NEURAL PROCESSING MODULE WITH INPUT ARCHITECTURES THAT MAKE MAXIMAL USE OF A WEIGHTED SYNAPSE ARRAY
    2.
    发明公开
    NEURAL PROCESSING MODULE WITH INPUT ARCHITECTURES THAT MAKE MAXIMAL USE OF A WEIGHTED SYNAPSE ARRAY 审中-公开
    模块具有输入神经处理架构AN突触矩阵极大EXPLOIT

    公开(公告)号:EP1247194A2

    公开(公告)日:2002-10-09

    申请号:EP99967712.3

    申请日:1999-12-30

    IPC分类号: G06F15/18

    CPC分类号: G06N3/063

    摘要: A neural processing module (100) is disclosed which combines a weighted synapse array (200) that performs 'primitive arithmetic' (products and sums) in parallel with a weight change architecture and a data input architecture that collectively maximize the use of the weighted synapse array by providing it with signal permutations (310) as frequently as possible. The neural processing module may be used independently, or in combination with other modules in a planar or stacked arrangement.

    INFRARED WIRELESS COMMUNICATION BETWEEN ELECTRONIC SYSTEM COMPONENTS
    4.
    发明公开
    INFRARED WIRELESS COMMUNICATION BETWEEN ELECTRONIC SYSTEM COMPONENTS 失效
    电子系统的组件之间的无线红外传输

    公开(公告)号:EP0852856A1

    公开(公告)日:1998-07-15

    申请号:EP95935157.0

    申请日:1995-09-27

    IPC分类号: H04B10

    CPC分类号: H04B10/11

    摘要: A wireless infrared pulse-transmitting system for communication with electronic components. The receiver (30) for such a system has a radically reduced current (and power) values, which permit a battery-powered receiver to remain on while awaiting transmitter signals. Use of several mosfet transistors operating in the subthreshold region minimizes power. Bandwidth requirements are met, in spite of the low power operation. In order to eliminate amplifier (40) saturation, with the accompanying problem of recovery time which slows the transmission process, clamping circuitry (46) is used to cause instantaneous shunting of the signals when a predetermined signal level is reached.

    FABRICATING ELECTRONIC CIRCUITRY UNIT CONTAINING STACKED IC LAYERS HAVING LEAD REROUTING
    5.
    发明公开
    FABRICATING ELECTRONIC CIRCUITRY UNIT CONTAINING STACKED IC LAYERS HAVING LEAD REROUTING 失效
    制作电子电路单元包括具有远期线路堆叠式芯片单据。

    公开(公告)号:EP0593666A1

    公开(公告)日:1994-04-27

    申请号:EP92916059.0

    申请日:1992-06-24

    IPC分类号: H01L21 H01L23 H01L25

    摘要: L'invention concerne un procédé et un produit appliquant des concepts avancés d'une technique en Z dans le domaine des boîtiers électroniques à haute densité de composants. Pour des tranches de silicium classiques contenant des puces (20), on suit des modes opératoires de modification qui créent des puces CI (38) possédant des conducteurs métalliques de deuxième niveau (30) sur la surface supérieure de la couche de passivation (52) (qui couvre le silicium d'origine (48) et l'aluminium ou autre métallisation (44)). Le métal des conducteurs de deuxième niveau (30) est différent de celui de la métallisation dans les circuits intégrés et permet une meilleure conduction électrique. Les puces modifiées (38) sont séparées des tranches (20) et sont ensuite empilées pour former des dispositifs à circuits intégrés multicouches (60). Une pile comporte un ou plusieurs plans d'accès (64). Après superposition, et avant application de la métallisation sur le plan d'accès, une opération d'attaque chimique enlève tout aluminium (ou autre matériau) qui pourrait interférer avec la métallisation formée sur le plan d'accès. Des plots métalliques (92) sont formés en contact avec les cosses des conducteurs de deuxième niveau (30) sur les puces empilées. Les plots et les cosses sont formés de la même matière métallique afin d'optimiser le rendement de conduction des connexions T.

    EP0465591A4 -
    6.
    发明公开
    EP0465591A4 - 失效
    EP0465591A4 - Google专利

    公开(公告)号:EP0465591A4

    公开(公告)日:1994-03-02

    申请号:EP90906503

    申请日:1990-03-23

    摘要: A multi-layer Z-technology module (20) having a two dimensional photodetector (24) mosaic is disclosed, in which the function of A/D signal conversion is accomplished in each on-chip channel. In order to satisfy the power and real estate limitations of the modules, a substantial part of the A/D conversion circuitry is located off-chip. Two devices are required in each channel on each chip (22), a precision comparator, and a storage register. These may be combined with an off-chip analog ram, and an off-chip digital ramp. Certain on-chip performance enhancements are disclosed, which can operate either in the analog mode or the digital mode. One such enhancement is compensating for the voltage offset of each comparator. Another enhancement is reducing the duty cycle of each precision comparator, in order to lower power requirements.

    摘要翻译: 公开了一种具有二维光电探测器(24)马赛克的多层Z技术模块(20),其中A / D信号转换的功能在每个片上信道中完成。 为了满足模块的功率和房地产限制,A / D转换电路的绝大部分位于芯片外。 每个芯片(22)的每个通道需要两个器件,一个精密比较器和一个存储寄存器。 这些可能与一个片外模拟RAM和一个片外数字斜坡相结合。 公开了某些片上性能增强功能,可以在模拟模式或数字模式下运行。 一种这样的增强是补偿每个比较器的电压偏移。 另一项增强功能是降低每个精密比较器的占空比,以降低功耗要求。

    HIGH-DENSITY ELECTRONIC MODULES, PROCESS AND PRODUCT
    7.
    发明公开
    HIGH-DENSITY ELECTRONIC MODULES, PROCESS AND PRODUCT 失效
    高密度电子模块,工艺和产品

    公开(公告)号:EP0385979A4

    公开(公告)日:1991-06-12

    申请号:EP88900618

    申请日:1987-10-20

    发明人: GO, TIONG, C.

    CPC分类号: H05K7/023

    摘要: A high-density electronic module (24) is disclosed, which is suitable for use as a DRAM, SRAM, ROM, logic unit, arithmetic unit, etc. It is formed by stacking integrated-circuit chips (22), each of which carries integrated circuitry. The chips are glued together, with their leads along one edge, so that all the leads of the stack are exposed on an access plane (28). Where heat extraction augmentation is needed, additional interleaved layers are included in the stacks which have high thermal conductivity, and are electrical insulators. These interleaved layers may carry rerouting electrical conductors. Bonding bumps (31 and 34) are formed at appropriate points on the access plane. A supporting substrate (26), formed of light transparent material, such as silicon, is provided with suitable circuitry and bonding bumps (38 and 42) on its face. A layer of insulation is applied to either the access plane or substrate face, preferably the latter. The bonding bumps on the insulation-carrying surface are formed after the insulation has been applied. The substrate face is placed on the access plane of the stack, their bonding bumps are microscopically aligned, and then bonded together under heat and/or pressure. A layer of thermally conductive (but electrically non-conductive) adhesive material is inserted between the substrate and stack. The substrate and stack combination is then placed and wire bonded in a protective container having leads extending therethrough for external connection.

    METHOD FOR FABRICATING MODULES COMPRISING STACKED CIRCUIT-CARRYING LAYERS
    8.
    发明授权
    METHOD FOR FABRICATING MODULES COMPRISING STACKED CIRCUIT-CARRYING LAYERS 失效
    用于制作包含堆叠电路层的模块的方法

    公开(公告)号:EP0204767B1

    公开(公告)日:1991-02-27

    申请号:EP85906123.6

    申请日:1985-11-20

    IPC分类号: H01L21/70 H01L21/90

    摘要: A method and related fixtures which permit formation of stacks of thin circuitry-carrying layers. The layers terminate in an access plane having a two-dimensional array of closely-spaced electrical leads. The method includes the steps of measuring the thickness of separate chips, selecting groups of chips having appropriate thicknesses, applying appropriate amounts of epoxy between adjacent chips, aligning the chips (and their electrical leads) in the direction parallel to their planes (i.e., the X-axis), and closing the cavity with an end wall which (a) exerts pressure on the stacked chips and epoxy in a direction perpendicular to the chip planes, and (b) establishes a fixed height of the stack in order to align the leads in the Y-axis. The final fixture provides a fixed-size cavity for confining the layers during curing of thermo-setting adhesive which has been applied between each adjacent pair of layers. An initial fixture is provided for accurately measuring the thickness of each layer under substantial layer-flattening pressure. An intermediate fixture is provided for wet stacking the layers prior to their insertion into the final fixture.

    MULTIPLEXER CIRCUITRY FOR HIGH DENSITY ANALOG SIGNALS
    9.
    发明授权
    MULTIPLEXER CIRCUITRY FOR HIGH DENSITY ANALOG SIGNALS 失效
    用于高密度模拟信号的多路复用器电路

    公开(公告)号:EP0116072B1

    公开(公告)日:1989-10-11

    申请号:EP83902618.4

    申请日:1983-07-25

    CPC分类号: H03K17/693

    摘要: A multiplexer circuit, for use with such signal sources as focal plane detector arrays (20), which contains a large number of parallel branches (24), each of which includes a transconductance MOSFET amplifier and a MOSFET switch of opposite channel polarity from the amplifier (40 and 42, respectively). The amplifier in each branch receives high impedance voltage signals originating from its individual detector (20) and converts them with high power gain into current signals which feed into the common output line (34) whenever the switch (42) in the same branch is turned on. The multiplexer branches, together with the multiplexer control logic, and other electronic devices, are all included on a single IC chip which provides CMOS logic.

    APPARATUS AND METHOD FOR FABRICATING MODULES COMPRISING STACKED CIRCUIT-CARRYING LAYERS
    10.
    发明公开
    APPARATUS AND METHOD FOR FABRICATING MODULES COMPRISING STACKED CIRCUIT-CARRYING LAYERS 失效
    用于生产模块,堆叠的集成电路包括在内。

    公开(公告)号:EP0204767A1

    公开(公告)日:1986-12-17

    申请号:EP85906123.0

    申请日:1985-11-20

    IPC分类号: H01L25 H01L21 H01L23 H01L27 H05K3

    摘要: Un procédé et des pièces correspondantes permettent de former des piles de fines couches porteuses de circuits. Ces couches se terminent par un plan d'accès ayant un agencement bidimensionnel de sorties électriques étroitement serrées. Le procédé comprend les étapes de mesure de l'épaisseur de puces isolées, de sélection de puces ayant l'épaisseur appropriée, d'application de quantités appropriées d'époxy entre des puces adjacentes, d'alignement des puces (et de leurs conducteurs électriques) dans un sens parallèle à leur plan (c'est-à-dire, l'axe X), et de fermeture de la cavité par une paroi terminale qui (a) exerce une pression sur les puces empilées et l'époxy dans un sens perpendiculaire au plan des puces, et (b) établit une hauteur fixe de la pile afin d'aligner les sorties selon l'axe Y. La pièce finale forme une cavité de dimensions fixes pour confiner les couches pendant le durcissement de l'adhésif thermodurcissable appliqué entre chaque paire de couches adjacentes. Une pièce initiale mesure avec précision l'épaisseur de chaque couche sous une pression considérable d'aplatissement de la couche. Une pièce intermédiaire sert à empiler par voie humide les couches avant leur introduction dans la pièce finale.