The present invention relates to a system for transmitting independent communication channels through a light-wave medium e.g. an optical fiber, and to the devices used to operate said system. It relates to systems employing "go -- no go" modulation of a coherent or incoherent light wave. This kind of modulation utilises, for example, "laser" diodes i.e. electro-luminescent diodes which emit within a narrow band of wavelengths, and it is possible to choose the wavelength as a function of the conditions of manufacture of the diode, for example by doping the semi-conductor material.

Those skilled in the art will be aware that the maximum data rate which can be carried by an optical fibre of given length, for a given communication error raten is limited by two main factors which are contradictory in nature:

-- THE APERTURAL ANGLE OF THE LIGHT BEAM WHICH THE FIBRE IS EFFECTIVELY CAPABLE OF PROPAGATING; THIS DEPENDS SOLELY UPON THE REFRACTIVE INDICES OF THE GLASS USED TO MANUFACTURE THE FIBRE; THE RATE IS THE HIGHER THE SMALLER SAID ANGLE IS; FOR EXAMPLE, FOR A GIVEN FIBRE 1 KILOMETER LONG, AN APERTURAL HALF-ANGLE IN AIR, OF 36°, MUST NOT BE EXCEED IF A RATE OF 10 MEGABITS PER SECOND IS TO BE ACHIEVED;

-- THE LIGHT POWER AVAILABLE AT THE FIBRE OUTPUT; THIS MUST BE IN EXCESS OF A THRESHOLD BELOW WHICH THE COMMUNICATION ERROR RATE IS IN EXCESS OF THE FIXED LIMIT. For example, for a permissible error rate of 10^{- 8}, the output detector must receive a minimum power of 1 microwatt peak, which requires that at the fibre input a power of 10 milliwatts is necessary if the fibre attenuation is 40 decibels.

The relationship:

P = K sin2 u                                          1.

where k is a constant characteristic of the light source, applies between the power P injected at the fibre input and the apertural angle u.

Thus, there is a limit on the power for a given rate, since the value of the rate fixes a maximum value of u; similarly, there is a limit on rate for a given communication error rate.

In accordance with the invention, there is provided a system for multiplexing an optic communication comprising at least one optical fibre, a multiplicity of electroluminescent diodes producing light radiation of different wavelengths, and a multiplicity of photodetectors, means for selecting said wavelengths, modulating means applied to said diodes, and demodulating means applied to said photodetectors.

The invention will be better understood and other of its features rendered apparent, from a consideration of the ensuing description and the accompanying drawings in which:

FIG. 1 illustrates an example of the invention;

FIG. 2 is an explanatory diagram;

FIGS. 3, 4 and 5 illustrate variant embodiments of the invention;

FIG. 6 illustrates a device for operating the system in accordance with the invention.

FIG. 1 illustrates a device with 10 communication channels (unidirectional) operating on the principle of the system in accordance with the invention. A fibre 1, for example 1 kilometer long, has an input face 11 at the "transmitting" end, and an output face 12 at the "receiving" end. At the "transmitting" end, 10 electro-luminescent diodes E₁₀, E₁₁, . . . E₁₈, E₁₉ are arranged in a parallel array. The two first and two last of these diodes, are the only ones to have been illustrated. For example aluminum-doped gallium arsenide diodes, of the double heterostructure type, whose different doping levels are such that 10 different emission spectral "lines" in the near-infrared, are obtained, will be used.

These spectra have been shown on a graph in FIG. 2, the graph plotting the luminous intensity I as a function of the wavelength λ in microns. There can be seen 10 "lines" S₀, S₁, . . . S₉, substantially all having the same highly elongated bell form, with a base measuring less than 50 angstrom units, the mid-height width being of the order of 20 angstroms units, and the axes of the lines being regularly spaced from 0.88 to 0.97 micron, in other words at 100 angstrom unit-intervals.

However, it is currently possible to manufacture band-pass and band-cut filters which are sufficiently selective in order to readily separate such spectral lines. These filters are designed in the form of assemblies of transparent plates of given thicknesses (of the order of half the wavelength).

Equally, it is known to manufacture selective mirrors which will reflect the radiation within a narrow frequency band (corresponding to the width of a line in FIG. 2), and which behave as transparent plates vis-a-vis other wavelengths.

Selective mirrors, orientated at 45° in relation to the axis F₁ X₁ of the entry face 11, centered at F₁, are arranged in a parallel fashion. The mirrors M₁₁ to M₁₉ selectively reflect the lines S₁ to S₉ of FIG. 2. These mirrors are preceded by a non-selective mirror M_1Q. Similarly, the mirrors M₂₀ to M₂₈, orientated at 45° in relation to the exit axis F₂ X₂ of the fibre, selectively reflect the lines S₀ to S₈ and the terminal mirror M₂₉ is a non-selective mirror. The terminal mirrors M₁₀ and M₂₉ could either be replaced by selective mirrors, which would mean an unnecessary extra expenditure, or be discarded provided that the light transmitter and receptor corresponding to the relevant channels, were suitably positioned.

In the device shown in FIG. 1 the diodes E₁₀ to E₁₉ are respectively supplied with voltage pulses by the input circuits C₁₀ to C₁₉ responsible for amplifying the power of the pulses for transmission, the pulses being received at the channel inputs V₁₀ to V₁₉, and for modulating the diodes E₁₀ to E₁₉. Circuits of this kind, well known to those skilled in the art, have been symbolically illustrated.

The diodes E₁₀ to E₁₉ are respectively arranged at the foci of lenses L₁₀ to L₁₉ whose axes are perpendicular to the axis F₁ X₁. The mirrors M₁₀ to M₁₉ are orientated in such a manner as to direct the reflected rays in the direction F₁ X₁. These rays are focussed at F₁ by a lens 101.

At the receiving end, similarly there is to be found a lens 102 whose axis is coincidental with the axis F₂ X₂ of the face 12 of the fibre 1, and which has a focal point F₂ arranged at the centre of said face. The mirrors M₂₀ to M₂₉ are arranged at 45° in relation to the axis F₂ X₂ in order to direct the rays from F₂, on to lenses L₂₀ to L₂₉ having axes perpendicular to F₂ X₂. At the foci of these lenses, there are located photodetectors R₂₀ to R₂₉ injecting into amplifier-demodulator circuits C₂₀ to C₂₉ of a kind well known to the person skilled in the art, and thus furnishing their signals to the channel outputs U₂₀ to U₂₉.

The operation of the device is as follows. The infrared radiation emitted by diode E₁₁ in the spectrum S₁, is selectively reflected by the mirror M₁₁ which passes the radiation substantially unattenuated, the radiation then passing the ensuing mirrors, for example M₁₈ and M₁₉, again substantially without attenuation. The same radiation, reconstituted by the fibre at F₂, is rendered parallel to the axis F₂ X₂ by the collimating lens 102, passing thence across the mirror M₂₀ to be selectively reflected by the mirror M₂₁ and finally focussed by the lens L₂₁ of the photodetector R₂₁ which supplies the circuit C₂₁.

It will be observed that since the mirrors M₁₀ to M₁₉ on the one hand and those M₂₀ to M₂₉ on the other, are arranged on the axis F₁ X₁ and F₂ X₂ in reverse order, the rays emitted and received at each end in fact pass through them in the same order. Because of this fact, the radiation components produced by each of the channels, have the same number of mirrors to pass and consequently experience the same overall attenuation which is not negligible despite the fact that it is the sum of very small attenuation levels.

By way of example of a practical embodiment, using two one kilometer long fibres each capable (in the absence of the device in accordance with the invention) of transmitting 10 megabits per second with an attenuation of 40 decibels and a communication error rate of 10^{- 8}, it is possible to build, using the terminal electro-optical equipment shown in FIG. 1, a 10-channel system capable of handling 10 times 10, in other words 100 megabits per second. The characteristics of the equipment used in the system, are for example as follows:

a. Diodes E₁₀ to E₁₉ (aluminium-doped gallium arsenide diodes):

-- threshold current 300 mA;

-- output power at 500mA: 10 mW;

-- beam divergence in the plane of the junction: 15°;

-- beam divergence in the perpendicular plane: 50°;

-- source dimensions: 80 microns by 0.5 microns.

b. Photodetectors (silicon PIN photodiodes):

-- photosensitive area: 1mm² ;

-- total capacitance (for reverse voltage of 20 V): 2pF;

-- dark current: 2 mA;

-- sensitivity at 8500 angstroms units: 0.5 microamps per microwatt.

c. Lenses L₁₀ to L₂₉ :

-- diameter: 1 cm;

-- focal length: 1.5 cm.

d. Selective mirrors:

-- pass bands: 50 angstrom units for the wavelength in question;

-- dimensions: 1.4 cm by 1 cm.

It should be pointed out that since an optical fibre is capable of simultaneous transmission in both directions, in FIG. 1 it is possible to exchange any diode E with the associated photodetector R, provided that the input and output circuits are also exchanged.

In a first variant embodiment, the diodes and the photodetectors for the different channels have been respectively arranged in the focal planes on the transmitting and receiving ends. In FIG. 3, the transmitting end of this kind of system has been shown. The receiving end, not shown, is strictly symmetrical. At the transmitting end, the diodes E_i and E_j emitting at the different wavelength i and j are arranged at different points on the straight line representing the intersection of the plane of the figure with the focal plane of a lens L₁ whose centre and focus are disposed in the plane of the figure together with the centre F₁ of the entry face 11 of the fibre 1.

The different beams emitted by the diodes E_i and E_j, give rise to beams made up of rays parallel to the rays I_i and I_j coming from the lens L₁. They pass through a dispersive element 30 which is for example a prism of a diffraction grating. The positions of the diodes E_i and E_j and the inclination of the elements 30 are calculated in such a fashion that two incident rays I_i and I_j give rise to two parallel (refracted or diffracted) rays R_i and R_j. In the case of the diffraction grating, these will for example be two first order diffracted rays. These rays are focussed by a lens L₂ on the face 11.

At the receiving end, the rays R_i and R_J collimated by a lens which is symmetrical with L₂, and then dispersed by an element symmetrical with 30, are finally focussed by a lens symmetrical with L₁, on the separate photodetectors.

In a second variant embodiment, the assembly of three optical elements L₁, 30 and L₂ FIG. 3, is replaced in FIG. 4 by a single holographic lens 40 which is a hologram in a thick medium. Considering the point F₁ which is the centre of the entry face 11 of the fibre 1, the rays emerging from the hologram 40 seem to stem from a single source. At the receiving end, the rays emerging from a symmetrical hologram, are focussed at different points depending upon the wavelength and it is at these points that the photodetectors pertaining to the different channels are arranged.

In a third variant embodiment, a single fibre can be used for both directions of transmission, thanks to the provision of the terminal equipment shown in FIG. 5 pertaining to an example involving six channels, three in either direction. At the end corresponding to the face 11, a glass plate 51 is arranged, having parallel faces, at either side of which there are disposed three diodes E₁, E₂, E₃ and three photodetectors R₄, R₅, R₆. At the end corresponding to the face 12, in a symmetrical arrangement there are encountered a plate 52, three diodes E₄, E₅, E₆ and three photodetectors R₁, R₂, R₃. In front of each diode and each photodetector there is a lens 53 having a focal point on the corresponding emissive or photodetecting face. Band-pass filters T₁ to T₆ centred in pairs on six different wavelengths (as in the case with the spectrum shown in FIG. 2 for example) are arranged respectively in front of the corresponding lenses 53.

The path of the light rays is as follows. First of all let us consider the trajectories of the light rays issuing from the diode E₁. They experience two refractions at the end corresponding with the face 11 and, at the end corresponding with the face 12 successively a refraction, five reflections and another refraction, before terminating at the detector R₁, thus in total four refractions and five reflections. It will readily be appreciated that in the arrangement shown in FIG. 5, for the other trajectories E₂, R₂, E₃, R₃ etc . . . , the same number of refractions and reflections is experienced albeit in a different order.

The result of this is that the six channels, vis-a-vis the transmitted radiation, will present the same total attenuation. If we assume that the faces of the glass plates 51 and 52 act as semi-reflective walls, then a total attenuation of the order of 25-30 decibels will be experienced in the case where there are three channels in either direction. In this variant embodiment, the system would seem to be limited to the operation of a small number of channels. It is possible to combine the equipment provided in the third variant embodiment, for a small number of channels, with the equipment provided for the embodiment shown in FIG. 1, which handles a larger number of channels. In this way, the communication facilities offered by the fibres, are only partially utilised in both directions.

If reliability of communication takes precedence over other considerations, then a bunch of several fibres will be used, the entry and exit faces of which are arranged side by side in the same plane. It is worth while combining this arrangement with the multiplexing system which forms the object of the present invention, in order to provide several channels or, more simply, a go channel and a return channel. Thus, in this latter case, we arrive at the device shown by way of example in FIG. 6. A diode E₁, radiating within a narrow frequency band, is arranged at the focus of a lens L₁. The rays leaving this lens form cylindrical beam passing a mirror M₁ of selective type arranged at 45° in relation to the beam and producing no substantial attenuation vis-a-vis the emission spectrum of E₁. A lens L₂, parallel to L₁ then concentrates the rays coming from E₁ and L₁ on the entry face 601 of a bunch of fibres 600. At the opposite end of the bunch, there are the lenses L₄ and L₅ between which there is disposed, at 45°, a selective mirror M₂ identical to that M₁. The lens L₅ concentrates the rays coming from the exit face 602, on the photosensitive electrode of a photodetector, R₁.

Because of the provision of the selective mirrors M₁ and M₂, the rays emitted by a diode E₂ arranged at the focus of a lens L₆ (disposed at 90° to the parallel lenses L₄ and L₅) are reflected by the mirror M₂ which has a normal reflective power in the emission spectrum of E₂. After passing the lens L₄, they converge on the face 602 which, vis-a-vis this spectrum, acts as entry face. In the same manner, the face 601 plays the part of exit face vis-a-vis the radiation from E₂, which is concentrated on the photosensitive electrode of a photodetector R₂ by the lens L₂, the mirror M₁ and the lens L₃ arranged at 90° to the parallel lenses L₁ and L₂.

Self-evidently, in the case of the device shown in FIG. 6, the emission spectra of the diodes E₁ and E₂ are chosen to have a much larger interval from one another than in the case of FIG. 2. For example, the interval will be of the order of 1400 angstrom units instead of 100 angstrom units as was the case in FIG. 2. The design of the selective mirror is facilitated by the magnitude of this interval ant the cost is also reduced. It should be pointed out that the device operates equally well with simple semi-transparent mirrors although the transmission efficiency is poorer.

The device shown in FIG. 6 is applicable in particular to the transmission of remote-control signals, for example in airborne equipment.