Abstract: An optical connector for device to fiber coupling is disclosed comprising a connector housing having waveguide channels formed therethrough. The connector housing is injection molded from a plastic material. These waveguide channels each include a first end and a second end. Attachment apparatus for an optical fiber is included at each of the first ends of the waveguide channels, and attachment apparatus for an optical device is included at each of the second ends of the waveguide channels. Waveguide core portions, positioned within the waveguide channels, are made from an optical plastic material.

Abstract: A collision detection method and apparatus is disclosed for detecting code rule violations in Manchester encoded data packets. The signals.delta..sub.0 =S.sub.2n -S.sub.2n-1.delta..sub.1 =S.sub.2n+1 -S.sub.2n.delta..sub.2 =S.sub.2n+1 -S.sub.2n-1.sigma..sub.0 =S.sub.2n +S.sub.2n-1.are determined at time nT where S.sub.2n-1 and S.sub.2n are the Manchester code bit pair representative of the nth bit in a data stream having a bit period T. Processing of these signals permits the detection of most collisions since the Manchester coding rules are violated more than half the time in the signal formed by the colliding packets. To ensure 100% collision detection, a preamble using an intentional violation of the Manchester code in the pattern 10 11 01 10 is advantageously used in each data packet and a collision is signaled upon detection of two or more coding violations.

Abstract: A star-coupler and a method for constructing the star-coupler employing injection molding technology are disclosed. The star-coupler may be constructed from a single injection molded block or a plurality of mating sections, and is supplied with a plurality of input ports and a plurality of output ports. Each input port as well as each output port is connected to at least one channel. The channels are filled with a fiber optical core material and the single block or mating sections are fabricated from fiber optical core material. Alternatively, a separate cladding material may be coated on the surfaces of the block or mating sections which define the channels. Optical signals are guided and confined by the core and cladding material from input ports to output ports along the channels.

Abstract: A fiber optic star-coupler with a biconical mixing element and method for inexpensively manufacturing the same are disclosed. The fiber optic star-coupler includes an input fiber optic bundle, an output fiber optic bundle and a biconical mixing element. Each of the bundles comprises a plurality of individual optical fibers. The biconical mixing element includes tapered regions and a central mixing zone to promote uniform mixing and equal distribution of optical energy among all of the optical fibers in the output fiber optic bundle. The biconical mixing element is provided with connectors at each of its two ends for coupling to other connectors, adaptors, and fiber optic bundles. Alternatively, the biconical mixing element may be formed by injection molding techniques. The optical fibers may be constructed from plastic or silica.

Abstract: A shadow-masking process for manufacturing low dark current photodetectors with low noise characteristics is disclosed. The process includes shadow-masking a semiconductor wafer by positioning a patterned shadow-mask on a surface of the wafer. The shadow-mask is patterned with, for example, circular openings or stripe openings. Layers, such as metallization layers to form metallic contacts or anti-reflection layers are deposited onto the wafer through the patterned openings in the shadow-masks. This shadow-masking process may be used in the production of any semiconductor device requiring patterned layers.

Abstract: A star-coupler and a method for constructing the star-coupler employing injection molding technology are disclosed. The star-coupler may be constructed from a single injection molded block or a plurality of mating sections, and is supplied with a plurality of input ports and a plurality of output ports. Each input port as well as each output port is connected to at least one channel. The channels are filled with a fiber optical core material and the single block or mating sections are fabricated from fiber optical core material. Alternatively, a separate cladding material may be coated on the surfaces of the block or mating sections which define the channels. Optical signals are guided and confined by the core and cladding material from input ports to output ports along the channels.

Abstract: A method and apparatus for connecting optical fibers is disclosed in which a housing is provided with a built-in alignment channel. An insert is inserted into the alignment channel and precisely and accurately maintains a fiber element concentrically within the alignment channel. A mating optical fiber is also introduced into the alignment channel. Means are provided for adjusting the separation distance between the two optical fibers. The housing is preferably injection molded.

Abstract: Through the method and apparatus of the present invention the collision of two packets can be detected with a high degree of accuracy under a variety of conditions. The preamble of each packet transmitted is modified to include collision detection data comprising an initial pulse having an amplitude sufficiently high to be detected even when attenuated and having a pulse width twice the normal data pulse width followed by a transceiver code of a predetermined length unique to each transceiver. Thus, in the event two packets collide, the collision may be detected by recognizing a second large pulse within a time period less than the sum of the packet duration and the gap between packets, by determining if a large initial pulse has a pulse width larger than the generated pulse width, and by determining if the transceiver code echoed back to the transmitting transceiver is equal to transceiver code stored in memory on the transceiver.

Abstract: Through the method and apparatus of the present invention the collision of two packets can be detected under a variety of conditions. The preamble of each packet transmitted is modified to include collision detection data comprising an initial pulse having an amplitude sufficiently high to be detected even when attenuated and having a pulse width twice the normal data pulse width followed by a randomly generated code of a predetermined length in which a predetermined number of pulses of the code are at a high binary data signal level. Thus, in the event two packets collide, the collision may be detected by detecting a second large pulse within a time period less than the sum of the packet duration and the gap, by determining if a large initial pulse has a pulse width larger than the generated pulse width, by determining if a subsequent large pulse occurs within the remainder of the packet and by determining if more than the predetermined number of pulses of the code have a high binary data level.

Abstract: A method is disclosed for implementing a token ring network on a token bus network. In accordance with the invention a token ring node is connected to each one of a plurality of token bus nodes such that the token ring node appears to the token bus node to which it is connected as station equipment and the token bus node appears to the token ring node as both the next token ring node and the preceeding token ring node in a ring network. A message is transmitted from a token ring node by forming a token ring frame and transmitting the token ring frame to the next token ring node which is the token bus node to which the token ring node is connected. The token bus node encapsulates the token ring frame that it receives in a token bus frame and transmits the resulting token bus frame to a destination bus node.

Abstract: A high efficiency edge-emitting LED is disclosed in which optical radiation produced in an active layer of the device leaves the device without being substantially absorbed by the material comprising the active layer. The device is constructed so that an optical waveguide is formed such that light produced in the active layer either leaves the device without any reflections or with a maximum of one reflection. The thickness of the active layer and the compositions of the active layer and confining layers are chosen so that electrical carriers are confined in the active layer while light produced in the active layer is not confined therein.

Abstract: An edge-emitting LED in which light produced in the active layer optically tunnels out of the active layer through a thin confining layer into a single guide layer which is composed of a material which does not absorb light energy. The guide and active layers are constructed so that they are surrounded by layers of lower index of refraction so that an optical wave guide is formed. To minimize light losses by absorption, the non-absorbing guide layer is preferably much thicker than the active and confining layers so that light produced in the active layer travels primarily in the non-absorbing guide layer before it leaves the device at the exit facet of the LED.

Abstract: A method and apparatus are disclosed for recovering data and clock information from high speed self-clocking data streams at rates of up to approximately 50 Mbps. The self-clocking data stream has a predetermined bit period and two voltage levels with negative-going and positive-going voltage level transitions. A pulse is generated for each voltage level transition in the coded stream. This pulse is applied to a non-retriggerable one-shot that produces an output pulse that is approximately the duration of one-half of a bit period. After the output pulse of the one-shot terminates, the one-shot is non-retriggerable for a period of time less than the duration of one-half of a bit period. The output pulse from the one-shot provides clock information derived from the coded bit stream. This clock information is provided to the clock input of a bistable storage device and the coded bit stream is applied to a data input to the storage device.

Abstract: A data communication receiver that provides very low pulse distortion at excellent sensitivity over a wide range of optical signal levels and simplified circuitry for the detection of data collisions. The receiver comprises an amplifier having a series of stages, all but one of which is directly coupled to the preceding stage and one of which is capacitively coupled by a capacitor that also functions as part of an RC differentiator. The output of the amplifier is a series of peak signals that are applied to a Schmitt trigger, the output of which is applied to a collision detection circuit.

Abstract: Method and apparatus for detecting when a collision occurs between two or more binary data signal packets at a single binary data signal receiver. Whenever a packet is transmitted, a large signal impulse is substituted for the initial bit of the packet. The amplitude of the impulse is such that it is significantly greater than the highest binary signal level transmitted. The receiver recognizes the impulse not as a data bit but as an indication of the beginning of a packet. Every packet has a known duration (or packet period) and is separated from the next packet by at least a known gap period. If two impulses are received within a period less than or equal to the sum of the packet and gap periods, a collision has occurred and a packet collision signal is generated.

Abstract: A heterojunction photodiode with improved wavelength-selectivity and risetime. The problem of short-wavelength diffusion-tail response is avoided by interposing between the window and active layers a barrier layer of higher bandgap than that of the window layer, which prevents high-energy photocarriers generated in the window layer from diffusing to the PN junction. In one embodiment, n-type substrate, active, barrier, and window layers are initially grown, and the window layer is coated with an opaque oxide. A window is opened in the oxide layer, and a p-type dopant is diffused heavily through the opening, through the window layer, and partly into the barrier layer. A PN junction is thus formed in the barrier layer, its depletion region extending through the remaining n-type region of the barrier layer and into the active layer, where photocarriers are generated by photons passing through the window-opening.

Abstract: The alignment of optical fibers is improved by mounting the optical fiber on a mounting block that has a sloping surface relative to an underlying support that bears both the fiber and a light source or photodetector. As a result, the alignment of the fiber and the light source or photodetector may be maintained in the vertical direction by the correct position of the fiber on the sloping surface of the mounting block. The alignment of the fiber in the horizontal direction is determined by the position of the mounting block on the support. Such an arrangement permits the weight of the fiber to be supported by the block while an appropriate cement is being cured to hold the fiber and block in place. Advantageously the mounting block is transparent to ultraviolet (UV) light and UV curable cements are used to secure the fiber to the mounting block and the block to the support.