Method for controlling optical signal/noise ratio, optical telecommunication method, telecommunication system, amplifier with optical active fiber, and optical active fiber

The essence of the process is to form the active fiber (1,6) with a high emission value curve which also has emission depression relative to the adjacent regions, and in the process eliminates or reduces this depression on the emission curve by selected additives are introduced into the fiber (1,6) in a predetermined amount.

The invention further relates to an optical telecommunication method, wherein at least one optical signal belonging to a particular band is generated, transmitted over a transmission line and amplified by a line amplifier (32, 45) comprising an active fiber (1,6), and the signal is received through a receiver (33, 49).

The essence of the method is that at least one of the line amplifiers (32, 45) is formed by an active filament (1, 6) formed by a fluorescent main impurity and a secondary impurity cooperating with the main impurity in the glass matrix of the active fiber (1,6). contaminants and adjusting the signal gain in the line amplifier (32.45) to an input power of -20 dBm or less, so that the gain differs by less than 1.6 dB from the gain in that wavelength band in an unfiltered arrangement.

The present invention also relates to an optical telecommunication method, whereby a signal belonging to a specific band is generated, transmitted on a transmission line, while amplifying it with a line amplifier (32, 45) comprising at least one active fiber (1,6), and receiver (33).

The essence of the process is that the active fiber (1,6) of the at least one active fiber (1,6) line amplifier is formed with a silica-based core and the active fiber (1,6) is a cooperating main fluorescent impurity. and a secondary pollutant such that the optical signal-to-noise ratio obtained with a reception width of 0,5 nm at the reception is not less than 15 dB at the wavelength range.

The present invention also relates to a telecommunications system comprising an optical signal transmitter (27, 28, 37-40), a receiver (33, 49), a transmission line interconnecting them, and at least two serial active fiber (1,6) line amplifiers (1,6). 32, 32 ', 45, 45').

The essence of the system is that the line amplifier (32,45) has a silica-based active fiber (6, 13) whose core is contaminated with at least a fluorescent main impurity and a secondary impurity such that by interaction of the two with the receiver (33, 49) a transmitted signal-to-noise ratio having a signal-to-noise ratio greater than 15 dB, measured at a 0.5 nm filter width, for signals in the band.

The invention further relates to an active fiber optical amplifier comprising a silica-based active fiber (1, 6) and an excitation source connected thereto, and a transmission signal coupling element to the active fiber (1,6).

The essence of the amplifier is that the core of the active fiber (1, 6) contains a cooperating main fluorescent impurity and a secondary impurity, and when transmitting two signals of different wavelengths in the same band, the maximum amplification difference between signals is less than -20 dBm or at equal power, less than 2.5 dB of active fiber without filter element (1,6).

The present invention is still an active fiber, primarily for optical amplifiers used in telecommunication systems, and having a numerical aperture greater than 0.15 and having a core comprising a cooperating fluorescent main impurity and a secondary impurity, and the emission curve in the predetermined band of the active fiber (1, 6), has a depression of less than 1 dB in relation to the emission value in at least one adjacent region when the light is transmitted to the active fiber (1,6).

The present invention relates to a method for controlling an optical signal / noise ratio, an optical telecommunication method, a telecommunication system, an optical fiber optical amplifier and an optical active fiber. The present invention is primarily applicable to wavelength divisional data (WDM) transmission.

In wavelength division multiplex systems (WDM), separate channels or transmission signals are transmitted independently along the line. The telecommunication line itself is preferably an optical fiber and the data transmission is performed by multiplexing within the optical wavelength ranges. The transmitted channels may be digital or analogue channels, which must be properly decoupled since each is assigned a specific frequency.

In such a transmission, the different channels must be substantially the same in terms of their parameters, neither being in a discriminated position relative to the other, either in terms of signal level or signal quality.

In the case of an amplifier arrangement, especially in the case of opti45 amplifiers, all channels must have the same transmission parameters. In order for a large number of channel transmissions to be feasible, the amplifier must operate in a wide frequency band.

Optical amplifiers can be configured by selecting appropriate fluorescent impurities and their parameters. An example of such a contaminant is erbium, which is present in the core of the optical fiber. Erbium can be excited by light-scattering energy, which will have a very high emissivity in the wavelength range within the minimum light attenuation range of silicon-based optical fibers.

When using an erbium-doped optical fiber, where erbium is excited today55

EN 216 228 Β, then a light signal is transmitted through the optical fiber having a wavelength corresponding to this high emission, the signal causes the erbium atoms to pass lower, and at the same time emits adequate light energy along the wavelength of the signal. This is how the signal is essentially amplified.

The decay of erbium atoms starts spontaneously from the excited state. This spontaneous decay produces a random emission which produces a "background noise" that is superimposed on an amplified signal corresponding to the excited emission.

The light excitation energy transmitted to the contaminated optical fiber can produce light emission of different wavelengths, which is also characteristic of the contaminating medium, and thus originates the fluorescent spectrum in the optical fiber.

In order to achieve maximum signal amplification and associated high signal-to-noise ratios with optical fibers as described above in optical telecommunication systems, the signal used is generally generated by a laser source and is selected at a wavelength corresponding to the maximum fluorescence spectrum of the optical fiber. corresponds to the band in question if the fiber contains the impurities already mentioned.

In the case of erbium-contaminated optical fibers, the emission spectrum has a relatively narrow peak. Its parameters depend on the glass into which erbium has been introduced as a contaminant and have a relatively high intensity spectral range of interest in the wavelength range close to the above-mentioned peak. Thus, it is desirable to use such optical fibers to form broadband optical amplifiers.

In the case of erbium-contaminated optical fibers, the emission spectrum is uneven. This unevenness makes it possible to achieve uniform gain across the entire frequency band.

Thus, in order to obtain a substantially straight line, i.e., line-of-line gain curve, the gain at various wavelengths should be as uniform as possible, eliminating noise from spontaneous emission, and performing appropriate filtering. Such silencing and filtering arrangements are described in EP 426222, EP 441211, EP 417 441.

However, in the aforementioned announcements, the behavior of the amplifier for wavelength division multiplex systems is not described, nor is there any indication of the system behavior of the various amplifiers connected in cascade.

The nature of the emission spectrum is highly dependent on the impurities present in the core of the optical fiber, as it also determines how to increase the reflectivity. U.S. Pat. No. 5,2,82079 discloses a fluorescent spectrum of an alumina and erbium-doped optical fiber having a less prominent peak shifted to lower wavelengths (up to 1532 nm) than a germanium-erbium-doped active saccade. Such an optical fiber has a numerical aperture (NA) of 0.15.

1-4 of ECOC '93 ThC 12.1. On the side, an optical fiber contaminated with aluminum and lanthanum, which has a very low sensitivity to hydrogen, is described for use as an amplifier. The aluminum filament optical fiber filament described herein has a numerical aperture of 0.16 and the aluminum filament and lanthanum filament filament has a numerical aperture of 0.3.

In ECOC '93 Tu Issue 4, pp. 181-184. A optical amplifier comprising optical fibers contaminated with erbium is described on page. The experiments were performed with optical fibers whose core was contaminated with aluminum, aluminum / germanium and lanthanum / germanium. Experience has shown that the best results were obtained with optical fibers contaminated with Al / La.

Electronics Letters, June 6, 1991, pages 1065-1067. This article describes an optical amplifier which also contains an erbium-contaminated optical fiber which is also contaminated with alumina and this alumina contamination allows for a larger and smoother gain profile. This article refers to the comparison of optical amplifiers with alumina, germanium, and erbium-contaminated optical fibers compared to amplifiers with lanthanum, germanium, and erbium-contaminated optical amplifiers. range is available with the latter.

In ECOC '91 TuPSl-3, pp. 285-288. page Al₂SHE₃-S₂ A type of optical fiber contaminated with erbium and lanthanum is disclosed. The purpose of this optical fiber was to achieve a higher reflection factor and to reduce inclusions containing erbium ions. The absorption and fluorescence spectra of the Er / La-contaminated optical fiber were similar to that of an erbium-contaminated Al₂SHE₃-S₂ optical fiber. The numerical aperture was also 0.31 and the erbium concentration was 23-10^l8cm 3.

In ECPC '89, Post-Deadline Papers, PDA-8, 33-36. In the September 1989 issue, an experiment involving twelve cascade-connected optical amplifiers using erbium-doped optical fibers was described. The arrangement was examined for a single wavelength of 1,536 pm and emphasized that the signal wavelength should be controlled with an accuracy of 0.01 nm for constant operation, considering that the BIT (Bit Error Rate) responds very quickly to changes in the signal wavelength.

U.S. Patent No. 5,117,303 discloses an optical telecommunication system that includes cascade-switched optical amplifiers and is calculated based on high signal-to-noise ratio in saturated mode.

The amplifiers described above were made of erbium-doped optical fibers having Al₂SHE₃-S₂ had a core,

EN 216 228 Β and filters were also applied. The calculated parameters were achieved at a single wavelength, since there were no incoming signals within a wide wavelength range.

In the present invention, it has been discovered that an appropriate combination of contaminants introduced into the core of an optical fiber enables the large numerical aperture to live within an emission spectrum that can be used as an optical amplifier and wavelength division multiplex system and to provide a single wavelength with one or more serially connected amplifiers.

The present invention relates to a method for controlling optical signal-to-noise ratio in a predetermined wavelength band for optical telecommunication systems, which is a telecommunication system.

- optical transmitters,

- optical receiver,

- and includes an optical telecommunication transmission line linking the optical transmitter and receiver, and

- at least one active fiber optic line amplifier is disposed along the transmission line.

The essence of the process is to form the active fiber with an emission curve having a high emission region that falls within a predetermined wavelength band and also has an emission depression relative to the adjacent regions within the wavelength band, and during the process eliminating or reducing depression by administering preselected amounts of additives selected in the active fiber.

Preferably, the predetermined transmission wavelength band is between 1530 and 1560 nm, more preferably between 1525 and 1560 nm.

Optical signal-to-noise ratio is preferably greater than 15 dB at 0.5 nm filter width.

In the process, it is preferable to use at least two serially connected active fiber line amplifiers.

It is advantageous to introduce two types of impurities into the active fiber, at least one fluorescent main impurity and at least one secondary impurity cooperating with the main impurity in the glass fiber matrix of the active fiber, thereby reducing depression to less than 1 dB. adjacent ranges in that band.

The impurity is oxide in the form of oxide and at least two additional secondary impurities.

Preferably the secondary impurities are germanium, aluminum and lanthanum in the form of oxide.

The invention further relates to an optical telecommunication method comprising the steps of:

generating at least one optical signal having a wavelength in a predetermined transmission wavelength band,

- transmitting the signal thus generated over an optical telecommunication line,

amplifying the optical signal at least once with a line amplifier comprising at least one active fiber, and

receiving the signal through a receiver (33, 49).

The process comprises forming at least one of the line amplifiers with an active fiber contaminated with a fluorescent main impurity and at least one secondary impurity cooperating with the main impurity in the fiber matrix of the active fiber and an optical fiber of a predetermined wavelength in the active fiber line amplifier. adjust the signal gain factor to an input power of -20 dBm or less, that the gain differs by less than 1.6 dB from the gain in that wavelength band, in an unfiltered configuration.

Yet another object of the present invention is to provide an optical telecommunication method comprising the steps of:

generating at least one optical signal in a predetermined wavelength band,

- transmitted via optical telecommunication line,

amplifying the optical signal at least once with a line amplifier comprising at least one active fiber, and

receiving the signal through a receiver.

The essence of this method is to form an active fiber of at least one active fiber line amplifier with a silica-based core, and to contaminate the active fiber with one of the main fluorescent pollutants and at least one secondary pollutant so that the optical signal is received during reception. / Noise ratio of less than 15 dB, measured at a wavelength of 0,5 nm, at a filter width of 0,5 nm.

This telecommunication method is advantageous if the optical signal / noise ratio measured at the receiver with a filter width of 0.5 nm is greater than 15 dB for signals of at least two different wavelengths simultaneously fed in the given band.

It is desirable to amplify the optical signal at least twice along the transmission line with the active fiber line amplifiers connected in series.

The wavelength band used is preferably between 1530 and 1560 nm, more preferably between 1525 and 1560 nm.

It is preferred that the active fiber of the active fiber line amplifier be contaminated with erbium as the main fluorescent impurity and at least two secondary impurities cooperating with the main impurity.

The secondary impurities are preferably aluminum, lanthanum, germanium, in the form of oxides.

The invention further relates to a telecommunications system which:

- a transmitter generating optical signals in a predetermined wavelength band,

- the buyer,

an optical telecommunication transmission line linking the transmitter and the receiver, and

- comprising at least two active fiber-optic line amplifiers connected in series along the transmission line.

HU 216 228 Β

The essence of the system is that it comprises at least one line amplifier formed by a silicon oxide based active fiber whose core is contaminated with at least one fluorescent main impurity and at least one secondary impurity such that the signal transmitted to the receiver as a result of the interaction of the two / noise ratio, which is a signal / noise ratio greater than 15 dB at a wavelength within the band, measured at a 0.5 nm filter width.

The telecommunication system is advantageous if the optical signal-to-noise ratio measured at the receiver, measured with a 0.5 nm filter, is greater than 15 dB for each signal fed to the signal in a given wavelength band.

It is preferred that the main fluorescent impurity in the active fiber be in the form of erbium, oxide, and the secondary impurities in the form of aluminum, germanium, lanthanum, oxides.

The predetermined transmission wavelength band of the telecommunication system is preferably between 1530 and 1560 nm.

It is also advantageous if at least three line amplifiers are connected in series along the transmission line and at least the line amplifier is formed with an active filament (6, 13) contaminated with aluminum, germanium, lanthanum and erbium in the form of suitable oxides.

The invention further relates to an active fiber optical amplifier which

- contains at least one silica-based active fiber,

- it also contains a scare source connected to that active fiber,

- an element which couples the optical excitation signal and one or more transmission signals in a predetermined wavelength band to the active fiber.

The essence of the amplifier is that the core of the active fiber contains at least one main fluorescent impurity and at least one secondary impurity, and the maximum amplification difference between the signals in the same wavelength band is less than -20 dBm, or less than 2.5 dB for active fiber without filter element, at an equal power.

Preferably, the main fluorescent impurity in the active fiber is in the form of erbium, oxide, and the secondary impurities are aluminum, germanium, lanthanum, in the form of suitable oxides.

Preferably, the active fiber optic amplifier has an emission curve of greater than 1 dB in the pre-selected depression-free band, and a depression of the emission curve of less than or equal to 0.5 dB, at least relative to the emission values ​​of one of the adjacent bands.

The preselected transmission band is preferably between 1530 and 1560 nm, preferably between 1525 and 1560 nm, and the numerical aperture of the active fiber used is greater than 0.15.

The active fiber optic amplifier preferably comprises at least two silicon oxide based active fibers coupled to a suitable source of excitation energy, and wherein at least one active fiber has at least one fluorescent main impurity and at least one secondary impurity cooperating with the main impurity equipped with a gain difference of less than -20 dBm between the amplifications of two signals of different wavelengths in the same wavelength band, An input power of less than 2.5 dB for active filaments without a filter element, or having an input power equal to or less than 2.5 dB.

The present invention also relates to an optical fiber for optical amplifiers used primarily in optical telecommunication systems.

The fiber optic fiber has a numerical aperture greater than 0.15 and is provided with a contaminated core comprising at least one fluorescent main impurity and at least one secondary impurity cooperating with each other and in a predetermined wavelength band in the active fiber. the emission curve has a depression of less than 1 dB when the light is transmitted to the active fiber, relative to the emission value in at least one adjacent range.

Preferably, the depression of the emission curve in the active fiber is less than 0.5 dB relative to the emission value in at least one adjacent region.

Preferably, the main impurity in the active fiber is erbium oxide, while the secondary impurities are aluminum, germanium, lanthanum in the form of suitable oxides.

It is preferred that the lanthanum content in the core of the active fiber (1, 6), expressed as oxide, is greater than 0.1 mol%, more preferably, the lanthanum content in the core of the active fiber (1,6), expressed as oxide, is greater than or equal to 0.2 mol%, most preferably the lanthanum content, expressed as oxide, in the core of the active fiber is greater than or equal to 5 mol%.

It is further preferred that the optical active fiber core has a molar ratio of lanthanum to germanium oxides of from 10 to 100 or a molar ratio of lanthanum to germanium oxides of about 50.

It may also be advantageous if the content of aluminum in the active fiber core, expressed as oxide, is greater than 1 mol%, more preferably greater than 2 mol%.

It is also preferred that the optical active fiber has an erbium content of 20-5000 ppm / mol expressed as oxide in the nucleus, more preferably 100-1000 ppm / mol.

It is also desirable to have a numerical aperture of the active fiber greater than 0.18.

The invention will now be described in more detail by way of exemplary embodiments in the accompanying drawings. The

First FIG. 4A is a view showing an embodiment of an optical amplifier according to the invention, FIG

Second Fig. 4A shows an optical amplifier provided with a comb filter;

Third Fig. 2A shows an experimental arrangement for determining the emission spectrum for different types of optical active fibers;

HU 216 228 Β

4th FIG. 3A shows the emission spectra of the different types of active fibers measured by the arrangement of FIG

5th 1A shows the gain curve of the amplifier of FIG. 1 at different wavelengths and at two different input power levels,

6th 2A shows the gain curve of the optical amplifier of FIG. 2 at different wavelengths and at three different input power levels, formed by the active fiber according to the invention, FIG.

7th 2A shows the gain curve of the optical amplifier of FIG. 2 at different wavelengths and at three different input signal levels, formed by the known active fiber,

8th FIG. 4A is a telecommunications system comprising cascade amplifiers with two input signals multiplexed along the same transmission line, FIG.

9th 8A shows the Bit Error Rate for different embodiments of the embodiment shown in FIG.

10th FIG. 4A is a telecommunications system implemented with different cascade line amplifiers where four signals of different wavelengths are multiplexed along the same transmission line; FIG.

11th 10A shows the power levels at the input of the first line amplifier of FIG. 10 with the active fiber of the present invention;

12th Figure 10 shows the power levels at the input of the second line amplifier for the layout shown in Figure 10;

13th Figure 10 shows the power levels at the input of the third line amplifier for the layout shown in Figure 10;

14th Figure 10 shows the signal levels displayed at the input of the fourth line amplifier, also for the arrangement shown in Figure 10,

15th 10A shows the input power levels of the preamplifier of the arrangement of FIG. 10, a

16th 10A shows the preamplifier input power level for the arrangement shown in FIG. 10 using a known amplifier.

The optical amplifier shown in FIG. 1 comprises an erbium-contaminated active fiber 1 and further includes an excitation laser 2 which is connected to the active fiber 1 via a dichroic coupler 3. At the input of the active fiber 1, an optical disconnector 4 is disposed, which is in the direction of the signal transmission, and an additional optical disconnector 5 is disposed after the active fiber 1.

It is desirable, but not necessary, for the dichroic coupling 3 to be located downstream of the active fiber 1, as shown in the flow direction. Thus, this dichroic coupler 3 generates a scintillating energy that flows in the opposite direction of the signal.

The optical amplifier itself contains a second erbium-contaminated active fiber 6 and an associated excitation laser 7 which is also connected to the active fiber 6 by a dichroic coupling element 8 and also generates countercurrent excitation energy.

After the active fiber 6, an additional optical separator 9 is arranged downstream.

The heat detector lasers 2 and 7 can be Quantum Well type excitation lasers with the following parameters:

- emission wavelength λ_ρ= 980 nm;

maximum optical output power_u= 89mW.

For example, lasers of this type are manufactured by LASERTRON Inc., 37 North Avenue, Burlington, MA (US).

The dichroic couplers 3 and 8 are preferably soldered active filament designs which are welded from single-mode active filaments at 980 nm and provide an output power change of less than 0.2 dB in the 1530-1560 nm wavelength range, depending on polarization.

The aforementioned 3,8 dichroic couplings are commercially available, for example, from GOULD Inc., Fiber Optic Division, Baymeadow Drive, Gelm Bumie, MD (US) and SIFAM Ltd., Fiber Optic Division, Woodland Road, Torquay, Devon. (GB).

For optical isolators 4, 5 and 9, polarization control is independent of transmission polarization, provides greater than 35 dB isolation, and has a reflectance of less than -50 dB.

This type of separator may be the MDL I-.15 PIPT-A S / N 1016 separator manufactured by ISOWAVE (64 Harding Avenue, Dover, New Jersey, US).

Fig. 2 shows an optical amplifier according to a further embodiment of the invention, the elements of which are largely identical to the embodiment shown in Fig. 1, except that after the dichroic coupler 3 and before the optical separator 5 there is a comb filter 10 positioned along a transmission line comprising an optical active fiber portion having two optically coupled cores configured for a given and selected wavelength. One core is disposed in one axis with the coupling optical active fiber, the other is disposed eccentrically, i.e., offset centerline, and its ends are cut off. This arrangement is described in more detail in EP 441211 and EP 417441.

The comb filter 10 itself is sized to attach to the offset core a wavelength corresponding to a portion of the amplifier's emission spectrum. This offset core is cut off at its ends, thereby ensuring that the wavelength signals entered here are scattered and cannot return to the first main core.

The parameters of the comb filter 10 with two cores may be, for example:

wavelength range in the second core BW (-3dB): 8-10 nm

- filter length: 35 nm

HU 216 228 Β

The maximum attenuation of the comb filter 10 thus formed is at the peak emission of active fiber.

Another 10-comb filter used in the experiments had the following parameters:

damping λ at 530 nm or 5 dB

damping Z_s 11 dB at 1532 nm.

The purpose of such comb filters 10 is to reduce the intensity of the signals in a given wavelength range primarily at the peak of the optical fiber emission so that the amplification curve of the amplifier is approximately the same at different wavelengths, i.e. "flat".

This requirement is particularly important in the wavelength transmission according to the present invention, where the gain parameters for each channel should be as similar as possible.

Various types of erbium-contaminated active AD fibers have been studied for the optical amplifier of the present invention and their composition and optical parameters are shown in the following table.

First spreadsheet

<td rowspan="2">Active thread</td><td colspan="2">A1₂SHE₃</td><td colspan="2">GeO₂</td><td colspan="2">La2O₃</td><td colspan="2">Er2O₃</td><td>SO</td><td></td>

<td>t%</td><td>(mol%)</td><td>t%</td><td>(% Mol)</td><td></td><td>(% Mol)</td><td>t%</td><td>(% Mol)</td><td></td><td>nm</td>

<td>THE</td><td> 4</td><td> (2,6)</td><td> 18</td><td> (11,4)</td><td> 1</td><td> (0,2)</td><td> 0,2</td><td> (0,03)</td><td> 0,219</td><td> 911</td>

<td>B</td><td> 1,65</td><td> (1,0</td><td> 22,5</td><td> (14,3)</td><td> 0</td><td> (0)</td><td> 0,2</td><td> (0,03)</td><td> 0,19</td><td> 900</td>

<td>C</td><td> 4</td><td> (2,6)</td><td> 18</td><td> (11,4)</td><td> 0</td><td> (0)</td><td> 0,2</td><td> (0,03)</td><td> 0,20</td><td> 1025</td>

<td>D</td><td> 4</td><td> (2,6)</td><td> 0</td><td> (0)</td><td> 3,5</td><td> (0,7)</td><td> 0,2</td><td> (0,03)</td><td> 0,19</td><td> 900</td>

where t% = average weight% of oxide content in the core mol% = molecular distribution of the oxide content in the core NA = numerical aperture (nl²-n2²)^1/2

X_c= cutting wavelength (LP11).

The composition was analyzed by microscopic sampling and scanning electron microscopy prior to drawing the AD strands.

The analysis was performed at 1300 discrete points spaced 200 µm apart.

The AD fibers were prepared by vacuum electroplating where the metal was galvanized into a quartz glass tube.

For AC fibers where the pollutant was germanium, the germanium was₂into a matrix by synthesis.

The introduction of erbium, alumina, and lanthanum into the core was accomplished by so-called solute addition, wherein the aqueous chloride solution of the contaminant was brought into contact with the core material before it had cured.

See U.S. Patent No. 5,282,079 for a detailed description of the contaminant input by dissolution.

Comparing the higher numerical aperture (NA) values ​​in fiber A with the other fibers B, C, D, the flow rate of the reagent used in fiber C (Al / Ge / Er) and germanium was not reduced.

Subsequently, lanthanum and aluminum were added to the solute contaminant, resulting in a higher reflectance of the core than expected, with the unexpected benefit of improving the gain and transmission parameters described below.

For the determination of the emission spectra of AD fibers, the experimental arrangement shown in Figure 3 was used and the emission spectra for one of the active fibers A, B, C and D is shown in Figure 4.

The geodetecting laser 11, which is a laser diode, was excited at 980 nm at 35 wavelengths. This signal was applied via a 980/1550 12 dichroic coupler to the 13 active fibers being tested. The emission spectrum of the active fiber 13 was measured with an optical spectrum analyzer 14.

All geolayer laser in the 13 active fibers generated approximately 60 mW of power. The length of the active fiber 13 is chosen so that it can be matched to the given excitation power to achieve proper amplification. Each of the fibers tested, which contained 45% of the erbium impurity, had to be approximately 11 meters long for proper reinforcement.

The choice of the various erbium content in the 13 active fibers and the appropriate length for reinforcement is obvious to those skilled in the art.

The optical spectrum analyzer 14 was a TQ8345 instrument manufactured by ADVANTEST Corporation, Shinjuku-NS Bldg. 2-4-1.

The measurements were made by excitation of the fibers at 55 980 nm and observed the spontaneous emission spectrum of the active fiber.

In Figure 5, curve 15 is the emission spectrum of fiber A, curve 16 is the emission spectrum of fiber B, curve 17 is the emission spectrum of fiber C, and curve 18 is the emission spectrum of fiber D60.

HU 216 228 Β

It is evident from Figure 4 that the emission peaks and maximum intensities of the fibers B, C and D are around 1532.5 nm. A relatively broad emission range can then be observed at 1560-1565 nm, which represents a second, highly flattened peak.

Comparing the curves 16 and 17 for fibers B and C, it can be seen that the higher alumina content of the fibers B and C increases the level of the high emission range. By replacing lanthanum with germanium, this is shown in curve 18 and refers to fiber D, so that an even higher level can be achieved in the 1535-1560 nm range than in previous cases.

In fibers B, C, and D, a region d between 1535 and 1540 nm was marked with some depression, which was located essentially between the first major peak and the second emission peak. In this depression range, the emission value is at least 2 dB lower than the maximum emission values ​​in the adjacent ranges, which means that the first major peak and the second peak are considered here. This difference is denoted by h and is plotted on curve 16 alone, but can also be identified on curves 17 and 18.

In contrast, curve 15 has shown experimentally that fiber A does not show significant depression in the ad region, or, in some cases, less than 0.5 dB.

Curve 15 also shows that the peak emission of fiber A is at a lower wavelength than that of fibers B, C and D, which is around 1530 nm, which also means that fiber A has a high emission level of 1520 nm. nm near the peak.

Fiber A was used for the optical amplifiers shown in Figures 1 and 2.

The first active fiber 1 was about 8 meters long, while the second active fiber 6 was 15 meters for the embodiment shown in Figure 1 and 13 meters for the embodiment shown in Figure 2.

Figure 5 shows the gain curves at different wavelengths for two different input power levels, in the case of the optical amplifier shown in Figure 1, while Figure 6 also shows the gain curves for the three wavelengths of the optical power shown in Figure 2. amplifier.

Curve 19 in FIG. 5 refers to an input power of -20 dBm, while curve 20 relates to an input power of -25 dBm in the case of the optical amplifier shown in FIG.

In FIG. 6, curve 21 refers to an input power level of -20 dBm for the optical amplifier shown in FIG. 2, curve 22 relates to an input power of -25 dBm, and curve 23 relates to an input power of -30 dBm. Articles 5 and 6 FIG. 6A is particularly noticeable when looking at the curves 19 and 21, which refer to power levels of -20 dBm, and which are of great importance in telecommunication systems for the use of a filter with or without a filter arrangement, containing aluminum oxide, germanium and lanthanum along with erbium, allows for a relatively straight gain curve, particularly in the wavelength range of 1536 and 1540 nm, moreover, the result can be achieved even without applying a filter.

In the case without the filter and at the input power of -20 dBm, the gain difference between the signals of different wavelengths was less than 1.6 dB while in the case of the filter, and also at the input power of -20 dBm the gain difference between the signals of different wavelengths was less than 0, 9 dB.

Figure 7 shows the gain curves for three different input signal levels at different wavelengths for an optical amplifier constructed according to the embodiment shown in Figure 2, using fiber C as the active fiber, which was Al / Ge / Er .

In Figure 7, curve 24 refers to an input power level of -20 dBm, curve 25 relates to an input signal level of -25 dBm, and curve 26 relates to an input signal level of -30 dBm.

It can be seen that at -20 dBm, the gain difference between signals of different wavelengths was about 2.1 dB.

Comparing fiber A (Al / Ge / La / Er) applied to an optical amplifier that has no filter, to an optical amplifier using fiber C (Al / Ge / Er) but equipped with a filter, that using fiber A gives a much smoother gain curve than fiber C.

The optical amplifiers shown in Figures 1 and 2 were performed for both fiber A (Al / Ge / La / Er) and fiber C (Al / Ge / Er) by transmitting the signals over long distances and optical amplifier was switched to cascade mode, ie in series. Such an experimental arrangement is shown in Figure 8. 8, a transmitter 27 generating two λ wavelength input signals to an input of a multiplexer 30 and an λ₂28 transmitters generating a wavelength input signal were connected.

A λ! = 1536 nm while λ₂= 1556 nm.

The output of the multiplexer 30 is led through an active fiber 29 to a power amplifier 32a, after which a damping member 31 is disposed downstream. The arrangement itself consisted of alternating line amplifiers 32, 32 ', 32 "and 32'" in series and spaced damping members 31, and finally, after the last damping member 31, the arrangement was connected to the receiver 33 via an optical demultiplexer 34.

The optical demultiplexer 34 coupled in front of the receiver 33 contained an interference filter having a bandwidth selected at 1 nm at -3 dB.

EN 216 228 Β, and was always used to select the wavelength.

The signals of the respective transmitters 27 and 28 produced by the lasers, not shown in the figure, were 0 dBm. The total multiplexer power in the 29 active fibers was also 0 dB, as a result of a decrease in the 3 dB connection power.

The multiplexer 30 is a "1x2 coupler" manufactured by E-TEK DYNAMICS Inc. (1885 Lundy Aven, San Jose, CA (US)).

The power amplifier 32a was a commercially available optical fiber optic amplifier and had the following parameters:

- Input power -5 - + 2 dBm,

- output power 13 dBm,

operating wavelength 1530-1560 nm.

The power amplifier 32 did not include a 10 comb filter.

The applicant's TPA / E-12 product was used. The amplifier contained fiber C (Al / Ge / Er), thus 1 active fiber contaminated with erbium.

The power amplifier 32a is designed to operate under saturation conditions, i.e., the output power depends on the excitation power as described in EP 439 867.

The first damping member 31 is connected in series to the line amplifier 32 having a total optical power of about -18 dBm.

The Va5 can be used as a damping member manufactured by JDS FITEL Inc. (570 Heston Drive, Nepean (Ottawa), Ontario (CA), Canada) and these 31 damping members are those with 30 dBes of attenuation for a 100km optical fiber .

The 32, 32 ', 32 "and 32'" line amplifiers are of the same design and each has a total output power range of +12 dBm_Η all of λ₂gain of approximately 30 dB at wavelength.

The transmitter 27 has a signal having λ₍ wavelength 1536 nm, directly modulated by a 2.5 Gbit / s signal generated by a DFB laser located in a SLX-1/16 Model SDH terminal and supplied by PHILIPS NEDERLAND BV, 2500BV Gravenhage (NL) produce.

Transmitter 28 has a signal having λ₂ had a wavelength of 1556 nm, a continuous signal (CW) generated by a DFB laser of type MG0948L3 with a power of 0 dBm produced by ANRITSU Corporation, 5-10-27 Minato-ku, Tokyo (JP), for example, in Japan. The optical demultiplexer 34 is an interference filter type TB4500 manufactured by FITEL, Inc.

First experiment

For the first experimental 32 line amplifier, 1 or 6 active fibers were used as fiber A (Al / Ge / La / Er)

First that is, the comb filter 10 was not used.

Second experiment

The second experimental line amplifier 32 was constructed as shown in Figure 2, using fiber A (Al / Ge / La / Er) as active fibers 1 and 6, respectively, and using a 10-comb filter.

Through the 33 receivers shown in Fig. 8, the bit error rate (BIR) was measured at various average reception powers, λ! = 1536 nm.

The measurement result is shown in Figure 9, where curve 35 refers to Experiment 1 and curve 36 relates to Experiment 2.

As shown in Fig. 9, although the gain curve of a single line amplifier 32 formed with a fiber A (Al / Ge / La / Er) formed by a comb filter 10 is substantially the same when the comb filter 10 is not used, , 32 'line amplifiers are cascaded, with a significantly higher bit error rate at 1536 nm when the reception power was the same.

Third Experimental layout

The second experimental set-up, shown in Figure 10, was used. In this case, four λ_Η λ₂, λ₃ and λ₄ wavelength transmitter signals 37, 38, 39 and 40 are routed through a pre-equalizer 43 to a multiplexer 42. A λ, 1536 nm, aX₂ 1556 nm, α₃ 1500 and AA₄ 1544 nm. The signals from the multiplexer 42 were routed through the fiber 41 to the power amplifier 44.

The signal level at the line input is set at the pre-equalizer 43. After the power amplifier 44, the signal was amplified by four serially connected line amplifiers 45, 45 ', 45', 45 '', with a damping member 46 between each output and the next input.

The damping member 46 connected to the line amplifier output 45 '' was led to an input 47 of the preamplifier output to the receiver 49 via a demultiplexer 48.

The signals were generated with a DFB laser at 1536 nm and directly modulated at 2.5 Gbit / s, which was part of the 49 receivers. For the 1556 nm wavelength, a continuous emission DFB laser manufactured by ANRITSU was used. The 1550 nm wavelength DFB laser is also a continuous emission type and is also manufactured by ANRITSU. An additional ECL laser with variable wavelengths was used and used₄= 1544 nm, the latter also of continuous emission type, such as HP81678A manufactured by HEWLETT PACKARD Company, Rockwell, MD (US).

Returning to FIG. 10, the pre-equalizer 43 is configured with four variable damping members 43 manufactured by JDS, for which each damping value is set depending on the optical power of the respective channels.

The multiplexer 42 connected to the output of the pre-equalizer 43 is an 1x4 bundle manufactured by E-TEK DYNAMICS.

The 44 power amplifier is a commercially available TPA / E-13 product, as previously mentioned.

The 45, 45 ', 45 "and 45'" line amplifiers are identical, each with a gain of about 30 dB, + 12 dBm total output power.

HU 216 228 Β

Line amplifiers 45 were designed as active fibers 1 and 6, respectively, using fiber A (Al / Ge / La / Er) as shown in Figure 1.

The damping members 46 connected to the outputs of the power amplifier 44 and the line amplifiers 45, 45 ', 45 "and 45"' are also of the same design, each damping approximately 30 dB, which corresponds to about 100 km of optical fiber.

The optical damping members are of the VA5 type, also manufactured by JDS FITEL.

The preamplifier 47 is a commercially available optical amplifier having the following parameters: gain 22 dB noise factor &lt;4.5 dB output power -26 to 11 dB operating wavelength range 1530-1560 nm.

The RPA / EF is commercially available and is manufactured by the applicant company. For line amplifier 46, C fiber was used as 1 or 6 active fibers which were Al / Ge / Er contaminated.

The function of the preamplifier 47 is that when the equipment receives a signal, its amplification is very low, e.g., -50 dBm, and then amplified before being transmitted to receiver 49, and the gain change is always such that the receiver 49 is fully rendered fit.

The optical demultiplexer 48 is a wavelength tuned Fabry-Perot filter having a bandwidth of 0.8 nm at -3 dB, which is substantially incorporated into the preamplifier 47 and is a portion thereof.

An experiment was performed when the FabryPerot filter was tuned to λ = 1536 nm, identified as a critical wavelength, and tuned with a pilot tone generated by transmitter 37.

The 49 receivers were an SDH terminal, type SLXX-1/16, commercially available and manufactured by PHILIPS NEDERLANDS BV, 2500BV in the Netherlands.

11-15. FIGS. 4A to 5B show the signals at the inputs of the line amplifiers 45,45 ', 45 ", 45' 'and the preamplifier 47 respectively.

The pre-equalizer 43 performs an initial equalization of up to 7 dB between the various channels as described below. is shown. The purpose of this pre-equalization is to compensate for the saturation limit at low wavelengths in the cascade amplifier arrangement.

The pre-equalization was performed by adjusting the optical signal-to-noise ratio (S / N) at the output of the preamplifier 47.

The successive line amplifiers 45, 45 ', 45 "show that the gain curve decreases in the lower wavelength ranges due to the saturation phenomenon described above, while the optical signal-to-noise ratio for each channel remains relatively high up to the output of the preamplifier 47. (S / N> 15 dB, Δλ = 0.5 nm).

A further experiment was carried out using as an amplifier the embodiment shown in Figure 2, where fiber C was used as active fibers 1 and 6, respectively, and comb filter 10 was used. In this experiment, it was observed that the signal power at the wavelengths of 1536 nm and 1544 nm was greatly reduced, and there was also a significant imbalance in the signal / noise conditions of the various channels, as shown in Figure 16. The 16th. FIG. 6A shows the power of the various channels, and a very strong junction is observed at the input of the preamplifier 47 at the channel of 1540 nm.

In this case, the pre-equalization prevents the asymmetry between the different channels, which in effect means that between the channels, which are primarily between 1535 and 1540 nm, would place one at a significantly disadvantageous position relative to the other. However, in this way, an acceptable signal-to-noise ratio for all signals in the wavelength range in question cannot be obtained that can be maintained. If pre-equalization was performed for the channels, the most favorable channel for the 1550 and 1556 nm would have a very large initial attenuation which would produce a very low signal-to-noise ratio (of the order of 8-10 dB). In this way, however, it would be virtually impossible to receive signals.

The better results obtained with 10 comb filters and Al / Ge / Er impurities 1 and 6 are due to the fact that the emission curve of fiber A is practically free of local minima and depressions, or its value is insignificant, especially in the ranges close to the emission peak, i.e. 1535-1540 nm.

In practice, it has been found that, when signals of different wavelengths are fed simultaneously to the active fibers 1 and 6, the depression or local minima of the emission curve at the wavelengths where the depression is present, amplify the signal to a lesser extent than in the adjacent regions. signals.

From the foregoing, it is obvious that the higher gain factor in these adjacent wavelength ranges removes gelling energy from the signal, so that the signal is saturated at a relatively low output level, i.e., after amplification, its level no longer depends on the input signal value, , and thus the difference in level between the various signs gradually increases.

For cascade 32-32 '' and 45-45 '' line amplifiers, this phenomenon increases step by step in wavelength transmission, precisely because of perceived response differences and cannot really be compensated for during pre-equalization or the like.

Experience has shown that this phenomenon is present at the signals at the depression points of the emission curve and is a consequence of the amplification competition for wavelength ranges near the depression wavelengths, but this phenomenon does not occur, or at least to a small extent, for signals within the usable signal range.

However, for these wavelengths, the emission value may, in absolute terms, be less than or equal to the degree of depression.

Surprisingly, the fact that even lanthanum impurities have been added to the Al / Ge / Er-contaminated fibers 1 and 6 has contributed to the reduction of local emission minima, although this is the case with Al / La / Er and Al / Ge / Er fiber data was not expected.

Both Al / La / Er and Al / Ge / Er-contaminated active fibers 1 and 6 have substantial minimum emission in the 1535-1540 nm range, thus, in the light of the above phenomena and knowing these active fibers, Al / Ge / La / Er fibers are not expected to exhibit particularly advantageous behavior, especially not in enhanced wavelength transmission.

Unexpectedly, we made a further discovery. It has been found that the peak value in the high emission ranges and the depression or at least the decreasing value of the adjacent ranges near this peak are responsible for the inadequate signal-to-noise ratio of the signals in the depression range, and 6 active fibers that can somewhat limit or reduce this depression, a problem with wavelength data transmission with one or more 32, or 45 line amplifiers.

Thus, the present invention utilizes active fibers 1 and 6 which have contaminants which exhibit an emission curve having a relatively high value over a given wavelength range but which are free of local depression or there is no local depression in the wavelength ranges near this range. This would create a very large gain difference for systems of different wavelengths in a given band if the signals of different wavelengths are multiplexed in the fiber. Thus, experience has shown that such optical amplifiers are well suited for telecommunication lines when at least two amplifiers are connected in series and the signals are transmitted in wavelength division multiplexing.

A further advantage of the present invention is that, in the above transmission systems, proper control of the signal-to-noise ratio can be accomplished not only by the use of filters or by appropriate reduction of the transmission band, thus excluding unfavorable wavelength ranges, by selecting and dosing the appropriate pollutant in the nucleus so that the emission curve is wide enough within the band. This means that it should be suitable for at least 1525-1560 nm, or at least 1530-1560 nm. In this case, no unwanted signal amplification occurs in one or more regions of the emission curve, although the peak emission is in that region.

This is explained by the fact that the presence of a higher emission near the depression ranges, or possibly an emission peak, and the presence of signals in these adjacent ranges results in a slight suppression of the signal amplification in the region near the depression ranges.

For emission curves or spectra whose peak value is relatively large in the range in question and which we use, preferably between 1525 and 1560 nm, the excited 1 and 6 active fibers should have emission curves with emission values ​​outside of that band. are also large enough to ensure that the signal is amplified in a given wavelength band. Such a range is generally defined between two end values ​​where the end values ​​are the values ​​at which the emission is 3 dB lower than that range. However, it is preferred that the emission in this range is approximately constant. This is the range that essentially corresponds to where good gain can be achieved.

By peak emission, we mean that the emission in a given wavelength range or wavelength is significantly higher than in the out-of-range wavelengths, which also means that it behaves differently for signals with wavelengths within and outside this range. fall.

Significant local depression is defined as the portion of the emission curve that has a second emission minimum within a given and used wavelength band, which is lower than either the emission value at any of the thresholds in the above wavelength band or a predetermined value at the emission value maximum. smaller in adjacent wavelength ranges. The main emission peak for erbium is lower than their depression minimum at both thresholds, and such as the secondary emission peak at higher wavelengths.

In the present invention, those predetermined depression values ​​that are greater than 0.5 dB, and in particular those greater than 1 dB, already have a noticeable effect.

It has been found that when line amplifiers 32, 32 'and 45, 45' are coupled together as a cascade amplifier, the application of a comb filter 10, which can reduce the intensity of the main emission peak by 45 'provides a substantially flat gain curve for line amplifiers without allowing the above phenomenon to be eliminated.

The application of the comb filter 10 to multiple cascade amplifiers in the lower wavelength ranges, where it has a midpoint or a center frequency, acts as a damping element, causing the depression range of the emission curve to be slightly broadened. This damping effect is added to the aforementioned saturation phenomenon and further disadvantages11

EN 216 228 Β for signals with a wavelength within this range of depression or a local minimum.

The use of other similar filters that can be used to attenuate or otherwise reduce emissions at the peak emission levels, as described, for example, in EP 426222, does not make a significant difference in eliminating these phenomena.

Preferably, lanthanum is present in the core of active fibers 1 and 6, respectively, in a concentration greater than 0.1 mol%, germanium is preferably greater than 5 mol%, and the ratio of germanium lanthanum is preferably about 50, but must be between 10 and 100.

By having lanthanum in the core of the active fibers 1 and 6, it is possible to introduce greater amounts of germanium and alumina as impurities, thus having a larger numerical aperture, generally greater than 0.18, preferably at least 0.2. value, which is a very significant advantage in that the gain of the gain is better and that the response signal can be kept more constant in a given band.

In addition, the presence of lanthanum allows us to increase the erbium content of the active fibers 1 and 6, respectively, without being associated with compaction. The erbium content is preferably selected in the range of 20-5000 ppm, more preferably, the erbium content is 100-1000 ppm.

The optical amplifier of the present invention has been described as 32 and 45 line amplifiers, however, the active fibers 1 and 6 of the present invention may also be used in preamplifiers and are particularly suitable for receiving and amplifying signals of very low intensity, e.g. before being transmitted to receivers 33 and 49, respectively.

Preferably, the optical amplifier is designed as a two-stage amplifier, and optionally one of these amplifiers is formed by the active fibers 1 and 6 according to the invention.

It will be readily apparent to those skilled in the art that the operating parameters are determined, and that the operating parameters are determined to determine the pollutant content so that the desired results are present in the output signal.

According to the description of the invention, one skilled in the art will design an active fiber 1 and 6 which contains a major impurity, preferably erbium, since it is well-suited for telecommunication systems and which fluoresces in the region of interest and a secondary, contains the pollutant that cooperates with the previous one, you can easily create, determine how much, and the proportion of pollutants to be fed to the active fibers 1 and 6, respectively, so as to produce appropriate changes in the emission curve, to provide an optical amplifier with appropriate parameters, and to provide a system where in the range the signal / noise ratio is good.

In the development of the present invention, the research was primarily conducted with erbium as the main impurity and fluorescent material, germanium, aluminum and lanthanum being present in the fiber as an oxide as a secondary impurity, since experiments have shown that in order to address this objective Min.

The description of the invention will also allow one skilled in the art to solve their own problems, which may or may not be the same as those of the invention, but will provide clear guidance on the various contaminants and their dosage.

One skilled in the art will further advise that the use of the individual pollutants, or combinations of individual pollutants, according to the present invention, unless otherwise applied, may not produce satisfactory results, particularly with regard to the signal-to-noise ratio.