专利汇可以提供CURRENT SUMMED OPTOELECTRONIC CROSSBAR SWITCH专利检索,专利查询,专利分析的服务。并且Le commutateur crossbar opto-electronique (50) permet de connecter sélectivement des signaux sur une pluralité de canaux de fibre optique (52) à un nombre donné de récepteurs de sortie (78a-h). Chaque canal à fibre optique est divisé en un nombre différent de fibres optiques (56a-h) qui sont agencées en un réseau constitué de plusieurs rangées et colonnes. Chaque rangée contient une fibre provenant de chaque canal. Un réseau d'addition de courant (58) est prévu pour chaque rangée et utilise avantageusement une pluralité de détecteurs sélectivement actionnables (60-74). Les détecteurs sont maintenus dans leur état de circuit ouvert via un multiplexeur CMOS (80) de manière à réduire au minimum la diaphonie entre les détecteurs adressés et les détecteurs non adressés. La conception du commutateur (50) permet également de réduire au minimum le nombre de composants requis.,下面是CURRENT SUMMED OPTOELECTRONIC CROSSBAR SWITCH专利的具体信息内容。
CURRENT SUMMED OPTOELECTRONIC CROSSBAR SWITCH
The U.S. Government has rights in this application pursuant to Contract No. N 6601-86-C-0110 with the Department of the Navy.
CROSS REFERENCE TO RELATED APPLICATIONS The subject matter of this application is related to U.S. Serial No. , filed December 21, 1987, entitled "Distributed Crossbar Switch" by Tangonan which is assigned to the same assignee as the present invention- which is incorporated herein by reference.
BACKGROUND OF THE INVENTION 1. Technical Field
This invention relates to switching devices and, more particularly, to techniques for switching a multiplicity of data signals in communication networks.
2. Discussion
Computer architectures now commonly include a large number of processors that execute programs concurrently. These so called parallel computing architectures typically require a communication network that allows efficient data transfer between processors, and between an individual processor and its associated memories. In many parallel processing systems it is desirable to permit many processors to transfer 1 information simultaneously. The use of optics to accomplish this task is particularly effective since optical systems can provide appropriate inherent parallelism, suitable band widths and low interaction between the separate beams in a linear medium. One of the most important optical systems is known as a crossbar network. A crossbar network permits a plurality of processors or memories to be dynamically interconnected in any selected configuration without
10 moving any existing interconnections.
FIG. 1 illustrates a known "star" crossbar network 10 that is disclosed in more detail in the paper authored by MacDonald et al., in Electron Lett. 14 (1978) page 502, entitled "Optoelectronic Broadband
15 Switch Array" which is hereby incorporated by reference. Briefly, network 10 includes four transmitters 12, 14, 16 and 18 which may be connected to one or more processors. The transmitters provide input signals which are coupled to optical fibers 20,
20 22, 24 and 26. The fibers are, in turn, fanned out and made incident on a plurality of detectors 28. Each of the detectors includes a photodetector 29 connected in series with a preamplifier 31. Each of the detectors operates to selectively convert light incident on its
25 associated photodetector to an appropriate electrical signal. The detectors are arranged in an array of rows and columns. Each of the detectors in a row are connected to a common output line 33. The outputs of the lines 33 are coupled for rebroadcast to suitable
30 output receivers labeled 30-36 in FIG. 1. Typically, optical fibers are used to transmit the signals from the crossbar network 10 to these receivers. Accordingly, an additional reconversion of the electrical signals on lines 33 to optical signals are
- - performed by optical repeaters 38-44. Typically, the selection of which detector to supply its output signal onto line 33 is made by bias control circuitry (not shown) . In many of the known crossbar networks it is necessary to provide a deactivating bias to all of the detectors which are not selected. As a result, the switching control is somewhat complicated and unwanted noise crosstalk between detectors has been experienced. In addition, each detector requires its own preamplifier circuitry. Thus, the necessity for these additional components limits the ability to fabricate large, compact crossbar networks.
SUMMARY OF THE INVENTION In accordance with the present invention, a crossbar switch is provided for selectively connecting signals from M optical fiber channels to N optical receivers. The switch includes a splitter for splitting each optical channel into N optical fibers. The fibers are arranged in an array having a plurality of rows and columns, each row containing fibers from each channel. A summation network is provided for each row of fibers. The network includes M selectively activatable detectors each receiving optical signals for a given fiber. Each detector provides an electrical output signal at its output if activated by a given bias signal when illuminated by an incident optical signal from its associated fiber. Addressing means coupled between a source of the bias voltage and the detectors are utilized to selectively activate a given detector by applying the bias voltage source thereto. All of the detector outputs are coupled to a common node which, in turn, requires only one amplifier means for amplifying the current delivered to the node from the selected detector. Accordingly, only one amplifier is required for each row of fibers thereby minimizing the number of components required for the crossbar switch.
BRIEF DESCRIPTION OF THE DRAWINGS The various advantages of the present invention will become apparent to those skilled in the art after reading the following specification and by reference to the drawings in which:
FIG. 1 is a perspective view of a crossbar switch of the prior art; and
FIG. 2 is a schematic diagram of a crossbar switch made in accordance with the teachings of the present invention.
DESCRIPTION OF THE PREFERRED EMBODIMENT
Turning now to FIG. 2, a crossbar switch 50 of the preferred embodiment of this invention is illustrated. It includes an 8x8 array of detectors for receiving outputs from a like number of optical fibers split from eight optical channels. In FIG. 2, for sake of simplifying the drawing, one optical channel 52 is illustrated which is split by suitable splitter means 54 into eight separate optical fibers 56a-56h. The other seven optical channels are similarly split and the outputs of the individual fibers are arranged in an array having a plurality of rows and columns. In this embodiment, there are eight rows and eight columns. Each row includes a substantially identical summation network 58a-58h. Since each of the summation networks is substantially identical in construction, a detailed description of only one of them will suffice.
Network 58a includes eight different photodetectors 60-74. Each of the photodetectors receives light from a fiber associated with a different optical channel. In the preferred embodiment, each photodetector 60-74 is a metal-semiconductor-metal (MSM) device fabricated from gallium arsenide (GaAs) .
MSM devices are preferred because they are efficient detectors with excellent on/off switching capabilities.
The excellent on/off ratio results from the absence of any photovoltaic effect at zero current or zero voltage. Preferably, the MSM devices are constructed of two integral, back-to-back Schottky barrier diodes formed by the deposition of an interdigitated tungsten cathode and anode deposited onto the surface of a <100> semi-insulating gallium arsenide substrate. Further characteristics of suitable MSM detectors are described in D.J. Jackson et al, "Monolithically Intregrable High
Speed Photodetectors", Proc. SPIE 716 (1987) 104; and
Wei et al, "Symmetric MOTT Barrier As A Fast Photodetector", Electron Lett 17, 688 (1981), both of which are hereby incorporated by reference.
All of the outputs of the detectors 60-74 are coupled to a common line or node 76. Provision is made for addressing selected detectors so that they may operate to switch a signal from one or more of the eight optical input channels to the output receiver connected to the crossbar row outputs 78(a-h). It has been found that advantageous results can be obtained if the nonaddressed detectors 60-74 are placed into a substantially open circuit condition while applying a bias voltage only to those detectors desired to respond to the light incident upon them. Normally, only one detector is addressed or activated at one time. This so called "zero-current" operation substantially simplifies the addressing scheme required while also minimizing crosstalk between detectors. To this end, a complimentary metal oxide semiconductor (CMOS) multiplexer 80 is employed to selectively couple a suitable bias source V„ B to one or more of the detectors 60-74. CMOS devices for the multiplexer 80 are preferred because of their extremely large input impedance when in the nonconducting state. When each detector is in the "off" or nonaddressed state, the gate-to-source potential of its associated CMOS device 82-96 is much greater than the pinch off voltage for the FET. Thus, the potential across the MSM detector 60-74 will be zero. Thus, due to the lack of photovoltaic effect characteristic of the MSM detector its photocurrent is also zero. Hence, when the CMOS device 82-96 is in the "off" state, the detector 60-74 response is zero independent of whether or not there is an optical signal incident thereon. This provides for a high degree of isolation between detectors which are in the "off" state and the "on" states. A photodetector can be switched to the "on" state by appropriately applying control signals to the control input lines 98-102 to the multiplexer 80. The combination of signals on these lines is decoded and used by the multiplexer 80 to provide the appropriate gate voltage on one of the CMOS devices 82-96 so that i will conduct and thereby connect the bias voltage V_ to the selected detector. In this embodiment, the desired photodetector is "addressed (or switched "on") by setting the gate voltage to zero.
Note that this circuit provides for superior low noise performance than earlier known architectures where the detection element is switched to ground potential when in the "off" state. In this latter situation, a small photocurrent can be generated by the photodetector since such a current will generate a small voltage. Accordingly, a small signal can appear at the output of the receiver which can interfere with the photocurrent generated by the photodetectors in the "on" state. Other noise sources are minimized in this configuration. The effects of thermal resistance noise are minimized by the high resistances in each of the "off" state detectors. The total capacitance of the detector array is lessened by the series combination of the detector and CMOS switch capacitances. Lastly, the effect of light modulated conductances on the diodes has a minimal effect on input noise because of the series conductance of the CMOS switch.
The developed photocurrent from the addressed detector is coupled to node 76 and fed to the input of a preamplifier 104 which serves as a current to voltage converter for generating a voltage output signal as a function of the electrical current from one of the photodetectors. Its output is coupled to an amplifier circuit 106 which serves to boost the preamplified voltage to a level suitable for use by the optical repeaters which are coupled to the output lines 78(a-h) in a conventional manner. It is important to note that only one amplification network 18 is required for each row in the crossbar network 50. Accordingly, the number of components has been kept to a minimum, especially as compared to the prior art device shown in FIG. 1 which requires a preamplifier for every detector.
The various advantages of the present invention will become apparent to those skilled in the art after having the benefit of reading the specification, studying the drawings and by reference to the following claims.
标题 | 发布/更新时间 | 阅读量 |
---|---|---|
一种交叉开关结构阻变存储器的优化方法 | 2020-05-12 | 462 |
一种驾驶模拟器方向盘力反馈系统 | 2020-05-17 | 867 |
数据处理器 | 2020-05-16 | 540 |
光电开关架构 | 2020-05-15 | 234 |
光电开关架构 | 2020-05-12 | 61 |
片上系统 | 2020-05-11 | 576 |
一种双千兆电口光纤收发器及应用其的通信系统 | 2020-05-15 | 532 |
COMPACT COGNITIVE SYNAPTIC COMPUTING CIRCUITS | 2020-05-18 | 920 |
비 휘발성 메모리 크로스바 어레이 | 2020-05-19 | 514 |
음의 값을 포함하는 벡터-매트릭스 곱셈 | 2020-05-15 | 638 |
高效检索全球专利专利汇是专利免费检索,专利查询,专利分析-国家发明专利查询检索分析平台,是提供专利分析,专利查询,专利检索等数据服务功能的知识产权数据服务商。
我们的产品包含105个国家的1.26亿组数据,免费查、免费专利分析。
专利汇分析报告产品可以对行业情报数据进行梳理分析,涉及维度包括行业专利基本状况分析、地域分析、技术分析、发明人分析、申请人分析、专利权人分析、失效分析、核心专利分析、法律分析、研发重点分析、企业专利处境分析、技术处境分析、专利寿命分析、企业定位分析、引证分析等超过60个分析角度,系统通过AI智能系统对图表进行解读,只需1分钟,一键生成行业专利分析报告。