WAVELENGTH CONVERSION AND REGENERATION SYSTEM AND METHOD IN OPTICAL FIBER TRANSMISSION
DESCRIPTION The present invention generally refers to communication systems, and in particular to optical fiber transmission systems. The scope of suchlike systems is to convert an optical signal, fiber- transmitted, pulse-modulated and of an assigned wavelength (e.g. covering the ITU- grid), on another wavelength of an assigned set. Concomitantly, such a system regenerates in 2R mode (Reshaping and
Reamplification) the pulse signal when degraded, e.g., by the noise accumulated over a long route due to the required systematic and noisy optical reamplification. Wavelength conversion has different scopes in optical fiber transmission systems, e.g. that of routing a signal from a link on a certain directrix to another link on a different directrix by wavelength selection, or that of using on the same link another wavelength when that of the signal is already engaged. A first category of wavelength converters, among those actually proposed and used for optical fiber transmission systems, exploits the nonlinear properties found in the semiconductor optical amplifiers (SOAs). A second category exploits the properties of the dielectric susceptibility of crystals (e.g. LiNbO3), and still others use the non-linearity of the dielectric susceptibility in optical fibers. Among the fiber converters produced to date, some use a Four Wave Mixing
(FWM). These shift the wavelength Is of the signal to be converted to the wavelength Is - 2(ls-lp), where lp is the wavelength of a pump signal. The latter is added in-fiber to the signal to be shifted to frequency, and it usually is of a high intensity, thereby determining the desired non-linear effect. However, the maximum shifting of the wavelength is limited to 2(ls-lp).
Moreover, there is no threshold effect, i.e. for low values of the signal its wavelength-shifted replica is reproduced linearly; therefore, noise accumulation on low levels is not prevented, thereby implementing a 2R function 'lacking on the zeroes'. In other known fiber systems it is used a so-called "multiwave mixing" in which the maximum shifting is greater than in the preceding case and equal to N(ls- lp). Such a shifting has generally been attributed to a cascade of nonlinear effects to which the higher multiple components generated would take part in. I.e., the multiple
components would be generated by repeated FWM effects; the first component, FWM-generated and then frequency-shifted, would act as a pump for a subsequent FWM, generating a replica at 2(ls-lp)+ (ls-lp)=3(ls-lp), and so on. However, the schemes and the solutions proposed to date are restricted in that both the signal to be shifted and the pump signal should be near the 'zero dispersion' point of the fiber, and therefore in that the shifted signal always be in the neighborhood of this point. The maximum shifting is attained for high-rate harmonic components, which accordingly are of progressively limited width and therefore subjected to a reduction of the signal/noise ratio. Object of the present invention is to solve said known-art problems, by providing a wavelength conversion method of an input signal to be shifted to a desired wavelength, comprising the steps of: acquiring said input signal; and coupling said input signal to a first 'pump' carrier signal (SP), generating a combined signal, characterized in that it comprises the steps of: - combining a second carrier signal to said combined signal, obtaining a resulting signal; and - extracting from said resulting signal an output signal corresponding to said shifted input signal (SI), wherein said second carrier signal has a wavelength equal to said desired wavelength. A further object of the present invention is to provide a wavelength conversion system of an input signal to be shifted to a desired wavelength, comprising: - means for acquiring said input signal; and means for coupling said input signal to a first 'pump' carrier signal, generating a combined signal, characterized in that it comprises - combining means, apt to combine a second carrier signal to said combined signal, obtaining a resulting signal; and, - means for extracting from said resulting signal an output signal corresponding to said shifted input signal, wherein said second carrier signal has a wavelength equal to said desired wavelength. The main advantage of the present invention lies in that the wavelength
conversion is attained by exploiting the non-linearity properties of an optical fiber and in that it can take place over a wide range of wavelengths, almost independently from the 'zero dispersion' point of the fiber itself, by virtue of the injecting of an additional carrier. A further advantage lies in that, according to the present invention, the 2R function occurs on the low levels as well as on the high levels of the signal, thereby preventing noise accumulation on said levels and hence with improved effectiveness on long-distance transmission systems where there be several regenerators in a chain. Further advantages, features and the operation modes of the present invention will be made apparent in the following detailed description of a preferred embodiment thereof, given by way of example and without limitative purposes, making reference to the figures of the annexed drawings, wherein: figure 1 is a block diagram of a 2R wavelength converter and regenerator system according to the present invention; figure 2 is a block diagram of a control subsystem for adjusting the power of the pump and/or the gain of the amplification chain of the input signal to be converted; figure 3 is a graph showing the pattern of the power on the converted signal as a function of the power of the signal to be converted at the input; figure 4 is a representative graph of the spectrum of the signals at the input of the system according to the present invention; figure 5 is a representative graph of the spectrum of an output signal of a known-art system; figure 6 is a graph showing the pattern of the error rate as a function of the received power; figure 7 is a representative graph of the spectrum of an output signal of a system according to the present invention; figure 8 is a comparative graph showing the pattern of the error rate as a function of the received power; figure 9 is a graph showing the waveform of the input signal; and figure 10 is a graph showing the waveform of the signal converted and regenerated according to the present invention. The present invention is scientifically based on aspects of theories that will briefly be described hereinafter and are corroborated by experimental results that will also be illustrated.
The proposed theory, unlike what is known to the art, envisages that the multiple wavelength generation be due to an angular modulation induced by the pump/input signal beating. This determines a variation in the refractive index of the constituent material of the fiber. Such a variation reproduces the pulse modulation of the input signal. In fact, under non-linearity conditions, the refractive index depends on the field intensity. Thus, also a second added signal undergoes the same phase modulation. Hence, its spectrum, made of an individual line at the input of the fiber, at the output thereof will have multiple components typical of an angular modulation. Over time, the individual component is present or absent according to the presence or absence of the signal to be converted, thereby reproducing the on-off modulation of the latter. According to this theory, the fiber coupling of the signal to be converted, having a frequency fs, and of the pump, having a frequency fp, determines a signal whose intensity reproduces the beating of the two signals. I.e., when the field of the pump is Ep=Pcos(ωpt) and that of the input signal is Es= Scos(ωst), their beating yields:
Ep+Es= S(cosωpt+cosωst)+(P-S)cosωp =
Therefore, its intensity |Es + Ep will have a sinusal pattern at a frequency (ωp-ωs).
As the refractive index n(ω), defined as:
wherein:
3 (3) 8k
depends, by χ(3) (dielectric polarizability non-linearity coefficient) on the field intensity, in the propagation the fields will be subjected to phase modulation at a
frequency fp-fs, determining not a single converted component (conjugated in case of FWM), but rather a plurality of components spaced in frequency thereamong by multiples of the difference fp-fs. These components have widths whose pattern, as a function of the power of the pump signal and of the input signal, reproduces the pattern of the Bessell functions corresponding to the order of the related harmonics. This is congruent with the interpretation of the phenomenon stemming from a phase modulation whose depth (modulation index) is dependant on the values Ps and Pp by means of the variation of the dielectric constant with the power. The present invention further highlights the above phenomenon, via the injecting of a second carrier on which the input signal is shifted. In fact, the method according to the present invention provides, besides from the steps of acquiring an input signal and of coupling said signal to a first carrier signal, the steps of combining a second carrier signal 2P to the combined signal originated by the coupling to the first carrier and of extracting an output signal SF corresponding to the input signal shifted to the desired wavelength. This carrier is injected in-fiber at a wavelength near that on which it is desirable to convert the signal, and that therefore in principle can be any one. Thus, the variation of the refractive index determines the phase modulation of the second carrier. Thus, the spectrum of the second carrier will consist of multiple components typical of an angular modulation. Over time, each individual component will be present or absent according to the presence or absence of the input signal to be converted, thereby reproducing the pulse modulation thereof. Moreover, by selecting as a frequency of shifting that of a harmonic of an order subsequent to the first, in addition to the conversion on any one wavelength also a good 2R regeneration function is attained. This is so since the components of order higher than the first, as for their width as a function of the power of the input signal, have the pattern of the Bessell functions of the corresponding order, and therefore have a threshold on the low levels of the input signal and a clipping effect for the high levels of the input signal. Thus, in a pulse transmission, the noise is compressed on low levels as well as on high levels, thereby reducing its accumulation over long fiber spans in the presence of plural regenerators. Referring now to figure 1, it shows a block diagram of a converter-
regenerator according to the present invention. Numeral 1 indicates a coherent electromagnetic field source, which may e.g. be a laser device at a wavelength near to the 'zero dispersion' point. Its emission is suitably amplified by a booster amplifier 2, thereby making a 'pump' SP useful for the wavelength conversion and 2R regeneration process. Means 18, 19 for acquiring is provided to inject the input signal in the system. At the output of the amplifier 2 there is provided means 3, 4 for coupling the pump SP to the input signal SI, to generate a combined signal SC. Said means for coupling comprises a polarization controller 3 designed to provide that the polarization state of the pump SP has horizontal and vertical components of equal amplitude in order to make the converting operation independent from the polarization state of the signal to be regenerated and converted, as it will be explained hereinafter. The input signal to be regenerated and converted is sent to an input of a directional coupler 19 apt to draw a small fraction thereof as a monitor signal SM for a subsystem 17 in charge of controlling the gain and the power of the amplifiers of the system. A second output of the coupler 19 is sent to an optical booster amplifier 18 on the signal- way. The output SI of the amplifier 18 goes to feed a directional coupler 4, at which it combines to the pump, generating a combined signal SC. The system according to the present invention comprises means 5, 6, 30, for combining the combined signal SC to a second carrier 2P, generated e.g. through a second laser source 13 at a predetermined wavelength in the neighborhood of which it is desirable to convert the input signal. From this combination a resulting signal is obtained. The means 5, 6, 30 for combining comprises a polarization splitter 6 having a first port a for the combined signal SC. In particular, the input signal SI could have any polarization state, whereas the pump SP, by virtue of the action of the controller 3, will always have horizontal and vertical components adjusted so as to have equal amplitude. Thus, for any polarization state of the signal SI to be converted, its Horizontal and Vertical components will combine to the homologous components of the pump. Accordingly, it is prevented that the input signal and the pump may be found on states orthogonal therebetween when injected in a subsequent fiber. In fact, in this case neither beating nor a subsequent nonlinear interaction would occur.
E.g., if the input signal and the pump were on polarization states orthogonal therebetween at the input of the polarization splitter 6, at its outputs "V" and "O" they would have components with the same polarization state, and therefore capable of interacting. At a second port b of the polarization splitter 6 it is injected the second carrier
2P. The present invention further provides means for extracting from the resulting signal an output signal SF corresponding to the shifted input signal SI. Such means for extracting comprises a circulator 5. The second carrier 2P is inputted to the polarization splitter 6 via the circulator 5, located upstream of the port b of the polarization splitter 6, apt to split a converted signal SO from the second carrier 2P. According to the present invention, the second carrier 2P is subjected, in its horizontal and vertical components, to a phase modulation by effect of the pump/input signal beating. Hence, compatibly to the dispersion features of the fiber, it will be possible to select its wavelength over a wide spectrum, contrarily to what has occurred to date in the known art, where the wavelength selected for the conversion is closely linked to that of the input signal. Moreover, as the power level of the second carrier 2P may be high, also the signal/noise ratio is improved with respect to the known art. The means for extracting further comprises a conversion ring 30 comprising non-linear fiber spans 7, 9, 11, adjustable dispersion equalizers 8, 10 and a polarization controller 12 which is apt to transform the vertical polarization state V at the output of the polarization splitter 6 into the horizontal polarization state O after the signal has crossed the fibers 11, 9, 7 and the equalizers 8 and 10. The combined signal SC (SI+SP) and the second carrier 2P onto which the wavelength conversion is to be performed are sent into this conversion ring 30. The improved conversion of the input signal takes place just in this ring. In fact, the varying of the refractive index due to the beating of the input signal and of the pump, determines an angular modulation thereof. Likewise, also an added signal (the second carrier 2P) undergoes the same phase modulation, as it meets the same variation of the refractive index, regardless from the polarization state. As described hereto, its spectrum, consisting of a single line at the input of
the fiber, will be made of multiple components typical of an angular modulation, at the outlet of the fiber itself. Over time, each single component is present or absent according to the presence or absence of the signal to be converted, thereby reproducing the on-off pulse modulation thereof. The signal of the second carrier 2P, phase-modulated by the variation of the refractive index, returns to the port b of the polarization splitter 6. The means for extracting further comprises a filter 14 being inputted the converted signal SO splitted by the circulator 5. The filter 14 is apt to split the component corresponding to the desired conversion wavelength. This component reproduces the intensity modulation of the input signal. It should be noted that the polarization splitter 6 and the controller 12 also prevent the input signal, the pump and their nonlinearity products from reaching the output, abating or even totally eliminating the noise produced therefrom, thereby improving the signal/noise ratio. This occurs by virtue of the fact that in-fiber the carrier 2P has a state of orthogonal polarization with respect to the pump and therefore its spurious products are blocked by the polarization splitter 6. Moreover, the signal/noise ratio increases also since the power value of the second carrier 2P can be made high without the occurrence of spurious nonlinear phenomena. A first control subsystem 17 is inputted a first control signal SA derived from the input signal SI by means of a second directional coupler 19. The control subsystem 17 has the task of outputting adjusting signals zl, z2 for adjusting the gain and the powers of the optical booster amplifiers 18 and 2, as a function of the power of the control signal SA, proportional to the input signal SI. In cascade to the filter 14 it is arranged a third directional coupler 15, apt to draw a second control signal SB from the converted and filtered signal SF and to sent it to the input tl of a second control subsystem 16. The second control subsystem 16 is apt to output adjusting signals xl, x2 for the automatic adjusting of the equalizers 8, 10 as a function of the power of the control signal SB. In particular, the control signal SB derived from the coupler 15 is detected by a photodiode 23 and preferably amplified. A controller 22, in order to maximize the signal SF, acts at subsequent times first on the variable dispersion equalizers 8 and
10, equalizing the group delays between the second carrier and the envelope of the pump/input signal combined signal at subsequent times, at first on the variable dispersion equalizers 8 and 10, then sending a control signal to the first control subsystem 17 for a consequential adjusting of the gains of the amplifiers. All these values are optimized on the basis of the Pin/Pout pattern at the desired harmonic component of conversion, as exemplarily indicated in figure 3. The pattern of this curve essentially depends on the kind of nonlinear fiber used, and it can be expressed as a table from which the subsystem 17 can duly derive the gain and the output power to be set on the optical amplifiers 2 and 18. Finally, at the output of the system of figure 1 the converted signal SF will be obtained.
EXPERIMENTAL RESULTS Solely for the sake of completeness and in order to better illustrate the operation of the present invention, it is reported the description of the experimental experience carried out. In practice, the experimental set-up yielding results that corroborated the above theoretical considerations is that reported in figure 1. Figure 4 is representative of the spectrum of the signals at the input of the system according to the present invention; in particular, it is reported the spectrum of the input signal SI, the spectrum of the first 'pump' carrier signal SP and of the second carrier 2P. According to the known art, i.e. without using a second carrier, the spectrum of the output signal obtained as a combination of the signal to be converted and of the pump would be that reported in figure 5. All the components generated derive from the phase modulation due to the variation of the refractive index by nonlinear effect. This modulation, as already stressed hereto, is generated by the beating between the pump and the input signal. Some conversion and regeneration tests were performed without using a second carrier, singling out some spectral components corresponding to as many channels of the ITU (International Telecommunications Union) grid for partitioning the spectrum, having a replica of the input signal on each of said components. The results of these tests are reported in figure 6, reporting the error rate (BER: bit error rate) on a logarithmic scale, as a function of the power of the input signal when the latter is a signal at 1 OGbps.
On the contrary, with the activation of a second carrier 2P a signal is outputted having the spectrum of figure 7. Apparently, said signal is substantially that of figure 5, with the addition of a signal replica around the second carrier 2P, and therefore in a 'free' zone of the spectrum, i.e. containing none of the original components. BER measurements were performed in this case as well. In figure 8 there are comparatively reported the BER measurements on homologous channels, with and without use of the second carrier, always with reference to an input signal at 10 Gbps. Apparently, with the preceding configuration without a second carrier, in order to obtain an error rate of about 10"10, i.e. a bit error every 1010 bits sent, at the input it is required a signal power of about -24 dBm. Advantageously, using the present invention, in order to obtain the same error rate a signal power of about -31.5 dBm suffices, i.e. a signal power about 8 times weaker with respect to the preceding case, not to mention that with the single-carrier configuration the conversion on the Ch 54 would not have been possible either. The waveform regeneration capability is expressed in figures 9 and 10. In figure 9 it is reproduced the noisy waveform at the input (the signal is at 2.5 Gbps) whereas in figure 10 it is reported the same waveform, regenerated and wavelength-shifted by means of the method of the present invention. The present invention has hereto been described according to a preferred embodiment thereof, given by way of a non-limiting example. It is understood that other embodiments may be envisaged, all to be construed as falling within the protection scope thereof, as defined by the annexed claims.