Abstract: A commutator-type data store (102) is serially connected to a variable length shift register (104). Data is first written in the commutator store at the incoming varying line rate. A predetermined time later, the stored bits are read out from the commutator-type store and stored in the shift register at a given rate. Finally, the shift register data is read out at a constant predetermined rate. The length of the shift register is controlled by a counter (110). The phase relationship between the commutator write and read cycles is monitored by logic circuitry (501, 502, 503, 504, 505, 506) within the control unit (106). If this phase relationship has increased by a preselected amount, the commutator readout rate is increased along with the length of the shift register. Alternatively, if the phase relationship has decreased by a preselected amount, the commutator store readout rate is decreased along with a decrease of the shift register length.

Abstract: The successful application of the reactive ion etching technique to the III-V compounds requires the use of the appropriate etch gas. We have found that a gas mixture comprised of either CCl.sub.2 F.sub.2 alone or in combination with one or more of the gasses: argon (Ar), oxygen (O.sub.2) and nitrogen (N.sub.2) will cleanly and effectively etch GaAs and InP and their ternary and quaternary alloys as well as AlGaAs and the oxides of GaAs. The effective ranges of relative flow rates of Ar, CCl.sub.2 F.sub.2 and oxygen are: Ar (0-83%), CCl.sub.2 F.sub.2 (8-100%), O.sub.2 (0-50%), and N.sub.2 (0-60%).

Abstract: The adaptive decision level circuit (FIG. 3) slices a digital signal with respect to a decision level disposed between the signal levels. A correction signal generated from the sliced signals (S, S) compensates for variations in the signal levels by maintaining the decision level at a constant position relative to the signal levels. The correction signal is generated by comparing (321) the integrated difference (on lead 305) between the sliced signal and its complement with the statistically expected value (312) of the difference between the sliced signal and its complement. In the preferred embodiment, the expected value is equal to one-half the signal swing of the sliced signal.

Abstract: A reduction in idle channel noise and crosstalk in a mid-riser-biased successive approximation encoder is achieved through the use of two polarity decisions. Encoder (401) assigns a sampled analog input signal to the closest one of a multitude of discrete signal levels or code steps. A comparator (109) provides this assignment by successive comparisons of the sampled signal with a series of reference signals (110). Each comparison produces a binary digit. The first comparison, or polarity decision, is not transmitted and instead coupled to feedback circuitry (401) to reduce any dc component in the analog input signal to substantially zero. A second polarity decision is then made using a non-zero offset reference signal (402) corresponding to an intermediate position on a code step, typically the midpoint. The non-zero offset reference signal is applied along with subsequent reference signals to the comparator to determine the closest code step.

Abstract: Improved input offset voltage compensation of an amplifier (12) is achieved through the use of a servo loop which is added to electronic switch and capacitor offset compensation circuitry disposed in the input (14) and feedback (20) paths of the amplifier. A voltage approximately equal to the offset voltage is stored on a feedback capacitor (24) by operation of input reset (34), feedback shunt (22) and feedback reset (26) switches. An error correction voltage is then generated within the servo loop (40) to adjust for the residual offset produced by the operation of the input reset and feedback shunt switches as well as the finite gain of the amplifier.

Abstract: A connector assembly (10) has been developed with an aligning coupler (300) and connectors (100, 200) which simplify the joining of a pair of mateable optical fiber termination structures (112, 212). The coupler (300) comprises a pair of rigid elongated members (310, 340) having inner surfaces (312, 342) and tapering outer end surfaces (328, 358). Resilient members (370) hold the elongated members together to define a fiber termination structure receiving orifice (380) which has a cross section normally narrower in one dimension than that of the termination structures. Due to the resilient members, the elongated members can separate to ease insertion of the termination structures and then come back together to urge precise-aligning outer surface means (122, 222) on each termination structure into registration against a precisealigning coupler reference surface (332) in the orifice. Thus registration is maintained until the connector assembly can be fastened.

Abstract: In a digital transmission system, disparity is defined as the difference between the number of binary digits at each code state. The generation of a block of binary digits having zero disparity possesses many advantages. The present invention discloses coding apparatus (100, 300) for converting a block of binary digits having random disparity into a zero disparity block, and vice versa. The coder accomplishes this translation by determining (103) the disparity of the random disparity block and then selecting (103, 104, 106, 107, 110, 111) a bit position which divides the block into two digit segments each having half this disparity. Inversion (115) of either digit segment generates a zero disparity block. For decoding, data representing the bit position selected is transmitted along with the zero disparity block. The binary digits previously inverted by the coder are then reinverted. This coding/decoding technique is adaptable to any block size having an even number of binary digits.

Abstract: An improved molded optical fiber connector is achieved using an insert (111) disposed about the optical fiber. During molding, filler material (112) pushes the insert along the optical fiber (106) until it abuts a mold surface (110) comprising a portion of the connector mating end. This sliding action reduces any bends in the fiber and controls the position and angle of the fiber relative to the mating end. Use of the insert is suitable for the molding of low-loss monomode or multimode optical fiber connectors.

Abstract: A digital logic circuit for eliminating spurious transient pulses (21, 32) from among genuine pulses (51) in a binary dial pulse stream. The dial pulse stream is coupled to the circuit input (46). Output (47) is prevented from matching the input unless the duration of the input change exceeds a first time interval (57). Moreover, if the input change exceeds the first time interval and the input change was from a first to a second predetermined binary signal, then the new output signal is held constant for a second time interval (56) regardless of any change in the input signal.

Abstract: An insert (21) is molded within an optical fiber connector (2,3) to provide more precisely defined and symmetric alignment surfaces (5,6). The insert may be used for connectors directly molded about an optical fiber or for field mountable connectors subsequently affixed thereto. During molding, the insert partitions the mold cavity into two portions. The first portion (53) defines the alignment surfaces and the second defines the balance of the connector. Filler material (55) fills the second portion first and then is conducted through the insert to symmetrically fill the first portion. This directed flow control also precludes bending of the optical fiber (32) during direct molding of a connector about the fiber.

Abstract: An optical switch is disclosed which relies only on a rotatable, reflective concave surface (3) to couple light between an optical source (1) and a preselected one of a plurality of optical receptors (6). The switch is bidirectional, i.e., the position of the source(s) and receptor(s) can be interchanged. Optical fibers may be used as the source(s) and receptor(s). The source, however, may also be chosen from such devices as lasers, light-emitting diodes or any other optical source known to those skilled in the art. Similarly, avalanche photodiodes or the like may be used as optical receptors.The above switch structure requiring only a reflective element is particularly suitable for switching light signals in telecommunication applications. The elimination of the need for a refractive element, such as a lens, provides a structure in which the output signal is independent of wavelength and index of refraction.

Abstract: A two-piece tool for circumferentially scoring an optical fiber (11). Retaining member (21) has a substantially flat end surface (26) beyond which a portion (13) of the optical fiber protrudes perpendicularly. The scoring ring holder (28) floatably mounts a scoring ring (30). When the retaining member is inserted into the scoring ring holder, the portion (13) of the optical fiber extends into the scoring ring. A preselected relative orbital motion between the retaining member and scoring ring holder causes the scoring ring to revolve about the optical fiber scoring a line about the fiber circumference.

Abstract: A semitransparent photolithography mask, i.e., one transparent to a portion of the visible light spectrum, is achieved by electron and ion bombardment of a positive photoresist. This bombardment is performed after the desired photoresist pattern is defined by the sequential steps of depositing, baking and developing the photoresist (14) on a light transparent substrate (15). In the preferred embodiment, the electron and ion bombardment is accomplished in a predominantly nitrogen atmosphere within a sputtering chamber using indium tin oxide as the sputtering target. As a result of the bombardment, the photoresist is converted into a material which is opaque to only the ultraviolet portion of the light spectrum.

Abstract: A phase-locked loop for PCM transmission systems has a phase comparator (20) which provides continuous regulation of the loop oscillator (30). When a PCM pulse signal is present, the phase comparator selectively compares the PCM and oscillator clock signal and generates an error signal which synchronizes the oscillator to the PCM frequency. During the absence of a PCM pulse, the phase comparator provides a high impedance output and oscillator synchronization is maintained by a loop filter (23). In the disclosed embodiment, the phase comparator is a dual section tristate gate (300).

Abstract: An apparatus and method for determining the tuned frequency of a digital repeater (2) solely through the use of external connections. A pulse burst (30) at a preselected transmission frequency, produced by generator circuitry (4), is inserted into repeater (2) to cause the timing circuit of the same to oscillate for a predetermined time interval (25). After burst (30) and during interval (25), a plurality of sequential sampling pulses (26, 27) is fed to repeater (2) and regenerated by same. This format of burst (30) followed by sampling pulses (26, 27) is preferably repeated a number of times. Counter (10) measures the average time interval between the regenerated sampling pulses and determines the tuned frequency.