摘要:A new configuration of high speed parallel multiplier is presented. In this configuration, Wallace tree is used to simplify the modified Booth’s algorithm. Carry save adders, suitable for NMOS implementation, are used. In the last level of two bits addition, Manchester type adder with fast carry propagation is used instead of ordinary carry look -ahead adder. A new NMOS realization of Manchester adder is presented. The total gate count of this high speed parallel multiplier is less than that of the ordinary multiplier with modified Booth’s algorithm. The time of multiplicationis also reduced. Furthermore, the configuration of CSA array is much more regular. It is especially suitable for VLSI layout design.
摘要:A new theorem of the invariance of the scattering matrix of n-port networks is proposed and proved. The theorem is useful in the analysis and design of electrical networks, and some of its applications are mentioned briefly.
摘要:A new method for computation of wire antennas is discussed. In case of symmetrical geometry of wire antennas or arrays with any voltage excitation, generalized even-odd mode expansion method can be used. This method and judicious applications of thinning technique will save about half computation time without sacrificing the specified accuracy. Some examples are given.
摘要:A single board computer control system for vapor epitaxial reactor is described, followed by a discussion of the structure and performance of the system. The mathematical model of the process and two design methods of the digital controller are given. A brief account of the control software development and the operation of this system is also introduced. The system has worked well on a production line in Beijing.
摘要:AR series obtained by observation, as they are mixed with observated noise, often reduce greatly both the resolution and the estimating performance of AR spectral estimation. The recursive algorithm of Householder transform is applied for AR spectral estimation by forming the high-order autocorrelation matrix, and an unbiased estimation of AR parameters is obtained. Thus, the noise pollution to AR spectral is overcome and a better estimation performance is ensured.
摘要:A receiver capable of overcoming severe intersymbol interference and speedily tracking High-Frequency(HF) channel fading is presented. This receivre, combines the pree-qualization(PE) with the nearly maximum likelihood sequence estimation,uses square root Kalman algorithm to jointly regulate the PE and overall channel parameters. The computer simulation shows that nearly optimal error-rate performence can be achieved with this receiver structure for serial high-speed digital transmission on HF channels.
摘要:Under the conditons that the signal and noise are two independent, mean squa-fe differentiable two-dimensional homogeneus Gaussian random fields and the mathematical expectation of noise is zero, the three important properties of mean squared difference template matching systems are proved, and the local accuracy estimation formulas of the contain-wise and pass-wise two matching manners are derived. When the integral distance, autocorrelation functions of signal and noise are known, the local accuracy for a system can be analytically computed and evaluated.
摘要:Discrete Hilbert trasformers designed by using complex wave digital lattice filters have advantages in accuracy and stability of characteristics, and in the amount of implementation hardware. In this paper, a new discrete optimization scheme for this structure is presented by introducing the orthogonal design method which is effectively modified and used together with direct search method. Examples and design tables with short wordlength coefficients are given.
摘要:This paper presents a transition function method which can be extended to solve the synthesis problems of multiple variable sequential circuits. An odd test codes state assignment and a difference weaght codes state assignment are also presented so that the synthesis of complex fail-safe sequential circuits is easy to be realized.
摘要:The helix-coupled vane slow wave structures are analysed. The formulas of dispersion characteristic and coupling impedance are given. The analysis takes into account the effect of the finite radial thickness of the helix. The results obtained agree better with the experiments. The effects of each dimensional variation on dispersion characteristics and coupling impedance are presented in a series of graphs on the basis of a large quantity of culculated results.
摘要:The distribution functions and statistical correlation properties of compound order filtering for binary sequence (i.i.d.) are investigated. Their applications to waveform analysis and image processing are also discussed.
摘要:Accurate solutions of elliptical and cylindrical striplines and microstrip lines are presented. The geometric figures of the elliptical and cylindrical striplines are conformally transformed into asymmetric and symmetric planar striplines, and the elliptical and cylindrical microstrip lines into planar microstrip lines, yielding the formulae for exactly calculating characteristic impedances of the six lines, of which one is new.
摘要:A new model of PLLs is given. Linear analysis of PLLs can be developed in a larger range with the new model. The analysis gives enough precision for engineering use for a PLL with arbitrary characteristic phase detector g (PPPPPP), which runs in whole positive slope range of g(ppppp).
摘要:Several methods of constructing orthogonal transforms are presented. By means of these methods, new orthogonal transforms having good properties in the spectrum domain can be generated from known orthogonal transforms. The spectra are combinations of window spectra of original orthogonal transforms, wherein the transforms generated by using H product have invariant property of power spectra for cyclic shift. All these methods are new tools for spectrum analysis and digital processing.
摘要:A systematic and comprehensive description of basic concepts, principles, techniques and application on the vectorization of serial arithmetic is given.
摘要:Polycrystalline GaAs films are grown successfully on insulating substrates by MOCVD technique. The film is flat and bright, close in texture, and has stoichiometry of GaAs. The poly-GaAs films were recrystallized with grain size changed from 200A’ to 40 μm when scanned by CW Ar+ laser beam. The schottky barrier diodes are fabricated on the leaser recrystallized GaAs film.
摘要:Using measurements of the thermopower and DC conductivities,the electrical properties of thin films of GaAs, GaP and InP which rendered amorphous by bombardment with rare gas ions arc investigated. It is shown that ion beam bombardment amorphization is an advantageous technique for studying and understanding the fundamental process in the amorphous compounds Semiconductors.
摘要:The microwave mixer with image recovery using dielectric resonators at X-band is presented. The design method, the real circuit and the experimental data for comparison are all discussed. The results show that the signal conversion loss is reduced by about 1dB, and the image to signal conversion loss ratio is more than 25dB. With simple circuit contruction and adjustment method, it is suitable for microwave reciver applications.
摘要:Using an silver head electric soldering iron, the alloy consisting of In, Au, Sb and Ni in the ratio of 91: 7: 1:1 is directly smeared on the silicon wafer, which is then annealed at 450℃, Good quality ohmic contact is obtained even when the silicon resistivity is as high as 600Ωcm. The specific contact resistivity of metal to N-Si is calculated as a function of doping concentration and compared with experimental results. This contacting technigue is successfully applied to the measurement of impurity compensation.
摘要:An improvement on G.A. Lang et al. s model of thermal fatigue in power transistors is presented, and corrective equations as well as analytical solutions are given. These results are also applicable for other electron devices with multiple-layer structures.
摘要:A new kind of adaptive least squares Escalator algorithm is derived, which has a computational requirement lower than the present one proposed by T. Kawase et al. in 1983, but with the same parameter learning capability.