JPH03296332A - Star coupler and optical communication network - Google Patents

Abstract

PURPOSE:To devise the system to be avoided of any modification for existing nodes or the like when extension of a node is required by adopting the configuration such that a signal transfer coefficient is selected to be zero between an input terminal and an output terminal in pairs to be connected to a same node. CONSTITUTION:Three directional couplers 2, 3, 4 are interconnected in a triangle shape to constitute one star coupler 1. The directional coupler 2 is formed by melting clad layers of two optical fibers to make the core layers of each optical fiber close to each other. Thus, the star coupler in which the diagonal components of a transfer characteristic matrix are all zero is formed. A combined number of nodes connected to the star couplers is increased/decreased by increasing/decreasing number of the star couplers. Thus, the expansion/ reduction of the optical communication network is facilitated.

Description

[Detailed description of the invention] [Industrial application field]

The present invention relates to star couplers that can be used in mutual connection, and to optical communication networks constructed using the same.

[Conventional technology]

Local area network (LAN)! ;! It is gradually becoming popular as a high-speed communication network between computers, workstations, etc. located relatively close to each other, and the most representative one is the one called "Ethernet" developed by Xerox. There is. FIG. 7 is a diagram showing an Ethernet network. In FIG. 7, 20 is a coaxial cable, 21 is a tap (branch point), and 22 is a node (terminal, station). Each node 22 is connected to the coaxial cable 20 at a tap 21. When the number of nodes 22 to be connected increases, a new tap 21 is provided and connections are made to it. In Ethernet, coaxial cables are used to configure a network, but with the advancement of optical fibers, attempts are being made to configure networks using optical fibers. With Ethernet, nodes can be added by installing taps one after another. However, if optical fiber is used instead of coaxial cable, it is difficult to configure a network similar to Ethernet because it is not possible to tap the cable. Therefore, a network has been proposed in which node transmission and reception are separated into separate terminals, and all nodes are distributed by star couplers (E, G. RAH5ON, IEEE TI?A11lSACrl
ONS ON C0MMυNICATIONS. VOL, C0M-26, NO, 7, JULY1978.
'Pibernet: Multimode 0pt
ical Fibers for Local Com
Puter Netw. FIG. 8 shows the above-mentioned proposed example of an optical communication network using a star coupler. In FIG. 8, 23 is an optical fiber, 22 is a node, 24 is a star coupler, and 25 is a terminal. Node The transmission line 22 is connected to the input side of the star coupler 24, and the reception line is connected to the output side. Fig. 9 shows a star coupler with the input side and the output side divided into two sides for easy understanding. Fig. 10 is a functional principle diagram of the star coupler 24.The 1! function of the star coupler is drawn by comparing it to an electric N path, and the input side of the star coupler is an OR circuit 24-1. The output of the OR circuit is equally distributed and output to each terminal on the output side. 25H is an input terminal into which signals from the nodes consisting of
25R is an output terminal paired with the input terminal (ie, an output terminal that emits a signal to the same node). When a signal is input to the input terminal 25H, the signal is divided equally into the number of output terminals and output from all the output terminals. For example, if the number of output terminals is 5, the input terminal 25H is
When a signal with a magnitude of "S" is input, a signal with a magnitude of S15 is output from each output terminal. The signal S15 is also sent back from the output terminal 25R to the node that sent the signal to the input terminal 25H. In other words, similar to Ethernet configured with coaxial cables,
This means that it has been possible to construct an optical communication network that has the property that a signal transmitted from one node is transmitted to all nodes (broadcast property). However, the star coupler described above simply divides the input signal into the number of output terminals and sends it out (such a star coupler is called a "passive star coupler"), so when the number of output terminals increases, the individual output signals It has the disadvantage that the size is small. In order to prevent the output signal from becoming small, a star coupler having amplification capability (such a star coupler is referred to as an "active star coupler") may be used as the star coupler. FIG. 11 shows an active star coupler. 1st! In the figure, 26 is an input terminal, 27 is a light emitting diode array,
28 is an amplifier, 29 is a waveform shaping circuit, 30 is a matrix circuit, 31 is a driver 32 is a light emitting diode array,
33 is an output terminal. An optical signal inputted from any of the input terminals 26 is converted into an electrical signal by a corresponding light emitting diode of the light emitting diode array 27. The electrical signal is amplified by an amplifier 28,
The waveform shaping circuit 29 shapes the signal so that it becomes clear whether it is high (hfgh) or low (low), and the signal is sent to the matrix circuit 30 . FIG. 12 shows details of the matrix circuit 30. This matrix is a square matrix in which the number of input terminals and the number of output terminals are the same. As understood from FIG. 12, when a signal enters from one input terminal, that signal is transmitted to all output terminals. Each output of the matrix circuit 30 is made into a desired size by a driver 31 and is manually inputted to a light emitting diode array 32.In the light emitting diode array 32, the electrical signal is converted back into an optical signal, and the output terminal 33 connects all nodes. can be conveyed to. Since the active star coupler has a built-in amplification function, the output signal will not become small even if there are many output terminals. FIG. 13 shows a star coupler in which corresponding input lines and output lines (that is, input lines and output lines connected to the same node) are placed close to each other. The star coupler 34 may be a passive star coupler or an active star coupler. C is a power line, D is an output line, and 5 and 6 are terminals. In the following description, when representing a star coupler, it will mainly be represented using such a diagram.

[Problem to be solved by the invention]

(Problem) However, in the optical communication network using star couplers, even if an attempt is made to add nodes to expand the network, the terminal pair (
There was a problem in that the number of pairs (input terminals and corresponding output terminals) could only be increased. (Description of Problems) (1) The number of terminal pairs included in a star coupler is determined, for example, 100 when a star coupler is manufactured. Therefore, by adding more nodes to 12
If you want to reduce the number to 0, use 120 star couplers.
The star coupler must be replaced with a star coupler having more than 200 terminal pairs (for example, 200 terminal pairs). In other words, we had to rebuild a larger network. This is not preferable because it requires a great deal of time and effort each time an attempt is made to connect nodes beyond the number of terminal pairs provided. (2) It appears that the number of connectable nodes can be increased by connecting star couplers one after another (that is, by making multiple connections) without replacing star couplers. However, this is not practical. Let me explain it. FIG. 14 is a diagram in which two star couplers are connected. A and B are star couplers, and 5 to 16 are terminals. The transmitting terminal 9 and the receiving terminal 10 forming one terminal pair of star coupler A are the receiving terminal 11 and the transmitting terminal forming one terminal pair of star coupler B. 12, respectively. Since star coupler A has 2 terminal pairs,
When star coupler A is used alone, the number of nodes that can be connected is 2. Similarly, when star coupler B is used alone, the number is n. However, if you connect star couplers A and B as shown in Figure 14,
By composing them, k+n-2 nodes can be connected. The reason why there are two fewer terminals than +n is that one pair of terminals on each side must be used to connect the star couplers. For example, if the star coupler A, B is a star coupler with Q00 terminal pairs, the composite star coupler can connect 198 nodes, creating a network that is almost twice as large as that of a single star coupler. can be configured. When it becomes necessary to increase the number of nodes, by connecting the next star coupler, it is possible to increase the number of nodes without having to reconnect the already connected nodes. However, when multiple connections are made as described above using conventional star couplers, the signal ``oscillates'' or ``ghosts'' occur, as will be explained in detail next, making it impractical. (3) Reason for causing oscillation The inventor devised to express the characteristics of the star coupler by a matrix, and by analyzing the determinant, the reason for causing oscillation will be explained. (3-1) Matrix representation of star coupler characteristics First, 1
A matrix representation of the characteristics of star couplers is shown. Assume that the star coupler 34 in FIG. 13 has n terminal pairs,
The input and output of each terminal pair are (x+, 3'+), respectively.
(xz, yt), .= (Xayn). One output, say y, is the sum of the components of all inputs X, ~x3 transmitted through star coupler 34 to terminal 6. If the transfer coefficient from each input terminal to terminal 6 is -mlj, then )' + -m r + ! +
+ ”' + m ) JX J+ −+ m r *
It can be expressed as X n. Similar equations are obtained for all outputs. 1 them
When expressed as one matrix, it becomes as follows. Since the matrix of m represents the transfer characteristic of this star coupler, it will be named the "transfer characteristic matrix" and will be represented by M. As for the specific value of m, assuming that the number of output terminals in Figure 1O is 4 and the input is equally divided and transmitted to each output terminal, the transfer coefficient m is 174 in all. . (3-2) When star couplers are interconnected FIG. 14 is a diagram in which two star couplers A and B are connected. It is assumed that star couplers A and B each have n terminal pairs. 5 to 16 are terminals, and two terminals drawn adjacent to each other (eg, terminals 5 and 6) represent a terminal pair. Output terminal 9 of star coupler A and input terminal 11 of star coupler B are connected, and input terminal 1 to star coupler
0 and the output terminal 12 of star coupler B are connected, thereby interconnecting the star couplers. Therefore, the input of terminal 10 is Xk, and the output of terminal 12 is y3.
Then, Xk''''y11. Also, if the output of terminal 9 is Y and the input of terminal 11 is X, then Y, -Xll. Signal oscillation is caused by the parts that interconnect the star couplers. Therefore, we will take out only this part and explain the reason for the oscillation. FIG. 15 is an enlarged view of the interconnections. A is the transfer coefficient from terminal 10 to terminal 9, and B and A are the transfer coefficient from terminal 11 to terminal 9.
This is the transfer coefficient to the terminal 12. Generally, star coupler A
, B represents the transfer coefficient from the first input terminal to the i-th output terminal as A ij + Bij, then y
, is 7 m = BlIIX I+ =・+ Baa−(X
@-1+ BII#l X @As shown in Figure 15,
! , -y, so y3-ΣB □X i + B
Similarly, for Yll, where m* Y h...■, substituting ■ into ■ gives F++-. The value of r is yl of the input coming in from the input terminal
It is nothing but the sum of the components that contribute to l. Note that the reason why yl is included on the right side of the equation representing y in equation (2) is that the changes in the signal over time have not been taken into consideration in the analysis up to now. Here, we will consider the relationship with time. Let y at time t be ya(t), and the signal is the 15th
If the time to complete one round of the loop in the figure is τ, then y, (t)−
r(t)+B,, Ahmy* (t-r) −・・■
becomes. The value after Ipchi, ya(t-τ) goes around once is y
a(t). If all inputs from terminals other than those of the interconnect are 0, then r(t)-0. In that case, )' * (t
) -B m*Am* )' a ( becomes -τ),
Once ym(t) is at its lowest, the ratio before and after one round of the loop, that is, the loop gain, is: y @ −r' 1B **A** Y m...■ ya(t-τ) Therefore, the noise If some signal is generated in the loop due to etc., oscillation occurs if l B ,, A*kl > 1, and IB□Akk! If it is -1, the signal of the same magnitude will continue to loop around the loop forever, and 0<IB,... If AIIkl<1, it will gradually attenuate. In all of the above cases, even after completing one round of the loop, there is a remnant of the signal, so this will be input again. This is not desirable. The case where the signal is not inputted again after going around the loop once is the case where the signal becomes 0 after going around the loop once. The condition that satisfies this is B□A,h*-0=.■.Therefore, B, .-0 or Adz-0 must hold true. If this holds true, the loop shown in FIG. 15 will be substantially cut off within star coupler B or star coupler A, and no oscillation will occur. Ao is a transfer coefficient from its own input terminal to its own output terminal in the star coupler AOk-th terminal pair. However, in the conventional star coupler, since the input is transmitted not only to the output terminals of other terminal pairs but also to the output terminal of its own terminal pair, Akk≠0. Similarly, B□≠
It is 0. Therefore, the condition (2) cannot be satisfied, and oscillations etc. occur. (4) Reasons why ghosts occur FIG. 16 is a diagram explaining the reason why ghosts occur. The symbols correspond to those in FIG. And A
ha, Adz, Akk, B +t, B
+ゎ, B1 is the transfer coefficient. When considering the output y1 of the terminal 15, the components from each input terminal are transmitted to the terminal 15 along the dotted line path in the figure6.For example, the input X1 from the terminal 5 is transmitted from the terminal 5 to the terminal 9 to the terminal 1.
It is transmitted along the path 1→terminal 15. The transmission coefficient in star coupler A on this path is A□, and the transmission coefficient in star coupler B is B8. Therefore, the signal values transmitted from the terminal 5 are A□, B, , XI. Except for one input, all other inputs are transmitted to terminal 15 through one path. However, the terminal 16 that is paired with the output terminal 15 that we are currently focusing on
Only the input from is transmitted to the terminal 15 through two paths. One is a component transmitted from terminal 16 to terminal 15. This value is B I I X I. The other is a component transmitted around star coupler A, which is the component transmitted through terminal 16.
It is transmitted along the path → terminal 12 → terminal lO → terminal 9 → terminal ll → terminal 15. This value is BlAkkBllI
xI. As a result, BIIXI +Bs+A**B
The value +*i is the component due to the input X. However, as can be seen from Fig. 16, B, , AmkB
laXl goes around star coupler A and connects to star coupler B
Since this is the value returned to BIIXI, it is delayed by the difference in propagation time compared to BIIXI. This delay causes ghosting. In FIG. 16, only two star couplers are connected, so there are two routes, but if three are connected, there are three routes. Therefore, as the number of interconnected star couplers increases by multiplexing the star couplers, the ghost becomes more severe. The present invention aims to solve the above problems! I!
This is the subject.

[Means to solve the problem]

In order to solve the above problem, in the present invention, the star coupler is configured such that the signal transmission coefficient between the input terminal and the output terminal that are paired to be connected to the same node is 0. . Moreover, in order to construct an optical communication network, it was decided to construct the star couplers as described above by connecting one pair of terminals to each other.

[For production]

According to the star coupler configured as described above, no signal is transmitted from the input terminals connected to the same node to the output terminals. Therefore, even if star couplers are connected to each other in order to increase the scale of the optical communication network, a signal sent from one star coupler to the other star coupler will not return. This makes it possible to eliminate oscillations and ghosts. Furthermore, it becomes possible to communicate between two nodes without the other nodes knowing the communication content (bidirectional communication). Furthermore, by using this bidirectional communication function, it is possible to easily detect collisions of signals that occur when simultaneously transmitted from two or more nodes.

【Example】

Embodiments of the present invention will be described in detail below with reference to the drawings. As mentioned above, when we investigated the cause of the occurrence of ghosts, we found that the cause was that the signal returned after passing through other interconnected star couplers. Therefore, in order to eliminate ghosts, it is sufficient to substantially prevent the return path from being formed. To do this, in the example of Figure 16, star coupler B
When the signal sent from terminal 10 to terminal 10 returns from terminal 9 to star coupler B, the value should be set to 0.Even if the value of 0 returns to star coupler B, This is because it is practically as if it never returned. The condition for making this happen is k11-0. When attempting to expand an optical communication network, it is uncertain which terminal pair will be used to interconnect star couplers, so this condition must be met no matter which terminal pair is used. There is a need. Therefore, it must be set as A,, -OA, -〇Amz"0. In other words, as shown below, the diagonal component of the transfer characteristic matrix of star coupler A is 0 (the terminal pair is (No signal is transmitted from the input terminal to the output terminal of the Since it is equivalent to A, it is necessary to satisfy the same conditions. If these conditions are satisfied, oscillation will not occur. Figure 3 shows the connection of star couplers that meet the above conditions. This is a diagram showing that ghosts do not occur if Since all are O, the value of the terminal 100 input from star coupler B returning to star coupler B through terminal 9 is always 0. Therefore, there is virtually no return, and the diagonal component of star coupler B also Since all of them are 0, the value transmitted from terminal 16 to terminal J5 is also always 0. If the star coupler of the present invention in which all diagonal line components are set to 0 is used, the same input can be transmitted through multiple paths. Since the diagonal line components are all O, a loop circulating between both star couplers as shown in Fig. 15 is not formed, so oscillation does not occur. Note that the fact that all diagonal line components are O means that in any terminal pair, the signal input from its own input terminal is not transmitted to its own output terminal.Originally, when its own node is Since there is no need for the device itself to receive the emitted signal, this does not impede the information transmission function.In fact, as described below, 2.
It provides a new feature that allows bidirectional communication between only one node. [Two-way communication] In FIG. 3, consider the case where input is made only from terminals 5 and 16. XI input from terminal 5 is transmitted to all output terminals other than output terminal 6 of its own terminal pair. Therefore, it is also transmitted to the terminal 15. XI input from the terminal 16 is transmitted to all output terminals other than the output terminal 15 of its own terminal pair. Therefore,
It is also transmitted to terminal 6. As a result, the signals received at each output terminal of the optical communication network are: (1) Signal at terminal 15...Xl (XI multiplied by the transfer coefficient) (2) Terminal 6...xl signal (x+ multiplied by the transfer coefficient) (3) Other output terminals...A signal resulting from interference between X and Xi. Since the node receiving the interfered signal does not know the contents of the signal, it does not mean that communication has taken place. However, between two nodes that have transmitted signals, the signals of the other party are transmitted to each other without fail. In other words, bidirectional communication is possible. This provides a new function that allows nodes to communicate with each other without the contents being known to other nodes. However, when transmitting from three nodes at the same time, the signals transmitted to any node are mixed signals, so bidirectional communication is not possible. As described above, by manufacturing a star coupler whose transfer characteristic matrix has all diagonal elements O and configuring an optical communication network using that star coupler, it is possible to prevent ghosts and oscillations from occurring, and also to Direct communication can be made possible. Next, a specific example of a star coupler having such characteristics will be described. FIG. 1 shows a passive star coupler according to the present invention. In FIG. 1, l is a star coupler, 2 to 4 are directional couplers, 2-1 to 2-4.3-1°3-2.4-1.
4-2 is the terminal, and S is the magnitude of the signal. One star coupler 1 is constructed by connecting three directional couplers 2.3.4 in a triangular shape. This star coupler 1 is a passive star coupler because it does not have the ability to amplify the input signal. The operation of the directional coupler 2 will be explained with reference to FIG. The directional coupler 2 is constructed by fusing the cladding layers of two optical fibers so that the core layers of the respective optical fibers are brought close to each other. In the directional coupler 2 having such a configuration, the terminal 2-1
When a signal of magnitude S is input to , a signal of magnitude S/2 is output from terminal 2-3.2-4 in the direction in which the signal advances, as shown by the solid arrow ( , but has the function of not transmitting to the terminal 2-2 in the opposite direction.The directional coupler 3.4 also has a similar function.This is shown in FIG. When combined in the following manner, it is possible to construct a star coupler in which the diagonal components of the transfer characteristic matrix are all 0, as desired by the present invention.When a signal of magnitude S is input to terminal 2-1, as shown in the figure As noted, the signal is transmitted while being halved each time it passes through the directional coupler. However, it is not transmitted to terminal 2-2, which is paired with terminal 2-1. , can be said about the inputs from all directional couplers. Therefore, the star coupler in Figure 1 corresponds to the case where all the diagonal elements of the transfer characteristic matrix are O, and it is the desired star coupler. Although the above is a passive type star coupler, the active type star coupler according to the present invention has a configuration that is different from the conventional active type star coupler shown in FIG. 11 by changing the internal configuration of the matrix circuit 3o. Fig. 6 shows a matrix circuit in the active star coupler of the present invention.A part 19 surrounded by a dotted line is a diagonal component part, but no diode is connected here (as in the conventional case). (see Figure 12), therefore, the diagonal elements of the matrix representing the characteristics of the matrix circuit are all 0.If we have the matrix circuit 30 as shown in 2,
The desired active star coupler is obtained. FIG. 4 shows an optical communication network constructed by multiple-connecting star couplers as described above. SC is a star coupler, and 17 is wiring. A portion where several star couplers SC are connected can be seen as a composite star coupler 18. The number of nodes to which the composite star coupler 18 can be connected can be increased or decreased to a desired number by increasing or decreasing the number of star couplers SC. Therefore, expansion and contraction of the optical communication network can be easily performed. However, when connecting star couplers to each other, care must be taken that they must not be connected in pairs of two or more terminals. FIG. 5 is a diagram illustrating the reason why star couplers should not be connected to each other using a plurality of terminal pairs. If a two-terminal pair is connected, a loop will be formed between the two star couplers, as indicated by the dotted line in the figure, and oscillation will occur. In addition, the signal transmission path is 2
It overlaps, and ghosts also live there. Note that bidirectional communication is possible in a network configured using the star coupler of the present invention, but if you decide not to do it because it is not necessary, use the bidirectional communication function to transmit the signal as follows. Collision detection can be performed. If each node receives a signal from another node at its receiving terminal while transmitting, it detects that the signals have collided. This is because, according to the star coupler of the present invention, one signal transmitted by itself is not transmitted to its own receiving terminal, and the received signal is a signal sent from another node. At this time, signals should be being transmitted from two or more nodes at the same time, and signal collision has occurred.

【Effect of the invention】

As described above, according to the star coupler of the present invention, no signal is transmitted from the input terminals connected to the same node to the output terminals. Therefore, even if star couplers are connected to each other in order to increase the scale of the optical communication network, a signal sent from one star coupler to the other star coupler will not return. Therefore, oscillations and ghosts that occur when conventional star couplers are connected are no longer generated. Therefore, when it becomes necessary to increase the number of nodes, it is possible to increase the number of nodes by adding star couplers without making any changes to the existing nodes. Furthermore, it becomes possible to communicate between two nodes without letting other nodes know the content of the communication (bidirectional communication). This is effective when communication with confidentiality is desired. Furthermore, if there is no need for two-way communication, collision detection can be easily performed using this function.

[Brief explanation of drawings]

Figure 1: Passive star coupler according to the present invention Figure 2: Directional coupler for optical communication Figure 3: Figure 4 showing that ghosts do not occur according to the present invention Figure: Optical communication network constructed by multiple connection of star couplers Figure 5: Figure 6 explains why star couplers should not be connected to each other with multiple terminal pairs: The present invention Matrix circuit in an active star coupler Figure 7 shows an Ethernet network while Figure 8 shows an Ethernet network.
Figure: An example of a proposed optical communication network using a star coupler Figure 9: A star coupler with the input side and output side separated into two sides for easy understanding Figure 10: Functional principles of a star coupler Figure 11...
・Active star coupler Fig. 12... Matrix circuit in active star coupler Fig. 13... Star coupler with corresponding input lines and output lines drawn close together Fig. 14... Two star couplers Figure 15 connecting the
Figure: Enlarged view of interconnection parts Figure 16: Diagram explaining why ghosts occur In the diagram, 1 is a star coupler, 2 to 4 are directional couplers, 2
-1~2-4.3-1.3-2.4-1.4-2.5~
16 is a terminal, 17 is a wiring, 18 is a composite star coupler, 1
9 is the diagonal component, 20 is the coaxial cable, 21 is the tap,
22 is a node, 23 is an optical fiber, 24 is a star coupler, 241 is an OR circuit, 25 is a terminal, 26 is an input terminal, 2
7 is a light emitting diode array, 28 is an amplifier, 29 is a waveform shaping circuit, 30 is a matrix circuit, 31 is a driver,
32 is a light emitting diode array, 33 is an output terminal, 34 is a star coupler, A and B are star couplers, C is an input line, D
is the output line.

Claims (2)

[Claims] (1) A star coupler characterized in that a signal transmission coefficient between an input terminal and an output terminal forming a pair to be connected to the same node is set to 0. (2) An optical communication network characterized in that the star couplers according to claim 1 are constructed by connecting one pair of terminals to each other.