Technical report, IDE1057, November 2010 Microwave Components Based on Magnetic Wires Master’s Thesis in Microelectronics and Photonics Sizhen Lan & Lian Shen School of Information Science, Computer and Electrical Engineering Halmstad University Copyright by Sizhen Lan and Lian Shen 2010 2 Microwave Components Based on Magnetic Wires By Sizhen Lan and Lian Shen M.Sc. Electrical Engineering, Halmstad University, 2010 Abstract With the continuous advances in microwave technology, microwave components and related magnetic materials become more important in industrial environment. In order to further develop the microwave components, it is of interest to find new kinds of technologies and materials. Here, we introduce a new kind of material -- amorphous metallic wires which could be used in microwave components, and use these wires to design new kinds of attenuators. Based on the fundamental magnetic properties of amorphous wires and transmission line theory, we design a series of experiments focusing on these wires, and analyze all the experimental results. Experimental results show that incident and reflected signals produce interference and generate standing waves along the wire. At given frequency, the insertion attenuation S 21[dB] of an amorphous wire increases monotonically with dc bias current, the empirical formula of attenuation I I with dc current is: A( I ) (8.686C d Z C ) ( R0 R1e 0 )[dB ]. The glass cover will influence the magnetic domain structure in amorphous metallic wires. Therefore, it will affect the circumference permeability and change the signal attenuation. It is necessary to achieve the impedance matching by coupling to an inductor and a capacitor in the circuit. The impedance matching makes the load impedance close to the characteristic impedance of transmission line. The magnetic wire-based attenuator designed in this thesis work are characterized and compared to conventional pin-diode attenuator. 3 Key Words Microwave component, Amorphous magnetic wire, GMI effect, Domain structure, Permeability, Impedance matching, Attenuator. 4 Acknowledgement With the help and support of many people, we could complete this thesis successfully. At first, we are honored to express our gratitude to our dedicated supervisor, Emil Nilsson. Through his guidance, we have gradually realized what we need to do for our paper. His erudition and preciseness has inspired us to write this essay with an open and positive mind. His interesting discussion with us was also informative and useful. We appreciate all his efforts. We also want appreciate our major teachers Håkan Pettersson, Jörgen Carlsson and Lars Landin. Their earnest, profession and enthusiasm inspired us deeply. We also would like to thank our parents for their hard work and support of our graduated studies from the very beginning of our postgraduate study. In the future study, we will maintain this enthusiasm and efforts. Lan sizhen Shen Lian Halmstad University Nov. 2010 5 Contents Abstract Acknowledgement 3 5 Chapter 1 Introduction 1.1 Microwave Components 1.2 Amorphous Wires 8 8 8 Chapter 2 Magnetic Properties of Amorphous Metallic Wires 2.1 Magnetic Properties 2.1.1 The Magnetic Domain Structure 2.1.2 Zero-Magnetostrictive Effect 2.1.3 Co-Fe-based Amorphous Metal Wire’s Magneto-Impedance Effect 2.2 Giant Magneto Impedance (GMI) Effect and Skin Effect 10 10 10 11 11 12 Chapter 3 Transmission Line 3.1 Descriptions 3.2 Transmission Line Equation 3.3 Characteristic Impedance 3.4 Reflectance 3.5 Impedance Matching 3.6 Standing Wave 3.7 Microstrip Transmission Line 3.8 Attenuation of Amorphous Metallic Wires 16 16 17 19 19 20 20 21 23 Chapter 4 Experiment 4.1 Wire Preparation 4.1.1 Amorphous Glass-Covered Wires 4.1.2 Glass Cover Removal 4.2 Experiment Procedure 24 24 24 25 25 Chapter 5 Analysis and Design 5.1 Analysis 5.1.1 The Characteristic Analysis of Signal Attenuation 5.1.2 Direct Current Effects on the Signal Attenuation 5.1.3 Amorphous Metallic Wire without Glass Cover 5.1.4 Impedance Matching Analysis 27 27 27 32 35 36 6 5.2 Design 5.2.1 Electronically Controlled Attenuators 39 39 Chapter 6 6.1 Conclusion 6.2 Future Work 44 45 Reference 46 7 Chapter 1 Introduction With the continuous advances in microwave technology, microwave components and related magnetic materials become more important in industrial environment. They are widely used in microwave communication, telemetry system, radar, navigation, biological medicine, artificial satellite, spacecraft etc. As the operating frequency of microwave devices further increases, the power capacity increases, the noise reduces, as well as the efficiency and reli ability improve, especially in the integration realization, which leads to new changes in microwave electronic systems. Amorphous materials especially for amorphous wires have a lot of significant benefits in their properties, such as high strength, high toughness, high magnetic permeability, and excellent chemical properties. It is of interest to know the prospects of amorphous wires for implementation as microwave components. 1.1 Microwave Components In the microwave systems, microwave components are the devices which achieve the directional microwave signal transmission, attenuation, isolation, filtering, phase control, transformation of waveform and polarization, impedance transformation [1]. Microwave components are the electromagnetic components which work in the microwave bands. In low-frequency electronic circuits, the most commonly used passive components are resistors, capacitors, inductors, transformers and so on. Similarly, in microwave circuits, passive components such as resistors, capacitors, inductors are also widely used. However, due to the increased frequency, the performance of these passive components will change a bit and some kinds of components used in low-frequency circuits can’t be used in microwave frequency bands. Owing to the research and development of microwave technology, such as asymmetric transmission lines, we could realize the functions of inductors and capacitors in microwave frequency bands. In order to constitute a microwave circuit with certain functions, it’s necessary to connect several passive microwave components, such as directional couplers, power dividers, impedance matching devices, microwave filters, attenuators, terminal loads, etc. In order to develop microwave components, it is of interest to find new kinds of techniques and materials. In this report, we introduce the properties of amorphous magnetic wires and analyze the experimental results, after these we design a new attenuator based on the amorphous wires. 1.2 Amorphous Wires Amorphous materials refer to the status of non-crystalline materials. If cooling the molten alloy rapidly, the structure after solidification presents the state of glass. Amorphous materials have many excellent properties such as high strength, high toughness, excellent magnetic properties and corrosion resistance. It seems that the magnetic properties of amorphous material are widely used and these materials can be made in a variety of shapes, such as films, ribbons, wires, powders [3]. In particular, amorphous wires have attracted a lot of interest from industry due to their potential applications in electronic systems. 8 Fig 1.1 Schematic representation of an amorphous glass-covered metallic wire Fig 1.1 illustrates the schematic representation of an amorphous glass-covered wire (AGCW). Here d is the diameter of the metallic (magnetic) core, and D is the diameter of wire with glass cover. These wires are obtained from the melt by the glass-coated melt spinning method. This method will be introduced in chapter 4. Amorphous magnetic wires with glass cover were prepared for the first time in 1974 by Wiesner and Schneider [4], but the wires didn’t show excellent reproducible performance, wherefore the interest in these wires decreased [5]. Due to the developing of theory and technology and many application potentials in the last several years, amorphous magnetic wires have been reconsidered. 9 Chapter 2 Magnetic properties of Amorphous Metallic Wires 2.1 Magnetic Properties Amorphous metallic wires are relative new materials which have attracted a lot of interest for basic research and potential applications in microwave components. These wires have a specific magnetic behavior which originates from their special magnetic domain structure due to the stress difference between the surface and center during preparation [6]. During the process of magnetization in longitudinal direction, the length of amorphous wires will change and the magnitude can be expressed as magnetostriction coefficient λ . If the length increases, the wires are called positive magnetostrictive materials such as Fe-based amorphous wires. If the length decreases, the wires are called negative magnetostrictive materials such as Co-based amorphous wires. Fe- and Co-based amorphous wires are very important amorphous metallic materials, because they have their own electromagnetic properties (e.g. high resistivity for Fe-based amorphous wires and high permeability for Co-based amorphous wires). 2.1.1 The Magnetic Domain Structure Some kinds of amorphous materials with low magnetic domain wall energy are isotropic. If shape and stress anisotropy exist, there will be higher domain wall energy. Different domain structures are observed in different types of materials. In the rapid cooling preparation process for amorphous wires, the cooling rates of surface and center of the wires are different. There will be a stress difference between the surface and center. When the direction of an external magnetic field is consistent with the stress direction, tensile stress will make the distance between atoms increase, and increase material magnetostriction coefficient λ . When direction of magnetic field is opposite to the stress direction, the compressive stress will make the distance between atoms decrease, and increase the absolute value of negative magnetostriction coefficient λ [8]. Because the amorphous materials do not have grain-boundaries, there is no grain-boundary resistance. In low magnetic field, Barkhausen jumps may occur due to irreversible displacements of domain walls in the process of magnetization. With the magnetic field increasing, all atoms in the domain wall reverse, only one or two Barkhausen jumps will complete the annexation displacement of domain wall [9][10], and this kind of amorphous metallic wire is a large magnetostrictive soft magnetism material. Fig2.1 (a) shows the domain structure of positive magnetostrictive amorphous wires (e.g. Fe-based wires). The outer shell domain is a radial domain with closure domains and the inner core domain with longitudinal magnetisation. Fig2.1 (b) shows the domain structure of negative magnetostrictive amorphous wires (e.g. Co-based wires). Because the surface anisotropy is circular and inner anisotropy is perpendicular to the axis of wire, the outer shell domain is circular and inner core domain is axial. 10 Fig 2.1 Domain structures of Fe-base (a) and Co-base (b) amorphous metal wire Different domain structures have different magnetic properties. By applying alternating current in the amorphous wire, there will be a circular magnetic field in the circumferential direction of the wire and wire will be magnetized in this direction. For positive magnetostrictive amorphous wires, stress distribution may result in radial easy axes in the outer shell, but for negative magnetostrictive amorphous wires, stress distribution will result in circumferential easy axes in the outer shell, so the circumferential magnetization is different, larger magnetization shows higher permeability and larger permeability gives higher magnetization [11]. For amorphous glass covered wires, because the glass and metallic core have different expansion coefficients, there will exist stress differences between glass and metallic core during the preparation, and the domain structure will be influenced by the glass cover. 2.1.2 Zero-Magnetostrictive Effect One of the main properties of amorphous magnetic wire applications is the zero-magnetostrictive effect. Zero magnetostrictive effect means that the material magnetostriction coefficient tends to zero. It makes the material have low magnetic domain’s wall energy, easy inversion of domain wall and high permeability. Co-Fe-based amorphous wire has zero magnetostriction coefficient [12]. It has high magnetic induction at low magnetic field. During the magnetization process, the length of Co-Fe-based amorphous wires with zero magnetostriction is constant and these wires are excellent soft magnetic materials. Under the condition of small magnetic field，these wires all have the characteristics of fast response and high stability [7]. 2.1.3 Co-Fe-based Amorphous Metalic Wire’s Magneto-Impedance Effect When high frequency current passes through the Co-Fe-based amorphous metallic wires, the impedance of the wires will change due to the so called skin effect. Under the external applied magnetic field, the skin depth and the impedance of the wires will change. This is called the magneto-impedance (MI) effect [13]. The amorphous wire with zero magnetostrictive effect will generate MI effect when coupled to alternating current. The induced electromotive force is directly proportional to the differential magnetic permeability along circumferential direction of wire. The sensitivity of MI effect is inversely proportional with the dynamic coercivity along circumferential direction of wire [14]. 11 2.2 Giant Magneto-Impedance (GMI) Effect and Skin Effect The giant magneto-impedance effect was observed by Machado et al. [15]. Here the impedance of soft magnetic amorphous metals undergo a large change under the condition of external magnetic fields. This effect is called “giant magneto-impedance” (GMI). Many researchers have recently been attracted by it because of its prospective applications in microwave components [16]. Some previous researchers found that when the magnetostictive coefficient of the amorphous wire is close to zero, and the frequency is larger than 10KHz, the GMI effect can be observed [17][18]. The common used circuit for GMI measurement is shown in Fig 2.2. Fig 2.2 Simplified circuit for GMI measurement When iac flows through the wire, it will induce a circular magnetic field Hac and the amorphous wire will be magnetized in the circumferential direction. The external field Hac will weaken the circumferential magnetization. When Hac increases, circumferential permeability decreases rapidly. Hence one can see that the circumferential permeability varies sensitively with external field is the major cause of GMI effect. When the signal frequency increases, the magnetic field will enhance, leading to high magnetization. The permeability could be considered as the function of frequency and field. When a dc bias current Idc is applied on the amorphous wires, it will induce a circumferential dc magnetic field Hdc. In this case, the asymmetry in GMI appears, due to the direction and magnitude of Idc. We know that, the main cause of skin effect is due to the eddy current generated within the amorphous wire. The dc bias current will weaken the eddy current and thus the skin effect will be reduced. Now it is very common to use the traditional framewore of electromagnetic theory to explain the GMI effect of amorphous wire. In the early reports on GMI phenomenon [19][20], people considered the origin of GMI effect as related to classical electromagnetic skin effect. When high frequency current flow through the conductor, the current distribution is not uniform in the conductor cross section owing to the skin effect and the current density is mainly concentrated in the surface of conductor. 12 In a magnetic field, the impedance of the magnetic materials can be written as follows: Z R eff ( , H ex ) jX eff ( , H ex ) Where is the angular frequency of the input, is the external magnetic field [21]. eff (2.1) is the effective permeability of materials, H ex Using Fourier analysis method to analyze the impedance of amorphous wire, we can get the mathematical model of the GMI effect. The impedance Z can be presented as follow [22]. Z= 1 R kr J 0 (kr ) J 1 (kr ) 2 dc k= (1 - j ) (2.2) δ Here Rdc = 1 σ (πr 2 ) is the direct current resistance per unit length, J 0 (kr ) and J 1 (kr ) are respectively the zero order and one order of Bessel functions, is the skin depth, r is the radius of wire. In the strong skin effect, we can get the impedance Z 1 1 (1 j ) 2 2r (2.3) In the long straight section of a wire, the constant current is uniformly distributed. The electrons flow through the wire in a manner similar to the way water flows through a pipe. This means that the path of any one electron essentially can be anywhere within the volume of the wire. Fig 2.3 A wire that is connected to a dc source When alternating current passes through the wire, due to the induction effect, the distribution of current density is not uniform in the conductor cross section. The more close to the conductor surface, the greater the current density. This phenomenon is called as "skin effect" [23]. When high frequency current flow through the wire, it can be considered that current flow through only in a thin surface layer of wire. This is equivalent to the wire cross-section decreases, resistance increases. 13 Fig 2.4 A wire that is connected to an ac source [24] For the alternating current, the self-induced electromotive force (EMF) will appear to resist the adoption of current in the wire. This EMF is in proportion to the size of the conductor cutting magnetic flux per unit time. With circular section conductor in Fig 2.4, the closer to the center of conductor, the greater self-induced EMF is generated by the external magnetic field. The closer to the surface of the conductor, the smaller influence can be affected by the internal magnetic field, and thus less self-induced EMF. As the self-inducted EMF increases with the frequency, skin effect will be more prominent and small skin depth will be obtained. These make the effective cross-sectional area smaller when the current passes through the conductor, thereby, causing the effective resistance larger [25]. There is another explanation of the skin effect which is related to the process of electromagnetic wave infiltrate into the conductor. When the electromagnetic wave infiltrate into the conductor, it will be attenuated due to the energy loss. The depth where the amplitude attenuation is e -1 times of the surface amplitude is called skin depth of electromagnetic field in the conductor. A simple example is the penetration of plane electromagnetic wave to semi-infinite conductor, the equation of skin depth is [26]: 1 f (2.4) Where f is the frequency, is conductor conductivity, is the permeability. The skin depth is inversely proportional to the square root of the three parameters. When the transverse dimension of conductors in a wire is smaller than the skin depth, the distribution of current density over the cross-section will become uniform and the resistance there will be similar to dc value. Determining the ac resistance of a wire from Maxwell’s equations is difficult for even the simplest cross-sectional shapes, such as circular wires [27][28]. However, if the skin depth is small compared to the dimensions of the cross section of the wire, then we could obtain very good results by assuming that the current is flowing in the skin depth of the outer surface of the wire. The approximate effective area of a circular wire shows in Fig 2.5 was previously determined as Aeffective r 2 r 2r 2 (2.5) The impedance of the wire is Z R j L , only consider R, we can get the resistance of wire with length l: R l Aeffective l 2r (2.6) 14 In extremely strong skin effect, the radius of the wire is much bigger than the skin depth ( r >> δ ), so in the formula (2.6), δ can be neglected. Fig 2.5 Effective area of current pass through the wire 15 Chapter 3 Transmission Line 3.1 Descriptions A transmission line is a kind of waveguide structure which can transmit electrical energy from one point to another. Although the input port can be connected directly to the output port, the input port is usually located some distance away from the output port. We then use a transmission line to connect the input and output port. The purpose of the transmission line is to transfer the energy with the lowest possible power loss. In order to achieve this, it is necessary to obtain special physical and electrical characteristics of the transmission line. In many electric circuits, the length of wires connecting the components can be ignored. There are several exceptions, one is that, if the interval of voltage is similar to the wavelength of signal in the wire, the length of wires should not be ignored and the wire should be treated as transmission line. A common empirical method is that, if the length of wire is greater than 1/10 of wavelength, it can be treated as a transmission line [29]. At this length, the phase delay, reflection and interference in the line need to be considered. Without using the transmission line theory, the microwave systems will have some unpredictable behaviors. There are several structures of transmission line in Fig 3.1. Various kinds of waveguides, which can transfer TE mode, TM mode, or mixed-mode, could be considered as generalized transmission line. We could use the viewpoint of the equivalent transmission line to analyze the distribution of the electromagnetic field along the propagation direction in waveguide. Fig 3.1 Example structures of transmission line [30]. 16 3.2 Transmission Line Equation Consider an uniform transmission line, the input port is connected to a sinusoidal signal source with angular frequency , and the output port is connected to a load impedance ZL. Suppose that the original point of coordinates is on the initial, we could get the complex voltage and current at z point is U(z) and I(z). After dz section, the voltage and current equal U(z)+dU(z) and I(z)+dI(z), respectively. As shown in Fig 3.2. Fig 3.2 Equivalent circuit of an element of a transmission line with a length of dz [31]. A short transmission line could be described by four lumped parameters: R, L, G, C, which represent the resistance in both conductors per unit length in Ω/m, the inductance in both conductors per unit length in H/m, the conductance of the dielectric media per unit length in S/m and the capacitance between the conductors per unit length in F/m, respectively. The incremental voltage dU(z) is generated by the distributed inductance Ldz and the distributed resistance Rdz, and the incremental current dI(z) is generated by distributed capacitance Cdz and distributed conductance Gdz. According to Kirchhoff's law, it is easy to write the equations dU ( z ) R jL I ( z )dz dI ( z ) G jC U ( z ) dU ( z )dz (3.1) Omit small higher-order, then: dU ( z ) dz RI ( z ) jLI ( z ) dI ( z ) GU ( z ) jCU ( z ) dz (3.2) Equation (3.2) is a first-order ordinary differential equation, and is also known as transmission line equation. It describes the variation of voltage and current on uniform transmission line of each infinitesimal section. According to the solution of this equation, we could obtain the expressions of voltage and current at any point on the transmission line, as well as the relationship between them. Therefore, equation (3.2) is the basic equation for uniform transmission lines. 17 Differentiate z on the both sides of equation (3.2), we could obtain d 2U ( z ) dI ( z ) dz 2 R j L dz 2 d I ( z ) G jC dU ( z ) dz 2 dz (3.3) Combine with equation (3.2), equation (3.3) could be rewritten as d 2U ( z ) 2 dz 2 R j L G jC U ( z ) U ( z ) 2 d I ( z ) R j L G jC I ( z ) 2 I ( z ) dz 2 (3.4) Therefore： R j L G jC (3.5) The general section of equation (3.4) is: U ( z ) A1e z A2e z z z I ( z ) A3e A4e (3.6) According to the first formula of equation (3.6) and equation (3.2), we could get I ( z) R j L ( A1e z A2e z ) 1 ( A1e z A2e z ) ZC (3.7) Where ZC R j L (3.8) Here, ZC is the characteristic impedance of transmission lines. Commonly refers to the transmission line wave propagation constant, a dimensionless complex number. The propagation constant could be used to describe the attenuation and phase shift of voltage and current traveling along the transmission line. Usually it is expressed as j Here, the real part (3.9) represents attenuation constant whose unit is dB/m and the imaginary represents phase shift constant whose unit is rad/m. For the non-loss line ( R G 0), we get 0 LC 2 (3.10) It explains that in the process of wave traveling there is no attenuation, and the wave moves with a wavelength of 2 radians of phase delay. Here is the permeability and is the permittivity. 18 3.3 Characteristic Impedance The ratio of voltage and current could be expressed as impedance or resistance. For a lossless transmission line, the current is in phase with the voltage and the impedance is re al. It is called the "Characteristic Impedance". It has no relationship with frequency, only depends on the line itself, such as the physical parameters and geometric dimensions. Usually, it is a complex constant. From formulas (3.5) and (3.8), the characteristic impedance can be expressed as follow. ZC R j L G jC (3.11) At high frequency，that is ωL >> R , ωC >> G , then ZC L C (3.12) A transmission line can be treated as a circuit composed of many inductance and capacitance. Here, the characteristic impedance can be considered as a real resistance. The energy generated from the generator will be stored temporarily in this effective resistor, and at some time later, the energy will be extracted and return to the generator, or convert to heat in a real resistance [32]. Because the transient distribution of electromagnetic field in cross section of transmission line is similar to the distribution of two-dimensional electrostatic field and static magnetic field, we can use static magnetic field and constant-current method to calculate the distribution of parameters C and L, and calculating the characteristic impedance ZC. Usually only calculate C, we can get the characteristic impedance by formula Z C = 1 νC , here v is the velocity of signal transmission. Therefore, we could further understand the characteristic impedance as a pure resistance, only related with the parameters of form, size of transmission line and the medium, but independent on frequency. 3.4 Reflectance When Z L ≠ Z C , here ZL is the load impedance coupled with the terminal, the incident wave which is transmitted to the load will generate reflected wave from the load to the source. On one point at the transmission line, the ratio of reflected wave voltage and incident wave voltage is the voltage reflection coefficient for this point, referred to as the reflection coefficient, usually it is a complex constant. For non-loss line, the reflection coefficient is Γ = Γ e jψ , and the mode Γ remains unchanged alone the line as the angle changes linearly. In the load (reflection points), the initial value of ψ and Γ is only related with the ratio of Z C Z L . The relationship between z and input impedance Z ( z ) = U ( z ) I ( z ) on z point in transmission line is ( z ) Z ( z ) ZC Z ( z ) ZC (3.13) When Z ( z ) |z L Z C , ΓL = 0 , there is only an incident wave transmitted to the load, no reflected wave from the load. Without any reflected wave, the transmission line is said to be impedance matched. 19 3.5 Impedance Matching Since a transmission line has impedance built in, one question is how does the impedance affect signals which transmit through a transmission line? The answer to this question mainly depends on the impedances of devices to which the transmission line is attached. If the impedance of the transmission line doesn’t equal the impedance of the load, the signals propagating through the line will only be partially absorbed by the load. The rest of the signal will be reflected back. Reflected signals are generally bad things in electronics. They represent an inefficient power transfer between two electrical devices [33]. The purpose of impedance matching is to transfer maximum power from transmission line to load. 3.6 Standing wave When the signal transmits through the transmission line, there will be signal reflect because of the impedance mismatching. The interaction of input signal and reflected signal will create a standing wave. Fig 3.3 shows a typical resulting standing wave pattern for a mismatching transmission line. The schematic is shown the amplitude of signal versus position along the transmission line [34]. Fig 3.3 Schematic of a typical resulting standing wave pattern for a mismatching transmission line A term used to describe the standing wave is the voltage standing wave ratio (VSWR), which is the radio of the maximum to minimum voltage. An ideal VSWR is 1:1, that is, the load impedance is equal to the characteristic impedance of transmission line, but it is almost impossible to achieve. Larger VSWR gives higher reflected power. For example, if VSWR equals to 1.25:1, the reflection power is 1.14% and the VSWR equals to 1.5:1, the reflection power is 4.06% [35]. 20 3.7 Microstrip Transmission Line The microstrip line is one of the most common types used in microwave circuit. It consists of a strip conductor and a ground metal plane separated by a dielectric medium [36]. The geometry of microstrip transmission lines is illustrated in Fig 3.4. A strip conductor of width W, and thickness T is printed on a grounded dielectric of thickness H and relative permittivity r . Three important electrical parameters for microstrip line design should be noticed, they are the characteristic impedance ZC, the guide wavelength g , and the attenuation constant . Fig 3.4 Schematic of single microstrip Fig 3.5 The EM fields are not contained entirely within a microstrip line but propagate outside of the line as well In a microstrip transmission line, as shown in Fig 3.5, the dielectric substrate half surround the conducting strip, the electromagnetic (EM) field lines are propagated in both the substrate and outside of the microstrip [37]. Both dielectric losses and conductor losses will introduce attenuation. The attenuation caused by the finite conductivity of the conductors is accounted for by the series resistance R, while attenuation caused by dielectric loss is modeled by the shunt conductance G in the distributed circuit model of the microstrip line. The separate attenuation constants are given by [38] αc = R G and α d = 2Z C 2YC (3.14) And the total attenuation is given by α = αc + α d (3.15) The following formulas give excellent results for the capacitance per meter of strip of width W at a height H above a ground plane, the air dielectric constant is 0 . Ca 2 0 ln 8H W W 4H , W W Ca 0 1.393 0.667 ln 1.444 , H H 21 W 1 H (3.16) W 1 H (3.17) Where C a is the capacitance of the unscaled air-filled line. The effective dielectric constant e e for the microstrip line is given by C Ca (3.18) The characteristic impedance is given by ZC L e 0 0 C C There is an easier way to find 0 0 1 e Ca (3.19) e . Schneider [38] deduced a quite simple formula for the effective dielectric constant of a microstrip line which is illustrated at below. 1 1 12 H e r r 1 2 2 W 1/2 F r , H 0.217( r 1) T (3.20) WH Where F r , H 0.02 r 11 W H for W H 1 and equals zero for W H 1 . 2 We evaluate the conductor loss is caused by finite conductivity of the microstrip and the ground plane. For the same microstrip line, the normalized series distributed resistance for the microstrip is RC [38]. RC LR Rs 1 1 4 W ln W 2 T W 10 H (3.21) The loss ratio (LR) is given by LR 1 W W LR 0.94 0.132 0.0062 H H W 0.5 H (3.22) W 10 H (3.23) 2 0.5 And the normalized series resistance RD of the ground plane is given by RD Rs W H W W H 5.8 0.03 H W (3.24) Rs is the skin-effect resistance which is equal to 1 (σδ ) , is skin depth and is the conductivity of wire. The total series resistance is the sum of RC and RD, so we can get c c ,microstrip c , groundplane RC RD 2Z C (3.25) In our experiment, because the attenuation caused by dielectric losses is very small, we only consider the attenuation caused by conductor losses [38]. 22 3.8 Attenuation of Amorphous Metallic Wires In our experiments, we will analyze the signal transmission and loss for the amorphous metallic wires. Therefore, we need to deduce a formula of signal attenuation for the amorphous metallic wires. Fig 3.6 The electromagnetic (EM) field lines of amorphous wire sample. Fig 3.6 shows the electromagnetic field lines of our sample. Comparing Fig 3.5 and Fig 3.6, we could find that the electromagnetic field lines distributed in the dielectric substrate is similar and the density of the electromagnetic field distributed in the middle of the substrate is larger than both sides. Therefore, based on the neglect of certain factors, by changing some of variables, we can directly apply the formulas of microstrips into our samples. Because the wire is circular, it can approximate W T d, according formula (3.21), then we get the normalized series distributed resistance for the wire is RC LR Rs 1 1 2 ln 4 W (3.26) The loss ratio is given by formula (3.22), (3.33). The normalized series resistance RD of the ground plane is the same with formula (3.24) The attenuation constant α c is given by A[dB] 20 L log e 8.686 L (3.27) -2 l The power in length l is p l p 0 e , where p(0) is the input power. The attenuation of the wire can be obtained as. c Rs LR 1 1 1 2 ln 4 Z c W H (W H 5.8 0.03 H W ) 23 (3.28) Chapter 4 Experiment 4.1 Wire Preparation 4.1.1 Amorphous Glass-Covered Wires The preparation of amorphous metallic wires with glass cover is based on high frequency induction heating melt spinning method. After a suitable temperature annealing, amorphous samples can be turned into amorphous metallic glass-covered wires. This method was initially used by Taylor in 1924 [39] and improved by Ulitovskiy [40]. The schematic diagram is illustrated in Fig 4.1. It mainly includes three parts: high-frequency induction heater, vacuum system, drawing system. Insert small pieces or powder of the metallic alloy into the pyrex tube after the tube is evacuated and the tube is subsequently filled with an inert gas. A high-frequency inductive coil is used to heat and melt the alloy. The glass wall becomes soft by the molten alloy. Due to the molten metal drop and the mechanical tension, the molten alloy and softening glass will be formed to the shape of glass-covered alloy wire. It immediately passes through the cooling liquid jet, and then the amorphous glass-covered alloy wire can be obtained. Fig 4.1 Schematic diagram of the glass-coated melt spinning process [41]. This is a modification of the Taylor method. It can easily produce the alloy systems with low wire forming capacity. The metallic melt stream will break into droplets before solidification. Since the presence of the glass cover, the molten metal and cooling liquid will not contact directly, this will drastically reduce the ability of melt stream break into droplets before solidification. It is not easy to break into droplets with the melt stream in glass cover which ensures a smooth cylindrical shape. So that we could obtain higher cooling rates, and produce amorphous wires more easy. Fig 4.2 shows two examples of amorphous metallic wires. 24 Fig 4.2 The SEM images of Pyrex-covered amorphous wire [42]. 4.1.2 Glass Cover Removal The amorphous glass covered wires have many potential applications in electronics industry. It offers a lot of advantages due to the insulating glass cover. However, for some special applications, it is necessary to remove the glass cover of the wires. As said before, the glass cover will affect the domain structure of amorphous wires. The magnetization of amorphous wires after glass removal will be different, and lead to different value of permeability. We could obtain a higher sensitivity of measured magnetic quantity [43]. The glass removal process can provide a strong effect on the magnetic and mechanical properties. In this thesis, we also study the magnetic properties of glass removal wires. There are two methods of removing the cover. The first is a chemistry method, which uses aqueous HF solution to dissolve the glass cover, a method utilized for the first time by Taylor [40]. It is necessary to control the reaction time accurately, otherwise the amorphous wires will be destroyed due to the strong HF solution. The other method is to use a tool to rub off the glass cover from the wire under a microscope. Using aqueous HF solution will be more complex and difficult. In our experiment, we use the second method to remove glass cover. 4.2 Experiment Procedure In experiments, we use a vector network analyzer to analyze signal transmission through amorphous wires. The vector network analyzer (VNA) is an essential tool which can be used to analyze the properties of electrical networks, especially those properties associated with the reflection and transmission of electrical signals known as scattering parameters and it is most commonly used in high frequency range. The range of operating frequencies is 10MHz to 12GHz. The entire experiment is shown as follows. Fig 4.3 Schematic structure of our experiments 25 The sample in the circuit can be described by a two-port network equivalent circuit, as Fig 4.4. There are 4 parameters Sij that are defined with the incident waves and reflected waves of input port and output port. S11, S12, S21 and S22 represent the input port voltage reflection coefficient, the reverse voltage gain, the forward voltage gain and the output port voltage reflection coefficient, respectively. Fig 4.4 Two-port network equivalent circuit of wire sample At port t 1, the input voltage and reflected voltage could express as V1+and V1 , and at port t 2, the input voltage is V2+ and V2 , then the relationship between the input voltage and reflected voltage could be expressed as V1 S11V1 S12V2 and V2 S21V1 S22V2 (4.1) Where S11 V1 V1 S12 V2 0 V1 V2 S21 V1 0 V2 V1 S22 V2 0 V2 V2 V1 0 (4.2) The return attenuation (S11[dB]) which represent the magnitude of reflected signals and insertion attenuation (S21[dB]) which represent the magnitude of transmitted signals : S11[dB] = 20log(S11 ) (4.3) S21[dB] = 20log(S 21 ) For the reciprocal network, |S12|=|S21|, for the symmetric network, |S11|=|S22|, for the non-loss network, |S 11|2+|S12|2=1. In this experiment, it is easy to know that the sample belongs to reciprocal network and symmetric network, but loss exists. Assume loss power is symbol ploss, and the loss coefficient η is ploss pin 0 1 (4.4) The loss attenuation Loss[dB] could be expressed as 2 2 Loss[dB] = 10log(η) = 10log(1 - ( S11 + S12 )) 26 (4.5) Chapter 5 5.1 Analysis In the process of signal transmission, there will be three kinds of signals. Because of the impedance mismatch, there will be reflected signal generated from the interface. Both wire and ground plane have finite conductivity and will exhibit some series resistance, and therefore lead to heat loss. When there is electric field across the medium, the alternating polarization of medium molecular and lattice collision will generate heat loss. The wire-field structure is semi-open, causing radiation loss. The heat loss and radiation loss constitute loss signal. The remaining part of the signal which transfers through the wire could be called the transmitted signal. Three aspects will be mainly considered in order to investigate the signal attenuation: The impact of frequency on signal attenuation, how direct current affects the signal attenuation, the comparison between glass removal wire and glass covered wire. 5.1.1 The characteristic analysis of signal attenuation Fig 5.1 (a) The attenuation of reflected signals (S11[dB]) with frequency at different dc bias currents, d 29 m, D 69 m, l 8.2cm glass covered wire (d is the metallic diameter and D is the diameter of wire with glass covered, l is the length of the wire 27 Fig 5.1 (b) The attenuation of loss signals (Loss[dB]) with frequency at different dc bias currents, d 29 m, D 69 m, l 8.2cm glass covered wire Fig5.1 (a) illustrates the changes of reflected signal which expressed by S 11[dB] with signal frequency at different dc bias currents. Generally, S11[dB] curves show a reduced trend as the frequency increases and have cyclical change. Different curves have the same periodic change with frquency. The explanation for the reason of cyclical change is careful in the next section. The amplitude of curves increases with the increased current and frequency. Fig5.1 (b) illustrates how the loss signal Loss[dB] changes with signal frequency at different dc currents. Generally for the wire, Loss[dB] curves show an increased tendency as the frequency increases. This is because high frequency will cause strong skin effect. This effect reduces the effective cross-sectional area of the wire, causing conductor loss to increase. Similarly, the curves show cyclical change with signal frequency. Loss[dB] increases with the increased frequency. As the signal frequency increases, the attenuation at different dc bias currents will tend to same constant value. 28 Fig 5.2 The attenuation of transmitted signals (S21[dB]) with frequency at different dc bias currents, d 29 m, D 69 m, l 8.2cm glass covered wire. Fig 5.2 shows the changes of transmitted signal S21[dB] with signal frequency at different dc bias currents. In low signal frequency range, S21[dB] curves appear great changes with increased dc bias current. With the increase in signal frequency, the influence of current to S21[dB] curves gradually weakens. When the signal frequency is higher than 10GHz, there is no change in S 21[dB] for different dc bias currents. The periodic variation of S 21[dB] curves will be more obvious as the dc bias current increases. In the absence of dc bias current, the trends of S 21[dB] curves with signal frequency can be expressed as Fig 5.3. We could see that the curve has a minimum value when signal frequency is 500MHz. Magnetization process consists of magnetic domain wall motion and domain rotation. At low frequency, the magnetic domain wall motion plays a major role on the magnetization process, while at higher frequency, the domain wall motion will be damped by the eddy current in amorphous wire and the domain rotation becomes more important. The increased signal frequency will increase the permeability. The increased signal frequency and permeability will enhance the skin effect and GMI effect, causing the impedance and attenuation increased. With further increase of signal frequency, the skin effect will not be enhanced any more, but the eddy current in amorphous wires will still damp the magnetization, causing a decline in permeability. This is not good for GMI effect. At extremely high frequency, the effective permeability will be reduced to a very small value, then the changes of permeability caused by magnetic field can be negligible, that is to say, the magnetic field doesn’t effect on the skin effect, resulting in the disappearance of GMI effect. When the dc bias current passes through, the current will obstruct the eddy current which is generated by high frequency, leading to inhibition of skin effect. 29 Fig 5.3 In the absence of dc bias current, the attenuation of transmit signals with frequency As known, all three kinds of attenuations show cyclical variation with frequency and the periodic frequency is near 1.6GHz. Under the condition that impedance matching, there is no reflected signal, and the magnitude of the voltage along the wire is constant, equal to | V1 |. When the impedance is mismatched, a reflected signal will be generated. Then the incident and reflected signals could produce interference and generate a standing wave along the wire. For a low-loss transmission line, the voltage at any point on the wire is given by V V1 e z V1 e z (5.1) Where j , Γ is the voltage reflection coefficient, by its definition, we can write V V So the magnitude of V is given by | V || V1 ||1 e2 z || V1 ||1 e2 l | (5.2) Where l = -z is the position distance from port 2 to port 1 in Fig 5.4 Let equal | | e j , then | V || V1 | e 2 l |1 | | e j ( 2 l ) | | V | e 2 l [(1 | |) 2 4 | | sin 2 ( l )] 2 1 2 (5.3) 2 l This equation illustrates the value of | V | fluctuating between maximum value | V1 | e (1 | |), when l 2 n and minimum value | V1 | e 2 l (1 | |), when l 2 n 2, here n is an integer. These results show that maximum voltage occurs when the incident and reflected waves add in same phase and the minimum voltage occurs when they add in opposite phase, as Fig 5.4. Maximum and minimum voltages appear when the interval is l 2 2, where is the wavelength. 30 Fig 5.4 the schematic diagram for the transmitted and reflected signals in the wire In Fig 5.2, wire length is l n l 8.2cm , using the formula (3.20), we can get the effective dielectric constant e 5.5 , Hence, f c0 3 10 8 2l e 2 8.2 102 5.5 0.78 109 [ Hz ] (5.4) This result is consistent with the curves. From this formula, we could also find that at a certain frequency, as the wire length changes, the signal attenuation will also show periodic variation. Fig 5.5 illustrates the changes of S21 with wire length at different frequencies. We could see that, somewhere between 4.2cm and 6.5cm, there will be a maximum value for S 21. Fig 5.5 Transmitted signal attenuation (S21[dB]) with wire length at different frequencies for d 15.6 m, D 47 m, I 10.5mA wire Consider the curve at frequency f 2.5GHz , we can get the wavelength 2.5GHz c f e 3 108 5.1[cm] 2.5 109 5.5 (5.5) l n 2 n 2.55cm , when n=2, and l=5.10cm, then the attenuation of S21 is maximum. Because of λ2.5GHz = 2 λ5GHz = 3 λ7.5GHz = 4 λ10GHz , which mean all the wavelength of them are multiples of 2.5GHz, therefore, using formula (5.5), all curves have the peak value l=5.10cm. 31 5.1.2 Direct current effects on the signal attenuation The process is shown in Fig 5.6. We control the dc current applied on both ends of the sample and try to find the relationship between current and transmitted signal attenuation. According to our requirements, if the control current has small change, the transmitted signal attenuation should have significant changes. There is some power consumption in wire at the process of current controlled. Obviously, lower dc power loss gives better amorphous wire transmission performances. The transmission performances of amorphous magnetic wires can be judged by several factors, so we should find a way to compare the performances that contain those factors. We introduce a figure of merit (FOM) as follows in order to compare transmission performances. Q A I 1 A 1 A AMax Max I p p (5.6) Where ΔA(I ) ΔI is the ratio of attenuation change to current change, and it describes the sensitive of attenuation to the current, ΔI Δp is the ratio of current vibration to power consumption vibration, AMax is the maximum attenuation. Then, we can judge the wires’ performances, higher Q-value gives better transmission performances. Here we can calculate the power consumption p. Fig 5.6 shows the schematic of current control circuit. Fig 5.6 Schematic of Current control circuit In the experiment, a series resistance (481Ω) is connected to the circuit. To calculate power consumption on samples, we should remove the power consumption in R. p UI I 2 R (5.7) Where U is the voltage, I is the current applied on the sample, R is the series resistance. We choose the wire with d 30 m, D 65 m, l 8.2cm as an example. Base on the experimental results, use the FOM, we get the Q-value of this wire at different frequency as follow table. Frequency 10MHz 100MHz 500MHz 1GHz 2.5GHz 5GHz 7.5GHz 10GHz 12GHz sample 4.23 5.87 6.34 12.78 10.25 6.65 3.88 1.74 1.23 Table 5.1 the Q-values for the amorphous wire with d 30 m, D 65 m, l 8.2cm. 32 From the above table, we could know sample have better performance at frequency 1GHz. We choose this position as the analysis point. Fig 5.7 shows the transmitted signal attenuation change with control current for different wires at frequency 1GHz. Fig 5.7 Control current and transmitted signal attenuation (S21[dB]) for different samples at 1GHz, d is the metallic diameter and D is the diameter of wire with glass covered. It is easy to see that S 21[dB] curves for four kinds of wires have similar trends. For a specific wire, in the absence of current, the attenuation values of S 21 are minimums. When increasing current, S21[dB] increases and the increment of S21[dB] decreases gradually. The reason for such change is that as the controlled current increases, the magnetic field around the amorphous wire will change, thereby weakening the magnetic permeability. The increase of skin depth leads to reduced amorphous wire resistance, in turn causing the increase of S21[dB]. By some calculations, we try to deduce empirical formula in line with S 21[dB] curve. Using formula (2.6) and skin-effect resistance, it’s easy to obtain the resistance R(I) which changes with current. R( I ) Rs ( I )l (5.8) (d ) Rs is the skin-effect resistance which is equal to 1 (σδ ) , is the skin depth, is the conductivity of the wire, l is the wire length, d is the metallic diameter. In the experiment, the frequency range is from 10MHz~12GHz, the skin depth is very small and produce a strong skin effect, d , formula (5.8) can be approximated as: R( I ) Rs ( I )l (5.9) d 33 Using formulas (3.27), (3.28) and (5.9) R( I ) Z c A( I ) 8.686C d [] (5.10) Where A(I) is a function between attenuation and control current. For a certain wire ，at a given frequency, ZC is the characteristics impedance, and LR 1 1 1 1 C [m ] 2 ln 4 H (W H 5.8 0.03 H W ) W Choose sample with d 29 m, D 69 m, l 8.2cm. For the sample, which the height of the 3 substrate H 1.5 10 m and relative permittivity r 4, through the formulas (3.16), (3.19), (3.20), (3.22) and (3.27), we could obtain Z C 153 and C 1.93 104 m-1, then R( I ) 9.7 A( I )[] (5.11) The relationship between resistance and current is shown in Fig 5.8. Using Matlab, we could find an appropriate but simple function as follows. R( I ) [ R0 R1e I I0 ][] (5.12) Where R0 , R1 , I 0 can be obtained from a certain curve, the unit of R0 , R1 are Ω, and I 0 is ampere. Combine with the formula (5.10), we could find the relationship between attenuation and current. A( I ) I 8.686C d ( R0 R1e I0 )[dB] Zc (5.13) For the sample, we could obtain: R( I ) 80 122e I 13.8103 [] and A( I ) [8.25 12.58e I 13.8103 ][dB] Fig 5.8 Impedance and Current for d = 30μm, D = 65μm,l = 8.2cm wire 34 For the other wires, get the empirical formula and the parameters are demonstrated as follows d, D (μm) l(cm) C (m-1) 15.6，47 8.2 8.2 20，54 29，69 30，54 8.2 8.2 ZC (Ω) R0(Ω) R1(Ω) I0 (10-3 A) 3.69x104 170 176 103 6.1 2.88 x10 4 164 139 57 4.8 1.99 x10 4 154 120 100 6.8 1.93 x10 4 153 80 122 13.8 Table 5.2 The parameters of the empirical formula of each wire 5.1.3 Amorphous Metallic Wire without Glass Cover As known, magnetic domain structure of amorphous metal wire is not only relevant to material stress, but also relevant to the surface state of wire. Because glass and metallic core have different expansion coefficients, there will be stress difference between them. The domain structures of amorphous glass covered wire and the wire after glass removal should be different. The wires without glass cover should have lower domain wall energy. At the same magnetic field, the wire after glass removal will be more easily to be magnetized. This means that the permeability of wire after glass removal is larger, so that the skin effect will be stronger and we could obtain larger impedance and signal attenuation. Therefore, it affects the circumference permeability of amorphous wire and further leads to changes in transmitted signal. Fig 5.9 The attenuation of transmitted signals (S21[dB]) with frequency at different currents for glass removal wire and glass covered wire The comparison of S 21 for glass removal and for glass covered wire is demonstrated in Fig 5.9. Generally, the curves have same trend of variation. In the absence of current and at a low signal frequency, the magnitude of attenuation S 21 for glass removal is smaller than for glass covered. As the current or frequency increases, S21 of glass removal and glass covered will approach. In order to compare the performance of the two curves, we use the formula (5.6). 35 Choose the variation value of power consumption p from 0 to 70mV. Table 5.3 shows Q-value for glass covered and glass removal wire at different frequencies. Frequency 10MHz 100MHz 500MHz 1GHz 2.5GHz 5GHz 7.5GHz 10GHz 12GHz Glass covered wire 5.00 6.40 5.91 9.19 10.59 7.93 3.40 1.50 0.73 Glass removal wire 11.20 15.37 18.82 25.01 18.89 9.41 4.41 1.50 2.06 Table 5.3 the Q-values for glass covered wire and glass removal wire From the table, we found that at low frequencies, Q-values of glass removal wires are significantly larger than that of the glass covered wire. That is to say, for glass removal wire, in this frequency range, current play a great role on regulation of transmitted signal. With frequency increasing, the two Q-values are gradually approaching. 5.1.4 Impedance matching analysis From the preceding analysis we saw that the reflected signal was found to be much greater than the transmitted signal，which is quite negative to signal transmission. As a result, we need to consider reducing the reflected signals. The prevailing method is to make wire impedance equal to transmission line characteristic impedance, so as to achieve impedance matching. Choose the metallic radius 30μm wire as a sample. When there is no current applied, its impedance is about 200Ω which is much bigger than transmission line characteristic impedance. The impedance mismatched causes strong reflection, so the main point is to make the load impedance as close to 50Ω as possible. In order to achieve impedance matching, we need to adjust the value of the load impedance Z 1 by coupling with inductor or capacitor in the circuit. As shown in Fig 5.10 Fig 5.10 Couple in capacitance and inductance, to achieve impedance matching For our sample with d 30 m, D 65 m, l 7cm , the substrate dielectric constant is 3, and thickness is 0.7mm. The condition of impedance matching is: Z1 1 L ZC Z0 C (5.14) For the chosen wire, the impedance of the wire Z 0 = 200Ω, and the transmission line characteristic impedance Z C = 50Ω. We then get the following relationship between inductance and capacitance: 36 L[ H ] C[ F ] 1104 [2 ] (5.15) Choose an inductor 744760118C produced by Wurth company, whose inductance is 18nH, the maximum current 500mA, maximum direct current resistance 200mΩ and self-resonant frequency 3.3GHz. According to formula (5.15), we use the capacitor with capacitance 1.8pF. Through approximate impedance matching, we obtained results shown in Fig5.11 Fig 5.11 The attenuation of three signals with frequency, I=26mA S21[dB] increases with signal frequency, reaches its maximum value at 640 MHz, then sharply declines with increasing frequency. It reaches its minimum value at 2.5GHz, after which it will increase slowly. A frequency of 640MHz thus makes the impedance matching better. In the frequency range from 640MHz to 2.5GHz, a series of inductance in the circuit will obstruct high frequency, and most of the signal is reflected. When the frequency is larger than the inductance self-resonant frequency, the inductance parasitic capacitance leaks the signal through it. 37 Fig 5.12 (a) Measurement of S11[dB], S21[dB] and Loss[dB] with controlled current at f=640MHz. Fig 5.12 (b) The load impedance in Fig 5.10 changes with controlled current at f=640MHz. Figure 5.12 (a) illustrates that signals change with the controlled current. S21[dB] increases with increasing dc bias current. When the current is larger than 10mA, it is larger than S11[dB], and the minimum value of S21[dB] is close to -4dB. Using the formula (5.10), we can get the experimental impedance of the combination with wire, inductors and capacitors as shown in Fig 5.12 (b). When current is larger than 26mA, the load impedance Z 1 is approaching to 50Ω which is equal to the transmission line characteristic impedance ZC and could be considered as impedance matching. The expression of theoretical load impedance in Fig 5.12 (b) can be written as R( I ) 50 76.7e I 7.6103 [] 38 Fig 5.13 the attenuation of S21 with the power consumption Fig 5.13 demonstrates S21[dB] changes with the power consumption in the wire. In the range 0-70mW (when the current ranges from 0-26mA), S 21[dB] obviously changes, but when power consumption is larger than 70mW, it doesn’t change significantly. Therefore, in the application, we can set the controlled current below 26mA. Using the formula (5.6), the Q-value is 17.8. 5.2 Device 5.2.1 Electronically controlled attenuators In a microwave system, it is necessary to couple in an attenuator in the system in order to control the transmitted power, thus leading to better performance. There are two kinds of attenuators: fixed attenuator and variable attenuator. For the fixed attenuator, the attenuation is constant and for the variable attenuator, it could be adjusted within a certain range. The attenuator is a two-port network, and the attenuation can be achieved in two ways. One is due to the internal dissipation in the network, part of power will radiate as heat, then attenuation be achieved. This kind of attenuation is called absorptive attenuation. At the network input port, the electromagnetic waves will reflect, causing the power attenuation. This is the other way to achieve the attenuation and we call it reflection attenuation. At present, the microwave attenuator has been developed in a variety of structures. In addition to attenuation resistor networks, electrically controlled attenuators based on PIN diodes also exist. In the following we will discuss the advantages and disadvantages of using the amorphous wire as electrically controlled attenuator component. The basic structure of the attenuator is shown in the following Fig 5.14 39 Fig 5.14 Schematic of amorphous wires electronically controlled attenuator. Where L=18nH, C=1.8pF, R=481Ω, I<26mA, the wire has the dimensions d 30 m, D 65 m and l 7cm. According to the analysis of Chapter 5.1.4，we concluded that a wire which is impedance matched by capacitors and inductors has better performance. From Fig 5.11 we know, that better transmission performance can be obtained at 640MHz. When the control current is 26mA, we could obtain impedance matching. From this we determine the attenuator work frequency range 500-700MHz. Fig 5.15(a) Insertion loss (S21 [dB]) vs control current 40 Fig 5.15(b) Return loss (S11[dB]) VS. Control current Both reflected attenuation and absorbed attenuation exist in the wire attenuator. When the current is 26mA, the circuit is closed to impedance matching. Then from Fig 5.15(a) we know, the insertion loss is -4.0dB to -5.0dB, from Fig 5.15(b), the return loss is -5.7dB to-7.7dB, the voltage standing wave coefficient (VSWR) is less than 3. Fig 5.16 The range of attenuation (the range between maximum value and minimum value of attenuation) when the control current I<26mA at different frequencies Fig 5.16 illustrates the range of attenuation, when the control current I<26mA, the magnitude of it is between 5dB-7dB. 41 Fig 5.17(a) Variation of attenuation vs current Fig 5.17(b) Variation of attenuation vs power consumption Fig 5.17(a) demonstrates the variation of attenuation with control current. Generally the variation 3 is in line with A(dB) 4 6.13e I 7.610 at 640MHz, and the range of control current is from 0-26mA. The power consumption is less than 70mW. 42 The previous results are in negative numbers, which actually are expressed as gain. For the attenuator performance parameters, we should use attenuation to express them, then we drop the minus sign as it is indicated by term loss in "Insertion Loss", "Return Loss" etc. So from the above information, we can summarize the performance data as follows: Parameter Frequency Min Max Unit Insertion loss (I=26mA) 500-700MHz 4 5 dB Characteristics of the attenuation 500-700MHz 5.2 7 dB 640MHz 4 10.5 dB 500-700MHz 5.7 7.7 dB Control current 0 26 mA Control Voltage 0 15 V Power consumption 0 70 mW Attenuation Return loss (I=26mA) Table 5.4 The specifications of wire attenuator According to the range of work frequency, we choose the voltage variable attenuator type SKY12328-350LF produced by Skyworks company as a comparison. The performance data sheet of this attenuator is shown in the table below. Parameter Frequency Min Insertion loss 0.5–1.0 GHz 1.1 Characteristics of the attenuation ® 0.5–1.0 GHz Typical Max Unit 1.3 dB 15.5 Attenuation 900MHz 0 Return loss 0.5–1.0 GHz dB 15.5 dB 15 18 dB Control current 15 100 μA Control Voltage 2.7 5.5 V 500 μW Power Consumption ®. Characteristics of the attenuation include insertion loss. Table 5.5 the specifications of voltage variable attenuator (Type: SKY12328-350LF), Z 0 = 50Ω unless otherwise noted Comparing the two attenuators, we can obtain the advantages and disadvantages of the wire attenuator. The advantages include a simple structure and a current controlled tunable attenuation. The disadvantages are small insertion loss, large return loss and VSWR, large control current, large power consumption, and a relatively small range of attenuation. According to the application, we choose the different wires or inductors and capacitors to adjust the attenuation value. These could be applied in narrow-band electronic system. 43 Chapter 6 6.1 Conclusion This thesis work focuses on an investigation of a current controlled microwave attenuator based on amorphous metallic wires. We discuss the magnetic properties of the amorphous wires, such as GMI effect and skin effect, as well as their relationship. In a series of experiments we invetigated the influence of an applied direct current, signal frequency and the amorphous wire length on reflected signals, loss signals and transmitted signal attenuation. There are two kinds of amorphous wires used in the experiments. One is amorphous wires covered with glass prepared by a high frequency induction heating melt spinning method. The other type is amorphous wires where the glass has been removed by a mechanical rubbing technique. The sample consists of an amorphous metallic wire, a copper ground plane and two connectors. Based on the similarity between our samples and conventional microstrip lines, expressions for the attenuation could readily be deduced. According to the experiment results, some conclusions can be made: 1) In the process of signal transmission, signals can be divided into reflected signals, loss signals and transmitted signals. The attenuation of these three signals exhibit the same periodic variation with frequency, and the period frequency is fT=1.6GHz. Reflected signals will exist due to impedance mismatch, and the incident and reflected signal will produce interference and generate standing waves along the wire. 2) A figure of merit Q A (p AMax ) is introduced where larger Q indicates better transmission performance. Best Q-value is achieved at 1GHz in the experiments. The transmitted signal attenuations have similar trends for all types of wires, and in the absence of current , the signal attenuation is lowest. With increasing DC control current the attenuation increases and attenuation increment decreases gradually. The empirical formula in line with this curve is A( I ) (8.686C d Z C ) ( R0 R1e I I0 )[dB ] . 3) The glass cover will influence the magnetic domain structure in amorphous metallic wires. Therefore, it will affect the circumference permeability and change the transmitted signal attenuation. At low frequencies (f<7.5GHz), the Q-value of wires with removed glass cover is larger than that of the glass covered wires. With increasing frequency, the two Q-values are gradually approaching each other. For most frequencies, the DC current for glass removed wires leads to a better control of the attenuation of the transmitted signal. 4) It is possible to impedance match the wire with a simple LC network also working as a DC-block. Under impedance matched conditions, the transmitted signal attenuation increases with signal frequency and reaches its maximum value at 640MHz. The minimum attenuation of S 21 is I close to 4dB. The theoretical impedance of wire is R( I ) 50 76.7e 7.610 []. 3 44 5) The magnetic wires can be used as microwave attenuators with tunable transmission. The frequency range spans from 500MHz to 700MHz. The circuit will achieve impedance matching when the current is 26mA. The insertion loss is from -4.0dB to -5.0dB, the return loss is -5.7dB to -7.7dB, the voltage standing wave coefficient VSWR is less than 3. The range of current is from 0-26mA. The power consumption is less than 70mW. In comparison between wire attenuators and the commercial voltage controlled attenuator SKY12146-321, our design offers simplicity and easily controlled attenuation. The disadvantage is the insertion loss, large return loss, large control current, power consumption, and the relative small adjustable range of attenuation. 6.2 Future Work On completion of the previous research, we propose the following further work. 1. Effects of external factors. To date, we have no comprehensive study about external factors such as stress on the wire, temperature, etc. The external factors will affect magnetic permeability, causing attenuation change. Because it is complex to control external factor, it wasn’t researched completely in this paper. 2. Composition of amorphous metallic wires. The amorphous metallic wire with the composition FeCoSiB is widely used, but it hard to say that it is the best one. The composition is the most important factor to decide the property of wire, and we can try to find a better composition which has better properties. 3. Improve the construction of microwave components. This thesis work focuses on the attenuator. The circuit for the attenuator is not optimized, impedance matching is not complete, and return losses are large. 45 Reference: [1] Development and Status of microwave components, Lan Ma, Electronic Products China, Page 39-41, May 2008 [2] Materials Science and Engineering for the 1990s: Maintaining Competiveness in the Age of Materials, National Research Council (U.S.). Committee on Materials Science and Engineering, 1989 [3] Amorphous Glass-Covered magnetic Wires for Sensing Applications, H. Chiriac, T. A. Óvári, Gh. Pop and Firuta Barariu, Sensors and Attenuators, 1996 [4] Magnetic Properties of Amorphous Fe-P Alloys Containing Ga, Ge and As, Wiesner. H and Schneider. J, Phys. Stat. Sol, 1974 [5] Preparation of amorphous wires, Ohnaka. I, Fukusako. T and Matui. T, J. Jpn Inst. Met, 1981 [6] Magnetic Amorphous Glass-Covered Wires, Horia CHIRIAC, Tibor-Adrian OVARI, National Institute of R&D for Technical Physics, Romania, 1999 [7] Applications of amorphous and nanocrystalline magnetic materials, J Magn Magn Materials, 1994, [8] Large Barkhausen Effect and Matteucci Effect in Cold-Drawn Amorphous Magnetostrictive Wires, Mohri K, et al, Journal of the Magnetics Society of Japan, 1984, 8:233 [9] Magnetic Materials, Wun-Fogle M, et al, IEEE Trans Magn, 1989, 25:3617 [10] Angiogenin cleaves tRNA and promotes stress-induced translational repression Yamasaki J, et al, IEEE Trans Magn, 1993, 29:2545 [11] Hiroshi Kimura, et al, Journal of the Magnetics Society of Japan, 1991, 15:475 [12] Domain Wall Propagation in Nearly Zero Magnetostrictive Amorphous Microwires, Chiriac. H, Ovari, Tibor-Adrian, Tibu, Mihai, IEEE Trans Magn, 2008Vol.44 [13] Thin film magnetic field sensor utilizing magneto impedance effect, YuJi Nishibe, Norikazu Ohta, Toyota Central Research Vol.35 No.4 2000 [14] Magneto-impedance effect in amorphous wires, Panina L V et al. Appl Phy Lett, 1994, 65 1189 [15] Non-linear electromagnetic systems: advanced techniques and mathematical methods, Volkmar Kose, J. Sievert, IOS Press ,1998 [16] Composites with large magnetostriction, Basheer Rafil A, Morelli Donald T, Delphi Technologies Inc, 2005 [17] Magneto-induced effect in amorphous wires, Mohri K, Kohzawa T, et al, 1[J] IEEE Trans Magn ,1992 ,28 :31501 [18] Giant magnetic field dependent impedance of amorphous FeCoSiB wire, Beach R S, Berkowitz A E, [J] 1 Appl Phys Lett ,1994 ,64 :3652 [19] Magnetization Processes with Orthogonal Fields in Amorphous Magnetostrictive Wires, Panina L V, Mohri K,et al.[J]. Sensors and Actuators A, 1997, 59:20 [20] Very large magneto-impedance in amorphous soft ferromagnetic wires, Rao K V, Humphrey F B, et al.[J]. J Appl Phys, 1994 46 [21] Giant magneto-impedance effects and their applications, Journal of Tianjin institute of technology, 2002 [22] Micro magnetic sensor based on the giant magneto-impedance effect of the amorphous wire, Sun ji, Deng jihao, Gao zhen, Song song, School of Aerospace Science and Engineering, Beijing Institute of Technology, Beijing 100081, China [23] Central Electricity Generating Board, Modern Power Station Practice. Pergamon Press, 1982 [24] Fields and Waves in Communications Electronics, Ramo, Whinnery, Van Duzer, John Wiley and Sons, 1994 [25] Engineering Electromagnetics Seventh Edition, Hayt, William Hart, New York McGraw Hill, 2006 [26] Fields and Waves in Communication Electronics, Simon Ramo, John R. Whinnery, Theodore Van Duzer, New York John Wiley & Sons Inc, 1965. [27] Magnetic field transducers based on the phase characteristics of GMI sensors and aimed at biomedical applications, Chwee Teck Lim and James C.H.Goh 13th international conference on biomedical engineering, 2008 [28] Electromagnetic compatibility handbook, Kenneth L. Kaiser, CRC, ISBN 0-8493-2087-9, 2004 [29] Electrical wave transmission, Pupin, M., U.S. Patent 1,541,845. [30] Modern Transmission Line Theory and Applications, L.N.Dworsky, New York John Wiley & Sons, 1979 [31] Microwave Electronic Circuit Technology, Yoshihiro Konishi, CRC, ISBN 0-8247-0101-1, 1998 [32] Introduction to microwave circuits: radio frequency and design applications, by Robe rt J. Weber, IEEE Microwave Theory and Techniques Society, 2001 [33] A Handbook for Wireless/ RF and High-Speed Electronics, Ron Schmitt, Elsevier science (USA), Newnes, ISBN 0-7506-7403-2, 2002 [34] Practical Electronics for Inventors, Paul Scherz, McGraw-Hill/TAB Electronics 1 edition ISBN-13 978-0-07-142581-6, 2004 [35] Practical RF Circuit Design for Modern Wireless Systems: Passive Circuits, Les Besser, Rowan Gilmore, Artech House microwave library, 2003 [36] Introduction to Electromagnetic and Microwave Engineering, Paul R. Karmel, Gabriel D. Colef, Raymond L. Camisa, Wiley-Interscience 1998 [37] A Designer's Guide to Microstrip Line, I.J. Bahl, D.K. Trivedi, MicroWaves pp.174-182, May 1977 [38] Foundations for microwave engineering, Robert E. Collin, Wiley-IEEE Press, ISBN 0-07-112569-8, 1992 [39] A Method of Drawing Metallic Filaments and a Discussion of their Properties and Uses, G.F. Taylor, Phys.Rev.23, 655-660, 1924 [40] A class of micromachined magnetic resonator for high-frequency magnetic sensor applications, Ulitovsky. A.V, Pribory. Tech. Eksper, 3,115, J. Appl. Phys 1957 47 [41] Glass-coated Fe-based amorphous alloy preparation methods of wires, Zhao Zhengjie, Yuan Jianzhong, Fang Yunzhang, Jiang Keyu, Yang Xielong, Journal of East China Normal University (Natural Science), 2002 [42] Glass-coated amorphous ferromagnetic microwires at microwave frequencies, Manuel Vazquez, Anne-Lise Adenot-Engelvin, Journal of Magnetism and Magnetic Materials 321, 2066-2073, 2009 [43] Amorphous glass-covered magnetic wires: Preparation, Properties, Applications, H. Chiriac and T. A. Óvári, national institute of R&D for technical physics lasi, Romania, 1996 48

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement