FR2688966A1 - System for very-long-distance transmission over optical fibre, with compensation for distortion on reception - Google Patents

Abstract

The field of the invention is that of very-long-distance digital transmission over optical fibre. The invention relates to a system for very-long-distance transmission of a digital signal between a sending station (21) and a receiving station (25), in which the said sending (21) and receiving (25) stations are linked by a monomode optical fibre (22) with negative chromatic dispersion at the operating wavelength of the system, with a length (LTot) of at least one thousand kilometres. According to the invention, the said receiving station (25) comprising means of compensating for the distortion which is due to the combination of non-linear effects and of the chromatic dispersion introduced by the said transmission line, the said compensation means carrying out a positive chromatic dispersion of the received signal, the amplitude of the said positive chromatic dispersion being a function particularly of the amplitude of the negative chromatic dispersion induced by the said optical fibre, as well as of the mean optical power on-line of the said signal transmitted over the said optical fibre.

Description

 Very long distance fiber optic transmission system with distortion compensation on reception.

 The field of the invention is that of very long distance digital transmission (several thousand kilometers) on optical fiber, in systems using optical amplification in line.

 In such very long-distance systems, one of the main limiting factors of the flow rate is the distortion induced by the transmission fiber.

 Indeed, if these distortions can be more or less neglected in conventional optical fiber systems, (providing links of the order of a few hundred kilometers), they become rather very disturbing in long-distance transmission systems.

 The invention relates to a transmission system on an optical fiber line making it possible to compensate, on reception, this distortion induced by the transmission line, in the case of transmission lines with a length of at least one thousand kilometers, such as than those used for transoceanic links.

 The distortion provided by the transmission fiber stems from the combined existence of two phenomena occurring in monomode fibers: chromatic dispersion and non-linear effects.

 The first phenomenon is chromatic dispersion. This phenomenon comes from the frequency dependence of the index of silica. n results in different propagation times depending on the operating wavelength. In general, the chromatic dispersion tends to widen the pulses of the digital trains and thus to create intersymbol interference.

 In usual fibers, the chromatic dispersion is nil around 1.3m and takes a positive value of about 17ps / nmlkm around 1.55, ut. It is also possible to use shifted dispersion fibers which are designed to have zero chromatic dispersion at around 1.55, ut.

 Very long distance systems (several thousand km) operate at 1.55 um. The excessive value of the chromatic dispersion of the usual fibers at this wavelength prohibits their use, and in this case we will use shifted dispersion fibers.

 It should be noted that the chromatic dispersion distortion effect strongly depends on the spectral components of the pulses: if a pulse has positive optical phase variations at its beginning and negative at its end, it will be greatly enlarged by a positive chromatic dispersion. The converse is true for negative dispersions.

 The second phenomenon concerns non-linear effects. The most important non-linear effect in a fiber is the Kerr effect. This effect, for example, is described in the KW document "Non linear effects in optical fibers and fiber devices" (nonlinear effects in fiber optics and fiber equipment). Blow and N.J.

Doran (IEEE Proceedings, June 87, pp. 138-144), translates a linear dependence of the silica index on optical power.

 It is very weak in the usual areas of operation of optical systems (distance less than about 400 km and power less than about 10 mW), but becomes significant for very high powers (of the order of 1 W) or for very large propagation distances at reasonable power (a few thousand kilometers in a system with periodic amplification).

 In the absence of chromatic dispersion, the Kerr effect induces phase auto-modulation of the optical pulse: the instantaneous frequency decreases at the beginning of the pulse and then increases at its end, in proportion to the derivative of the optical power. This induces a broadening of the spectrum and a spectral composition which favors a large broadening for negative chromatic dispersions.

 The distortion provided by the transmission fiber must therefore be considered as a combination of chromatic dispersion (first phenomenon) and non-linear effects (second phenomenon).

 The combination of these two effects can be described by a nonlinear equation with partial derivatives in distance and in time known under the name of nonlinear Schrödinger equation, whose resolution is discussed in particular in the book "Non linear fiber optics" (optics of non-linear fibers) of G.

Agrawal (Academic Press).

The numerical study of this equation shows that there are two qualitatively very different behaviors according to the sign of the chromatic dispersion (D)
* First case: D> 0. In this case, we observe phenomena
of modulation instability. The pulses "burst" into impulses
very short after 1000 to 2000 km and the optical spectrum
expands considerably, which may pose problems of
optical bandwidth.

* Second case: D <0. There is no modulation instability and the
impulses keep some integrity while the spectrum
expands almost monotonically during the course of propagation
keeping reasonable widths. However, the impulses
broaden temporally, which creates interferences
these inter-symbols. These interferences become very annoying,
example as soon as the chromatic dispersion exceeds 0.05ps / nmlkm
in absolute value for flow rates of SGbit / s over 6000 to 8000 km.

 The most effective case is of course the second case, that is to say the case of negative chromatic dispersion. But to operate very long distance systems in negative dispersion, the chromatic dispersion values of the fibers used must necessarily be very low.

 Any method of compensation of the two phenomena at the origin of the distortion induced by the fiber (chromatic dispersion and nonlinear effects) is therefore of great interest since it makes it possible to overcome the disadvantage of the low values imposed on chromatic dispersion. Indeed, through the use of a method of compensation for the distortion provided by the transmission fiber, two strategies can be envisaged.

 In a first strategy, for given characteristics of negative chromatic dispersion of the transmission fiber, the compensation makes it possible to increase the product of the online throughput by the range of the link.

 In a second strategy, for a fixed in-line flow and a fixed range, the compensation makes it possible to use line fibers having constraints on the more relaxed chromatic dispersion characteristics. These fibers are easier to manufacture industrially and to sort for the constitution of an underwater connection for example.

 Several methods of compensating for the distortion introduced by the transmission line on optical fiber are already known.

 A first known method, for example described by H. Gnauck et al. in the article "Dispersion penalty reduction using an optical modulator with adjustable chirp", OFC 1991. Post-deadline paper No. 17, consists in using the emission an optical amplitude modulator having a variation of optical instantaneous frequency (11clirrp "in Anglo-Saxon) adjustable.

 This first method has been proposed only to partially compensate for chromatic dispersion penalties on conventional optical fiber, in short-distance transmission systems (a few hundred kilometers).

 Indeed, it was not a question of compensating the nonlinear effects, completely negligible in this case, but only the chromatic dispersion.

 By tailors, this first known method relates specifically to conventional optical fibers (with zero chromatic dispersion around 1.3 μm). However, in very long distance transmission systems (several thousand kilometers), the excessive value of the chromatic dispersion of the usual fibers prohibits their use, and shifted dispersion (ie chromatic dispersion) fibers are generally used. none around 1.55 clam).

 Finally, the compensation takes place on transmission and therefore in advance, and not on reception, on the actually disturbed signal.

 It is therefore clear that this first known method is not suitable for very long distance transmission systems and does not concern a system for compensating for distortions on reception.

A second known method of compensation, for example described by
T. Koyama et al. in the article "Compensation for non-linear pulse distortion in optical fibers by employing technical prechirp" (Compensation of non-linear impulse distortion in optical fibers by the use of a precautionary technique), ECOC 91. WeC7- 2 p. 469, consists of compensating, on transmission, the self-modulation of phase generated by the non-linear effects.

 For this purpose, the pulses are subjected to a continuous scanning of the optical frequency, during the bit time, by over-modulation of the emission laser.

 In this case it is used that the optical frequency of a semiconductor laser is a function of its current. But only certain laser structures can thus have a high frequency modulation efficiency, that is to say a significant variation of this optical frequency, for a small variation of the control current and therefore of the optical power.

 This second known method has the disadvantage of requiring a particular synchronous over-modulation of the useful digital train on the transmission laser used. Therefore, only certain types of lasers, those capable of offering a significant variation of the optical frequency for a small variation of the control current, can be used in this second method.

 In addition, this method is applicable only for a particular type of binary coding, namely the online RZ code.

 Moreover, as in the case of the first method described, the compensation takes place on issue and not on reception.

 Finally, this second method only compensates for nonlinear effects.

 Other compensation methods are known that apply to correct on reception the distortion due to the combination of strong positive chromatic dispersions and non-linear effects (and in particular "chirp" of the laser, the "chirp" the laser being a parasitic effect consisting of a variation of the instantaneous optical frequency during the duration of a pulse, especially at the beginning and at the end thereof).

 These other methods use different structures (Fabry-Perot filter for example) which have the disadvantage of only partially correcting the distortions.

 In particular, the non-linear effects are not compensated by these known filters or equalizers. However, the passage of a transmission line of a length of a few hundred kilometers to a line longer than a thousand kilometers leads to no longer neglecting the non-linear effects.

 Indeed, these nonlinear effects become as important, or even more important than chromatic dispersion. Therefore a specific solution must be found to solve this new problem.

 The invention particularly aims to overcome these disadvantages of the state of the art.

 More specifically, an object of the invention is to provide a system capable of compensating for the distortion induced by a transmission line having a length of at least one thousand kilometers on monomode optical fiber with negative chromatic dispersion.

 Another objective of the invention is to provide such a system that compensates for the combination of the two phenomena induced by the transmission fiber: the negative chromatic dispersion and the non-linear effects, and not just one or the other of these. phenomena.

 The invention also aims to provide such a system, which can be implemented regardless of the type of laser used on transmission.

 Another object of the invention is to provide such a system that is simple to implement, reliable and low cost.

 These objectives, as well as others which will appear subsequently, are achieved according to the invention by means of a very long distance transmission system of a digital signal between a transmitting station and a receiving station. said transmitting and receiving stations being connected by a monomode optical fiber with negative chromatic dispersion at the operating wavelength of the system, having a length of at least one thousand kilometers, said optical fiber inducing negative chromatic dispersion and non-linear effects, the combination of said nonlinear effects and said negative chromatic dispersion distorting the digital signal received by said receiving station, said receiving station comprising means for compensating said distortions, said compensating means producing a positive chromatic dispersion of said received signal, the magnitude of said positive chromatic dispersion being in particular, the amplitude of the negative chromatic dispersion induced by said optical fiber as well as the average line optical power of said signal transmitted on said optical fiber.

 The invention therefore relates specifically to very long distance transmission systems. According to the invention, a monomode optical fiber with negative chromatic dispersion is used, which seems to be the only one suitable for long distances.

 According to the invention, the magnitude of the positive chromatic dispersion for performing the compensation is not only a function of the magnitude of the negative chromatic dispersion induced by the optical fiber, but also takes into account the average optical power in line.

 Taking into account the average power in line makes it possible to take into account the non-linear effects. Indeed, nonlinear effects, which are specific to very long distance links, are a function of this average optical power line.

 In other words, while the known methods consist in compensating the negative chromatic dispersion induced by the fiber by producing a positive chromatic dispersion whose amplitude is exactly equal to that of the negative chromatic dispersion, the invention proposes to compensate the distortions due to the combination of negative chromatic dispersion and nonlinear effects induced by the fiber, by producing positive chromatic dispersion whose amplitude is not equal (except in special cases) to that of negative chromatic dispersion and is function of the optical power.

 Moreover, in short-distance transmission, the mean optical power in line does not intervene in the calculation of the compensation of distortions, it can advantageously be increased so as to also increase the signavbruit ratio, and thus reduce the error rate binary (TEB).

 On the contrary, in long-distance transmission, the average optical power in line intervenes in the calculation of the compensation of the distortions since the nonlinear effects are a function of this optical power. In the case of the invention, the value of the optical power therefore results from a compromise: it must not be too high, so that the non-linear effects are reduced, or too low so that the ratio of the signallessbruir is sufficiently important .

 After propagation in a fiber with negative dispersion and inducing more or less important non-linear effects, a pulse tends to widen.

This enlargement notably causes intersymbol interference.

 In the system according to the invention, the received pulses are compressed on reception by causing them to undergo a chromatic dispersion of positive sign.

This compression eliminates inter-symbol interference and thus improves the quality of the transmission.

 It should be noted that the realization of a positive chromatic dispersion on reception allows, on the one hand, and in an easily understandable way, a compensation of the negative dispersion due to the transmission fiber, but also on the other hand, and of much more surprising way, compensation for non-linear effects.

 Indeed, it is known in the field of optical transmission that non-linear effects occur during long-distance transmissions. It is also known that, because of these nonlinear effects, the optical spectrum at the transmission line output is greatly enlarged, the expansion factor being dependent on the optical power.

 However, according to the conventional ideas, any means inducing a high chromatic dispersion produces significant distortions over a wide spectrum.

Those skilled in the art are therefore clearly encouraged not to use means producing a strong chromatic (positive) dispersion in an attempt to reduce the distortions due to the combination of chromatic dispersion (negative) and nonlinear effects induced by the fiber. transmission.

 The invention thus goes against these ideas conventionally accepted since it specifically advocates the use of such means achieving a strong chromatic dispersion (positive) on a received signal having a broad spectrum due to non-linear effects.

 In other words, the invention provides a particularly simple solution to the problem specific to long distance transmission that represents the joint compensation of negative chromatic dispersion and nonlinear effects, showing that, contrary to the ideas conventionally accepted in the field of the optical transmission, the use of well-chosen positive chromatic dispersion means can allow the return of an exploitable corrected signal.

 Advantageously, said amplitude of the positive chromatic dispersion is also a function of the electrical bandwidth of said receiving station as well as of the digital signal emitted by said transmitting station.

 In this way, compensation for the distortions induced by the combination of negative chromatic dispersion and nonlinear effects is further improved.

 Advantageously, said digital signal is coded according to the NZ or NRZ binary format.

 In a first advantageous embodiment of the invention, said compensation means comprise at least one optical fiber section with chromatic dispersion positive at said operating wavelength.

Advantageously, said compensation means comprise at least one optical amplifier and at least two fiber sections with positive chromatic dispersion, each of said optical amplifiers being interposed between two consecutive fiber sections.
In this way, the attenuation of the fiber is compensated and the receiver receives sufficient power to ensure its operation.

 Advantageously, said compensation means comprise at least two sections of fiber with positive chromatic dispersion chosen in a set of fiber sections of at least two different lengths, so as to adjust said amplitude of the positive chromatic dispersion.

 In this way, the evaluation of the amplitude of the positive chromatic dispersion to be applied, which generally requires the use of intensive and heavy digital simulation computer programs, can be carried out simply by a test / error system on a link installed.

 This also makes it possible to determine the optimal compensation when the characteristic parameters of the transmission system (in particular the amplitude of the negative chromatic dispersion induced by the transmission fiber, the electric bandwidth of the reception station, the transmitted signal (modulated or not). in phase / frequency in addition to the amplitude modulation for example), the average optical power in line, ...) are not perfectly known before the establishment of the link.

 In other words, the correction fiber length is first roughly evaluated, then the addition or removal of sections of different lengths makes it possible to obtain the length of the fiber corresponding to the optimal compensation.

 Advantageously, a first part of said compensation means is inside the receiving station and a second part of said compensation means is outside the receiving station.

 Thus, by placing a first portion of the compensation means (i.e., a portion of the correction fiber length) out of the receiving station, space is saved in the receiving station.

 In addition, by placing a second part of the compensation means inside the receiving station, it is easy to intervene along the length of the compensation fiber by adding or removing small fiber sections. Similarly, the optical level problems are easily managed by arranging optical amplifiers in the receiving station.

 In a second advantageous embodiment of the invention, said compensation means comprise at least one set of at least two diffraction gratings mounted so as to induce a positive chromatic dispersion.

 This second embodiment is slightly more complex than the previous one, but on the other hand it makes it possible to obtain results of superior quality.

 Advantageously, the receiving station then comprises means for adjusting the spacing and / or the inclination of at least one of said diffraction gratings.

 Finally, it is quite possible to couple the two embodiments described above.

Other features and advantages of the invention will appear on reading the following description of a preferred embodiment of the invention, given by way of indicative and nonlimiting example, and the appended drawings, in which:
FIGS. 1a to 1c show curves of variation of the
optical power and the instantaneous optical frequency in
function of time, the three figures respectively corresponding
the same impulse during the next three successive stages
of a transmission with a system according to the invention:
* emission of the pulse,
* reception of the impulse at the end of the line,
* compensation of the pulse after reception.

- Figures 2 to 4 each show a simplified diagram of a mode
distinct embodiment of a transmission system according to the invention
the compensation means comprising respectively:
* a section of positive chromatic dispersion fiber,
* sections of chromatic dispersion fiber positive and
of different lengths between which are intercalated
optical amplifiers,
* two diffraction gratings.

 In a long-distance digital transmission (several thousand kilometers) on optical fiber using line optical amplification, the transmitted optical pulses undergo a distortion. This distortion is due to the fiber itself and comes from the combination of non-linear effects (Kerr effect in particular) and chromatic dispersion in the transmission fiber.

 Studies of the combination of these two effects (nonlinear effects and chromatic dispersion) have shown that it is interesting to operate very long distance systems in negative chromatic dispersion.

But the very small values of chromatic dispersion required make it difficult to achieve the fiber line.

 The transmission system, which will be described in detail later, includes means for compensating non-linear effects as well as negative chromatic dispersion, both types of phenomena being induced by the fiber of the transmission line.

 Thus, by retaining a transmission fiber having given negative chromatic dispersion characteristics, the system according to the invention makes it possible to increase the product (online throughput) * (range of the link).

 Or, according to another strategy, while maintaining a fixed in-line flow and a fixed range, the system according to the invention makes it possible to use line fibers having constraints on the chromatic dispersion characteristics which are looser and therefore easier to manufacture industrially. .

Figures la to lc each have two curves:
a first curve of variation of the optical power PO as a function of time, represented in solid lines;
a second curve of variation of the instantaneous optical frequency F1 (corresponding to the derivative with respect to the time of the phase) as a function of time t. This second curve is represented in dotted line.

 Figure la corresponds to an emitted pulse. The pulse has a width Lo. The instantaneous frequency FI is constant. This pulse is transmitted through a very long distance fiber (length greater than 1000 km) of negative chromatic dispersion.

 This same pulse received at the end of the line is shown in FIG. After propagation in a negative chromatic dispersion fiber, on the one hand, and having significant non-linear effects on the other hand, the pulse has widened (length LI greater than Lo) which causes intersymbol interference. This pulse has an instantaneous IF frequency (derived with respect to the time of the phase) quasi-linear, of slope corresponding to the chromatic dispersion incurred. Its spectrum has expanded because of non-linear effects. The lowest frequencies are at the beginning of the pulse and the highest frequencies at the end of the pulse.

 According to the invention, this pulse is compressed on reception by subjecting it to a chromatic dispersion of positive sign. Figure 1c shows a pulse having undergone such compression. The low frequencies have been slowed (10), and the high frequencies have been accelerated (11). This compression eliminates inter-symbol interference and thus improves the quality of the transmission.

 Note that this compensation does not give exactly the initial impulse. Due to nonlinear effects, the optical frequency decreases at the beginning (12) of the pulse and then increases at its end (13). However, this is of no consequence from the point of view of the digital signal.

 The pulse obtained near compensation is as short or shorter than the pulse emitted. This allows to find in reception a rate of opening of the eye equal or better than the emission.

 Figures 2 and 3 show two variants of a first embodiment of the transmission system according to the invention.

 The simplest transmission system to implement is shown in Figure 2.

The transmitted signal is composed of a pulse train. each of these pulses (as shown in FIG. 1a) is generated by a transmitting station 21. This pulse propagates in a fiber segment 22 of length
Ltot (in kilometers) and average chromatic dispersion per kilometer Dmoy (in ps / nm / km) at a transmission wavelength such that Dmoy is negative.

 After propagation, the pulse will have undergone a negative dispersion D (in ps / nm) with D = Dmoy * Ltor. The pulse then has the characteristics of a pulse as shown in FIG.

 According to the invention, the transmission system comprises, on reception, a positive chromatic dispersion fiber making it possible to compensate for the distortions due to the combination of negative chromatic dispersion on the one hand and non-linear effects on the other hand.

If, on reception, a correction fiber having an average positive chromatic dispersion per kilometer DCor (in ps / nm / km) is available, a segment 23 of this LCor length fiber can be produced.
The amplitude of the positive chromatic dispersion induced by this segment 23, namely LCor * D ,,, must be such that the distortions are compensated.

This amplitude of the positive chromatic dispersion is a function of several parameters, and in particular of:
the amplitude of the negative chromatic dispersion induced by the
segment 22 of transmission fiber,
- the average optical power in line of the transmitted signal on the
fiber segment 22,
- the electrical bandwidth of the receiver 24 of the station
reception 25,
the signal transmitted by the transmitting station 21.

In this first embodiment, it is considered that all these parameters are well defined. In this case, the amplitude of the chromatic dispersion
LCOR of the correction fiber segment 23 to be used can be calculated, for example, using numerical simulations taking into account all the parameters involved.

 There is thus a device providing a chromatic dispersion of opposite sign to that of the line and to remove the distortions. The receiver receives a pulse having the characteristics of a pulse as shown in FIG. 1c (suppression of inter-symbol interference and improvement of the quality of the received and corrected signal).

 The receiving station 25, situated downstream of the transmission fiber 22, comprises the correction fiber 23 as well as the receiver 24.

Of course, for the compensation to be applicable, the LCor section must not be too long compared to Ltot. It is therefore extremely interesting that
ID <> r)) I Dm0y I. If, for example, IDco / Dmoyl = 100, the length LCor will only correspond to 1% of the Ltop length, which seems reasonable from a production point of view.

 The transmission system shown in Figure 2, however, seems relatively difficult to implement as is. Indeed, it does not take into account a number of problems.

 Thus, the correction fiber, especially if it is long, has some attenuation. If this attenuation is too great, we can not afford to add the fiber as it ultimately, because the receiver would not then receive sufficient power to ensure its operation.

 An alternative to this system is shown in FIG. 3. In this variant, the length LCor is divided into n smaller sections 11, 12, ... 1n, between which optical amplifiers 311 to 31n are interposed to maintain the level of power.

 In this way, a pulse generated by a transmitting station 32 propagates in a fiber Ltot length and undergoes a negative dispersion D. Inside the receiving station 33, the pulse is then propagated by n sections of fiber length 11,12, ... ln ensuring a total chromatic dispersion L> r * Dcor. The optical amplifiers 311 to 31, interposed between the sections of fiber length 11, 12, ... ln compensate for the attenuation due to the correction fiber composed of n sections. The pulses transmitted to the receiver 34 therefore do not exhibit intersymbol interference.

 In addition, the parameters of the transmission line may not be defined before the line is laid. For example, it may not be possible, at first glance, to have exact knowledge of Dmoy because of manufacturing uncertainties, or because Dmoy varies from one link to another. Similarly, the average optical power in line is sometimes impossible to accurately simulate.

 It is therefore necessary to be able to adjust LCor at each link. To do this, it is interesting to divide the correction fiber very short sections compared to Lr expected. The addition or removal of these elementary sections during installation then makes it possible to obtain the optimal length with an accuracy that is all the better as the elementary sections are small compared to the expected LCor.

 If one still wants to improve the accuracy of the compensation, it is possible to resort to a system according to the principle of "boxes of weight". One then uses on the one hand longer lengths which give an approximate compensation and, on the other hand, smaller lengths which make it possible to adjust the compensation to its optimum value.

 Finally, there is the question of the physical location of the correction fiber. This can be in a line cable on an end section, or inside the receiving station.

The first solution, which consists of placing the correction fiber in a line cable, that is to say outside the receiving station, has the advantage of saving space in the receiving station, but presents two disadvantages:
- it will be very difficult to make adjustments in case of error
on rent (especially in the case of an underwater connection where the cable
is in the sea),
- any repair on the cable in the terminal section is likely to
modify LCor and thus degrade the compensation.

 The second solution, which consists in placing the correction fiber inside the receiving station, seems more flexible. This second solution is interesting especially if one does not know the exact compensation length or if this length changes as a result of changes on the line. It is indeed very easy to intervene on the compensation fiber length by adding or removing sections of small fiber. Similarly, the optical level problems are easily managed by arranging optical amplifiers in the receiving station.

 A third solution is finally to combine the two previous solutions: part of the correction length is in a terminal section cable and another part, possibly adjustable, in the receiving station.

 An exemplary embodiment of this first embodiment gives excellent results. The system, in this example, is a transoceanic link 8,000 kilometers long, at a rate of SGbit / s and an operating wavelength of 1.55 μm. The line consists of dispersion dispersed dispersion fiber sections Dmoy negative at 1.55, ut.

 Between these sections are intercalated, every 40 km, optical amplifiers that compensate for losses.

 Working at a wavelength of 1.55, ut makes it easy to obtain the highly positive dispersion fibers mentioned above. It suffices to use usual fibers whose chromatic dispersion is about 17 ps / nm / km at 1.55 .mu.m.

 Taking, for example, the case Dmoy = -0.1 ps / nm / km, the pulses then undergo a total negative dispersion during the propagation of: Dmoy * 8000 km = -800 ps.

In this case, the inventors have found that the correction fiber length to be used for the compensation is equal to 50 km. It should be noted that if only the negative chromatic dispersion induced by the transmission fiber were taken into account, the length would be equal to:
- (- 0.1 x 8000) / 17 = 47 km.

 It is clear in this example that the compensation of the combination of the non-linear effects and the negative chromatic dispersion is different from a simple compensation of the negative chromatic dispersion (compensation of the type used for short-distance transmissions).

 Depending on the case, the length of the correction fiber may be greater or less than the calculated length taking into account only the negative chromatic dispersion of the transmission fiber.

In summary, the characteristics of the transmission system are as follows
- length of the connection: 8000 km,
- bit rate: 5 Gbit / s,
- wavelength of use: 1, 55 μm,
- NRZ coding,
- Optical amplifiers at a distance of 40 km each
excess noise of 6 dB,
- transmission fiber:
* attenuation 0.2 dB / in,
* chromatic dispersion: -0.1 ps / nm / km,
- signal level at the amplifier output: -3dBm,
- fiber correction:
* length: 50 km,
* chromatic dispersion: 17 ps / nm / km.

The following results are obtained:
- without compensation: there is a penalty of 2dB on the opening of
the eye in reception. This penalty is related to the phenomena of
propagation and leads to an error rate of about 1C9;
- with compensation the cancellation of the penalty of opening of the eye
achieves an error rate better than 10-12. If we
increases the value of the positive chromatic dispersion of
correction (D greater than 800 ps / nm), one can even obtain a
improvement of the eye diagram (+ 1db) compared to the
diagram on issue.

A second exemplary embodiment has also given very good results. This second embodiment differs from the first only on the following characteristics
- length of the link: 7960 km,
- transmission fiber:
* chromatic dispersion double compared to the first
example: -0.2 ps / nmlkm,
- fiber correction:
* double length compared to the first example: 100 km
The following results are obtained:
- without compensation: there is a penalty of 5.5 dB on the opening
eye leading to an error rate of about 10-5;
- with compensation: the cancellation of the penalty of opening of the eye
achieves an error rate better than 10-12.

 Figure 4 shows a second embodiment of the transmission system according to the invention. The system includes a transmitting station 41 generating optical pulses. These optical pulses are transmitted by a fiber 42 to a receiving station 43. The transmission fiber 42 has a total length Ltot and a total negative chromatic dispersion (-D).

 The receiving station 43 comprises a reception photodiode 44 as well as an arrangement based on diffraction gratings 45 and 46. This assembly is intended to induce a positive chromatic dispersion that makes it possible to compensate for the negative chromatic dispersion (-D) of the fiber 42 as well as the non-linear effects induced by this same fiber 42.

 The general principle of this assembly is described in the work of G.P.

Agraval "Non linear fiber optics", Academic Press 1989, chapter 6. However, in this document, the assembly is specifically intended for the compression of ultra-short pulses. This is in no way to compensate for distortions in a transmission system.

 In other words, according to the invention, this is the adaptation of known means for pulse compression to a totally different domain, namely very long distance transmission, and for a different application, namely compensation of distances, and more precisely the compensation of distortions resulting from the combination of the negative dispersion and the non-linear effects induced by the fiber.

 Impulses propagated in the fiber 42 have their spectra broadened after propagation. They pass through a first lens 47. The optical beam 48 issuing from this first lens 47 is then sent through two diffraction gratings 45 and 46. The pulses are compressed during a double pass through these two diffraction gratings. The optical beam is reflected on a first mirror 49 and traverses a second time the two diffraction gratings 45 and 46. This double passage makes it possible to give the optical beam a circular cross section (section become ellipsoidal because of the first pass) and to multiply the compression factor by two.

 The first mirror 49 is tilted slightly to separate the incident beam from the reflected beam. A second mirror 410 (shown dashed and located in a plane raised with respect to the plane of the figure) receives the beam reflected by the first mirror 49 and deviates, without introducing any additional losses, to a second lens 411. The photodiode receiver 44 finally receives compressed pulses and without inter-symbol interference.

 Thus, by calculating the spacing and inclination between the two diffraction gratings, it is possible to obtain an induced chromatic dispersion that makes it possible to compensate for the distortions exactly.

 If we do not know exactly the chromatic dispersion of the connection, it is possible to modify the spacing and the angle between the two diffraction gratings by mechanical translation-rotation of these two networks. This makes it possible to adjust at best the positive chromatic dispersion of compensation.

 In this compensation arrangement based on diffraction gratings, the passage in beam in the air probably causes significant losses. It can then be predicted to precede (or follow) this compensation arrangement of an optical amplifier so that the pulses arrive with the correct level on the receiving photodiode.

 This second embodiment is a little more difficult to implement than the first but it makes it possible to obtain a more precise compensation.

 It is therefore also possible to provide a system successively comprising, on reception, an optical correction fiber according to the first embodiment (and which makes it possible to obtain a somewhat broad compensation), and a diffraction grating arrangement according to the second embodiment of FIG. realization, (and which is intended to refine the compensation a little broad already carried out).

Claims (10)

 A system for very long distance transmission of a digital signal between a transmitting station (21; 32; 41) and a receiving station (25; 34; 43), characterized in that said transmission and transmission stations are connected by an optical fiber (22; 42) with a negative chromatic dispersion single mode to the operating wavelength of the system, of a length (Lrot) of at least one thousand kilometers, said optical fiber (22; 42) inducing negative chromatic dispersion and nonlinear effects, combining said nonlinear effects and said negative chromatic dispersion distorting the digital signal received by said receiving station, and said receiving station (25; ; 43) comprises means for compensating for said distortions, said compensating means producing a positive chromatic dispersion of said received signal, the amplitude of said positive chromatic dispersion being a function of and the magnitude of the negative chromatic dispersion induced by said optical fiber as well as the average line optical power of said transmitted signal on said optical fiber.  2. System according to claim 1, characterized in that said amplitude of the positive chromatic dispersion is also a function of the electrical bandwidth of said receiving station (25; 34; 43), as well as of the digital signal emitted by said radio station. 'program.  3. System according to any one of claims 1 and 2, characterized in that said digital signal is encoded according to the binary format NZ or NRZ.  4. System according to any one of claims 1 to 3, characterized in that said compensation means comprise at least one optical fiber section with chromatic dispersion positive at said operating wavelength.  5. System according to claim 4, characterized in that said compensating means comprise at least one optical amplifier (311 to 31 and at least two sections of fiber with positive chromatic dispersion, each of said optical amplifiers being interposed between two consecutive fiber sections .  6. System according to any one of claims 4 and 5, characterized in that said compensating means comprise at least two sections of positive chromatic dispersion fiber selected from a set of fiber sections of at least two lengths (11 to ln), so as to adjust said amplitude of the positive chromatic dispersion.  7. System according to any one of claims 4 to 7, characterized in that a first part of said compensation means is inside the receiving station (25; 34; 43) and a second part of said means compensation is outside the receiving station.  8. System according to any one of claims 1 to 3, characterized in that said compensation means comprise at least one set of at least two diffraction gratings (45,46) mounted so as to induce said positive chromatic dispersion.  9. System according to claim 8, characterized in that it comprises means for adjusting the spacing and / or inclination of at least one of said diffraction gratings (45,46).  10. System characterized in that it comprises firstly compensation means according to any one of claims 4 to 7 and secondly compensation means according to any one of claims 8 and 9.