Contents:
Check copyright status Cite this Title Heteromagnetic microelectronics: Author Ignatiev, Alexander A. Other Authors Lyashenko, Alexander V. Microsystems of Active Type, by Alexander A. Lyashenko of JSC Research Institute Tantal in Russia, offers a very detailed and specialized account of the author's research and development of heteromagnetic materials and devices. The book is based on original material from the author's programs of designing heteromagnetic microsystems. Polyvalent, multiple parameter magneto-semiconductor microsystems are described and the book reports on extensive experimental and theoretical results of research in a range of frequencies up to GHz.
For the first time the direction of satisfying criteria, and burst technologies, which can make a subject of discovery, are discussed in great detail. This book is intended for post-graduate students and researchers specializing in the design and application of heteromagnetic materials and devices.
Contents Machine generated contents note: Spectra of Regular and Noise Signals 1. Regimes of Low and Middle Power Levels 1. Regimes of High Power Level 1. Multifunctional Properties of Powerful Heteromagnetic Oscillators 1. Properties of Structures with Ferrites of Different Magnetizations 2. Structures with Ferrite KG-8 2. Structures with Ferrite KG 2. Generalization of Experimental Data 3. Control Over Energy and Spectral Characteristics 3. Control Over Characteristics of Spectral-Pure signals 3. Structures with Various Orientations in a Magnetic Field 3.
Structures with Ferrites of Various Magnetization 3. Structures with Ferrites of Various Magnetization 4. Generalization Control Characteristics in Generative Structures 4. Structure Characteristics with Various Orientations 4. Structure Characteristics with Various Magnetizations 4. Physical Mechanisms of Heteromagnetic Interactions pt.
Basic Model Equations 5. Modeling of Complicated Regimes Contents note continued: Multicircuit Model of a Multifunctional Heteromagnetic Oscillator 6. Methods of Finalizing Equivalent Parameters of Transistor 6. Equivalent Circuit of a Multifunctional Heteromagnetic Oscillator 6. Oscillating Modes of Subharmonic Constituents 6.
Regimes of Pseudonoise Signals pt. General Data on Programs 7. Method for Determination of Transistor Parameters 7. Magnetoelectronic Elements of LPL 7. IV Applied Aspects 9. Influence of External Factors 9. Estimation of Static Load 9. Strength of Beam-Type Bonds Contents note continued: Strength of Glue Fixation 9. Strength of Screw Connection 9. Resistivity to Dynamic Forces 9. Resistivity to Pressure Changes 9. Resistivity to Temperature Excitations 9. The spectral line computed for the HMG model is narrower than that observed experimentally.
It is explained that only noise properties of the passive elements and transistor have been considered in the model. Technical noises caused by the influence of feed 5 Heteromagnetic Oscillator Fig. For their account, pulsation generators in a loading mode of 2—10 mV were placed in the circuit of ideal power sources. In view of the instability of power sources the spectral line width increased up to the values observed experimentally Fig. Our researches have shown that the structures consisting of several elementary transistors at modeling can be replaced by one transistor with the corresponding parameters that considerably simplifies calculations.
By experimental determination of the spectral lines width of signal harmonics of the generator in classical multiplication modes1 their broadening was observed at transition to higher harmonics. The results of calculation of spectral line widths for the third harmonic of HMG signal on a frequency of 1, MHz are presented in Fig. In comparison with the results in Fig. The results of measurement and calculation of generation frequency and spectral line width depending on the number of the harmonics are shown in Table 5.
Analyzing the results of calculations and experimental data, it is possible to draw a conclusion on their good conformity that confirms the correctness of our choice of the equivalent HMG circuit and determination of its parameters. The analysis was made with the use of standard harmonic analysis. However, at such an approach it is impossible to solve the problem of stability of the obtained stationary solutions of the equations describing the generator and to pass to consideration of more complex operating modes, including irregular and complex multifrequency modes with modulation of the envelope of high-frequency oscillations.
Such modes exist in HMG, as follows from experiments, both in the presence of ferrite microresonators built in the magneto transistor and in their absence. A method of solution of ordinary differential equations describing its equivalent circuit was applied to description of such operating HMG modes. For simplification of analysis the full equivalent circuit of the generator was separated into external and internal ones.
For derivation of the equations describing the investigated generator, a method of central potentials with operational recording of capacitor and inductive conductances was used. Three units are connected to each other by active resistances and their potentials are related algebraically. For assemblages amount we shall make the equations, including that for units the sum of all flowing currents is equal to zero, 5 Heteromagnetic Oscillator i. Let us make the following substitutions: The currents in these units are described by the equations dIc 1.
Potentials 'c ; 'b ; 'e can be set as external conditions or be calculated as the solution of 5. The transistor is described by a set of 17 nonlinear ordinary differential equations. The external equivalent circuit of the investigated generator is presented in Fig. In experiments for realization of effective interactions with the transistor structure, the FMCR was placed directly above the emitter area of the transistor.
At a distance from the base, the multiple-parameter microfield interaction decreased. Therefore, at modeling of HMG a nonlinear magnetic-field-controlled contour Lf ; Cf ; Rf was placed in the equivalent circuit in the emitter area. The influence of the current emitter on the base current due to heteromagnetic interactions was described by elements Lf ; Lm and connection M between them. The equivalent circuit of HMG is presented in Fig. Let us transform the equations to the form as in 5.
The investigated equation set is complex enough, its solution demands significant computer time. No analysis of generation modes on the basis of the bifurcation theory has been made here. This parameter changes due to the magnetic field value in the HMG structure. A source modeling technical noise of the generator was placed into the collector circuit. The voltage of this source was set equal to 10 mV. The spectral power density Fig.
The structure of these domains essentially depends on the material, shape, and sizes of the sample, an external magnetizing field [18]. Nonlinear effects in the ferrite are shown from power levels of the order of 0: The resonant frequencies of FMCR depend on the saturation magnetization of the ferrite, the field of anisotropy, the orientation in an external constant magnetic field, the kind of polarization of the high-frequency magnetic field exciting oscillations of the magnetization vector in the ferrite.
Modeling of interaction of FMCR with a highfrequency magnetic field of a semi-conductor structure in unsaturated nonlinear modes was conducted: Therefore, FMCR was considered as a concentrated oscillatory system without regard to wave effects. In the structures with monoaxial anisotropy or a cubic crystal the domain structure is [17, 18] presented in Fig. The ferrite sample is divided into sites with opposite magnetizations in the neighboring domains.
In the model accepted by us, the domain structure is homogeneous and two domains with different magnetization orientations participate in the interaction with HF fields. At analysis of FMCR in unsaturated modes it is necessary to consider magnetization oscillations in domains and domain border oscillations simultaneously [43]. For a single domain A. In a more general case, the number of normal oscillations can exceed three. All the oscillations are coupled. Coupling between the 6. Thus, in the heteromagnetic generator, FMCR is modeled by a multiconnected nonlinear oscillatory system.
The frequencies of oscillations of the domain walls and the normal domain oscillation frequencies in the trivial case can be estimated theoretically [17]. This estimation was used as an initial approximation of the parameters of the equivalent circuit of the ferrite sample. The equivalent parameters of the oscillatory contours were estimated on the basis of the results of our research of the absorption spectra of the used ferrite samples in the working ranges of a magnetic field H 0 at various levels of high-frequency power. The equivalent parameters of HMG were optimized by the technique presented below.
The halfwidth of the ferromagnetic resonance line its nominal value determined the parameter of equivalent GB product. At changes of the bias field H0 for each oscillatory contour Fig. Due to the interactions of the HF magnetic fields of FMCR — hYIG and the HF magnetic fields of the emitter and base currents hbe , additional inductances coupled with each other and with the multicoherent equivalent contours appear in the base and emitter circuits of the transistor.
These multicoherent equivalent contours simulate nonlinear ferromagnetic oscillations and domain wall oscillations in the ferrite sample. Equivalent circuits of a multipurpose powerful HMG are presented in Fig. No direct current power elements of the transistor are shown in this figure. The elements Cg and Z are external in relation to the heteromagnetic transistor. The transistor is a component of the heteromagnetic structure that can be used in various generating, mixing, intensifying, and other modes.
All the unknown parameters of HMG were divided on static and dynamic ones. The static parameters were determined by means of measurements of the input and output characteristics families at direct current. The dynamic parameters of HMG were determined from experiments with UHF signals in various generating modes from the dependences of the signal generation frequency on the collector voltage at various displacement voltages on the base of the transistor.
The parameters of the equivalent circuit of HMG were optimized by the least squares method with the usage of groups of equivalent parameters and experimental characteristics. A group contained the parameters most essentially influencing a given family of characteristics. Let a family of characteristics of the heteromagnetic structure be set by the functions f1. For example, for static output characteristics these are the dependences of the collector current Ic1 ;: Optimization of the parameters was carried out by the minimum distinction of the areas between the experimental and calculated dependences summarized over all the curves of a family Fig.
Further, piecewise-linear approximation of the experimental and calculated curves was made and the total area of the trapezes was calculated Fig. From the family of target static characteristics of the KTB transistor resulted in Fig. The parameter Rb2 has the minimal error at U D 65 V.
This result is caused by some specificity of measurements of target characteristics when the emitter current is fixed. The parameter Re1 can be found from the family of input characteristics of the transistor presented in Fig. Therefore, at optimization, these parameters, together with the parameters Nf ; NR ; NE ; NC entering into the diode exponents, were united in a group to be processed first of all. The dependences of the error functions on the parameters VA and VB , which enter in expression for source of current, are presented in Figs.
The parameter VA to act the immaterial part and appeared at inverse of transistor switching-on. One can see that the parameters IKF and BR cannot be found from the family of target characteristics. The static equivalent parameters of the HMG model were sought by the gradient descent method. For calculation of the static characteristics, the set of nonlinear algebraic equations derived from 6. At a wrong choice of such groups, a loss of the physical sense of the optimized parameters a negative resistance, etc.
At the first stage, the method of gradient descent was applied in the space of these parameters. The parameters of a foreign analog of the used transistor were taken as an initial approximation. The second group included the parameters Rb2 ; Rc2 ; Re1 , and Rbb setting the base, collector, and emitter resistances. After application of the method of gradient descent consistently on the three groups of parameters, the optimization was finished with minimization of the error function in the full space of the required parameters of the model.
For determination of the dynamic parameters of the equivalent circuit, the family of experimental dependences of the generation frequency on the collector voltage was used at fixed displacement voltages on the base. The results of measurements of these dependences at various voltages on the base UBC D 11;: This family of characteristics was used for optimization of the dynamic parameters of HMG. From the figures, it follows that all these dynamic characteristics have expressed minima, which allows their optimization.
The parameter XCJC entering into the equations for the barrier capacities does not influence the error function Fig. The inductances of the terminals were taken as nominal data of the transistor. Thus, the technique of finding the 26 parameters of the transistor model has been developed. It confirms parametrical frequency modulation. In both the modes on one crystal the following were observed: The indicated modes in heteromagnetic structures would be described by multicoherent parametrical interactions3 in the ferrite and transistor subsystems possessing different nonlinear properties, including nonlinear resonance in the ferrite subsystem.
Therefore, for expansion of our analysis over the whole frequency range toward the area of the maximum subharmonic components with m D 1; 2;: Let us present the transistor as a nonlinear element with active and reactive components. This simplifies the equation set, allows the usage of the method of slowly varying amplitudes and to obtain a number of important analytical and numerical results confirming the effects observable in our physical experiments. The equivalent scheme of HMG is presented in Fig. The nonlinear elements Rt , Ct are determined by the transistor with a feedback in a dynamic mode.
On the nonlinear reactance L0 , the high-frequency current and voltage are shifted by phase, and this shift depends on their amplitude values. At small amplitudes of oscillations, the relation between the basic and subharmonic types of fluctuations in FMCR is weak. Neglecting this relation we shall Fig.
In this case, the voltage on the nonlinear element will be equal to the sum of the voltages on the contours with the corresponding factors of transformation. Using the method of nodal potentials [45], we derive a set of two second-order differential equations: Integrating the right-hand sides of 6. Despite the little nonlinear reactive conductance, their account leads to occurrence of qualitatively new modes in the system, including stochastic ones. Let us lead the derived equations to a dimensionless form. Let us pass from the equation set 6. The physical sense of the factors in 6. Studying of stationary points is important for understanding of the dynamics of the system investigated.
Generally, determination of analytical stationers for 6. Let us consider the first case. The stationary solution of 6. A zero third own value of 6. Note that the derived relationship is fair for the case of no mismatch between the adjustment frequencies of the contours. If the mismatch of the contours is not so great, it is possible to limit only to the third own value of the matrix 6.
The bifurcational curve6 in this case is set by the equation k. The energy exchange between the contours in 6. Having multiplied the first equation in 6. The oscillatory contours are not connected. No periodic energy swapping between the contours occurs. In practice, both situations show to some extent, depending on the voltage of displacement of the transistor. In the second case, the energy exchange between oscillations may have an oscillatory character, therefore, it is of interest for further research.
The modes of self-oscillations in 6. In this case, 6. For existence of the second stationary solution 6. Exploring the stability of the stationary solution7 6. The bifurcational diagram of the stationary solutions of set 6. The self-modulation frequency in the p vicinity of the bifurcation straight line will be!
In area 4, the values of parameter k are great and not realized in practice. Thus, within the limits of our model and our analysis of the stationary solutions of 6.
These are the following modes: It occurs owing to compensation of mismatch of the adjustment frequencies of the contours because of the influence of the own nonlinear capacity of the transistor. The parameter of excitation is fixed k D 3. From our analysis of the curves in Figs. The value of significantly depends on the displacement voltage of the transistor, therefore, at its change the existence range of multifrequency and stochastic oscillations will be reduced with an increase of the GB product of the electrodynamic system, which was observed experimentally.
Investigate the evolution of limiting cycles in the 6. The oscillations of the amplitude envelopes of the basic and subharmonic oscillations are in antiphase. At a distance from the bifurcation curve the amplitude of oscillations, which become relaxation ones, increases. This explains the smaller average intensity of oscillations on the subharmonic frequency.
Hitting the point B Fig. Thus, under the influence of external fluctuations, the movement of the representing point becomes chaotic.
If the movement occurs far from the separatrices, the external noise leads to insignificant peak and frequency modulations Fig. Despite the global nonconservativeness, the system contains an area in the phase space where the behavior of the trajectories is not rough, which explains generation of noise-type signals.
The equality is satisfied: The cascade of doubling period bifurcations. The spectrum is rich with harmonic components which slowly decrease with an increase in frequency. The self-modulation of microwave oscillations in the case of regular movement is close to rectangular. Further, a next doubling bifurcation arises to determine the parameters of the three-order equidistant frequency spectra Fig. Such a sequence of transition to noise-type fluctuations was observed in our physical experiments with HMG. The spectra in Fig.
The spectrum in Fig. Spectra similar to Fig. Transition to chaotic fluctuations through doubling of the period in this system is not unique. Modes of transition to noise-type fluctuations at which no spectra of higher orders appeared were experimentally observed. At changes of the operating parameters e. Similar changes of the modes of oscillations were observed in the investigated model of HMG as well. With an increase in the parameter of mismatch between the contours, the stochasticity in the system vanishes as a result of the return bifurcations of period doubling. In the field of high values of the parameter of mismatch 0: The distance between the neighboring components of the frequency grid in Fig.
Thus, at changes of the operating parameters the displacement voltage, the generator-loading coupling voltage, etc. Within the framework of the investigated HMG model it is possible to explain the occurrence of the following facts observed experimentally: The obtained results agree with our experimental data. The effect of interaction with the subharmonic oscillation is one of the mechanisms of occurrence of controlled equidistant frequency spectra and noise-type fluctuations in HMG.
For powerful heteromagnetic structures, thermophysical analysis of nonstationary and stationary thermal fields for magnetotransistors as a rectangular and as a multilayered cylinder was made. The crystal is connected to the terminals of the transistor with boil soft wires; besides, a powerful transistor can include matching circuits. These additional elements boil soft wires, the case, the transistor terminals, and matching circuits bring essential distortions in the work of the semiconductor crystal and should be contained in the equivalent circuit of the transistor.
Experimental families of the static characteristics of transistors serve as the initial data for calculation. The algorithm of the program is based on optimization of the static parameters of the models of transistors for the best calculation and experimental data fit.
The traditional technique of determination of the equivalent parameters of both linear and nonlinear models of transistors is based on carrying out numerous complex measurements with the usage of such vector circuit analyzers as, for example, NA. In the programs [53], optimization algorithm of the additional reactive elements of the transistor the inductances of boil soft wires and transistor terminals, matching element capacities is realized.
As input data for the program, reference data on the transistor its boundary frequency, gain factor on the working frequency, matching capacities, inductance of the terminals, or experimentally measured S parameters in 1 Agilent TechnologiesTM USA with a frequency band of 10 MHz— GHz, building-up of frequencies up to ,, and 1, GHz. The program [53] allows simulation of the bipolar and FET transistors with a good accuracy.
The system requirements of the program are as follows: The basic class of CGummelPoon contains the following functions, methods, and variables: AreaTrapeze x1, x2, y1, y2, y3, y4 — Calculation of the trapezoid area with the vertex coordinates x1, y1 , x2, y2 , x1, y3 , x2, y4. It is used for calculation of the error function. Calculate — Calculation of a curve family for the current values of the parameters par []. The variable type defines the coordinate of the residual vector: ERRFunction — Calculation of the normed error function for the current values of parameters par[].
In variable BestRes[], the found point is returned. Init — Initialization of the global variables. MotionToCurve x0[], i , type — Calculation of one the i th curve movement along the curve with the initial values x0[]. The variable type defines the positions of the calculated points: CalculatCurves[] — Calculated curves. Funs — The reference to the calculated function defines a set of equations for calculation of the transistor model.
MeassCurves[] — Experimental curves. NActPar — A quantity set in measurements varies in experiment. NConstPar — A quantity fixed in measurements. The global functions are as follows: CalculateInit type — Initialization of the variables for calculation of a certain family of curves. Each parameter is described as: The fields are separated with a symbol of tabulation.
Information on the experimentally measured points of the static characteristics is set in text files with arbitrary names. One file contains information on the experimental points of only one curve of the family of characteristics. The file format is: Here [Family identifier] the bipolar transistor can accept the following values: Set all the measured curves in text files 2. List all these files in the file FileIn. Set in the file parameters. To make sure that the experimental and calculated curves are displayed correctly. To calculate the error functions for each parameter.
To exclude the parameters not affecting the error function from optimization. To select for the remaining parameters their initial and final range borders to have the minimum of the error function within. To select for each parameter its maximal step to make the plot of the error function for the given parameter smooth enough. To start searching the minimum of the error function. After finding the minimum it is necessary to correct the initial values of the parameters in the file parameters.
If the error function has not reached a minimum by any parameters, restart the optimization steps 2—7. The found values of the parameters can be used for modeling bipolar and field transistors in CAD programs. As an initial approximation of the model, the known parameters of the powerful BFR92 transistor were taken. The calculation was made on the basis of families of the static input and output characteristics. As a second test task for the program [53] the reactive parameters of the equivalent circuit of the bipolar KTB transistor have been calculated.
The input data were: The initial values of the parameters and optimization results for the MPSA92 transistor are presented in Table 7. The experimental and calculated families of output characteristics for the MPSA92 transistor are shown in Fig. Unification of the model is necessary for simplification of further optimization by various classes of parameters and compatibility with various CADs; therefore, for further modeling, the equivalent circuit2 shown in Fig.
The field transistor connected in the circuit with a common source Fig. For calculation of the S parameters of the transistor, the common-source circuit is used.
The form of the frequency dependence of the S parameters is essentially influenced by the effects determined by reactive elements of the equivalent circuit. The parameters of the most significant reactive elements are shown in Table 7. Families of static characteristics and S parameters describe transistor processes weakly related to each other by some parameters and their various physical nature.
For optimization we shall divide the error function into its components, each correspondent to some family. Generally, the characteristics have the following functional dependence on its parameters: The iterative method of optimization essentially raises the productivity and stability of optimization because of decreasing the dimension of phase parameter space. A program is developed for determination of the parameters of field transistors with the aid of optimization by static characteristics and S parameters.
It contains schemes for calculation of characteristics families, an equivalent circuit of the field transistor, and a set of parameter optimizers by the error function 7. The schemes are as follows: In the latter case, time expenses for calculation essentially increased owing to slow convergence of the algorithm outside of the range of physically adequate values of parameters.
The used technique of construction of a computer model of the transistor allows an effective machine-focused algorithm for development of a heteromagnetic field transistor to be designed. The parameters of the model are shown in Table 7. The flowchart of the computer program of calculation of the parameters of the model of EHF field transistors on the basis of its commercial prototype with locking to the working frequency range and the maximal output capacity is shown in Fig.
A model of the nonlinear amplifier with a band filter of lower frequencies at output was used. The input parameters are as follows: Comparison of the gain factors, the maximal output capacity within various subranges of the frequency range for the HEMT-1 amplifier and its model are collected in Table 7. For modeling, the parameters of dispersion of the base HEMT transistor were used. The basic requirements to MECE are as follows: The topology of coupling elements see Table 7.
The circuit shown in Fig. The three-pole elements in these figures represent investigated MECEs. The choice of the corresponding coupling element is made by changing the NET field value in [58]. The oscillatory contour simulates FMCR. The contour parameters are defined by 7. By the algorithm from [55], the frequency dependencies of the transfer factors for various magnetic fields H0 Fig.
The analysis of planar MECEs of various types in a real time mode is made strictly electrodynamically. The calculation program allows changing the topology Fig. The algorithm was used in programs of calculation of generation modes of regular and quasi-noise signals at high levels of continuous and pulse power in the VHF, UHF, microwave frequency, and EHF-ranges: The various MECE types considered above had some restrictions at advancement into the frequency ranges from 10 up to GHz, namely, increased introduced losses, a decreased decoupling level.
In this connection, modified MECEs have been simulated, which allows obtaining effective interaction in the microwave and EHF frequency ranges. On frequencies above 20 GHz the decoupling level considerably decreases Fig. To increase the decoupling in a frequency up to GHz it is necessary to minimize the grounding hole resistance, which may cause certain technological difficulties.
The algorithm of calculation of the parameters of magnetoelectronic coupling elements was used in development of a program of analysis of powerful bipolar magnetoelectronic transistors in the microwave frequency and EHF ranges. Separate cells of the compound transistor are incorporated in parallel that essentially reduces its input resistance. With the purpose to increase the 7. The task of modeling of the powerful transistor is divided into modeling of separate transistor crystals and of the whole transistor assembly in view of the inductances of the boiled conductors and the capacity of the matching condenser.
As initial data, experimental families of the S parameters and of static characteristics of the bipolar transistor were used. Each cell of the 5-section bipolar transistor Fig. The calculated parameters of one cell of the compound transistor are as follows: The initial static characteristics for one section of the compound transistor were calculated from the experimental static characteristics of the 5-section transistor, thus the terminal current was divided between all the sections equally, and the terminal voltage of each section was accepted equal to the terminal voltage of the whole transistor.
At parallel—parallel connection of four-pole elements the Y -matrix of the circuit can be obtained by addition of the Y -matrixes of all its components. Let us pass from the Y -matrix to the S -matrix for the compound transistor by means of p p 7. For one transistor structure of the compound transistor, the matrix is Y0 D Y ; N 7. For calculation of the S -matrix of one transistor structure of the compound transistor we shall use the following expressions: Therefore, for analysis the model in Fig.
The current carrying system of the active element is presented by a piece of a two-wire line with a length 2R with a distance between conductors a and their diameter d. Near to the conductor located at x D 0, at a distance h C R a ferrite sphere of a radius R was placed. For calculations the equations obtained in 5. Research of the characteristics of a powerful HMG in the mode of regular signal generation was made by the method of harmonic balance. The following were investigated: The equivalent circuit of the powerful HMT on the 2TAc transistor includes five structures and allows getting pulse output power up to W see Fig.
The resonant frequency is determined from! The elements Re0 , L0e , R0b , L0b , Rk0 , L0k are determined by the conductors of the terminals and the transistor assembly on the plate. Ckb is the parasitic capacity formed at installation of a semiconductor structure in the case of the transistor. Z is the load resistance.
Coordination of the low output resistance of the transistor with the load is carried out with a P-shaped filter of their elements Lf , L0f , Cf. The capacities Ce and Cb are formed by the contact platforms of the emitter and base terminals of the transistor. The inductance Le is formed by a piece of an asymmetrical microstrip line. For reorganization of the generator by frequency, the parameters of the oscillatory systems in the emitter and base areas simultaneously changed.
Time realization of the oscillation amplitude of a powerful MES on the 2TAC transistor in the mode of continuous power generation is presented in Fig. For achievement of optimum values of the output power and DE of the generator at frequency changes, the elements of the equivalent circuit in the circuits of the emitter, base, and loading were tuned simultaneously. The other parameters of the circuit are taken from Table 7. The spectrum of oscillations is presented in Fig. The mechanism of their occurrence is investigated in 6.
The distribution of SPDN in the mode of noise-type oscillations with an integral output power W is presented in Fig. It is accompanied by a frequency change in the autogenerating mode. It corresponds to a relative reduction of frequency by 0. In HMG, introduction of compensation circuits of temperature drifts of frequencies with the use of a microprocessor control system is possible.
A powerful HMT consists of several coordinated transistors, connected in parallel. Effective frequency reorganization by magnetic field was reached at inclusion of an FMCR into each transistor. Owing to the small in comparison with the powerful HMT sizes of the spherical FMCR there was no possibility to provide effective communication between all the transistor structures and the microresonator.
By the use of various microresonators in each transistor structure there is a mismatch of these structures because of a deviation from the face value of characteristics of the used microresonators.
For this reason, in powerful HMT it is necessary to use a ferrite monocrystal as a plate overlapping all the transistor structures. Technologically, it is possible by the use of a planar HMT design. The range of reorganization in the investigated powerful generators is much less than in low-power HMGs. It is explained by the necessity of simultaneous fine tuning of the matching circuits of the powerful transistor in HMG at changes of the frequency of generation by a magnetic field. Thus, in a powerful HMG the following are possible: The principle of cascade connection of elementary transistor cells is used for increase of the output power of a field HMT.
A nonlinear model of the active area of the transistor considering the total impedance of the electrodes was applied. For modeling of cascade connection, the elementary cells of the transistor were considered quasi-identical. This assumption essentially limited the number of the parameters necessary to solve the optimization problem. Updating of the transistor model is necessary for simplification of optimization by various classes of parameters and compatibility with various CADs.
Therefore, further the equivalent circuit shown in Fig. It includes a coupling element, a ferrite structure, a powerful cascading field transistor in the form of a three-pole element an amplifier, Fig. The results of calculation of the frequency characteristics of the powerful FMT for several magnetic induction values are shown in Fig.
The dependencies of the gain factor of FMT on the magnetic induction are shown in Fig. A model of summation of the power of nonlinear amplifiers with a strip low-frequency filter at the output was used. At the first step, the program inputs its initial data: At the third step, the frequency and constant magnetizing field increase, achievement of the higher border of the frequency range is checked, and the process of calculation either repeats with the new frequency or finishes.
The results of calculation of the frequency dependence of the module of the transfer factor and output power are presented in Table 7. In this case, the temperature field of the semiconductor crystal can be presented as superposition of both the stationary and pulse components of temperature [59,61]. The stationary component of temperature is determined by the time-average thermal flux over all the elements of the HMT design due to heat conductivity to the environment and is calculated by means of the thermal scheme method [58].
The stationary temperature components Fig. In a rectangular N -layer semiconductor structure Fig. In each layer, there can occur bulk thermal emission with a uniform distribution over thickness and any distribution in the plane of this layer. On the top and bottom surfaces, the conditions of convective heat exchange with the environment are set. Any thermal fluxes from the other surface of the structure were neglected. The general problem equations for the devices shown in Fig. The general problem has been solved analytically by finite integral transformations, which allows calculation of stationary temperature fields in planar and bulk microcircuits, field transistors, semiconductor lasers, etc.
A problem of calculation of stationary temperature differences in multilayered axisymmetric cylindrical objects with local bulk thermal emission Fig. Its mathematical formulation looks like: By means of the obtained general solutions of the problem, stationary thermal modes in the corresponding constructive elements of HMT were modeled at some preset forms of the functions of heat sources and heat exchange conditions.
The pulse modes of HMT were described by the heat conduction equation: The spatial—temporal dependence of the volume density of thermal emission should be specified in each case. The nonstationary temperature field of a powerful HMT in its pulse mode has been calculated at the following simplifications: The structure whose temperature field has been calculated is shown in Fig.
The initial data for the program [62] are as follows: The flowchart of the program is shown in Fig. Monotonous warming up of the silicon semiconductor crystal with the following data: The temperature conductivity of silicon is a D 0: The number of thermal sources is 1.
The established value of the thermal resistance of the object is 7. All the parameters of the modeled structure are shown in Fig. A program has been developed [63] for analysis of the thermal modes of active semiconductor structures with a planar technology, and design elements of a similar geometry. The limiting value of the number of layers is 5. The initial data are as follows: The analytical solution of test task 8.
The stationary thermal resistance of a semiconductor crystal with the parameters from test task 8. The insignificant difference of this result from that of the previous test task is explained by the heat source having no thickness it is superficial while in the previous example the source possessed a small thickness of 0. The dependence of thermal resistance RT of a semiconductor structure on the location of a local thermal source with the following: The sizes of the crystal are A D 10 mm and B D 10 mm.
The number of layers is 2.
The thickness of the layers are h1 D 2: The program in [64] contains the following blocks: The work of the program of calculation of the stationary thermal resistance of the constructive elements of a powerful HMT in the form of a multilayered cylinder is illustrated on an example of the thermal resistance of the own case of a KT transistor. The thermal resistance of a ceramics-copper case with the following initial data: The effective radius of a heat source is 0.
The thickness of the ceramic layer is 2. The thickness of the copper layer is 1. The results of calculation are as follows: The thermal resistance of a two-layered structure considered in test task 8. The equivalent radius of the structure R D 5: The sizes of a source are a D 0: Modes of multipurpose generation of regular semi-noise and noise signals of a raised level of continuous and pulse power in the VHF, UHF, microwave, and EHF ranges are analyzed.
Modern principles of design of various types of frequency synthesizers are considered. Special attention is given to design of multipurpose operated frequency synthesizers on magnetotransistors in a frequency range up to GHz, including the modes with a pseudorandom working frequency and phase manipulation of a noise-type signal on the basis of heteromagnetic structures with discrete phase shifters. The results of physical researches of autogenerating magnetosensitive microcircuits and calculation methods of the basic characteristics for determination of small values of the magnetic induction vector with a raised accuracy and spatial resolution are given.
Ways to decrease the noise factor in the amplifying cascades on magnetotransistors are investigated and ways of their designing for frequency ranges up to 40 GHz are suggested. The basic elements of magnetotransistors of various types at low and high power levels, including low-power amplification circuits up to GHz and signal transformations up to 1, GHz are considered.
The results of our development of the know-how of field magnetotransistors of low mW and high W power levels are presented. Basic nonlinear effects in magnetotransistors and their role at power restrictions are discussed. The results of our theoretical research of the two-domain model of spherical microtransistors in unsaturated modes are presented. Chapter 9 Influence of External Factors 9. External influences can be subdivided into two groups, namely, those determined by natural — meteorological, climatic — factors and artificial factors. The conclusion about the stability of HMT to external influences is made on the basis of corresponding experimental tests.
Preliminary theoretical analysis of the stability of HMT to external influences is important to estimate how optimal the design is. The following types of mechanical EEFs can be discerned: At vibrations and impacts, distributed loadings are acted on the HMS design elements, and the peak values of the resultant forces are determined by the weight of the element and its peak acceleration.
These forces, depending on their direction, aspire to shift or tear off the element from its place. A preliminary theoretical estimation of the stability of HMS to mechanical influences was made on the basis of the quasistatic approach [66]. Deformations and pressures in the gauge elements were calculated under static loadings and were compared with the strength of the materials.
For analysis of dynamic influences, a correction factor named as the dynamic factor [66] to reduce the effective strength has been introduced. The investigated elements of the HMS design are schematically shown in Fig. In a brass HMS case, an assembly plate with two GaAs elements a two-cascade amplifier on the basis of a field transistor with a Schottky barrier and a feedback element to FMCR as a standard ferrite sphere was placed.
The cover of the gauge was fixed with screws. The semiconductor crystals were soldered to the assembly plate. The plate was fixed with screws to the corresponding groove of the gauge case. The cover had an aperture for a brass screw, which has a constant magnet at its end in the form of a washer. Such a representation is shown in Fig. The resulted estimations are true provided that the durability of the interfaced HMS elements surpasses that of the connecting layer solder.
Each beam is microwelded onto the corresponding contact platforms. Such welded possess a high durability and hold overloads of 16; Therefore, consider the probability of a break of the block from its lateral bend. The mechanical properties of glues widely vary. The adhesive properties of ETP-2 glue are resulted in Table 9. Take the radius of a glue drop to be R D 0: Then the volume of glue without a ferrite sphere Fig. The total weight of the glue drop-spherical microresonator system is The basic demerit of epoxy resins is their high TEC see Table 9.
Therefore, temperature changes in HMS essentially intensify development of defects in glue layers. Such defects originally arise owing to glue shrinkage at polymerization. In glue hardening, the residual solvent creates porosity in the bulk of glue, which is another cause of the occurrence of internal tensions and a decrease in the durability of glue joints. With growth of the glue layer thickness, the number of defects increases, and the durability of joints falls.
Usually, it is recommended to limit the thickness of the glue layer to values from 0. All the elements of the case, including the screws, are made of brass. The force to break the screw of fastening is determined by the external acceleration: The weight of the case the brass density is 8: The thickness of the case cover in the places of its fastening to the case is 0. For this, the acceleration to break the basic material is considerably smaller as follows: When a body undergoes a dispersed external force, a value of kd D 2 is used for estimation.
Therefore, for estimation of the stability of the HMS elements to a short-term below 2 ms single impact, it is necessary to halve the value of critical accelerations obtained earlier. At repeated mechanical impacts, a residual deformation is accumulated in HMS. Therefore, the stability was estimated not by the tensile strength but by the yield stress. Besides, it is necessary to consider that under a sine wave loading the sign of stress constantly varies. Therefore, for theoretical estimation of the HMS stability to sine wave vibrating influence, it is possible to take advantage of repeated impact data, having toughened them twice.
It is fair provided that the own frequency of the system noticeably differs from the frequency of the external sine wave influence. Let us estimate the own frequencies of the polycoric assembly plate of HMS of the sensor. The own frequency of mechanical oscillations of the plate is determined by r g 1 fown D: Doing similar calculations for the thin walls h D 2 mm of the cases of HMS, we will get fown.
The upper estimate of such a force is 2. Hence, it is possible to approve that the case of the HMS sensor is not sensitive to external pressure differences. In the case of a sharp increase in the pressure up to three atmospheres 2, mm Hg , this statement remains true. A change of the ambient temperature will result not only in a change of the thermal mode of the HMS sensor in the conditions of its functioning and a respective alteration of its working parameters, but also in possible occurrence of significant thermotensions in the joint places of heterogeneous elements.
Hence, the interfaced elements of the design should be coordinated by TEC. At a thermal emission power of 30 mW, the own overheat of the amplifier of the gauge above the ambient temperature will be 2. In exclusive situations, it is necessary to provide a system cure, an active semiconductor structure of the sensor maintaining the working temperature at the required level. Let us consider a situation with a sharp change of the ambient temperature in the conditions of storage transportation of the sensor and estimate the time of its cooling or heating.
Within the limits of our estimation, we shall consider that the conditions of heat exchange with the environment on the surface of the gauge will not vary eventually and that the thermophysical characteristics of the material of 9 Influence of External Factors the case of the gauge also do not vary. Then the volume-average temperature will vary in time under the law [67] " T. Generally, the factor of convective heat exchange depends on both the surface and the temperature difference between the surface and the environment.
In connection with that the temperature of the sensor changes in time exponentially asymptotically coming nearer to a finite value , we accept for the moment of time of acceptance by a temperature body T1 with such value at which the temperature of the sensor is equal to 0: In our example, only the passive elements of the sensor were considered.
It is connected with that the stability of the active semiconductor module of the sensor to mechanicoclimatic and temperature influences was specified by its manufacturer and, consequently, was considered as known. Protection against other kinds of influences dust, aggressive environments, mould fungi, etc. The problem of a high stability of the design of the gauge to EEF should be solved by an integrated approach in view of the conditions of its prospective operation, electromagnetic compatibility, and the stability to other influences, for example, to ionizing radiation.
The EC of radioelectronic equipment is its ability to function jointly and simultaneously with other devices having their electromagnetic properties, under the possible action of electromagnetic interferences and not creating inadmissible interferences to other radioelectronic equipment [68, 69].