been used to analyze the electro-magnetic field of electric machines and applied to obtain the power loss in ferrites whose properties have been treated as linear. This paper presents the SNM including ferrite's non-linear magnetic properties, which are magnetic saturation, hysteresis and dynamic magnetic losses. In addition, the dielectric features of Mn-Zn ferrite are taken into account. The simulated results are closely in accordance with the experimental values and demonstrate dimensional resonance caused by high-frequency excitations.

In the design of dc-dc converters, there is an ongoing effort to develop devices with higher power density and efficiency. A forward method is widely used as a circuit topology for applications with low output voltage and high output current. However, the fact that an output choke is necessary inhibits miniaturization and cost reduction. To resolve this problem, attempts have been made to integrate the output choke with the transformer. For example, a forward active clamp topology using a four-legged core has been proposed. This circuit operates as a hybrid of forward and flyback converters. The magnetic flux distribution in a transformer core used in this method has been clarified, but a design method has not been established. By means of an analysis, it was found that the magnetic flux in the leg, which acts as a output choke, has both dc and ac components, which are in a trade-off relation. This paper proposes a detailed transformer design methodology that takes account of this relation. When the core of a transformer was designed using the proposed method, it was possible to reduce the footprint of the magnetic device in comparison with that of a conventional converter.

Efficiency improvement and cost reduction are important requirements for the design of DC-DC converters. Zero voltage switching is one good solution for increasing efficiency, and an active clamp circuit is useful for implementing it effectively. A cost reduction could be achieved by devising a new transformer that functions not only as a conventional transformer but also as a DC reactor that reduces the switching device peak current. However, no design methodology for such a transformer has yet been established. This paper proposes a novel method that simulates the operations of a DC-DC converter and estimates the magnetic flux distribution in the new transformer. The magnetic circuit of the transformer is transformed into its equivalent electric circuit with two ideal transformers. Circuit simulation results agree fairly well with experimental ones.

With the goal of developing an artificial heart, a linear motion magnetic actuator has been produced and applied to a pump. The actuator is composed of two ring-shaped Nd-Fe-B magnets that face each other, the outer and inner diameters and thickness of which are 63mm, 17mm and 3mm, respectively. Rotating one of the magnets circumferentially via a motor produces reciprocating forces on the other magnet. The pumping motion is performed using the power of the reciprocating magnet. Because the actuator uses no windings, no copper loss, which causes heat generation in the body, is generated. The output powers of the actuator and the pump reached 1.03W and 0.65W, respectively.

A cross-regulation phenomenon in a multiple-output fly-back converter is analyzed with a circuit simulator and three-dimensional magnetic field FEM software, to determine how the coupling factors of a multiple-winding transformer affect it. The simulated transformer output voltages and their external characteristics coincide with the experimental results. A novel algorithm for designing the configuration of transformer windings is proposed, using circuit and magnetic field simulators.

Using two small Nd-Fe-B magnets (0.7 min x 0.7 min x 2 mm), a novel magnetic actuator, the length of which is 20 mm, has been developed. The actuator swims in water in parallel to the magnetic field applied. Its swimming velocity performance was experimentally evaluated with three parameters: dc bias magnetic field intensity, and the amplitude and the frequency of the ac magnetic field. A maximum velocity of 37 mm/s was obtained.

In this paper the FEM based approach is used to analyse the electromagnetic field distribution in a ferrimagnetic core with rectangular cross section. This numerical method provides a way to accurately visualise the field distribution. The iron losses which consist of magnetic loss, dielectric loss and eddy current loss are calculated separately based upon the electromagnetic field distribution. From these calculated results, it is found that at low frequency the magnetic loss is the major contributor to the total losses, but when frequency reaches the MHz range the dielectric loss becomes significant and can't be neglected. The calculated results are also compared with measured results and good agreement between the two values is achieved.

A micro-generating system for implanted medical devices is proposed, The system is composed of a micro-generator a high-ratio gear, a metal magnet and exciting coils. The micro-generator is driven by a metal magnet to which the low frequency magnetic field is applied by the exciting coils. A model experiment and its results are shown and magnetic field designing for a pacemaker is carried out.

A novel electric power supply system to the batteries of low power consumption devices implanted, without the periodical operations, is proposed. The system is composed of a small generator, a high-ratio gear, a metal magnet and exciting coils which make a low-frequency rotating field in a human body.

Finite element analysis using the dynamic magnetic loss parameter was carried out, in order to evaluate the power loss of Mn-Zn ferrite cores. The accuracy of the numerical analysis was verified by comparing the numerically computed power loss with the analytical solutions at 100 kHz, 200 kHz, 1 MHz, and 2 MHz. The frequency characteristics of the computed core loss coincide with the measured ones. Finite element analysis taking account of the dielectric characteristics of Mn-Zn ferrites clarified the mechanism of the power loss dependence on the core dimensions. The analysis showed that the displacement current in a core causes a magnetic field opposite to the exciting magneto-motive force in a high-frequency region.

Iron loss measurement of ferrite cores up to MHz range is carried out. Taking the DC magnetic hysteresis, the eddy and displacement currents into account, the electromagnetic fields and iron losses in the cores are computed. As a result, increasing the exciting frequency, the computed iron loss becomes considerably smaller than the experimented one. In order to explain the deference between them, a new magnetic field component yielding a dynamic magnetic loss is assumed and added to the magnetic field of the DC magnetic hysteresis. This assumption is verified by evaluating the iron loss frequency characteristics for different size cores of the same ferrite material.

Iron loss measurements of Mn-Zn ferrite cores up to the megahertz range are reported. Taking the de magnetic hysteresis, the eddy, and displacement currents into account, magnetic and electric field distributions in the cores are computed with the cylindrical coordinates and Bessel functions. The computed iron loss due to the magnetic and electric fields is compared with the experimental value at different exciting frequencies. It is noted that the computed iron loss becomes considerably smaller than the experimental at high frequencies. In order to explain the difference between the computed and experimental iron losses, a new magnetic field component yielding a dynamic magnetic loss is assumed and added to the magnetic field intensity of the de magnetic hysteresis. This assumption is verified by evaluating the iron losses in different size cores composed of the same ferrite material. Displacement current distribution in a ferrite core depends on the cross-sectional area of the magnetic flux path, which brings about the dependence of the frequency characteristics of the iron loss upon core size.

This paper describes the power loss characteristics of Mn-Zn and Ni-Zn ferrite cores in the high-frequency range. Two kinds of Mn-Zn and an Ni-Zn ferrite material were used to verify the proposed magnetization model. To analyze the iron loss, we propose an equivalent circuit of ferrite cores, whose circuit elements are derived from measured material constants. Analysis of the power loss, showed that the dielectric loss and magnetic loss are dominant in the high-frequency region. Ferrite cores having a large relative permittivity have a considerable skin effect, which accounts for the complex frequency characteristics of the iron loss in Mn-Zn ferrites.

This paper presents a novel objective function in order to design electromagnetic devices. The finite element method taking the open boundary condition into account is used for magnetic field computation. Comparison of the results obtained with the novel objective function and the least squares shows that the former is superior to the latter.

Previously, me have proposed the sampled pattern catching (SPM) method to analyze the inverse problems in magnetostatic fields. Because of current dipole angle subdivisions. our SPM method essentially requires a considerable CPU time. In order to remove the current dipole angle subdivisions. we propose here a novel formulation of the SPM method employing the cylindrical and spherical coordinate systems. Applications of this faster SPY method to the magnetocardiogram (MCG) and magnetoencephalogram (MEG) identify the heart defect and brain activated parts, respectively.

The recent developments of SQUID flux meters make it possible to measure the magnetic field caused by the neural activity in a human brain. The field distribution measured on the human head surface provides highly important information of the human brain activities. The problem of obtaining the current signal flow in the brain reduces to the inverse problem, the governing equation of which is written in an integral form. le propose a generalized correlative analysis method based on the Cauchy-Schwarz relation in order to solve the system equation derived from the governing equation. As a result. we have succeeded in estimating current signal flows in the human brain.

In this paper, we propose a novel formulation for the crack identification Problem in metallic materials. In this formulation. cracks are regarded as the equivalent field or potential sources due to the discontinuity of conductivity at the crack positions. This means that crack identification problems are reduced to the inverse problems of searching for equivalent sources. The system equation of the inverse problem. derived by discretizing the integral equation, is successfully solved by the sampled pattern matching method. In consequence. fairy good results are obtained even in the case of plural defect problems.