Showing papers on "Stopband published in 1976"

Iterative technique for designing non-recursive digital filter non-linear phase characteristics

Abstract: Iterative techniques based upon Remez's method can be used to design non-linear digital filters having small attenuation ripples in the stop band without appreciable distortion arising in the pass band.

Launching and/or receiving network for an antenna feedhorn

Abstract: A launching and/or receiving network is disclosed capable of coupling radio signals in a first frequency band, e.g., 2 GHz into and/or out of an existing antenna system without perturbing the signals being transmitted in other frequency bands, e.g., 4, 6 and/or 11 GHz. The first frequency band signal is launched through an evanescent mode waveguide filter which is both coupled to the flared sidewall of a feedhorn through an H-plane, T-junction, and provides a very broad stopband for the signals in the other frequency bands being launched in the feedhorn. To minimize mode conversion for the signals in the other frequency bands, a dummy evanescent mode waveguide filter is connected at one end thereof to the feedhorn facing the first frequency band launching network and at the other end thereof to a matched load through a waveguide section. A second launching and/or receiving network can be similarly coupled to the feedhorn in a plane normal to the first network to permit a different orthogonally polarized beam, at the first frequency band, to be simultaneously transmitted and/or received by each of the two networks.

Anti-aliasing pre-filter circuit for semiconductor charge transfer device

Abstract: An on-chip semiconductor pre-filter device is designed with a prescribed low frequency passband (0 to f) and a prescribed stopband (F±f), with a frequency roll-off therebetween. The pre-filter performs a weighted averaging of signal samples taken from an input electrical signal at a suitable sampling frequency and converts the weighted average into an equivalent charge packet for input by injection into a semiconductor charge coupled device (CCD) driven on the same chip by a clock of frequency F. In this way, the "reflected" parts (in band: F±f) of the frequency spectrum of the CCD are suppressed from the input to the CCD, thereby suppressing undesired aliasing in the CCD.

Design of minimum noise digital filters subject to inequality constraints using quadratic programming

Abstract: A numerical method is presented for designing digital filters. The method allows one to minimize the mean-square error or noise power over some intervals of frequency, while simultaneously constraining the maximum error in other intervals of frequency. Thus, for example, one can minimize noise power from a stopband of frequencies while constraining signal fidelity in a passband of frequencies by limiting the maximum passband deviation.

Transfer Function of Maximally Flat Group Delay Low-Pass Filters with Equal-Ripple Attenuation in the Stopband and Flat Attenuation in the Passband

Abstract: A general class of transfer functions of maximally flat group delay low-pass filters with equal-ripple attenuation in the stopband and flat attenuation in the passband is investigated. Filter transmission zeros are calculated and listed in a table.

Comments on "On the approximation problem for recursive digital filters with arbitrary attenuation curve in the pass-band and the stop-band"

Abstract: This correspondence contains a demonstration of the effectiveness and validity of the Fletcher-Powell optimization technique as opposed to the algorithm presented in the above paper by Dubois and Leich.

Radar m.t.i.: a comparison of 2nd-order digital-filter realisations when the stopband is very narrow

Abstract: Design of recursive digital m.t.i. filters becomes difficult when the cutoff frequency is as low as 1/100 p.r.f., because the filter may have large internal gain, generate high noise levels and have a response which is sensitive to parameter errors. The results of a study of 2nd-order sections are described and two realisations which minimise the word length are presented.

Design procedure for improving the usable bandwidth of an MTI radar signal processor

Abstract: A technique is presented for improving the bandwidth throughout which a radar return can be detected while keeping the ground-clutter power rejection below an acceptable level. The processor employs a high-pass digital filter which has been designed in accordance with the McClellan-Parks algorithm with the pass-band lower edge selected in an optimum fashion when the other filter design parameters are specified. Conventional approaches to the design of a moving-target indicator (MTI) filter such as the three-pulse canceller and the covariance design method typically improve the signal-to-noise ratio over less than 50% of the range between dc and the pulse-repetition frequency whereas this technique extends the usable bandwidth to 80% or more. A design example is included which utilizes parameter values from a C-band radar and demonstrates the interactive influence of such filter parameters as the number of weights, passband ripple and bandedge, and stopband attenuation and cut-off.

Crystal band‐stop filters with improved spurious resonance behaviour

Abstract: It is a well known fact that piezoelectric band-stop filters manifest not only the desired stopband at the fundamental resonant frequency of the crystal, but also additional stopbands at its spurious resonant frequencies. The greater the degree to which such parasitic attenuation peaks are to be suppressed, the more complex is the required circuitry. The most common crystal band-stop filter circuits can be considered as all-pass sections in which a capacitance is ‘disturbed’ over a narrow bandwidth by a resonant circuit. If this ‘disturbance’, possibly after conversion by means of known one-port equivalents, consists of a high impedance series resonant circuit with the inductance Ls in parallel to the entire lattice arm R/jx1, the 3 dB bandwidth of the stopband becomes Δω0= R/Ls(1+x) in the loss-free case. Sometimes the dual representation of the ‘disturbance’ is more appropriate. With a finite crystal Q-factor of Qq, the insertion loss of a desired or parasitic stopband assumes the finite maximum value ǎ ∼ In (1+QqΔω0/ω∞). The width and height of the possible attenuation peaks are given in formulas, normalized curves and worked examples for seven band-stop filter circuits of increasing complexity, each with a single crystal.

Analysis and Design of a Millimeter-Wave Bandpass Filter with a Stopband in the Specified Higher Frequencies

Abstract: The characteristics of a rectangular waveguide bandpass filter (BPF) are not well known in the high-frequency range where higher order modes can propagate. In this paper, the frequency response of a BPF, composed of symmetric inductive windows, is investigated. The range of frequency under investigation covers the region where not only the TE/sub 10/ mode but also higher modes can propagate in the waveguide. A window with incident waves of various modes is first characterized by a scattering matrix. The matrix elements for various modes are obtained in a closed form by means of the variational method. The overall characteristics of the BPF are then obtained by calculating the product of these matrices. A design of a BPF with a stopband in the specified higher frequencies is proposed. Prototype production proved the validity of the theoretical investigation.

Evaluation of general-purpose CCD transversal filter

Abstract: The fabrication and evaluation of a general purpose, 32- stage, split electrode, CCD transversal filter, will be described. Highpass and lowpass filters are implanted on storage electrodes of the filter. A 30-dB stopband rejection and 1-dB passband ripple has been realized.

Synthesis of Two Classes of Acoustic Surface-Wave Filter Tap Weights

Abstract: Techniques are presented for the synthesis of interdigital acoustic surface-wave filter tap weights so as to give transfer functions having a Chebyshev stopband attenuation characteristic with a specified minimum attenuation, along with a single-peaked or a Chebyshev double-peaked passband characteristic. The techniques first give the gain zero locations, and as a result are very flexible. Linear phase or minimum or maximum phase-shift characteristics are obtainable. The transfer function can be factored into two sets of zeros which can be realized as two separate anodized transducers which, when operated with a multistrip coupler, will give the desired overall transfer function. Also, it is shown that the class of designs having a double-peaked passband can be realized in the form of an unanodized phase-reversal transducer (PRT) in cascade with a second transducer having very smooth anodization and no phase reversals. The fact that such designs have such smooth anodization with relatively few small taps can help in obtaining precision performance.