Showing papers by "NEC published in 1978"

Dynamic programming algorithm optimization for spoken word recognition

Abstract: This paper reports on an optimum dynamic progxamming (DP) based time-normalization algorithm for spoken word recognition. First, a general principle of time-normalization is given using time-warping function. Then, two time-normalized distance definitions, called symmetric and asymmetric forms, are derived from the principle. These two forms are compared with each other through theoretical discussions and experimental studies. The symmetric form algorithm superiority is established. A new technique, called slope constraint, is successfully introduced, in which the warping function slope is restricted so as to improve discrimination between words in different categories. The effective slope constraint characteristic is qualitatively analyzed, and the optimum slope constraint condition is determined through experiments. The optimized algorithm is then extensively subjected to experimental comparison with various DP-algorithms, previously applied to spoken word recognition by different research groups. The experiment shows that the present algorithm gives no more than about two-thirds errors, even compared to the best conventional algorithm.

Textural Features Corresponding to Visual Perception

Abstract: Textural features corresponding to human visual perception are very useful for optimum feature selection and texture analyzer design. We approximated in computational form six basic textural features, namely, coarseness, contrast, directionality, line-likeness, regularity, and roughness. In comparison with psychological measurements for human subjects, the computational measures gave good correspondences in rank correlation of 16 typical texture patterns. Similarity measurements using these features were attempted. The discrepancies between human vision and computerized techniques that we encountered in this study indicate fundamental problems in digital analysis of textures. Some of them could be overcome by analyzing their causes and using more sophisticated techniques.

A Highly Stabilized Low-Noise GaAs FET Integrated Oscillator with a Dielectric Resonator in the C Band

Abstract: A GaAs FET integrated oscillator stabilized with a BaO--TiO/sub 2/ system ceramic dielectric resonator provides a high-frequency-stabilized low-noise compact microwave power source. The newly developed ceramic has an expansion coefficient and dielectric constant temperature coefficient that offset each other and result in a small resonant frequency temperature coefficient. A stabilized oscillator output of 100 mW with a 17-percent efficiency and a frequency temperature coefficient as low as 2.3 ppm//spl deg/C are obtained at 6 GHz. FM noise level is reduced more the 30 dB by the stabilization. The dynamic properties of the oscillator and resonator are precisely measured to determine equivalent circuit representations. A large-signal design theory based on these equivalent circuit representations is presented to realize the optimal coupling condition between the oscillator and stabilizing resonator. The stabilized oscillator performance is sufficient for application to microwave communications systems.

Light emission and burnout characteristics of GaAs power MESFET's

Abstract: In order to obtain information on the field distributions and weak places of GaAs power MESFET's, the light emission and pulse burnout characteristics were investigated. Their dependence on the drain structure was also studied. The light emission occurs at the drain edge of the active region when the gate bias is zero volt, slightly changing its place depending on the drain structure. The drain edge of the epitaxial layer was also damaged by the pulse burnout experiments at zero gate bias. When the gate bias is increased, the light emission at the drain edge decreases rapidly until the gate Schottky breaks down and begins to give the light emission at the gate edge. The pulse burnout experiments suggest that the burnout initiates inside of the active or buffer epitaxial layer when the gate bias is near the pinchoff voltage.

High Dose Ion Implantation into Photoresist

Abstract: The characteristics of heavily ion‐implanted photoresist films were studied in relation to types of photoresist, ion species, accelerating energies, and dose levels. By high energy, high dose ion implantation it was observed that the optical transmission of the resist film was exceedingly decreased and the resist becomes more mechanically, thermally, and chemically resistant. Several experimental data indicated that these results are due to the change of photoresist to disordered graphite. As an application of this ion‐implanted resist, a new photomask fabrication process is developed.

Improvement of the drain breakdown voltage of GaAs power MESFET's by a simple recess structure

Abstract: Dependence of the drain-to-source breakdown voltage on the drain structure of GaAs power FET's was investigated. It was found that the drain breakdown voltage is improved by a simple recess structure without surface n+contact layer. This is due to relaxation of the field at the drain region by increase of the thickness of the active epitaxial layer. The GaAs MESFET with this simple recess structure could be operated up to 24 V. There was no explicit difference in the microwave properties of both recess structure devices with and without the n+contact layer. As a practical device, an output power of more than 3 W with 4-dB gain is obtained at 6.5 GHz from this simple recess and cross-over structure GaAs FET.

Digital transmission system for television video signals

Abstract: A digital transmission system for television video signals of the type which comprises a transmitter and a receiver, the transmitter having a plurality of input terminals for receiving respective television video signals to be transmitted, a plurality of encoding units for encoding the respective video signals into digitized video signals, and a multiplexers for multiplexing the respective digitized video signals in a time division manner, and the receiver having a demultiplexer for receiving the multiplexed video signal from the multiplexer and for separating the same into said digitized video signals, and a plurality of decoding units for decoding the digitized video signals into said respective television video signals, is disclosed.

GaAs dual-gate MESFET's

Abstract: Performance of GaAs dual-gate MESFET, including high-frequency noise behavior, was analyzed on the basis of Statz's model. Under the design considerations developed from the analysis, fabrication and characterization of a prototype device were carried out. The present analysis was confirmed to reproduce satisfactorily the performance observed. Minimum noise figure and associated gain observed in the device with two 1-µm gates were; 1.2 dB and 16.7 dB at 4 GHz, 2.2 dB and 16.3 dB at 8 GHz, and 3.2 dB and 12.6 dB at 12 GHz, respectively. More than 35-dB gain controllability was also obtained at 8 GHz.

An Improved Method for Digital SSB-FDM Modulation and Demodulation

Abstract: A new method for digital implementation of SSB-FDM modulation and demodulation is presented, which utilizes the FFT algorithm to reduce the multiplication rate. The hardware realization of the presented method is shown to be the simplest; it consists of only an FFT processor and a set of complex bandpass filters which operate at the same rate as the input baseband sequences; no signal conversions are required prior to FFT processing. Simple and practical approaches are presented for design and implementation of the EFT processor and the set of complex bandpass filters. Design examples, which are made for implementation of a TDM-FDM translator at the supergroup level, show that the proposed method can attain the minimum multiplication rate among methods previously reported.

Delay compensator for a television signal

Abstract: Delay compensation apparatus for minimizing the delay differences between a video signal transmitted through a plurality of frame synchronizers, which delay the video signal, and a normally undelayed audio signal, the audio and video signal making up a complete television signal. Combinations of discreet delay intervals are imposed on the audio signal such that delay differences between the audio and video signal are minimized to the extent that the negligible remaining delay difference does not degrade the reproduced television picture.

Reliability Evaluation and Failure Analysis for Multilayer Ceramic Chip Capacitors

Abstract: Reliability evaluation tests for multilayer ceramic chip capacitors mounted on a hybrid IC were implemented. Failure modes, failure mechanisms, and drift of characteristics were analyzed. Humidity acceleration as well as voltage and temperature accelerations were investigated to estimate the chip capacitor reliability. The chip capacitor reliability proved to be high and adequate for the service period of the communication system.

Printed wiring board comprising a conductive pattern retreating at least partly in a through-hole

Abstract: In a printed wiring board having an outwardly flaring through-hole, a conductive pattern is formed on the inwardly directed surface of the hole so as to reach a principal surface of the board only at that circumferential portion of the hole where the pattern is continuous to a pattern portion, if any, formed on the principal surface. The pattern is formed in the through-hole by positive use of that portion of a photoresist film formed on the principal surface to cover the hole which partly protrudes into the hole.

Recursive digital filter comprising a circuit responsive to first sum and feedback sign bits and second sum sign and integer bits for detecting overflow in the second sum

Abstract: In a recursive digital filter for dealing with data words given by two's complement representation in a common word format comprising a sign bit, an integer bit, and a predetermined number of fractional bits, an overflow detect and correct circuit is supplied with simultaneously produced sign bits of bit-serial first sum, feedback, and second sum data words and with the integer bit of the second sum data word and detects overflow in the second sum data word to produce, for use in the circuit, an overflow detect pulse indicative of presence or absence of overflow. In either event, the circuit produces an overflow-free data word for use in the filter. When overflow is detected, the circuit produces a polarity decision pulse that decides polarities of the overflow-free bits. Otherwise, the circuit determines the overflow-free bits directly by the corresponding bits of the second sum data word. Use is preferred of a first and a second timing signal which specify time slots for the sign bit and a prescribed bit, respectively, of each serial data words and which are for directly deciding the overflow-free fractional bits by the polarity decision pulse and for producing a second polarity decision pulse for direct decision of polarities of the overflow-free sign and integer bits. A memory having a plurality of memory areas is preferred for production of the overflow detect and the polarity decision pulses and of the overflow-free bits.

Light branching device

Abstract: A light branching device includes a light-focusing transmission body having a refractive index which decreases in proportion to the square of the distance from the plane of its optical axis, and two end faces perpendicular to the optical axis and a substantial light path length equal to an integral multiple of approximately one-half of the pitch at which the light propagates. The ends of a plurality of optical fibers are disposed adjacent to the end faces so that light emanating from one of the optical fibers is dispersed primarily in a plane parallel with the optical axis plane and impinges upon the end of at least one of the other optical fibers for transmission therethrough.

Discharge lamp operating system

Abstract: A discharge lamp is ignited in every half cycle in its operating system including a discharge lamp operating circuit provided with a low frequency alternating current power source, a single winding type current limiter. The discharge lamp is connected to the power source through the current limiter and a series circuit including a high voltage output generator is connected in parallel to the discharge lamp. The high voltage output generator operates during the lamp operation for reigniting the discharge lamp. The voltage of the low frequency alternating current power source is set to less than the required reignition voltage of the discharge lamp during its operation, whereby the lamp current stabilizer size is minimized. Further, a filament preheating circuit is arranged to use current derived from the high voltage generator. The filament preheating circuit is combined with this operating circuit and so is a time delay for assuring a stable operation.

Numerical Experiments on Trajectory Instabilities in Dissipative Systems

Abstract: It is confirmed by computer experiments that a probabilistic approach is necessary for description of the dynamics of dissipative systems with unstable trajectories, although given equations of motion are deterministic. How initially close points in phase space do spread with time is examined. From this it is shown that the conditional probabilities can be defined naturally by introducing the concept of coarse graining in phase space and time, and more­ over they satisfy a definition of a Markov process. The effect of external random forces is investigated by adding the Langevin forces to the equations of motion. It is suggested that the statistical properties of unstable dissipative systems are determined mainly by the inherent stochasticity due to the nonlinearity of the systems, unlike equilibrium systems whose statistical properties are governed by external random forces coming from contact with a heat reservoir.

Adaptive speech signal detector

Abstract: An adaptive speech signal detector for use in a 4-wire telephone channel performs an adaptive threshold value setting operation depending on the channel noise level on a transmitter-side channel to detect a speech signal present at the transmitter. The adaptive operation of the speed signal detector is inhibited, however, if the signal level at the related receiver-side channel becomes higher than a preset value. This permits the use of the adaptive speech signal detector with DSI (digital speech interpolation) systems without malfunction due to the operation of an echo suppressor.

Band-pass filter

Abstract: A bandpass filter is disclosed which includes a low-pass filter and an analog delay circuit in which the delay time is set by a clock signal. THe analog delay circuit samples the voltage of the low-pass filter capacitor and delays the sampled voltage according to the clock signal. The output of the analog delay circuit is then used to equalize within a certain period of time the voltage of the low-pass filter capacitor. The analog delay circuit may consist of an integrated circuit BBD (Bucket Brigade Device) or CCD (Charge Coupled Device).

Influence of substrate current on hold-time characteristics of dynamic MOS IC's

Abstract: The charge hold-time characteristics of three-transistor-type MOS memory cells were investigated. It was observed that, at first, the holding-node voltage V/SUB G/ decreased slowly, due to junction leakage current. However, when V/SUB G/ reached a value determined by the sense-inverter dc supply voltage and external load resistance, V/SUB G/ fell rapidly. The onset value of the rapid drop coincided well with the value of V/SUB G/ that caused substrate current to flow in the sensing inverter. It is suggested that an extremely small amount of substrate current I/SUB sub/ arrives at the holding node and that positive feedback between V/SUB G/ and I/SUB sub/ causes the rapid drop of V/SUB G/. A simple analysis was pursued to estimate the transport probability /spl alpha/ of the substrate current I/SUB sub/ arriving at the holding node. It was also observed that the hold-time characteristics of neighboring physically independent cells were degraded by substrate current generated by one cell. The transport probability between these holding nodes and the driver transistor generating the I/SUB sub/ was estimated as a function of the separating distance.

Reduction in minority carrier storage effect by fluorine ion implantation damage

Abstract: Damage is produced in p-n diodes by fluorine ion implantation to reduce minority carrier storage effect. The switching time, reverse leakage current, and I-V characteristics were investigated for annealing temperature between 450°C and 650°C. The accelation energy is 130 keV and doses are 1013-1015/cm2. Annealing causes restoration in switching time, but leakage current increases with annealing temperature rise for doses more than 1 × 1014/cm2. The best diodes indicate 1.5-order reduction in switching time and 10 nA in reverse leakage current. These properties, caused by implantation damage, are retained after long-cycle annealing at 450°C and are expected to be stable under practical use. These diodes can be obtained by annealing at 450°C and they furnish satisfactory diode performance.

A small-signal analytical theory for GaAs field-effect transistors at large drain voltages

Abstract: Small-signal microwave performance of GaAs field-effect transistors (FET's) at large drain voltages is investigated, using a new analytical model which takes into account the carrier drift-velocity reduction and saturation due to electron upper valley scattering, and the extension of the depletion layer towards a drain side. Both of these play important roles in FET operation at large drain voltages. Small-signal y-parameters are calculated and the transit time effect which occurs in high-frequency operations is shown explicitly. Equivalent FET circuit elements are derived from the obtained y-parameters. Their dependence on device physical parameters, as well as on dc bias conditions, is calculated. The theoretical results are compared with the measured small-signal characteristics of a practical power GaAs FET and a reasonable agreement between them is obtained.

Container for storage or transportation of semiconductor chips

Abstract: A container for semiconductor chips is disclosed. The container includes an annular ring having a first diameter and an annular lower casing having a second diameter, greater than the first diameter. The annular lower container houses the annular ring. A plastic film, holding the semiconductor chips, is stretched over an annular opening in the annular ring. An upper container mates with the lower container to form an air-tight housing for the annular ring. The upper container is shaped in such a manner that the upper container fixes the annular ring in place but does not contact the semiconductor chip.

A new technique to minimize the EPROM cell

Abstract: This paper presents a new technique to minimize the UV Erasable, Electrically Programmable, Read-Only Memory (EPROM) cell based on N-channel double polysilicon gate MOS. The new technique features a fully self-aligned floating gate structure which reduces the EPROM cell size/bit to almost that of presently most advanced mask ROMs. A new fabrication process and the experimental results for both programming and erasure are presented. The dependence of characteristics on various device parameters are discussed with emphasis on the difference from conventional devices.

GaAs power MEWFETs with a simplified recess structure

Abstract: A simplified recess structure to reduce the drain breakdown voltage of GaAs power FETs will be covered. The structure, with an internal matching network has delivered an output power of 7W at 5.6GHz.

Fault location system for a repeatered PCM transmission system

Abstract: A PCM fault location system for repeaters and a code generator used therefor employ fault location codes generated on the basis of pseudo-random codes. The PCM fault location system transmits the codes through a repeatered transmission line, and the fault location codes including repeater transmission errors are received by the repeaters. A component of the repeater-designation frequency signal is extracted from the received fault location codes by means of a particular repeater-built-in fault locating circuit, and the thus extracted repeater-designation frequency component is received at a transmitting point of the fault location codes. The code generator employed at the transmitting point includes first means for obtaining codes which have an appropriate and sufficient difference in an average number of occurrences of digits "1" and "0" and are constructed using at least one series of pseudo-random codes, second means for controlling the state of the codes obtained by the first means in accordance with a repeater-designation frequency signal, and third means for performing differential code-conversion of the codes obtained by the second means.

System for suppressing sideband components of angle-modulated wave

Abstract: A system for suppressing a predetermined part of the sideband components of an angle-modulated wave in a frequency-division multiplex system is disclosed. The system branches the received angle-modulated wave into two angle-modulated waves, one of which is applied as the signal input of a phase modulator. The other of the two branched angle-modulated waves is demodulated to baseband. The demodulated signal is filtered, adjusted in level and phase and then supplied as the inversely modulating signal input of the phase modulator. The output of the phase modulator is therefore a wave which can be reused for transmission of other sideband components. The demodulated and filtered wave is provided as another output of the system.

Thin-Film Capacitors Made from TaN Films

Abstract: Thin-film capacitors were prepared from sputtered tantalum nitride thin films (TaN capacitors). The dielectric was formed by anodic oxidation and the counterelectrode was deposited by evaporation of nichrome-gold. The TaN capacitors excelled the usual TM capacitors in various properties. The temperature coefficient of capacitance was 129 ppm/°C and the dissipation factor was 0.0014 at 1 kHz. In addition, TaN capacitors showed high endurance to heat treatment. The capacitance change was only -0.02 percent after 350°C heat treatment for 2 minutes. Reliability tests were performed which indicated a failure rate of 0.1 FIT at 40°C, 6 V. The capacitance change was within ±0.4 percent after 3000 hours at 125°C. The TaN capacitors were applied to make highly stable and highly precise filter circuits such as multifrequency receivers for touch tone telephones now in commercial use.