Abstract: A detachable fiber optic connector assembly consists of two major components, a cable terminator and a connector unit. The cable terminator includes a collar which is glued or crimped onto the cable jacket adjacent the stripped end of the cable and a cylindrical housing which is trapped on the cable by the collar. The connector unit includes a short length of permanently-retained fiber with one exposed end. This unit further includes a splice sleeve for receiving the stripped end of the cable fiber and the other end of the permanently-retained fiber. The cable terminator and the connector unit have latches, such as BNC components, for detachably coupling the connector unit to the cable terminator.

Abstract: Operating parameters for a multiple facet holographic disc scanner are adjusted on real time basis as a function of the actual diffraction efficiency of the facet region aligned with the impinging laser beam. The operating parameters are adjusted to achieve optimum signal levels independent of facet characteristics. Diffraction efficiency is inversely related to zero order beam power for a given level of input beam power. A photodetector measures the zero order beam power. The output of this photodetector is used to control video amplifier gain and/or semiconductor laser current.

Abstract: This interface circuit between a bidirectional line (Lg) and a telephone exchange includes a high frequency transformer (Tr) whose secondary (N2) is connected to said line through a detector circuit (DET) and whose primary (N1) is connected to the switching network of said telephone exchange and to a high frequency carrier source. The circuit further includes means for adjusting the impedances seen across the points where the telephone line is connected to the exchange, so as to meet specified values applicable to the particular location of the exchange. The impedance adaptation means include means (A1) for amplifying the line current variations seen at the center tap (K) of the high frequency transformer (Tr), addition means (.SIGMA.1) for adding the output from (A1) to the voice signal to be sent to the telephone line (Lg), and second amplification means (A2) responsive to the output from (.SIGMA.1) to drive said center tap (K).

Abstract: For use in a teleconference system, a method of reducing line width variations, which are assumed to be the result of undersampling of the original image. Bilevel data received at a remote site is written into a pair of two-dimensional storage arrays. Data stored in the first array is examined row by row to identify different length-dependent classes of black horizontal pel runs. White or black pels in the second array are overwritten with gray pels at specific pel locations as a function of run length to reduce variations in length. The first array data is then examined on a column by column basis. Black vertical runs are classified and the second array is overwritten with gray pels using the same rules as were applied to horizontal runs. The trilevel second array data may be used to drive a display monitor or may first be subjected to further image processing.

Abstract: A line adapter for permitting the transfer of information between terminals and a bus connected to the central control unit CCU of a communications controller. The line adapter includes a microprocessor associated with a control memory and a scanner. A random access memory is split into zones, one zone being assigned to each line. During an initiating phase, the zones are personalized under the control of the microcode by sending initiating parameters (which are characteristics of the lines) to the memory, through an asynchronous path bus and external registers. Exchanges of temporarily-stored data between the CCU bus and the lines are carried out through cycle steal operations between the control memory and the random access memory.

Abstract: A technique for enhancing a displayed image originally stored at a density of 1 data bit per pel. The stored data is scanned to locate unit steps, one-row or one-column steps in binary values. Whenever a unit step is found, the distance from the step to the nearest transitions on both sides of the step are found. The distances are used to access lookup tables which assign 4 bit gray scale values to limited numbers of pels on both sides of the unit step. The assigned gray scale values produce a gradually-darkening string of pels extending through the step. When the entire image field has been scanned, a maximum 4 bit pel value is assigned to any unmodified pel having an original binary value of 1 and a minimum pel value is assigned to any unmodified pel having an original binary value of 0. The visual quality of the displayed image is enhanced by the selective introduction of intermediate gray scale values for pels adjacent unit steps.

Abstract: Low loss optical couplers for an optical data bus include a first component having permanent input/output connections to the bus and a reference face capable of reflecting energy from the input to the output. If optical energy is to be coupled into or out of the bus, a second component is secured to the first component. The second component has input and/or output optical connections and a reference face which contacts at least part of the reference face of the first component and which permits optical energy to be transferred between the two components. Frequency-selective encoders and decoders using active optical devices are also disclosed.

Abstract: A self-focussing coupler for optical fibers includes a transparent body having the configuration of a paraboloid of revolution. One or more optical fibers are connected to the body at or near the point of focus. A flat face is formed at a normal to the axis of symmetry. A coupler assembly is formed by bringing the faces of two couplers together. Different optical components (beam splitting plates, bandpass filters, etc.) may be placed in an interface gap and bypass fibers may be added to alter the functions of the coupler assembly.

Abstract: An optical scanner in a retail store checkout stand includes a single item detector upstream of the scanner window. Output signals from the item detector and from symbol recognition logic are used to open and close a shutter in the beam path of the scanner laser. The shutter is opened and a first time out sequence is started when the beam path is interrupted. A second, short time out sequence is restarted each time a valid symbol is recognized. A third time out sequence of intermediate length supersedes the first time out sequence when the beam path is cleared. The laser shutter is closed and the symbol is transferred to a terminal or controller upon the completion of any active time out sequence or upon recognition of a predetermined number of identical symbols, whichever comes first.

Abstract: High voltage tolerant FET circuits are characterized by the use of shield structures surrounding source/drain diffusion pockets, with the shields tied to apropriate potentials, which in some cases is the associated gate potential. Some embodiments use enhancement mode devices which however have implanted channels underlying the shield structures. Operation of several embodiments is achieved near the snap-back limits by the use of a clamp to maintain potential drop below this limit. High voltage switching at heavy loads is achieved by a voltage divider providing appropriate gate potentials to the load carrying FETs.

Abstract: In decoding a UPC-type label, a partial decode and majority vote technique is employed to select final values for ambiguous characters (i.e., 1-7, 2-8) appearing in the label. During the partial decode phase, character-identifying values are established for every character but a common value is assigned to any ambiguous character in a given subset of such characters. An ambiguity-resolving value is additionally assigned to an ambiguous character each time it appears. Through algebraic summation of the ambiguity-resolving values for each character position occupied by an ambiguous character, the ambiguous character which appears a majority of times in that position is finally selected.

Abstract: A switching network is comprised of a plurality of identical chips and a network (control) processor. Each of the chips is a novel intelligent switch and includes both a cross point array as well as apparatus to control the cross point array in response to a set of multi bit commands or to formulate a response to a query concerning status of the switching array. The number of command lines connecting network (control) processor to each of the array chips may be as few as two. The cross point array includes a cross point device (switch) for each inlet-outlet combination. The total number of cross points in the network is lower than that dictated by a CLOS network.

Abstract: Production quantities of a multi-element holographic scanner disc are made by optically replicating a silver halide master disc one element at a time in a dichromated gelatin film. The dichromated gelatin film swells during processing. The swell is monitored during production by determining the shift in the angle of the Bragg surfaces within the gel. The angle of the replicating beam for each element is changed from that of the original reference beam to establish a Bragg angle at exposure which will be tilted to the proper angle after swelling in order to maximize the diffraction efficiency of the element at the original reference beam angle.

Abstract: A scan pattern for a bar code scanner is produced by the combination of a coherent light source, a rotating disc with holographic facets and an array of fixed, beam folding mirrors. Light reflected from a bar code label follows a retroreflective path to a photodetector within the scanner. To reduce variations in the amount of light collected through different facets, the areas of the facets are varied as a function of beam elevation and beam path length.

Abstract: A technique for compensating for systematic errors in printing or reading bar codes is disclosed. Label decoding always begins with a center character and proceeds one character at a time toward a margin. When bar-space pair measurements indicate an ambiguous character (which can only be fully decoded using bar width measurements) error correction is based on the characteristics of the adjacent, previously decoded character.

Abstract: A scanning system includes a pair of buffer memories for alternately accumulating strings of signals which may contain bar coded label data. Candidate selection logic circuits examine incoming signals to generate a candidate signal when a string of signals satisfies gross logical tests. A control counter responds to the incoming signals and to the candidate signal to select an alternate buffer memory only when the active memory is fully loaded and contains at least part of a label candidate. The contents of the fully loaded memory are transferred to a processor for further analysis while incoming signals are routed to the alternate memory.

Abstract: A bar code label candidate is framed in a stream of scanner-generated data by a circuit which counts the width of each pulse and which accumulates a fixed number of pulse width values on a rolling basis to provide overlapping basic unit signals. Each basic unit signal represents a potential N-pulse character. Basic unit signals occurring at N intervals are compared in a comparator circuit to determine whether predetermined character-indicating relationships exist. Character-indicating signals are accumulated in registers. Outputs from these registers are decoded to determine whether a label candidate (appropriate type and number of apparent characters in appropriate time sequence) has been found.

Abstract: Communication control functions are performed by an integrated adapter implemented as microcode resident in host CPU storage. The integrated adapter shares a common high speed bus with other CPU facilities. A high speed bus adapter provides an interface between the common high speed bus and low speed line adapters. Communication controlling commands and register structures are described.