GB2172165A - Optical signal power divider - Google Patents

Abstract

An optical signal power divider has X (P1-P6) input ports and N(P1'-P6') output ports; and coupling means (11-16) for optically coupling each input port with N-Z (Z =2 in Fig. 2a) output ports whereby in use the signal power at each input port is divided between the respective N-2 output ports. In another example (Fig. 3 not shown), additional couplers can be provided so that each input port is connected to N-1 output ports. <IMAGE>

Description

SPECIFICATION Optical signal power divider The invention relates to optical signal power dividers for example for use in multiple access networks such as wavelength division multiple access networks.

Multiple access techniques based on wavelength multiplexing may be used to provide wideband transparent channels on a passive power dividing optical network. In a network with passive power dividers, traffic separation between different communication routes across the network is achieved by allocating to each transmitter an exclusive communication channel defined for example by means of an exclusive wavelength, time slot, or code. The number of channels available is thus limited which leads to a limit in the number of users who can communicate with the network. Furthermore, such networks have a minimum insertion loss of 101og10N (dB), where N is the number of output ports. This minimum can be achieved if the network is based on the transmissive star topology, where one fibre is used for transmission and one for reception.

In accordance with one aspect of the present invention, an optical signal power divider comprises X input ports and N output ports; and coupling means for optically coupling each input port with N-Z output ports where N-Z > 1 and Z > O, whereby in use the signal power at each input port is divided between the respective N-Z output ports.

We have realised that a significant saving, up to 50%, in the number of communication channels required for a given number of input and output ports can be achieved by removing the restriction that each input port is coupled with each output port.

Preferably, Z= 1. In this case, it can be arranged that the output port with which an input is not coupled is the output port connected to a receiver associated with the transmitter connected to the input port. We have realised that a given transmitting/receiving station does not generally wish to receive back its own transmitted signal. There is therefore no signficant loss in communication ability. Not only would such a network be more channel efficient, but in the case of wavelength division multiple access networks, it would also allow wavelength reuse since the return channel could use the same wavelength. Thus, if the new network is used to distribute wavelength division multiple access in (WDMA) channels for example, we can achieve a 50% saving in wavelengths, since each bidirectional channel needs only one wavelength.

A still simpler optical signal power divider is obtained if Z=2.

This simpler power divider comes very close to the power divider in which Z= 1 in that once again a 50% saving in the number of channels is achieved but in addition there is a saving in the numbers of couplers provided by the coupling means. The only compromise in this particular divider is that for each input port there is a unique output port (or"mate") in addition to its own corresponding output port with which the input port cannot communicate directly. This is thought to be only a minor problem since the probability of wanting to communicate with the "mate" could be arranged to be low. The traffic on networks where a path exists to all terminals is rarely uniformly distributed. When such a path is required, the connection could be achieved via an idle station or some other port.If the "mate" ports are not used then the effective size of the network is halved but the potential of channel reuse is retained.

In addition the insertion loss, in dB's, of the simpler network in which Z=2 is given by 101og10(N-2), and is therefore less than that of a conventional network (ie. lOlog N).

It should be understood that the channels can be defined by wavelengths time slots and codes.

Conveniently X=N although with suitable coupling means such as Y couplers X could be different from N.

Typically, the coupling means comprises a plurality of M-way couplers for example two port (M=2) directional couplers, each coupler having M input ports and M output ports. In this case, the number of output ports N will be given by: N=M"+M where n is an integer. There will be M-l "mates" per input port and the insertion loss is 101og10(N-M) dB.

If the coupling means can be arranged topologically into logM(NM) columns each containing N/M couplers then conveniently the i th port in the j th column, denoted (i,j) where j=n, can be connected according to the following expression: (i,j)([i + M'[i + M - X]modM]modNi+ 1). (1) where the terms in [] are calculated modulo M or N as indicated and x=O, 1 . . M- I. There are a large number of possible arrangements which are equivalent to the expression (1).

The coupling means may additionally or alternatively comprise a plurality of Y couplers.

Although examples of Z=1 and Z=2 have been described, other examples with Z > 2 are possible.

Some examples of optical signal power dividers according to the invention will now be described and contrasted with a prior art example with reference to the accompanying drawings, in which: Figures large are transfer matrices for a conventional power divider and for two power dividers according to the invention respectively; Figure 2a is a block diagram of a first example of a power divider according to the invention; Figure 2b is a block diagram of a second example according to the invention; Figure 2c is a.modification of the Fig. 2a example; and, Figure 3 is a block diagram of a third example according to the invention.

The transfer matrix for a conventional power divider is shown in Fig. 1a. This illustrates that each input port P1-P6 is connected with each output port P1,-P6, with the power of each input signal being equally divided between each output port. In this case N=6.

A first, or ideal, example according to the invention is defined by the transfer matrix in Fig.

1 b. This indicates that each input port P,-P6 is connected to a respective group of five output ports with the input power equally divided between each of the five output ports. For example, the input port P1 is connected to the output ports P2,-P6,.

As has been explained above, a significant reduction in insertion loss can be achieved by reducing still further the number of output ports which are connected with each input port. Fig.

1c illustrates a second example according to the invention in which each input port P1-P6 is connected to four or N-2 output ports. For example, the input port P1 is connected to the output ports P2,-P5, with the input power equally divided between each of the output ports. In this example, the mate of the input port P1 is the output port P6'.

It is assumed in all these examples described that the input port P, and the output port Pal are associated with a transmitter and receiver respectively of the same station. Thus, it does not matter that Pa is not coupled with Pa' since a transmitting/receiving station will not want to receive its own signals.

An example of a physical arrangement of the power divider defined in Fig. 1c is illustrated in Fig. 2a. This example comprises six two port directional couplers 11-16 arranged in a maths of two columns and three rows. These couplers are passive couplers so that each of their input ports is permanently connected to each of the output ports. For example, in the directional coupler 11 the input port P, is connected to each output port labelled 1, 2 as is the input port P6. Power is divided between each output port of the coupler 11. Preferably, power is equally divided between each output port but in practice this cannot usually be achieved.

The couplers 11-16 are connected together by optical waveguides such as optical fibres. Each input port P,-P6 will be connected to respective optical signal transmitters while each output port P1,-P6, will be connected to respective optical receivers. Pairs of transmitters and receivers will be associated in respective stations. As can be seen in Fig. 2a, half of the connections between the couplers 11-16 are straight and this may lead to savings in the quantity of splices if the couplers are made in cascades of log2 (N-2). The remaining connections are offset 2"-' places, where n=j in Fig. 2a. The output ports of the transmission network are shifted cylically, M places (M=2 in this example) in the -i direction relatively to the input port locations.This simple property results from the interconnection rule described by equation (1) above. This equation also indicates how the Fig. 2a example can be extended to any size of network.

Fig. 2b illustrates a 10 port network (N=10) based on the same principle as the Fig. 2a example using two port directional couplers. In this example, each input port Pa cannot communicate with its own output port Pat and also cannot communicate with one other output port, its mate. For example, the input port P1 cannot communicate with the output ports P1, and P10,.

The couplers used in the networks so far described are two port couplers, for example X couplers. Examples may also be developed using alternative forms of coupler, for example Y couplers. Fig. 2c illustrates a modification of the Fig. 2a example in which the couplers 11-16 have been replaced by Y couplers 11'-16'. The number of input ports has been halved to three and each input port Pl-P3 is connected with each of the two output ports of the Y couplers 11'-13'. The couplers 1 1'-13' are connected in a similar way to the couplers 14'-16' as in the Fig. 2a example, each of the couplers 14'-16' being connected to respective output ports P1,-P3,. It should be noted in this example that the only output port with which an input port cannot communicate is its own output port.

In order to eiiminate the mates in the M=2 network of Fig. 2b a pair of additional two port directional couplers 17, 18 are positioned between each of the send and receive mates. This is illustrated in Fig. 3 for the mates PI, P10, and P10, P1,. In general, a total of N extra couplers will be required to eliminate all mates and produce a transfer characteristic of the form shown in Fig.

1b. The additional couplers are arranged to transfer a fraction 1-r of the power to the mate and a fraction r to the network shown in Fig. 2b, where r=(N-2)[- 1 +[1 + 4/(N - 2]11/2 (2) In the limit of large N, r tends to (N-3)/(N-2). The insertion loss of the modified network is 10log10(N-2)/r2 which, in the limit of large N is approximately 10log10(N-1).

It will be seen in Fig. 3 that input port P1 communicates via the coupler 17 with a coupler 19 as in Fig. 2b but a fraction 1 -r of the signal on the input port P1 is added to the multiplexed receive signal from the coupler 20 and is fed to output port P,d.

Claims (10)

1. An optical signal power divider comprising X input ports and N output ports; and coupling means for optically coupling each input port with N-Z output ports, where N-Z > 1 and Z > O, whereby in use the signal power in each input port is divided between the respective N-Z output ports.

2. An optical signal power divider according to claim 1, wherein Z=2.

3. A power divider according to claim 1 or claim 2, wherein X=N.

4. A power divider according to any of the preceding claims, wherein the coupling means comprises a plurality of M-way couplers, each coupler having M input ports and M output ports.

5. A power divider according to claim 4, wherein M=2.

6. A power divider according to claim 4 or claim 5, wherein the M-way couplers are arranged in a form which is topologically equivalent to log,(N--M) columns each containing N/M couplers with the i th port in the j th column being connected in accordance with the following expression: (i,j) ,((i+M[i+M x]modM]mOdNif 1) (1) where the terms in [j are calculated modulo M or N as indicated and x=0,1 . . M-1.

7. A power divider according to any of claims 1 to 5, wherein the coupling means comprises a plurality of Y couplers.

8. An optical signal power divider substantially as hereinbefore described with reference to any of the examples shown in Figs. 1b, 1c, 2a, 2b, 2c and 3 of the accompanying drawings.

9. A multiple access optical network including at least one optical signal power divider according to any of the preceding claims.

10. A multiple access optical network according to claim 9, the network comprising X transmitting terminals connected to respective ones of the X input ports of the power divider; and N output terminals connected to respective ones of the N output ports of the power divider.