EMI电源滤波器的插入损耗分析

在一般EMI滤波器的共模和差模等效电路的基础上,分析了源阻抗和负载阻抗对滤波器插入损耗的影响.提出了共模插入损耗和差模插入损耗的计算方法,推导了滤波器插入损耗与阻抗关系的表达式,并且对这一关系作了仿真分析,仿真结果验证了理论计算和分析的正确性.     

General Synthesis of Quarter-Wave Impedance Transformers

This paper presents the general synthesis of a radio frequency impedance transformer of n quarter-wave steps, given an "insertion loss function" of permissible form. This procedure parallels that of Darlington for lumped constant filters by providing the connection between Collin's canonical form for the insertion loss function and Richards' demonstration that a reactance function may always be realized as a cascade of equal length impedance transformers terminated in either a short or open circuit. In particular, it is shown that insertion loss functions of the form selected by Collin are always realizable with positive characteristic impedances, and that the synthesis procedure, for maximally flat and Tchebycheff performance, involves the solution, at most, of quadratic equations. In addition, this procedure permits the proof of Collin's conjecture that, for his insertion loss function, the resulting reflection coefficients are symmetrical. Finally, closed expressions are given for the coefficients of the input impedance of a given n section transformer in terms of the n characteristic impedances and vice versa.     

A Dissipative Coaxial RFI Filter

A lossy coaxial filter has been developed to protect electroexplosive devices from accidental detonation by stray high-power RF fields. The characteristics of a coaxial transmission line, filled with dielectromagnetic material, are analyzed in terms of various proposed filter configurations. The design concept for the prototype filter is chosen so as to provide a maximum stop-band attenuation. Measured results are presented to show that the filter provides an insertion loss of about 90 db from 5 Mc to at least 10 kMc. Furthermore, it is shown that the use of a lossy filter guarantees a minimum value of insertion loss regardless of the terminating impedances.     

Equivalent representation of broad-band transmission-line transformer with vector impedances

A method of analyzing broad-band Transmission-line transformers from their open- and short-circuit vector impedances is presented. From a set of open- and short-circuit impedance measurements, the input impedance and insertion loss of the transformer when terminated in any arbitrary load impedance are calculated directly.     

Two Worst-Case Insertion Loss Test Methods for Passive Power-Line Interference Filters

The feasibility of two methods for determining the worst-case insertion loss of power-line interference filters is demonstrated. The first method makes use of injection and detection probes (transformers) which introduce an extremely small impedance (L ?50 nH, R ? 3 m?) into the test circuit. The second method does not use injection and detection probes, thus eliminating their problems. Both methods maintain the high Q of the test circuit and permit interfacial resonances to occur between the power-line and filter interface impedances. A conventional filter and a lossy filter were measured by both methods. Large resonances, including a negative insertion loss at 150 kHz, were measured in the stopband of the conventional filter. No resonances were detected in the stopband of the lossy filter. Measurements with the two test methods differed by only a few dB.     

New Dimensions in Measuring Filter Insertion Loss

The MIL-STD-220A 50-ohm resistive method of measuring filter insertion loss using buffer networks has successfully masked the deficiencies of filters for many years. The method of measuring filter insertion loss proposed herein overcomes all the shortcomings of previous standards. It can measure insertion loss with full-rated line currents flowing (up to 400 amperes), without being adversely affected by the heavy currents and without affecting the performance of the system. It can be clamped around filter lines at installations without disturbing the circuits under test and thus can measure filter performance working into realistic load and source impedances. Filters can be evaluated using load and source impendance conditions selected as preferred from tenths of an ohm, to conjugate matched, to complex-resistive-reactive combinations, and to 50 ohms as desired. Any type of filter can be tested from lumped element to absorptive, active, or ceramic. Of major importance is the fact that this technique can also evaluate in the passband where filter insertion loss and gain is of great interest. These advantages are combined in a system that is capable of completing measurements quickly.     

Matching of an impedance to a feeder with the help of Γ sections consisting of reactive LC elements with finite Q factors

Relationships for calculation of the elements of ?? sections, which consist of reactive LC elements with finite Q factors and are used for matching of the load impedance to the resistance of a generator are derived. Recommendations on the choice of a particular type of the ?? section out of eight possible design versions for an arbitrary position of the point representing the load impedance on the Smith chart are given. It is shown that, in the case of a high level of the load standing-wave ratio (more than 100), the dissipative insertion loss of a matching ?? section with imperfect elements experiences substantial oscillations, which depend on the phase of the reflection coefficient and the relationships between the Q factors of reactive LC elements. As an example, the results on matching of small electric and magnetic monopoles with lengths of (0.2?C0.9)??/4 to a circuit with a characteristic impedance of 50 ?? are presented.     

On the equivalent generator voltage and generator internal impedance for receiving antennas

In the circuit model of receiving antennas, it is commonly assumed that there is no coupling between the load and the equivalent generator. Then the equivalent generator voltage is assumed to be the open-circuit voltage and the generator internal impedance is assumed to be the input impedance in the transmitting mode. In this investigation, these voltages and impedances are computed numerically for cylindrical antennas. From the numerical results, it is found that these two assumptions are not correct individually. However, it is found that the uncoupled-circuit model can yield correct results for the load current. Further, the validity of the uncoupled-circuit formula of the load current is proved by using the reciprocity theorem. Thereby, it is concluded that the uncoupled-circuit model can yield the correct result for the load current, but the open-circuit voltage and the input impedance are not at all identical to the equivalent generator voltage and the generator internal impedance, respectively.     

A Simple Model for Weakly Coupled Lossy Transmission Lines of Finite Length

A simple formula for coupling between two parallel lossy wires of finite length above a conducting plane is derived. Weak coupling is assumed. The first wire is driven by a source at one end and terminated with an impedance at the other end. The second wire is terminated with an impedance at each end. The formula is shown to agree well with the results of a previously reported experiment. For low frequencies, the formula reduces to the familiar inductive and capacitive coupling terms plus a term which is contributed by wire losses. For higher frequencies, the lossiness of the wires is shown to be a dominant factor in the coupling mechanism if the undriven wire is terminated by large impedances and the driven wire is terminated by low impedances.     

Single-pole double-throw CMOS switches for 900-MHz and 2.4-GHz applications on p/sup -/silicon substrates

A 900-MHz single-pole double-throw (SPDT) switch with an insertion loss of 0.5 dB and a 2.4-GHz SPDT switch with an insertion loss of 0.8 dB were implemented using 3.3-V 0.35-/spl mu/m NMOS transistors in a 0.18-/spl mu/m bulk CMOS process utilizing 20-/spl Omega//spl middot/cm p/sup -/ substrates. Impedance transformation was used to reduce the source and load impedances seen by the switch to increase the power handling capability. SPDT switches with 30-/spl Omega/ impedance transformation networks exhibit 0.97-dB insertion loss and 24.3-dBm output P/sub 1dB/ when tuned for 900-MHz operation, and 1.10-dB insertion loss and 20.6-dBm output P/sub 1dB/ when tuned for 2.4-GHz operation. The 2.4-GHz switch is the first bulk CMOS switch which can be used for 802.11b wireless local area network applications.     

Variationally computed antenna impedances and accuracy of resulting current distributions

The difference between a stationary and a nonstationary impedance formulation is shown to become zero if a complex current distribution produces zero tangential electric field along the metallic surface of the antenna. This difference between the two impedances is used as a criterion for establishing the accuracy for a given approximate current distribution. Illustrative examples consider dielectric-coated antennas in free space and insulated antennas in a magnetoionic medium.     

Input filter design for multiple-module DC power systems

When multiple DC/DC power converters are operated from a DC bus with a nonzero source impedance, undesirable interactions can occur between an individual regulator and the input impedance of the other regulators on the bus. Consequently, criteria for input filter design in the presence of a significant source impedance are developed, which, when used in conjunction with already-known input filter criteria, permit the input filter to be designed so that each regulator operates reliably. Proper filter design tends to decouple the negative regulator impedances from the bus, leaving only the passive input filter impedances to affect the other converters. These filter impedances appear in parallel with the source impedance and reduce the overall source impedance. Hence, the use of multiple modules on the same bus actually improves the performance of the individual regulators. An example, the buck current mode controlled power converter, is examined in detail. Extensive experimental evidence is presented to verify the analytical results     

The Quarter-Wave Transformer Prototype Circuit

A quarter-wave transformer not only changes impedance levels, but will also behave as a band-pass filter. In practice, however, band-pass filters are usually required to terminate in equal input and output impedances. They often consist of several direct-coupled cavities, which are similar to transformers whose impedance steps have been replaced by reactive obstacles. It is demonstrated how one can synthesize a quarter-wave transformer, and then "distort" it to obtain a direct coupled cavity filter with a predictable performance. This is illustrated and confirmed by numerical examples. The method is particularly convenient to use in reverse. The quarter-wave transformer prototype is easily derived for a direct-coupled cavity filter which has already been designed by another approximate method, and thus gives an independent evaluation of its performance. If necessary, the filter can then be redesigned, as illustrated in this paper.     

Generalized Theory of Impedance Loaded Multiconductor Transmission Lines in an Incident Field

A general theory is advanced for determining the currents in the load impedances of an N-conductor isolated transmission line excited by an electromagnetic field with the electric vector directed parallel to the wires. The number of impedance loads in the circuit is 2N. An impedance is connected in series with each conductor at its ends. At each end of the transmission line the impedances emanate from a common node. There is no requirement that the conductors be of the same radius, be equally spaced, or lie in a common plane; however, their axes must be parallel. Evidently the cross section of the line must be sufficiently small in terms of the wavelength that transmission line theory applies. Numerical values for the load currents In a three-conductor model are given. Scattering from end loaded multiconductor transmission lines is considered. It is shown that for configurations lacking geometrical symmetry such problems become arduous if not solved by computer.     

On numerical convergence of moment solutions of moderately thick wire antennas using sinusoidal basis functions

Wire antennas are solved using a moments solution where the method of subsectional basis is applied with both the expansion and testing functions being sinusoidal distributions. This allows not only a simplification of near-field terms but also the far-field expression of the radiated field from each segment, regardless of the lengthL. Using sinusoidal basis functions, the terms of the impedance matrix obtained become equivalent to the mutual impedances between the subsectional dipoles. These impedances are the familiar impedances found using the induced EMF method. In the induced EMF method an equivalent radius is usually used in the evaluation of the self-impedance term to reduce computation time. However, it is shown that only for very thin segments that the correct equivalent radius is independent of length. When the radius to length ratio (a/L) is not small, an expansion for the equivalent radius in terms ofa/Lis given for the self-impedance term. The use of incorrect self-term, obtained by using a constant equivalent radius term, is shown to be responsible for divergence of numerical solutions as the number of sections is increased. This occurrence is related to the ratio ofa/Lof the subsections and hence becomes a problem for moderately thick wire antennas even for a reasonably small number of segments per wavelength. Examples are given showing the convergence with the correct self-terms and the divergence when only a length independent equivalent radius is used. The converged solutions are also compared to King's second- and third-order solutions for moderately thick dipoles.     

Insertion loss concepts

The term "insertion loss" has become such a popular one and is applied to so many different concepts that it is becoming unsuitable for precise use. Certain other concepts connected with insertion loss such as attenuation and mismatch loss, are in need of clarification. A logical set of definitions and equations are presented as a suggested approach to more precise terminology. The effect of complex characteristic impedance on these equations is given in an Appendix. Specific suggestions are given for changes in the usage of terms, which if acceptable and adopted would improve the situation.     

Insertion Loss of Mismatched EMI Suppressors

The insertion loss (IL) of EMI suppressors operating in systems with mismatched load and source impedances is studied. The design of an EMI suppressor is difficult because the actual impedances connected to it are often unknown. It has been recognized for a long time and by many authors [1]-[7] that the IL defined in a 50-ohm system (MIL-STD-220) is rather meaningless for general application. It may happen that the line voltage delivered to systems changes strongly after the insertion of the suppressor [8]. The IL under operating conditions may be quite different from that predicted. The source and load impedances are the variable parameters; for variations of one of them, the maximum and minimum IL's are calculated. Some general guidelines are also given in order to improve the perfonnance of suppressors.     

Impedances of an Elliptic Waveguide (For the /sub e/H/sub 1/ Mode)

The power-voltage, power-current and voltage-current impedances for the elliptical waveguide for the fundamental mode (/sub e/H/sub 1/ mode) are obtained by two different methods. The first method consists of using the exact fields inside a perfectly conducting elliptical pipe. Numerical results were obtained by numerical integration of the integrals involving Mathieu functions by the Gaussian Quadratures method by a digital computer. In the second method approximate fields which satisfy the boundary conditions were used. By this approximate method, actual expressions for the impedances are obtained as a function of minor to major diameter ratio with no need of numerical integration. The actual expressions for the impedance obtained by the approximate method give the impedance for elliptical waveguide within six per cent. On the basis of comparison with the exact numerical solution the expressions for the approximate impedance give the impedance of elliptical waveguide within three per cent if they are scaled by 1.03.     

Computer Design of Filters with Lumped-Distributed Elements or Frequency Variable Terminations

A method for realizing prescribed insertion loss characteristics is presented which is applicable to mixed distributed lumped parameter systems, as well as to transmission line structures which operate into resistive frequency variable terminations. The method utilizes scattering matrix renormalization and is implemented by a computer program. Examples given include capacitor and inductor loaded transmission line sections, as well as a TEM filter terminated in TE/sub 10/ wave impedances. The insertion loss characteristics have equal ripple passband behavior.     

Excitation of a Coaxial Line Through a Transverse Slot

A rocket with removed access plate is simulated by a section of coaxial transmission line with a transverse elliptical slot cut in its sheath. The internal circuit consists of two arbitrary impedances in series with the inner conductor at its ends. The object is to find the currents in these impedances when the cylinder is illuminated from the outside by an electromagneticfield that enters the aperture and excites the internal circuit. The problem is solved by application of the reciprocal theorem. The current in a dipole antenna is determined when this is inthe far field maintained by the slotted coaxial line when driven by a generator in series with one of the load impedances. The field in the aperture is replaced by equivalent electric and magnetic dipoles. The reciprocal theorem gives the current in the load impedance when the distantdipole is driven. A numerical example is given.