WO1989004567A1

WO1989004567A1 - Broadband local area network

Info

Publication number: WO1989004567A1
Application number: PCT/US1988/003200
Authority: WO
Inventors: Joseph Paul Preschutti
Original assignee: Amp Incorporated; Broadband Networks, Inc.
Priority date: 1987-11-02
Filing date: 1988-09-19
Publication date: 1989-05-18
Also published as: EP0339078A1; CA1300239C; KR890702348A; JPH02502061A; US4970722A

Abstract

An easy to install and easy to expand broadband local area network (100) is disclosed. The network has a star-type architecture based upon a fixed-gain, fixed-loss approach to amplification and a star-feeder approach to providing user connection to the network.

Description

-1-

BROADBAND LOCAL AREA NETWORK The present invention relates to a local area network (LAN) and, more particularly, to a broadband local area network using coaxial cable which is easy to design, easy to install, and easy to expand.

The term broadband refers generically to any wideband transmission medium. Broadband coaxial communications systems, as the name implies, transport ^• multiple information signals using coaxial cable. In use for video for over thirt years, broadband coax has only recently found widespread use for the carriage of voice and data communications. The large bandwidth available allows large numbers of services to be carried on a single cable. An understanding of broadband is most easily obtained by contrasting it with baseband communications. In a baseband system, the information signal (i.e. voice, data or video) may be connected directly^* to the cable. In this manner the service has access to the full transmission channel, but only one service may utilize the cable at any given time. Broadband utilizes a technique similar to that employed by all radio and television broadcasters, i.e. RF^' (radio-frequency) modulation. The baseband information is first modulated onto an RF carrier before being applied to the cable. By using many different carriers, multiple signals can be handled on the same transmission path.

The transportation medium in a broadband system is coaxial cable. A typical broadband coax system supports carrier frequencies from 5 MHz to 450 MHz. Many services may. use the same cable simultaneously by operating each on a different carrier.

As with any transmission medium, the RF signals travelling along the cable are attenuated, or lose their power. Unless the system is relatively short, RF amplifiers are necessary to boost the signals back to their original levels. Broadband amplifiers used in LANs will oftentimes be referred to as bidirectional amplifiers. To provide two-way communications on a broadband network, the available bandwidth of the cable is split in two. The two resulting bandwidths are referred to as the outbound and inbound bandwidths. These two bandwidths carry the information signals in opposite directions. A typical midsplit broadband amplifier, for instance, passes 1-120 MHz in one direction (inbound) and 150-450 MHz in the other (outbound) . This provides full-duplex communications on a single cable.

Broadband communications technology similar to that developed for the Cable Television (CATV) industry is often used in local area networks. Conventional broadband coaxial cable local area networks are generally arranged in what is known as a "tree-and-branch" configuration utilizing equipment and technologies developed for CATV systems. Such a network has a headend, which illustratively has a central computer facility. A signal to be distributed to locations throughout the facility leaves the central computer facility on a "trunk" cable. The "trunk" cable characteristically has no user devices connected, and is intended to transport the signals carried on the system to _a general area where "users" are to be located. Most systems require "trunk amplifiers to amplify the RF signals carried on the "trunk" cable at various locations. In addition, the trunk cable can be split over two or more paths, "branching" out to several general areas where users are located.

User devices are connected to the system via a "tapped feeder" cable. This tapped feeder portion of the network emanates from "bridging" amplifiers that isolate the "trunk" from the "tapped feeder" portion of the system. In some cases "terminating trunk" or "terminating bridger" amplifiers are used to delineate the point at which a "tapped feeder" portion of the system begins and the "trunk" portion of the system ends. Tapped feeder cables are generally arranged as a serial connection of 5 several relatively short (compared to the length of trunk cable) lengths of coaxial cable and tap units. In addition, the tapped feeder cable can be split into two or more cables, providing several portions of serially connected taps and cables. Each user device such as a , ^' 10 ^■ computer terminal is connected to a tap by means of a drop cable of limited length.

Problems with such conventional CATV equipment and architecture in, broadband local area networks include a high degree of complexity in network design and component

15 specification, inflexibility for growth and rearrangement which causes network costs to be artificially high, relatively reduced reliability and fault correction time due to the relatively high number of serially connected devices, and relatively higher installation and ⁰ maintenance costs associated with the nature and complexity of conventional CATV hardware used to implement these networks.

Accordingly, it is the object of the invention to provide a LAN which is easier to design, install, ^{5 •} maintain, rearrange and expand than prior art LANs as well as being of lower cost than prior art LANs.

The present invention is a local area network which •. is simple to. design, install, maintain, rearrange and expand. The simplicity is achieved by a combination of

³⁰ elements, which taken together, provide a "tinker-toy" or "building block" arrangement which, as a package, can be configured to form a LAN which meets certain required system specifications, such as IEEE 802.7 BROADBAND LOCAL AREA NETWORK RECOMMENDED PRACTICES and General Motors

³⁵ Manufacturing Automation Protocol (MAP) . The LAN makes use of an inventive means of providing for R signal amplification and RF signal balancing identified herein as the "fixed-gain, fixed-loss" technique, which can be used to simplify design, installation and maintenance of the inventive system's counterpart to the conventional prior art trunk portion of the LAN system. The inventive system also includes an inventive method of providing user outlet connections, identified herein am the "star-feeder" technique, which can .meet required standards of system performance and simplify design, installation and maintenance of the inventive system's counterpart to the conventional prior art "tapped-feeder" portion of the system. The inventive LAN also includes an inventive arrangement for interfacing the "fixed-gain, fixed-loss" technique of broadband RF amplification with the inventive "star-feeder" technique.

FIG l schematically illustrates a LA in accordance with an illustrative embodiment of the present invention. FIG 2 schematically illustrates the "fixed-gain, fixed-loss" bidirectional broadband amplification system concept.

FIG 3 schematically illustrates a node suitable for use in the network of FIG 2.

FIG 4A, 4B and 4C schematically illustrate remote outlet clusters in accordance with the present invention. FIG 5 schematically illustrates a variable cable simulator used in the line balancer.

FIG 6 schematically illustrates a hub for use in the network of FIG 2. 1. Overview of the Inventive Local Area Network

Turning to FIG 1, the local area network of the present invention is schematically illustrated. The local area network 100 of FIG 1 permits bidirectional communication among a plurality of network user devices such as personal computers, video devices, automated manufacturing devices and other devices. The LAN 100 also enables bidirectional communication between the LAN user devices and external devices, networks, and data bases.

The network 100 of FIG 1 comprises a central hub 102. Illustratively, the hub includes a plurality of inputs 104 which are designed for applicable standards as a network ^■interface for connecting devices to the network. Examples of such standards are the IEEE 802.7 and the General Motors. MAP specification. • Devices 106 which may be connected to the network include for example translators and remodulators for systems supplied by Allen-Bradley, Ungerman-Bass and Concord Data as well as other networks such as the Chipco Ethernet Network and the IBM PC Net. The hub also includes a plurality of output ports 108. τh_e output ports operate at RF signal power levels required to implement the fixed-gain, fixed-loss approach to broadband system design incorporated in the inventive network.

At least some of the output ports 108 are connected to the nodes 110. Illustratively, the hub 102 and each node 110 may be separated by up to 20 decibels of cable loss at 450 MHz. A node 110 may include a line balancer 111 which provides the proper amount of attenuation to achieve the fixed-gain, fixed-loss approach to broadband system design incorporated in the inventive network. Through use of the line balancer, system design is simplified to the point of making a single calculation: the decibel loss of the coaxial cable at the highest outbound frequency, illustratively 450 MHz, and then specifying the proper line balancer for that particular location. This compares to four calculations and selection or adjustment of up to eight components in the conventional local area network.

A node 110- may include two node expansion ports 109 to drive two other nodes. Thus, for example, the node designated 110' in FIG 1 is connected to the node 112. The node 112 is connected to the node 114, which in turn is connected to the node 116. All node-to-node connections are provided by the node expansion ports. Illustratively, the node-to-node separation may be 10 decibels of cable loss at 450 MHz.

In addition, a node may have two output slots, each of which can have-inserted one of tw© different types of oμtput devices, either a drop panel 118 or an .output array 119. A node may drive a plurality of direct connections to user devices through use of the drop panel 118 which may have a plurality of outputs designed for applicable standards as a network interface for connecting devices to the network, for instance IEEE 802.7 or General Motors MAP specification. Illustratively, each direct drop connection from a node to a user connection device may be up to 120 feet of RG-6 type coaxial cable. This compares to typical allowable drop lengths of 50 feet of RG-6 coaxial cable for prior art networks. A node may drive a plurality of remote outlet clusters 120. The remote outlet clusters are connected to the output array 119. The remote outlet clusters may drive a plurality of direct connections to user devices 122 with said applicable standards applying. Illustratively, each direct drop connection from a remote outlet cluster 120 to a user connection device 122 may be up to 75 feet of RG-6 type coaxial cable. This compares to typical allowable drop lengths of 50 feet of RG-6 coaxial cable for prior art networks. Depending upon the distance between the node and the particular remote outlet cluster 120, the remote outlet cluster 120 may be a 2, 4, 6 or 8 port device. Illustratively, the remote outlet cluster and the node are connected by RG-11 cable. If the separation between the node and the remote outlet cluster is up to 75 feet of RG-11 cable, the 8-port device is used. If the separation is between 75 and 140 feet, the 6-port device is used. If the separation is between 140 and 210 feet, the 4-port device is used. If the separation is between 210 and 300 feet, the 2-port device is used. 2. Fixed-Gain, Fixed-Loss Amplification

Broadband. systems used in both CATV and LANs utilize bidirectional amplifiers which are based upon a system concept which solves the problem of. variable loss between amplification locations by providing a variable gain amplifier. The means by which the variable gain amplifier is realized is by inclusion o.f several plug-in and hard-wired circuits which accomplish the task of tailoring the gain versus frequency characteristic of the amplifier to counterbalance the variable loss versus frequency characteristic of the transmission path between amplifiers. . This variable gain amplifier concept lends itself well to the CATV environment, where the system size usually is sufficient to allow for a staff of well trained technicians to maintain the system, and wherein there is little or no need for two-way communications equipment, since most CATV systems operate broadband in a transmit mode only. However, when this technology and system. concept is applied to local area networks, which are usually significantly smaller in size, and which are universally operating in a two-way mode with significant bandwidth used for outbound as. well as inbound signal transmission, the prior CATV art causes problems.

To overcome these problems, an inventive^' means of providing two-way broadband amplification for local area networks is shown in FIG 2. The means for providing two-way broadband amplification is designated herein as the "fixed-gain, fixed-loss approach". The bidirectional amplifying unit 200 comprises two sections, the line balancer 202, and amplification module 203. The combined loss of the variable cable length 204 with the gain of the bidirectional amplifying unit 200 is illustratively designed to provide a unity gain simultaneously in both inbound and outbound directions. That is, outbound signals transmitted into the system at input 205 of the variable length of coaxial cable appear at the output 206 of the bidirectional- amplifier 203 at the same signal strength while inbound signals transmitted into the system at output 206 of the bidirectional amplifier 203 appear at the input 205 of the variable length of coaxial cable 204 at the same signal strength. Further, the decibel gain of the bidirectional amplification module 203 when measured between input port 207 and output port 206 is equal in both outbound and inbound directions.

The line balancer 202 comprises two separate . sections, a fixed cable equalizer 208 and a variable cable simulator 209. The variable cable simulator 209 is a continuously adjustable circuit which can simulate the decibel attenuation characteristics of a variable length of cable, illustratively shown to be 0 to 5 decibels at 450 MHz with Cable attenuation characteristics simulated between the frequencies 5 to 450 MHz. The fixed cable equalizer is a fixed attenuation circuit which provides an attenuation characteristic which varies with frequency in a manner which, when combined with a specified length of coaxial cable, will provide a combined flat attenuation characteristic across a specified bandpass, in this case 5 to 450 MHz.

The fixed-gain, fixed-loss amplification system for broadband and local area network operates as follows: Illustratively, the gain when measured between port 207 and port 206 of the bidirectional amplifier 203 in FIG 2 is a fixed value such as 25 decibels for both inbound and outbound signals. This represents the "fixed-gain" portion of the system. The "fixed-loss" portion of the system includes the combined attenuation of the length of σoaxial cable 204 and the line balancer 202, the sum total of which should equal the fixed gain of the amplifier 203 (i.e. 25 decibels) for all frequencies between 5 and 450 MHz. Illustratively, the minimum attenuation of the line balancer is 5.0.dB at all frequencies between 5 and 450 MHz when the. variable cable simulator circuit is adjusted to the minimum loss position. Thus, the maximum allowable decibel length of cable is 20 db at 450 MHz. Illustratively, the variable cable simulator is capable^' of. simulating the loss characteristics of up to five decibels of coaxial cable at 450 MHz at all frequencies between 5 and 450 MHz. Thus, in order to provide for system designs with potential cable lengths of 0 to 20 dB at 450 MHz between amplifier locations, it is desirable to provide four different line balancers, shown as a list of equipment blow.

Each such line balancer includes a different equalizer 208 capable of providing a different amount of fixed attenuation. The operation of the line balancer 202 of FIG 6 is now discussed in more detail. Illustratively, assume that the variable cable simulator circuit 209 has a flat attenuation characteristic of 3-..0 db for all frequencies between 5 and 450 MHz when adjusted to the "zero dB cable simulation" position for any line balancer. For the LBE-5 listed above, then the attenuation characteristic of the cable equalizer circuit would be 17.0 dB at 450 MHz and attenuation characteristics at frequencies below 450 MHz are such that the combined attenuation of a 5 db length of coax at 450 MHz plus the cable equalizer circuit would be equal to 22.0 dB at any frequency between 5 and 450 MHz. Therefore, if a five dB cable length at 450 MHz is required by the system design and an LBE-5 is used with the adjustment of the variable cable simulator set to the "zero dB' position, the loss of the cable plus the loss of the LBE-5 is exactly 25 dB across the entire frequency range of 5 to 450 MHz. Now, assume that a system designer must use a 3 dB length of coaxial cable at 450 MHz between bidirectional amplifiers instead of the aforementioned 5 dB length of coaxial cable. By selecting the LBE-5 (selecting one component) and adjusting the variable simulator circuit control to simulate 2.0 dB of coaxial cable (adjusting one component) the loss of the signal path between cable input 205 of FIG 2 and bidirectional amplifier port 207 has again been set to exactly 25 dB for all frequencies between 5 and 450 MHz, because the variable cable simulator adds attenuation to the circuit path in a manner identical to coaxial cable attenuation versus frequency characteristics. Thus, the task of providing a unity gain system for both outbound and inbound signal paths has been accomplished in both directions simultaneously by selecting a line balancer according to the above table. The following table illustrates the losses of the cable equalizers for the set of line balancers in this example. Cable Loss Line Balancer Cable EQ loss Total LBE loss § 450 MHz dB @ 450 MHz db dB @ 450 MHz with variable simulator § Zero 5 LBE-5 17.0 20.00

10 LBE-10 12.0 15.0

15 LBE-15 7.0 10.0

20 LBE-20 2.0 5.0

The inventive means of implementing two-way broadband amplification through utilizing the "fixed-gain, fixed-loss" technique for local area networks is simple and easy, to design, install and maintain. Simplicity of design is achieved because the system designer needs to have knowledge of only one parameter, e.g. the coaxial cable loss at 450 MHz. Then, for a given system, this one calculation can be used to select the proper line balancer. The reason for this is that cable losses are maximal at the highest frequency, i.e. 450 MHz. Thus, if sufficient amplification is provided to compensate losses at this frequency, then the system will perform satisfactorily throughout the desired frequency range. Simplicity of installation results due to the need to install only one line balancer per bidirectional amplifier and adjust only one component, the variable simulator circuit.

Simplicity in maintenance is due to the separation of the line balancer from the bidirectional amplifier. The line balancer circuit is generally a passive circuit and has a much greater reliability than the bidirectional amplifier, which consumes power, is more complex, and has active semiconductor devices which are more prone to failure than the passive circuits used to implement the lins balancer. Thus, the most probable fault in a station is the failure of the bidirectional amplifier, which, for the purposes of this inventive method, is a module separate from the line balancer. Thus, replacement of failed bidirectional amplifiers does not require readjustment of the line balancer. This compares to adjustment of components required in replacement of the variable gain amplifiers used to implement prior art systems. Furthermore, the line balancer described in this inventive approach can be used with any type of broadband system, regardless of the split of the bands. That is, the inherent nature of the line balancer concept is to provide an accurate and complete adjustment across the entire frequency band of interest. The split of the bandpass is then only a function of the type of bidirectional amplifier used, thereby dramatically reducing the number of line balancers needed by the equipment manufacturer in order to provide a line of equipment used in sub-split, mid-split, high-split or dual cable systems. The same line balancer would be used whether the system was sub-split, mid-split, high-split or dual cable. This simplifies the manufacture, inventory and distribution of the product line for the equipment supplier.

3. Star Feeder Concept For User Connections

A purpose of the invention network 100 of FIG 1 is to provide for ease of design, installation and use of a broadband LAN. Instead of providing a serially connected succession of short cable lengths and tap units that provide user connection ports, and instead of requiring four different calculations of path loss to determine if the network will meet specified parameters, which causes difficulty l design, the inventive network is based upon a design principle which requires that only knowledge of the path loss at the highest frequency needs to be determined. This, in turn, directly translates into a knowledge of a maximum length of cable, given a cable type. In turn, the system design parameters for the user connections can then be limited to a trivial calculation of distance from a particular user device location to the user connection port.

The '^Star-Feeder" concept disclosed here is based upon a design concept which, allows^' the only variable in system design to- be the length of coaxial cable being used. Assuming that other components in the broadband network are properly designed, if the path loss variations inherent in the design of the network are due only to coaxial cable variations, then the maximum variation in path loss, which will be experienced by any signal on the system will be at the highest design frequency of the system. It follows that, given the above assumptions and conditions, that if the path loss is within specification at the highest frequency, then path loss variations at any lower frequency will be smaller in magnitude and do not have to be calculated.

The "star-feeder" concept disclosed here is based upon two different types of user connection devices. The first is ^' a direct user drop panel connection device located in the node (118 of FIG 1) , and a second is a device capable of providing a plurality of user connection outlets at a location remote from the note (120 of FIG 1) . This latter device has been designated herein the "Remote Outlet Cluster". - For the case of the direct user drop panel located at the node, an attempt is made, to take full advantage of the allowable variation in path loss design value to provide as long a drop connection cable as possible. For the case of the IEEE 802.7 or MAP system specifications, for instance, this would indicate that the design value for the outbound path loss at 450 MHz be specified as 3 dB below the nominal system design value for signals transmitted from the hub to the user device receiver, while the design value for the inbound path loss might also be 3 dB below the nominal system design value for signals transmitted by the user devices. ^"Therefore, a direct connection with a nominally short length"of cable would result in system performance requirements for path loss being achieved. The other limit of cable length is where the- loss of coaxial cable used for the drop connection is at the maximum permissible value for the particular system path loss specification. For example, the IEEE 802.7 or MAP system specifications permit the path loss to be a maximum of 3 dB above nominal values. ι_n this case, the allowable cable drop connection might be 6 dB at 450 MHz. . That is, the drop panel path loss added to the 6 dB loss of the coaxial drop connection would result in a path loss 3 dB above nominal for outbound signals at 450 MHz. However, since the cable loss is less than 6 dB for all frequencies below 450 MHz, the path loss must then be within required specifications.

In the LAN of the present invention, a direct drop panel (e.g. the 118 of FIG 1) located at a node is capable of providing a maximal variation i -the allowable drop connection cable length. Thus, for example, a node might have a direct drop .panel which comprises a conventional 16-port splitter within a system which provides- for the path loss at the highest frequency of the system to be 3 dB below nominal allowable path loss and at which the path loss at the lowest frequency of the system is 3 dB below nominal allowable path loss. In this manner,, for example, sixteen user devices in a MAP or IEEE 802.7 system can be serviced using RG-6 type cable within a radius of 120 feet of a node location. This compares with the 50 foot service radius of conventional local area networks. However, a system which relies totally upon the use of direct drop connections from a node which also contains a bidirectional amplifier would require the LAN utilization of a large number of nodes, which would result in a LAN design that would be more costly than the conventional LAN. - It is desirable in the inventive - network to have a means for providing a plurality of user connections at locations remote from the node locations. However, it is also desirable to maintain the simplicity of design inherent in the inventive LAN.

Therefore, the network of FIG 1 indicates the use of remote outlet cluster devices 120 which are intended to provide, a plurality of user connection outlets at a location remote from the node. Use of such remote outlet clusters' provide significant advantages. Configuring a system' which uses only the loss of the coaxial cable as a path loss variable results in an easy to design system. Further, such devices can be deployed in a manner which does not depend upon other such devices, such as in a star configuration. Thus, the remote outlet cluster devices enable a LAN which is easy to design, install, maintain, expand and rearrange.

In order to accomplish the results stated above, the inventive network utilizes remotely located devices which are identified as remote outlet clusters 120 in FIG 1. Since a ser connection at a remote outlet cluster comprises two separate lengths of coaxial cable, one to connect the remote outlet cluster to the node and another, the "drop connection" to connect the user"device to he Remote Outlet Cluster, the inventive system shares the allowable variation in cable loss equally between these : two cable sections. Illustratively, the connection between the node and the Remote Outlet Cluster is allowed to. vary 3 dB and the drop connection is allowed to vary 3 dB. In this manner, up to 75 feet of direct drop coaxial cable is allowed from a remote outlet cluster to a user device.

In order to optimize the number of user outlets and also provide for considerable flexibility in their 5 deployment, a plurality of different remote outlet clusters may be utilized.

The first is an 8-port Remote Outlet Cluster (see FIG 4A) which comprises a conventional 8-port splitter circuit 210 and a 1 dB fixed attenuator 211. The conventional

10 8-port ^' splitter ideally has an attenuation of approximately 9 dB, therefore, the 8-port Remote Outlet Cluster will have 10 dB of path loss for both inbound and outbound signal paths. In addition, this 8-port Remote Outlet Cluster would be allowed to be connected to a

¹⁵ signal path with art allowable variation in coaxial cable loss of 6 dB at 450 MHz. Thus, a node might comprise a plurality of outlets capable of driving remote outlet clusters, said node outlets having a path loss of 13 dB below the. nominal specified path loss of the system in

²⁰ both inbound and outbound signal paths. Therefore, the location of the 8-port user outlet splitter might vary from 0 to 3 dB of coaxial cable from the node and have a drop connection which might vary from 0 to 3 dB. The path loss of a user device _.connected- to the system woul then

²⁵. be a minimum of 3 dB below nominal for the case of a very short length of cable being used for both the connection of the remote outlet cluster to the node and also for the user device to the remote outlet cluster, while the maximu path loss would be 3 dB above nominal for the case

³⁰ of both cable lengths being at the maximum 3 dB permissible value at 450 MHz. Both inbound and outbound path losses for all other frequencies below 450 MHz would then be less than the required maximum values.

The second device is a 4-port remote outlet cluster

-^{> ~J} (see FIG 4B) which comprises a conventional 4-port splitter 212 and a circuit which equalizes coaxial cable 213. The conventional 4-port splitter ideally has an attenuation of approximately 6 dB, while the coaxial cable equalizer circuit nominally has an attenuation of 1 dB at 5 ^• 450 MHz. Since it would be desirable for all -remote outlet clusters to be connected to similar outlets at the node, and since in the- paragraph above the node has been defined as comprising a plurality of outlets having a path loss of 13 dB below the nominal specified path .loss of the • 10 system in both inbound and outbound signal paths for driving remote- outlet clusters, and since it would be desirable for the allowable drop connections from all remote outlet clusters in the system to have similar specifications, it follows that the application of the ■ 15 4-port remote outlet cluster is for locations which are a. ^■ minimum of 3 dB of coaxial cable loss from the node and for a maximum of 6 dB of coaxial cable loss from the node. It also follows that the attenuation characteristics of the coaxial cable equalizer in the 4-port remote outlet 20 cluster should be such that when the loss of 3 dB of coaxial cable at 450 MHz is added to the loss of the cable equalizer, that the resultant combined losses of both should be 4 dB at all frequencies from 5 to 450 MHz.

The third device is a 2-port remote outlet cluster 2⁵ (see FIG 4C) which comprises a conventional 2-port splitter^' 214 and a circuit 215 which equalizes coaxial cable. The conventional 2-port splitter ideally has an . attenuation of approximately 3 dB, while the coaxial cable equalizer circuit nominally has an attenuation of 1 dB at ³⁰ 450 MHz. Since it would be desirable for all remote outlet, clusters to be connected to similar outlets at the node,- and since in the paragraph above the node has been defined as comprising a plurality of outlets having a path loss 13 dB below the nominal specified path loss of the 5 system in both inbound and outbound signal paths for driving remote outlet clusters, and since it would be desirable for the allowable drop connections from all remote outlet clusters in the system to have similar specifications, it follows that the application of the 5 2-port remote outlet cluster is for locations which are a minimum of 6 dB of coaxial cable loss from the node and for a maximum of 9 dB of coaxial cable loss from the node. It also follows that the attenuation characteristics of the coaxial cable equalizer in the 2-port remote outlet ■10 cluster should be such that when the loss of 6 dB of coaxial cable at 450 MHz is added to the loss of the cable equalizer, ^' that the resultant combined losses of both should be 7 dB at all frequencies from 5 to 450 MHz.

In summation, a "star feeder" system has been defined

15 whic provides for direct drop connections from a node in the broadband star network and which also provides for a plurality of remotely located outlet clusters connected to the node. The inventive system design concept of the "star feeder" and components defined to implement the

20 system provides a means which allows a very simple system design, depending only upon a trivial calculation of distance between devices in the system, which allows ease of installation. Illustratively, only three "taplike" devices are required to implement the system and only

25 ^• areas which have known users might be initially installed. The system provides for growth, by providing a node which is capable of a plurality of direct user connections and a plurality of user outlet splitters to be connected. 4. Description of the Network Nodes

³⁰ One of -the- nodes 110 of FIG 1 is shown in greater detail in FIG 3. Illustratively, the nodes 110 of FIG 1 is connected directly to the hub 102 via lines 130. The node 110 includes a line balancer 202 (see FIG 2) which can balance a variable -amount of cable loss

³5 depending upon how far the node is located, from the hub (or from another node) and a "fixed-gain" bidirectional amplifier module 203 (see FIG 2) which is used to implement the "fixed-gain, fixed-loss" approach to the broadband system design. Since the network of FIG 1 is bidirectional, the bandwidth of the network is divided into two bands. One band is used for outbound (i.e. away from the hub) transmission and the other band is used for inbound (i.e. towards the hub) transmission. Thus, within the bidirectional amplifier 203, the outbound transmission path includes the high-pass filters 235, 236 and amplifier 237. The inbound transmission path includes low pass filters 238, 239 and amplifier 240.

The node also contains a user output array area 250 which can accept two types of panels intended to implement th "star-feeder" concept. One of each type is shown illustratively as a 16-port drop panel 251 (which is a conventional 16-way splitter plus a fixed attenuator) and an 8-port output panel 252 (which is a conventional 8-way splitter) for providing connections to remote outlet clusters (see FIG 6) . For example, a node might have two 16-port drop panels 251 installed in the user output array area 250, used exclusively to connect direct drop cables to user devices, providing up to 32 direct connections, or a node might have two 8-port drop panels 251 installed in the user output area 250, used exclusively to connect to remote outlet clusters or a combination of one of each type of panel.

The node also contains a node expansion port 253 which is a conventional 2-port splitter that is used to connect other nodes to the system by expanding upon the node already installed.

FIG 2 shows the line balancer 202 in greater detail. As discussed above, the line balancer 202 enables a node to be located a variable distance from the hub or another node. The line balancer 202 comprises fixed cable loss equalizer 208 and variable cable simulator 209. The variable cable simulator simulates the load of an adjustable amount of cable. This adjustment is performed by adjusting the position of a single control on a -circuit .element which comprises a dual-ganged potention eter. The loss of the fixed cable equalizer provides a predetermined loss which enables the combined loss of a selected amount of cable plus the loss of the cable equalizer to equal a fixed amount of flat attenuation across the entire frequency band of interes of the system. The sum total of the loss of the cable between hub and node plus the loss of the line balancer will be equivalent to a flat attenuation across the frequency and of interest. The cable simulator, which illustratively comprises a bridge T variable attenuator is shown in more detail in FIG 5.

The variable cable simulator is a variable bridge-tee circuit shown in FIG 5. R7 and R8 are the tee resistors and are 75 ohms. Rl and R2 form a ganged variable resistor pair that combined provide proper resistance so that the input and output driving point impedances are 75 ohms at any setting of the ganged resistor pair. Cl and R6 in conjunction with R9 and L3 limit the range of potentionmeter. R3 and R10 in conjunction also limit the range of the potentionmeter. R5, LI, C2 and Rll provides proper high frequency (above 150 MHz) responses for the simulator. R4, L2 and C3 provide proper operation at lower frequencies (5 to 150 MHz) .

In order for the node to be attached to other nodes in the system, a node expansion port 253 of FIG 3 is provided. In order for the fixed-gain, fixed-loss approach to be implemented, it is necessary for a unity gain to be established in the signal path between the output 255 of the amplification module 203 of the node 110 of FIG 3 and the corresponding point in the adjacent node attached by way of the node expansion port 253. Since it is desirable to provide only one bidirectional amplifier in the node to provide amplification of signals for ^" connection to both the user connection portion of the system and also to feed other nodes in the system, most of the power being delivered by the amplification module 203 in the outbound direction, for instance, is being supplied to the user connections via the directional coupler 260 which forms part of the RF balancer 254 of FIG 3. Thus, the path loss for implementing the fixed-gain, fixed-loss concept in the node system is determined by the allowable path loss between node-to-node connections via the signal path from RF output balancer input 255 to the output of node expansion port 253 to the input of the line balancer 232 of the next node. Thus, the gain specification of the amplification module 203 is equal to the sum of the loss of the line balancer plus the loss of the path from output 255 to node .expansion port 153 plus the allowable spacing between nodes. This gain relationship determines the unity gain relationship in the fixed-gain, fixed-loss system for node-to-node connections. In order to implement the system it is necessary to provide line balancers which will produce a fixed, flat attenuation in this signal path for all frequencies. For example, if the path loss for the node connection path including directional coupler 260 and node expansion port 253 is 12 dB and the line balancer loss is 5 dB at the zero adjust position and the spacing between the node-to-node connections is 10 dB, then the gain of amplification module 203 must be equal to the sum of these losses, or 27 dB.

The hub to node connection has a different relationship which determines the gain of the bidirectional amplifier in the hub, as well as determines the specification for the line balancer used in nodes connected to the hub. In order to implement the fixed-gain, fixed-loss concept at the hub, it is necessary to provide a flat attenuation characteristic (not necessarily restricted to the unity gain required for the node-to-node path or equal to the node-to-node path loss) between the output array -286 of the hub shown in FIG 6 and a directly connected node. For example, it might be desirable to have 20 dB hub-to-node spacing. - Assuming a line balancer minimum loss of 5 dB, this in turn requires that the path loss between the hub output array 286 of FIG 6 and output of the line balancer in the attached node be 25 dB flat attenuation across the band of interest and also that the operational outbound RF signal levels at the outp t array of the hub be at levels lOdB above those at the output of the node expansion port 153, as well as requiring that the operational inbound RF signal levels at the output array of the hub be at RF signal levels which are 10 dB below those of the node expansion port 153. Thus, the implementation of the fixed-gain, fixed-loss approach to bidirectional amplification in the broadband.network establishes firm and fixed relationships between the specifications of all of the devices required to perform the implementation in a manner not found in conventional prior art systems. Implementation of a specific set of required network parameters, such as IEEE 802.7 or MAP requires that proper gain and operational signal levels are established simultaneously in an integrated manner for all devices used in the network.

The node also contains an RF output balancer 254 which provides for not only splitting and recombining of RF signals for both inbound and outbound transmission paths, but also contains circuitry to balance the requirements of the network regarding standard path loss for the user connections implemented with the "star-feeder" section of the system with the system requirements for implementing the "fixed-gain, fixed-loss" system design approach.

Before describing the RF output balancer in more detail, the problem regarding the need for the RF output balancer will be briefly explained. The inventive network being described utilizes the fixed-gain, fixed-loss system concept in- order to simplify design of the equivalent of the "trunk" section of a conventional network. The inventive network being described also is accomplishing the task of providing a simple and easy implementation of user connections to the system via the "star-feeder" concept and implementation. The inventive network requires that this user connection portion of the system provide specific path loss specifications at the direct drop panel 251 and the 8-port 252 output array depending upon the specifications of the type of broadband network to be implemented. In the inventive network, it is necessary to add a new circuit, the..RF output balancer 254 to simultaneously be able to provide an optimized "fixed-gain, fixed-loss" amplification system and connect the amplification section of the system to the simple and easy to install "star-feeder" user outlet portion of the system which is also optimized for performance.

The RF output balancer circuit 254 of FIG 3 includes an input port 255, two user output array ports 256, and 257 and a node expansion port output 258. A conventional directional coupler 260 is connected to input port 255 which provides a lower path loss to the user connection portion of the system and provides the rest of the power to output 258 which is used to provide a path for signals to be connected to other nodes in the system. Thus, the outputs of directional coupler 260 are connected to node expansion port 253 and to the input of a diplex filter, comprising high pass filter 261 and low pass filter 264. The outbound transmission path from the input of the άiplex filter comprises the high pass filters 261 and 262 and the conventional splitter 66= The outputs ©f splitter 266 are connected to .user output array ports 256 and 257. The inbound transmission path from user output "port 256 (and 257) includes the splitter 266, low pass filters 263- and 264, and attenuator 265. In this, manner, the inbound transmission path for the signals originating at the user connections are attenuated to provide for proper balancing of inbound signal levels generated by users located at nodes connected to the node expansion port 253. The value of attenuator 265 is predetermined and not user selectable.

The 16-port drop panel 251, in conjunction with the RF output balancer, is provided for direct drop connection^'s to user devices from the node as described in the "star-feeder" concept. In this implementation the 16-port drop panel comprises a fixed attenuator and a conventional 16-port splitter. This allows either an 8-port output panel or a 16-port drop panel to be located in a node by connecting to the RF output balancer.

It is the node structure which gives the network 100 of FIG 2 its great simplicity and flexibility. First, because of the use of the fixed-gain bidirectional amplifier, which requires no adjustments and can be replaced in the system without requiring that any adjustments be performed. In addition, the use of the line balancer concept, which simultaneously provides for balancing of the path loss between amplifiers in the system in both inbound and outbound directions with a single variable adjustment. In addition, the network may be expanded from a relatively small size by adding additional nodes or additional user outlet splitters to_. already existing nodes without affecting the path loss at existing user connections. In addition, the network is capable of implementing both the simplified bidirectional amplification system of the fixed-gain, fixed loss concept while simultaneously providing a "star-feeder" user . connection system through use of the RF output balancer, which is predetermined and not required to be adjusted by the system operator. In addition, the network is capable of providing a longer drop length from the 16-port drop panel than from conventional prior art networks. '5. Description of the Network Hub

The network hub 102 of FIG 1 is shown in greater detail in FIG 6. Illustratively, the hub includes an output array 286 comprising a conventional 8-port splitter having the outputs 108. Each output 108 of the output array 286 might be connected to a node via coaxial cable. The hub also illustratively includes an input array 280 comprising a conventional 16-port splitter. The inputs to the input array are designated 106 in FIGs 1 and 6. As was previously discussed, the input array is intended to provide a means for connecting user devices at the "headed" of the broadband transmission medium with the system connection specifications for the medium intended to meet required standards such as IEEE 802.7 and MAP. The number of hub user connections can be expanded from the original 16 user connections 106 to 64 user connections via lines 282, which are provided for attachment of three additional 16-port input arrays similar to input array 280. This is accomplished by providing the combination circuits comprising of conventional 2-port splitters 283.

The hub also includes a hub driver module 284, which is a fixed-gain bidirectional amplifier intended to perform the amplification function needed at the hub to implement the fixed-gain, fixed-loss approach to system ^' design. The fixed-gain bidirectional amplifier 284 located at the hub differs from the fixed-gain amplifiers 203 of FIGs 2 and 3 located at the nodes in that the node fixed-gain amplifiers must have equal gain in both inbound and outbound transmission directions in order to allow for the connection of other identical nodes to the system via the node expansion ports 253 of FIG 2. In order to implement the fixed-gain, fixed-loss concept at the hub it is necessary that the gains of the- outbound^'and inbound signal paths provided by the hub driver modular be selected independently to simultaneously provide the correct and proper^■ user connection transmit and receive levels at the input array 280 while providing correct and proper transmit and receive levels at the output array 286.

In this manner, that is by predetermining the exact amount of RF gain required by both inbound and outbound signal paths for the hub, which connects a plurality of user connection devices and provides for the amplification and distribution of these signals in a manner which exactly provides the proper levels for implementation of the fixed-gain, fixed-loss approach to broadband system design, it is possible to provide a hub for use in the inventive network which requires no adjustments by the user. This facilitates design, installation and use of the network.

The number of direct cond^uctions to nodes may be increased from eight to sixteen via line 290, which is derived from a conventional splitter 291. The method for adding extra outputs to the hub is to add an additional hub driver module and output array connected to line 290. 6. Conclusion An easy to install and easy to expand broadband local area network having a star-type architecture based upon a "fixed-gain, fixed-loss" approach to bidirectional broadband amplifications system and a "star-feeder" approach to providing user connections throughout the system has been disclosed. The embodiments of the invention include a hub, node, line balancer, user outlet splitters and the integration of the required parameters of these elements into a system. Finally, the above- described embodiments of the invention are intended to be illustrative only. Numerous alternative embodiments may be devised by those skilled in the art without departing from the spirit and scope of the following claims.

Claims

CLAIMS ;

1. A device (200) for balancing losses in a coaxial transmission path (204) of a network having inbound and outbound signal paths operating in inbound and outbound frequency ranges, respectively, said device comprising: a fixed-gain bidirectional amplifier (203) having equal gain in the inbound and outbound signal paths, a variable coaxial cable simulator circuit (209) for simulating the loss of a variable amount of coaxial cable over said inbound and outbound frequency ranges, and an equalizer circuit (208) for equalizing the loss of a fixed amount of coaxial cable, whereby the sum of the losses 'provided by the coaxial transmission path .(204) , the equalizer circuit (208) , and simulator circuit (209) can be adjusted to be equal to the gain of the fixed gain amplifier (203) across said inbound and outbound frequency ranges.

2. The device of claim 1 wherein said circuit means comprises a bridge T variable attenuator.

3. A device for balancing losses in a transmission path (204) of a network capable of transmitting signals in an inbound and an outbound direction, said device comprising: a coaxial cable equalizer circuit (208) , and a variable simulator circuit (209) for simulating the loss of a variable amount of coaxial cable, the combined loss of said transmission path, said equalizer circuit and simulator circuit being adjustable to provide a predetermined fixed loss across the inbound and outbound bandwidth of the network.

4. A broadband local area network (100) comprising: a central hub (102) a first plurality of nodes (110) connected to the hub by variable length coaxial transmission paths, each of said nodes (110) in said first plurality comprising a bidirectional amplifier (203) having a fixed gain across an inbound and outbound frequency range and a line balancer (202) including a variable cable simulator circuit (209) for simulating the loss of a variable amount of coaxial cable and an equalizer circuit (208) for equalizing the loss of a fixed amount of coaxial cable, and a plurality of remote outlet clusters (120), each remote outlet cluster being connected to one of said nodes (110) by a coaxial transmission path, each remote outlet cluster comprising a fixed attenuation means (211, 213, 215) and an RF splitter (210, 212, 214) having a plurality of outlets for connection to user devices (122), the number of outlets depending on the length of the coaxial transmission path between the remote outlet cluster and the node (110) to which it is connected.

5. The network of claim wherein said network comprises a second plurality of nodes (112) , each node in said second plurality being connected to a node of said first plurality by a coaxial transmission path and including a particular line balancer (202) depending on the length of said connection to a node (110) in said first plurality.

6. A combination for use in a local area network (100) for transporting inbound and outbound signals in inbound and outbound frequency ranges, said combination comprising: a node (110') having a plurality of outlets a plurality of remote outlet clusters (120) , each remote outlet cluster being connected to a node outlet by a first coaxial transmission path, each remote outlet cluster comprising a splitter (210, 212, 214) having a plurality of user outlets for the connection of user devices (122) , each user device (122) being connectable to a user ©utlet by a second coaxial transmission path, the combined attenuation of said first and second transmission paths associated with each user device being 5 in a range determined by an allowable path loss variation and an insertion loss of the splitter (210,212,214) comprising the corresponding remote outlet cluster (120) .

7. The combination of claim 7 where the combined attenuation-of said first and second transmission paths

10 associated with each user device (122) at the highest frequency of said inbound or outbound frequency ranges being between (-A+3)ZdB and ((-A+3)Z + X)dB where ZdB is an insertion loss of a splitter having two outlets at said highest frequency, XdB is a maximum allowable path loss at 15 said highest frequency, and 2 A is the number of user outlets of the corresponding remote outlet cluster where A is 1, 2 or 3.

8. The combination of claim 7 wherein each second transmission path has an identical predetermined length.

²⁰ 9. A node (110) for use in a local area network comprising: a variable cable simulator circuit (209) , a fixed cable equalizer circuit (208) , a fixed gain bidirectional amplifier (203) , and ⁵ ■ an rf splitter (260) for splitting an output

(255) of said amplifier into a first path for node-to-node connections and a second path for user connections, said second path including a first output circuit means (251) for enabling user devices to be directly connected to said

³.⁰ node and a second output circuit means (252) for enabling the attachment to said node of one or more remote outlet clusters (120) each of which enables a plurality of user devices (122) to be connected to said network at a location remote from said node (110) . 35 10. The node of claim 9 wherein said second path further includes diplex means (261,262,263,26-4) for dividing said second path into an inbound path and an outbound path, said inbound path including fixed RF attenuation means (265) .