CN1638630A - Communication system and sub-band amplifying device using depressurized optical fiber amplifier - Google Patents

Abstract

A communication system and a sub-band amplifier using a depressed type optical fiber amplifier including an optical sub-band amplifier (212) for amplifying at least a part of S-band and C-band or L-band, the amplifier including a first portion (212A) and a second portion (212B) for amplifying long and short wavelength bands, respectively, the second portion (212B) including a short pass optical fiber (10), a core (12) doped with an active material such as erbium, a depressed inner cladding (14) and an outer cladding (16), the refractive index profile of the second amplifier generating a loss comparable to a high gain in the long wavelength band and a loss much smaller than a positive gain in the short wavelength band.

Description

Communication system and sub-band amplifying device using depressurized optical fiber amplifier

technical field

The present invention generally relates to a communication system and a sub-band amplifying device for amplifying optical signals. The wavelength bands of optical signals cover short-wave bands such as S-band and long-wave bands such as C-band. Communication system and sub-band amplification device.

This application relates to Application No. 10/095,303, filed March 8, 2002, Application No. 10/186,561, filed June 28, 2002 and Application No. 10/346,960, filed March 8, 2003 U.S. Patent dated 17th.

Background technique

The development of erbium-doped fiber amplifiers has successfully solved the problem of optical signal amplification for long-distance transmission. Erbium-doped fiber amplifier consists of a section of silicon fiber doped with ionized atoms (Er ³⁺ ) of pure erbium element in nature in the core. The fiber is pumped with a laser with a wavelength of 980nm or 1480nm. The pumped erbium-doped fiber is optically coupled to the delivery fiber to combine the input optical signal with the pump signal from the erbium-doped fiber. An isolator is generally required at the input and/or output to prevent light reflections that would turn the amplifier into a laser. Early erbium-doped fiber amplifiers provided 30 to 40dB of gain in the C-band extending between 1530-1565nm and a noise figure below 5dB. Recently, erbium-doped fiber amplifiers have been developed to provide similar performance to C-band in the L-band (1565 to 1625nm).

Of most interest is the development of a broadband amplifier capable of amplifying shorter wavelength optical signals spanning the C-band and L-band and the so-called "S-band" or "short-waveband". Although not currently identified, the S-band is believed to cover wavelengths between approximately 1425nm and 1525nm. The typical S-band gain observed in erbium-doped fiber amplifiers is limited by several factors, including incomplete inversion of active erbium ions and spontaneous emission or excitation from the high-gain peak around 1530 nm. Unfortunately, there are currently no effective means for suppressing spontaneous emission at 1530 nm or longer in erbium-doped fiber amplifiers.

The prior art provides various types of waveguides and fibers that can be used to fabricate erbium-doped fiber amplifiers. Most waveguides are designed to prevent coupling of light inserted into light exiting through the device, such as decoupled evanescent waves (tunneling), scattering, bending losses, and leaky mode losses. A general study of these devices can be found in papers such as "Radiative Leaky Mode Losses in Single-Mode Optical Waveguides with Depressed-Index Cladding" by L.G Cohen et al., published in the IEEE Journal of Quantum Electronics QE-18, No. 10, 1982, pp. 1467-72. U.S. Patents 5,892,615 and 6,118,575 describe the use of a W-shaped fiber similar to that described by L.G Cohen, the QC fiber, which suppresses unwanted frequencies to obtain high output in cladding-pumped lasers power. As discussed above, such fibers typically leak light at long wavelengths and are more sensitive to bending than other fibers.

The higher loss and lower gain over the entire S-band makes the choice of fiber and fiber shape more difficult when manufacturing S-band erbium-doped fiber amplifiers. In fact the problem is so severe that the prior art describes an Erbium-doped fiber amplifier with internal filters inserted between its parts to make an S-band Erb-doped fiber amplifier. For example, Ishikawa et al. disclosed a method in their "Novel 1500nm Band Erbium-Doped Fiber Amplifier with Discrete Raman Amplifier" "ECOC-2001 Final Paper", by stacking five levels of silicon-based erbium-doped fiber Amplifier and four filters to suppress spontaneous emission to fabricate S-band Erbium-doped fiber amplifier. In the experimental setup of Ishikawa et al., each erbium-doped fiber amplifier was 4.5 meters long. The absorption of each rejection filter at 1.53 μm is about 30 dB, while the insertion losses of each rejection filter at 1.48 μm and 0.98 μm are about 2 dB and 1 dB, respectively. The pumping mode is bidirectional, using a wavelength of 0.98μm to maintain a high inversion, D≥0.7 (D is the inversion rate). The front and rear pump powers are the same, and the total pump power is 480mW. Ishikawa et al. show a maximum gain of 25dB at 1518.7nm with a gain slope of 9dB.

This approach is relatively complex and cost-effective, as it requires five erbium-doped fiber amplifiers, four filters to suppress spontaneous emissions, and high pump power. Furthermore, each spontaneous emission suppression filter used in the method of Ishikawa et al. produces an additional insertion loss of 1-2 dB. In this way, the total additional insertion loss is about 4-8dB.

In US Patent 6,049,417, Srivastava et al. describe a broadband optical amplifier using a sub-band structure. This amplifier first divides an optical signal into several independent sub-band optical signals, and then each sub-band optical signal passes through each branch of the optical amplifier in parallel. Each branch is designed to be an optimal transmission channel for the optical signals of their respective sub-bands. In one embodiment, people such as Srivastava have described that in order to obtain the amplification of the S-band, there are a plurality of S-band erbium-doped fiber amplifiers and a plurality of gain equalization filters arranged in a plurality of branch paths, and these gain equalization filters Set between each S-band Erbium-doped fiber amplifier. In this embodiment, other sub-bands are provided with erbium-doped fiber amplifiers for amplifying C-band and L-band respectively. Unfortunately, the S-band branch of this broadband amplifier suffers from similar drawbacks to the amplifiers previously discussed by Ishikawa et al.

Another approach to provide S-band amplification that has attracted attention is fiber amplifiers (TDFAs) that incorporate thulium as the active medium in the core of a fluorinated fiber optic. For example, see "Gain of WDM Signals for 1.48-1.51nm Wavelength Range" published by Tadashi Kasamatsu et al. in IEEE "Photonic Technology Index", February 2001, Vol. 13, No. 1, pp. 31-31 Floating, Dual-Wavelength Pumped Thulium-Doped Fiber Amplifier". Good optical properties have been obtained using thulium-doped fiber amplifiers, but this property is only possible using complex, non-standard and/or expensive pumping configurations. Furthermore, thulium-doped fiber amplifiers suffer from the problems inherent in fluorinated fibers as host materials, namely high fiber cost, poor reliability, and difficulty in splicing with standard silica fibers used in optical amplification systems elsewhere.

Optical amplifiers (such as Erbium-doped fiber amplifiers, Thulium-doped fiber amplifiers, Raman amplifiers, semiconductor amplifiers, etc.) have various uses in communication networks. Its most important use is to compensate for transmission loss (fiber loss accumulated during transmission of tens or hundreds of kilometers). In this case, a typical amplifier is called an in-line amplifier. In-line amplifiers must deliver low to moderate optical power per optical channel (typically 0.1-10 mW), and must also exhibit low noise figure and good gain flatness in WDM networks. The latter two requirements arise from the cumulative effect of long cascaded amplifiers connected over hundreds to thousands of kilometers of long optical fiber.

Optical amplifiers are also used as preamplifiers. A typical application of a preamplifier is to improve the sensitivity of a receiver, and its use is well known in the art. Usually a preamplifier is placed in front of the signal receiver to boost the signal strength (optical power) above the (electronic or thermal) detector noise level. The preamplifier is primarily used to amplify one or a small number of optical channels, and while it does not need to operate at high power or have a flat gain curve, it must exhibit an excellent noise figure.

Optical amplifiers are also used as power amplifiers. It is well known in the art that power amplifiers are used to provide high optical power. Usually they work with relatively high input signal strength (such as saturation), have a good signal-to-noise ratio, and therefore don't need an excellent noise figure. Usually there is no need to have a particularly high gain. A large number of WDM channels today use power amplifiers, even when moderate power is required per channel. Power amplifiers are also used ahead of long/lossy connections to pre-compensate for expected losses.

Finally, optical amplifiers have many other uses in communication networks. Such as: increasing the power before or after splitting the signal into multiple parallel output signals, power compensation for lossy network modules such as cross-connects and switches, providing high enough optical power to pump nonlinear devices such as optically driven Optical switches or optical wavelength converters.

From the foregoing, it would be an advance in the art to provide a broadband amplifier that amplifies optical signals spanning the S-band, C-band, and L-band and exhibits high efficiency in the S-band. It would be an improvement, inter alia, to provide a broadband amplifier which amplifies S-band optical signals without the use of many filters and which takes advantage of a minimum number of pump sources. It would be another advance in the art to provide an optical communication system capable of amplifying S-band signals using erbium-doped fiber amplifiers. It would be a particular advance to provide a low-cost S-band erbium-doped fiber amplifier for use in such optical communication systems to achieve low-cost preamplification, power boost, and in-line amplification.

Contents of the invention

The main purpose of the present invention is to address the above-mentioned shortcomings of the prior art and provide a broadband amplifier for optical signals spanning long-wave bands such as C-band and/or L-band and short-wave bands such as S-band.

A particular object of the present invention is to provide a broadband amplifier which enables the use of erbium-doped fiber amplifiers with efficient pump configurations.

It is a further object of the present invention to provide a broadband amplifier using a minimum of components, especially in the S-band.

Another object of the present invention is to provide a method of designing a broadband optical fiber amplifier.

Still another object of the present invention is to provide an optical communication system capable of transmitting and amplifying S-band signals. In particular, optical communication systems use an erbium-doped fiber amplifier (EDFA) to amplify signals containing the S-band in a controllable manner to perform preamplification, power boosting, and in-line amplification.

Numerous advantages of the present invention will be apparent from the following description.

The objects and advantages of the present invention are obtained in that the sub-band amplifying device is provided with a first part for amplifying the long-wave band and a second part provided with a fiber amplifier for amplifying the short-wave band. The fiber amplifier in the second part has a core with a core cross-section and a refractive index n ₀ . The core is doped with active substances. A depressed cladding is wrapped around the fiber core, which has a depressed cladding section and a refractive index n ₁ , and a second cladding is wrapped around the depressed cladding, which has a second cladding section and a refractive index n ₂ . A pump source is set to pump the active material to a high population inversion D, so that the active material presents a positive gain in the short-wave band and a high gain in the long-wave band. The core cross-section, depressed cladding cross-section and refractive indices n ₀ , n ₁ and n ₂ are selected to produce a roll-off loss curve with respect to the cut-off wavelength λ _c that produces at least comparable high gain at long wavelengths Compared with the loss, the loss in the short-wave band is much smaller than the positive gain. In a preferred embodiment the active material is erbium, ie the fiber amplifier is a first erbium doped fiber amplifier (EDFA).

Short and long bands can be selected according to the application. For example, in telecommunications, the short band may be chosen to include at least a portion of the S-band, and the long wave band may be selected to include at least a portion of the C-band. When using this band selection, the cut-off wavelength λ _c is set to be the crossover wavelength between the S-band and the C-band, which is about 1530 nm. Of course, the long-wave band can include a wider wavelength range, for example, it can include at least a part of the L-band.

In a preferred embodiment, in the band amplifying device designed to amplify the long-wave band, the first part has a second erbium-doped fiber amplifier, so that both the first and second parts can use the second and first doped fiber amplifiers respectively. Erbium fiber amplifiers are used to amplify optical signals. Moreover, the first and second erbium-doped fiber amplifiers can share a common pump source to provide pump radiation. For example, the common pump source can be a laser diode providing 980nm pump radiation.

Sub-band amplification devices can be designed in many different ways. In one embodiment the first and second parts of the sub-band amplifying means share an overlapping part. In this embodiment, the overlapping portion may contain a second erbium-doped fiber amplifier for amplifying long-wavelength optical signals. Alternatively, the first and second parts may be separated from each other to form multiple independent branches of the device.

According to the method of the present invention, the sub-band amplifying device is used to amplify the optical signal spanning from the short-wave band to the long-wave band. Co-pumping and/or counter-pumping the erbium-doped fiber amplifiers is convenient when the first and second parts of the device use erbium-doped fiber amplifiers to amplify the optical signal. The co-pumping and/or anti-pumping of the Erbium-doped fiber amplifier can come from the same pumping source or multiple independent pumping sources.

The present invention further provides an optical communication system employing a signal source to provide a signal at an S-band wavelength, or simply S-band. The communication system includes an optical fiber carrying the S-band and one or more optical amplifiers, typically in the form of optical fiber amplifiers. A fiber amplifier has a core defined by the core cross-section and the refractive index _n0 . Erbium is doped into the core of the fiber amplifier to amplify the signal. The fiber amplifier has a depressed cladding surrounding the core and a second cladding surrounding the depressing cladding. The depressed cladding has a depressed cladding cross section and a refractive index n ₁ , and the second cladding has a second cladding cross section and a refractive index n ₂ . Fiber amplifiers have a pump source for pumping erbium contained in the fiber core to a high energy level relative population inversion D; in this case, erbium amplifies the signal. In particular, pumping causes Erbium to exhibit a positive gain in the S-band and a high gain in the long wavelength bands longer than the S-band, ie the C-band and L-band. The core cross-section, depressed cladding cross-section and refractive indices n ₀ , n ₁ , n ₂ are chosen to obtain at least comparable losses at long wavelengths with high gains and much less than positive gains at S-band .

In optical communication systems, fiber amplifiers can be used as preamplifiers, power amplifiers, or in-line amplifiers. The optical communication system may be a wavelength division multiplexing communication system, such as Dense Wavelength Division Multiplexing (DWDM). The wavelength division multiplexing system has a communication optical fiber, which is typically a long section or long span of communication optical fiber, which is located between the WDM multiplexer and the WDM demultiplexer of the wavelength division multiplexing system. The multiplexer in the wavelength division multiplexing system multiplexes multiple information-carrying signals and sends them from the transmitting end to the communication optical fiber. The demultiplexer in the wavelength division multiplexing system demultiplexes the signal arriving at the receiving end of the communication fiber. When the fiber amplifier is used as a preamplifier, it is generally installed after the demultiplexer of the wavelength division multiplexing system. Optical fiber amplifiers can also be installed between multiplexers and demultiplexers in wavelength division multiplexing systems, and used as power amplifiers and online amplifiers. Of course, the fiber amplifier can also be used as a preamplifier in this position.

In multiplexed systems, such as wavelength division multiplexed communication systems, the signal source typically includes lasers from a laser array. The pump source providing the population-inverted pump radiation of erbium ions may be any suitable pump source. For example, the pump source is a laser diode emitting pump radiation around 980nm. It may also be a pump source providing pump radiation around 980 nm or other erbium pump bands, which have shorter wavelengths than those contained in the S-band.

The detailed description and preferred and alternative embodiments of the present invention are described below with reference to the accompanying drawings.

Description of drawings

Fig. 1 is the schematic diagram of depressing optical fiber of the present invention and its conduction mode and non-conduction mode;

Fig. 2 is a typical refractive index distribution curve diagram of the optical fiber of Fig. 1;

Fig. 3 is the schematic diagram of x of various different parameters ρ as proportional S function;

Fig. 4 is the roll-off loss curve figure of the suitable erbium-doped fiber amplifier that suitably selects the core refractive index obtained among the present invention;

Fig. 5 is the axial view of the S-band erbium-doped fiber amplifier that the present invention adopts;

Fig. 6 has a S-band erbium-doped fiber amplifier and a C-band erbium-doped fiber amplifier and has a schematic diagram of a sub-band amplifying device of an overlapping portion;

7 is a schematic diagram of a sub-band amplifying device with non-overlapping first and second parts;

Figure 8 is a schematic diagram of using an erbium-doped fiber amplifier in an optical communication system;

Fig. 9 is a schematic diagram of using an erbium-doped fiber amplifier as a preamplifier at the receiving end of an optical communication system;

Fig. 10 is a schematic diagram of a part of an optical communication system using an erbium-doped fiber amplifier to amplify a sub-band of the S-band.

Detailed ways

Comparing FIG. 1 and FIG. 2 , the present invention can be well understood by reviewing the generation principle of the roll-off loss curve of the depressed or W-shaped optical fiber 10 . 1 is a schematic illustration of a partial cross-section of an optical fiber 10 having a core 12 surrounded by a depressive cladding 14 . The depressed cladding 14 is surrounded by a second cladding 16 . The core 12 has a circular cross-section, and the cross-sections of the depressed cladding 14 and the second cladding 16 are also circular. The range of region I corresponding to the core 12 is 0≤r≤r ₀ ; the ranges of regions II and III occupied by the depressed cladding 14 are r ₀ ≤r≤r ₁ and r≥r ₁ . The core 12 has a refractive index n ₀ , the depressed cladding 14 has a refractive index n ₁ , and the second cladding 16 has a refractive index n ₂ . The figure above the partial cross-section of the optical fiber 10 explains that the average refractive index profile 20 in the optical fiber 10 is defined as a W-shaped profile. The optical fiber 10 is a single-mode optical fiber in this embodiment.

An active substance 18 is doped in the core 12 of the optical fiber 10 . The active material 18 is an excitation medium, such as rare earth ions or any other lasing material that exhibits high gain at long wavelengths and positive gain at short wavelengths. In particular, when pumped to a high relative inversion D, the high gain of the active material 18 in the long-wavelength band results in amplified spontaneous emission (ASE) or lasing that reduces the population inversion of the active material 18 and thus reduces the short-wavelength band Positive gain, resulting in the inability to effectively amplify short-wavelength optical signals.

A fiber amplifier may contain any suitable active medium in its active core. For example, the active core can be doped with neodymium, erbium or thulium ions. When erbium is used, the fiber amplifier is an erbium-doped fiber amplifier, and in a preferred embodiment the cut-off wavelength λ _c is located around 1525 nm. Thus the erbium-doped fiber amplifier is pumped by radiation provided by a pump source with a pump wavelength of around 980nm. In this case, an erbium-doped fiber amplifier can be used to amplify signals falling into the short-wave region within the S-band.

Another example is the incorporation of thulium into fused silica fibers. Although thulium's typical gain is believed to be at 1.9 microns, and that is indeed where its gain peaks, the wavelength range over which gain can be obtained extends from 1.5 microns to 2.1 microns. The typical pump wavelength for thulium is 0.78 microns. Thulium can also be pumped by light at a wavelength of 1.48 micrometers, although this requires very high light intensities, perhaps as high as 100 milliwatts. A light intensity of 100 milliwatts at 1.48 microns is easily obtained by pumping at around 500 milliwatts at 1480 nm and nearby wavelengths using commercially available high quality diodes. Another good pump wavelength is at 1530nm, where high power sources above several watts are available.

The gain cross-section and higher laser level lifetime of thulium ions are similar to those of erbium ions, making them convenient for fabrication of 1.5 micron amplifiers. In this way, the threshold for gain is similar - requiring a few milliwatts of pump power.

Thulium ions can be applied to the shortwave end of its gain region in exactly the same way as erbium ions. By pumping with a strong pump source (about 30 mW), population inversion can be obtained even at short wavelengths. However, there is overwhelming hyperfluorescence around 1.9 microns before high gain is achieved at short wavelengths such as 1.6 microns.

If the fiber is designed with a fundamental mode cutoff between 1.9 microns and the short wavelength at which it is desired to operate, and if the cutoff is chosen such that the increase in loss at long wavelengths exceeds the increase in gain due to the higher cross section, the fiber Can be used to make useful amplifiers for short wavelengths. This technique makes it possible to fabricate useful amplifiers in the wavelength region between approximately 1.6 microns and 1.8 microns. Due to the high transmissibility of telecommunication fibers in this region, high expectations are expected for amplifiers operating in this region.

Figure 2 shows a W-shape 20A obtained using common manufacturing techniques. It is sufficient for the purposes of the present invention that the radially varying refractive index of the core 12 have an average value equal to n ₀ . It is also sufficient to depress the refractive indices of the cladding 14 and of the second cladding 16 on average to the values n ₁ and n ₂ . The average refractive index n ₀ of the core 12 is slightly higher than the refractive index n ₁ of the depressed cladding 14 and the refractive index n ₂ of the second cladding 16 . Appropriate values of the refractive indices n ₀ , n ₁ , n ₂ and radii r ₀ , r ₁ , r ₂ are chosen to obtain defined conduction characteristics of the optical fiber 10 as required by the present invention. In particular, the curve 20 is designed to have a fundamental mode cut-off wavelength _λc , so that at wavelengths smaller than _λc , fundamental mode light is obtained in the fiber core 12, and fundamental mode light at wavelengths λc or longer is at Losses are in the second cladding layer 16 for short distances. This is accomplished by properly designing the W-curve 20A.

The fundamental mode cutoff wavelength λc of the optical fiber 10 is a wavelength at which the fundamental mode (LP ₀₁ mode) changes from low loss to high loss in the core 12, ie, is cut off in the core 12. First, the fundamental mode cut-off wavelength λc of the optical fiber 10 is set according to the law of selection of the cross section and the refractive indices n ₀ , n ₁ , n ₂ obtained from Maxwell's equations. Using the weakly guided approximation (which is valid when the indices of refraction of the core 12 and cladding 14, 16 are all close to each other), the Maxwell vector equations can be replaced by scalar equations. The scalar ψ represents the strength of the transverse electric field in the fiber. For more information, see, for example, G. Agrawalr's "Nonlinear Fiber Optics" (Academic, San Diego, 1995), D. Marcuse's "Optics of Light Transmission" (Van Nostrand, Princeton, 1972), and D. Marcuse's ""Optoelectronic Waveguide Theory" (Academic, New York, 1974).

For convenience, we define the following parameters:

$\begin{array}{cc}{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="0"\; label="u"\; filenames="A0380559800251.png"\; state="\{\{state\}\}">u</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}<span\; class="notranslate">\; =</span>\sqrt{{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}& {\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="u"\; filenames="A0380559800251.png"\; state="\{\{state\}\}">u</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; =</span>\sqrt{{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span>\end{array}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

The scalar field ψ within the fiber 10 satisfies the wave equation, the solutions of which are Bessel functions and modified Bessel functions. For the fundamental mode supported by fiber 10, the scalar field ψ inside the core 12 is:

ψ＝J ₀ (kr) 0≤r≤r ₀ (region I) (2)

Among them, κ is the undetermined eigenvalue, J ₀ is the zero-order Bessel function.

In the subpressed cladding 14, the scalar field ψ is:

ψ＝AK ₀ (βr)+BI ₀ (βr) r ₀ ≤r≤r ₁ (region II) (3)

Among them, A and B are undetermined constants, ${\mathrm{\&beta;}}^2=({\mathrm{<figure-callout\; id="0"\; label="u"\; filenames="A0380559800251.png"\; state="\{\{state\}\}">u</figure-callout>}}_0^2+{\mathrm{<figure-callout\; id="1"\; label="u"\; filenames="A0380559800251.png"\; state="\{\{state\}\}">u</figure-callout>}}_1^2){(2\mathrm{\&pi;}/\mathrm{\&lambda;})}^2-k^2$ K ₀ and I ₀ are modified Bessel functions. Here λ is the wavelength of light in vacuum.

Inside the second cladding 16 we get:

ψ＝CK ₀ (γr) r≥r ₁ (4)

where C is another constant, ${\mathrm{\&gamma;}}^2={{\mathrm{<figure-callout\; id="0"\; label="u"\; filenames="A0380559800251.png"\; state="\{\{state\}\}">u</figure-callout>}}_0^2(2\mathrm{\&pi;}/\mathrm{\&lambda;})}^2-{\mathrm{<figure-callout\; id="2"\; label="k"\; filenames="A0380559800251.png"\; state="\{\{state\}\}">k</figure-callout>}}^2.$ A, B, C and κ are found using boundary conditions, which require that ψ and its first derivative are continuous at _r0 and _r1 .

It can be seen that the fundamental mode cutoff wavelength λc is the wavelength when γ=0. (See Cohen et al. IEEE J. QUANT. Electron. QE-18 (1982) 1467-1472).

For more convenience, we define the following parameters:

$\begin{array}{ccc}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 2</span>}{\mathrm{<span\; class="notranslate">\; \&pi;u</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="0"\; label="r"\; filenames="A0380559800251.png"\; state="\{\{state\}\}">r</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}}{{\mathrm{<span\; class="notranslate">\; \&lambda;</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}}& \mathrm{<span\; class="notranslate">\; \&rho;</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="u"\; filenames="A0380559800251.png"\; state="\{\{state\}\}">u</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; /</span>{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="0"\; label="u"\; filenames="A0380559800251.png"\; state="\{\{state\}\}">u</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}& \mathrm{<span\; class="notranslate">\; the\; s</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="r"\; filenames="A0380559800251.png"\; state="\{\{state\}\}">r</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; /</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}\end{array}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 5</span>}<span\; class="notranslate">\; )</span>$

Now, if the parameter x is determined, the cut-off wavelength λc can also be determined. Those skilled in the art can calculate this definite value by means of algebra, because the parameter x is the root of the following equation:

ρJ ₀ (x)k ⁰¹ (ρx)I ₁ (ρsx)-ρJ ₀ (x)I ₁ (ρx)k ₁ (ρsx)

-J ₁ (x)k ₁ (ρsx)I ₀ (ρx)-J ₁ (x)I ₁ (ρsx)k ₀ (ρx)＝0 (6)

There are three points to consider about the parameter x. First, x does not exist for all values of s and ρ. If ρ=1, and $\mathrm{the\; s}\&le;\sqrt{2}$ , then there is no x that satisfies equation (6). This means that all wavelengths can be conducted in the core 12 in this case. Equation (6) has a solution discriminant as:

s ² ≥ 1+1/ρ ² (7)

Second, x cannot be too small in practical applications. This is because according to equation (5), the parameter x is proportional to the radius r ₀ of the core 12, and the radius of the core 12 should be large enough to allow light to be easily coupled into or out of the core 12. (The smaller the core 12, the stronger the non-linear effect produced, which is often a disadvantage.) Therefore, since x=2πu ₀ r ₀ /λc, preferably x≥1. This means that ρ≥0.224 or expressed in terms of refractive index as $\sqrt{({\mathrm{no}}_2^2-{\mathrm{no}}_1^2)/({\mathrm{no}}_0^2-{\mathrm{no}}_2^2)}\&Greater\; Equal;0.224.$

Third, it is evident from Figure 3 that for larger values of s, the value of x hardly changes with s. This is very beneficial in the field of fiber parameter space, because the s error caused by processing defects has little influence on the value of the fundamental mode cut-off wavelength λc. So the law s≥1+1/ρ can be conveniently used, or expressed in terms of refractive index as:

$\frac{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}{{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="0"\; label="r"\; filenames="A0380559800251.png"\; state="\{\{state\}\}">r</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; +</span>\sqrt{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; /</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 8</span>}<span\; class="notranslate">\; )</span>$

The above-mentioned law is used to guide the selection of the cross-section and refractive index of the core 12, depressing cladding 14 and outer cladding 16 in setting an appropriate fundamental mode cut-off wavelength λc. First, you can pre-select λc, such as the wavelength near 1530nm, and then select u ₀ and r ₀ as arbitrary values. Based on these choices one can compute x from Equation 5 and conveniently compute x ≥ 1 (otherwise the previous choices can be adjusted). Then, find the appropriate values of s and ρ according to Equation 6. The range of values for s and ρ will yield the desired λc. Typically, all p values are greater than 0.224. Also, the laws of Equation 8 are used to further narrow the range of suitable s and p values.

Finally, the values of s and ρ have an additional restriction. That is, they must be selected so that the core 12 of the optical fiber 10 has a sufficiently large loss, such as 100 dB/m or even 200 dB/m or higher, at the wavelength λ≥λc. In order to generate loss at the wavelength λ≥λc, a fiber mode having light of the wavelength λ≥λc is required.

Equations (2), (3) and (4) set the fundamental mode when λ<λc. When λ>λc, the function ψ is an oscillatory function rather than an exponentially decaying function within the second cladding layer 16 . Therefore, when λ>λc, equation (4) is changed to:

ψ＝CJ ₀ (qr)+DN ₀ (qr) r≥r ₁ (region III) (9)

Among them, N ₀ (also known as Y ₀ ) is the zero-order Newman function, ${\mathrm{<figure-callout\; id="2"\; label="q"\; filenames="A0380559800251.png"\; state="\{\{state\}\}">q</figure-callout>}}^2=k^2-{\mathrm{<figure-callout\; id="0"\; label="u"\; filenames="A0380559800251.png"\; state="\{\{state\}\}">u</figure-callout>}}_0^2{(2\mathrm{\&pi;}/\mathrm{\&lambda;})}^2,$ C and D are undetermined constants.

There are two key terms to note about those modes where λ > λc. First there are 5 unknowns (A, B, C, D, and κ) and 4 boundary conditions (ψ and continuity of dψ/dr at r ₀ and r ₁ ). The equation is unrestricted: κ can be chosen between 0 and $(2\mathrm{\&pi;}/\mathrm{\&lambda;})\sqrt{{\mathrm{<figure-callout\; id="0"\; label="u"\; filenames="A0380559800251.png"\; state="\{\{state\}\}">u</figure-callout>}}_0^2+{\mathrm{<figure-callout\; id="1"\; label="u"\; filenames="A0380559800251.png"\; state="\{\{state\}\}">u</figure-callout>}}_1^2}$ any value in between. Thus, for each λ>λc, there is a continuous state, and the corresponding κ can have a continuum of hazy values, which is very different from the case when λ<λc, where the four unknowns (A, B, C and κ) Determined by four boundary conditions, resulting in the discrete eigenvalue κ corresponding to each λ<λc having a unique value.

Second, the modes determined by equations (2), (3) and (9) are characteristic modes of the fiber; such as W-type fiber; however, these modes are not consistent with the physically realized situation. The result of Equation (9) includes both incident waves and outgoing waves, but actually only outgoing waves exist (at wavelength λ>λc, the light originally propagating in the core 12 radiates out).

However, the patterns determined by equations (2), (3) and (9) can be used to estimate losses at those wavelengths greater than λc. First, for a given wavelength λ, find the value of κ that minimizes C ² +D ² . This corresponds to the mode with the longest lifetime in the core. (An analysis can be made between the wave equation of the scalar ψ in the fiber and the quantum mechanical wave equation of the particles in the potential energy well. The results of quantum mechanics can then be borrowed. See David Bohm, "Quantum Theory", Dover 1989, Chapter 12, 14 -twenty two.)

Second, once κ is found using the method described above, the outgoing wave can be calculated from equation (9). These outgoing waves reasonably determine the loss from the core 12 into the second cladding 18 even though no incoming waves are present. Along the length of the fiber 10, these outgoing waves will attenuate those beams propagating in the core 12 for which λ<λc. Assuming that the beam power is P, the power change along the distance Z of the optical fiber 10 is expressed by the following equation:

The loss is given by the coefficient Λ, which is approximated by

The loss Λ, whose unit is m ^-1 , can be converted into β in dB/m by the relation:

β=10log ₁₀ (e).∧ (12)

The term "loss" here refers to radiation that leaks out of the core 12 into the second cladding 16 . In fact, the radiation is not actually lost in the fiber 10 but remains in the second cladding 16 . Sometimes this is useful. On the other hand, light from the second cladding layer 16 can be outcoupled or absorbed as desired.

Another calculation method for loss involves the calculation of the complex propagation constants of the leaky modes of the fiber 10 . Leaky modes are discussed in Chapter 1 of Dr. Aarcuse's "Theory of Dielectric Optical Waveguides" (Education, New York, 1974). The loss is related to the imaginary part of the leaky mode complex propagation constant. The complex propagation constant, which is equivalent to the complex effective refractive index, can be calculated by using software commercially available, for example, from the Optiwave Corporation of Neapean, Canada.

Sometimes people prefer to solve the mode of a given fiber by using numerical calculation methods instead of using the Bessel function method of approximating the curve described above, because in fact the fiber does not have an ideal step index like the profile 20 shown in Figure 1 The distribution curve, but different from it, the refractive index distribution curve obtained in practice is as shown in the curve 20A of FIG. 2 . In particular, the most common method of manufacturing single-mode optical fibers today is the Metal Organic Chemical Vapor Deposition (MOCVD) process, which typically leaves an index depression in the center of the core 12 . The numerical solution of refractive index as a function of radius can be more easily regarded as the actual variation than the above method.

Since the actual refractive index varies slightly as a function of radius (see profile 20A) when equation (11) is used to estimate losses, the indices n1, n2, n3 will generally be the average indices of profile 20. The refractive index n also need not be radially symmetric. If the cross-section of fiber 10 is described by polar coordinates r and θ, the refractive index can depend on angle θ and radius r. That is, n=n(r, θ). For example, such an asymmetric fiber can have predictable polarization maintenance.

Here the fiber is expected to have a fundamental mode cutoff wavelength λc. Let R be a radius large enough that the refractive index at R actually reaches the equilibrium value n2. As follows, the optical fiber 10 will have a fundamental mode cut-off wavelength λc (see B.Simon, Ann. Physics. 97 (1976), page 79 of the third world country):

$\underset{\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; \&pi;</span>}}{<span\; class="notranslate">\; \&Integral;</span>}}\mathrm{<span\; class="notranslate">\; d\&theta;</span>}\underset{\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; R</span>}}{<span\; class="notranslate">\; \&Integral;</span>}}\mathrm{<span\; class="notranslate">\; rdr</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; no</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; r</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 13</span>}<span\; class="notranslate">\; )</span>$

Note the distribution curve given by Fig. 1, Equation (13) becomes:

$\mathrm{<span\; class="notranslate">\; \&pi;</span>}{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="0"\; label="r"\; filenames="A0380559800251.png"\; state="\{\{state\}\}">r</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="0"\; label="u"\; filenames="A0380559800251.png"\; state="\{\{state\}\}">u</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; \&pi;</span>}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="r"\; filenames="A0380559800251.png"\; state="\{\{state\}\}">r</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="0"\; label="r"\; filenames="A0380559800251.png"\; state="\{\{state\}\}">r</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="u"\; filenames="A0380559800251.png"\; state="\{\{state\}\}">u</figure-callout></span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; \&le;</span>\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 14</span>}<span\; class="notranslate">\; )</span>$

This is the same as equation (7) above.

The fundamental mode cutoff wavelength λc is the maximum wavelength where the eigenmodes located in the region I exist. For example, the loss at wavelengths longer than the cut-off wavelength λc can be determined by (i) solving for modes that are indeterminate but include both incident and outgoing waves, (ii) finding for each wavelength the mode with the smallest outgoing intensity, and (iii) Use the outgoing intensity to estimate the loss. As discussed above, those skilled in the art can also use other methods to calculate the loss. In general, an optical fiber 10 with a desired fundamental mode cutoff wavelength λc and loss can thus be designed by adjusting the profile n=n(r,θ), which is equivalent to adjusting the core 12, depressing the cladding 14 and the second cladding 16 cross-section and refractive index.

The above laws enable those skilled in the art to set the fundamental mode cut-off wavelength λc by selecting r ₀ , r ₁ , n ₀ , n ₁ and n ₂ . This selection of r ₀ , r ₁ , n ₀ , n ₁ and n ₂ provides a suppressed distribution of amplified spontaneous emission over the length of the fiber 10 and results in clusters of loss curves (with respect to wavelength) with different roll-offs. Therefore, as discussed below, for the purposes of the present invention, additional suppression must be set depending on the choice of r ₀ , r ₁ , n ₀ , n ₁ and n ₂ .

Referring back to FIG. 1 , superimposed on the mean refractive index profile 20 is the intensity profile of the waveguided fundamental mode 22 at the first wavelength λ1 < λc. The first wavelength λ1 is located in the short-wave band, ie the S-band. The fundamental mode 24 which is no longer guided by the fiber 10 is also superimposed on the refractive index profile 20 . Mode 24 is located at the cutoff wavelength λc. Also shown is the intensity distribution of an additional mode 26 that is not propagated by the fiber 10 and that exhibits an oscillating intensity distribution outside the core 12 and depressing cladding 14 . The radiation of mode 26 has a second wavelength λ2, which is longer than the cut-off wavelength, λc<λ2, and is located in long wavelength bands, such as C-band and L-band.

Figure 4 shows a gain curve 44 for the case where the active material 18 is erbium pumped to a high relative inversion D. The S-band is marked with parameter 42, and the long-wave band is marked with parameter 46. The intersection wavelength λcross between the S-band 42 and the long-waveband 46 is also marked. The gain curve 44 exhibits high gain in the long-band 46 and positive gain in the S-band 42 . In particular the high gain in the long wavelength band 46 includes a gain peak 48 at 1530nm which is very close to the crossover wavelength λcross.

In this embodiment, we make the cut-off wavelength λc just at the gain peak 48 in the long wavelength band 46 by selecting the cross-section and radius of the core 12 , the depressed cladding 14 , and the refractive indices n ₀ , n ₁ , and n ₂ . Also the roll-off loss curve 38 at the cutoff wavelength λc at the high gain peak 48 in the long wavelength band 46 is obtained by selecting the refractive index n0 of the core 12 . And the roll-off loss curve 38 is chosen to produce at least comparable losses in the long-band 46 for high gains, while producing much less loss in the S-band 42 than for positive gains. The roll-off loss curve falls off rapidly or has a large positive slope to the left relative to those wavelengths below the cutoff wavelength λc, so that the roll-off loss curve 38 falls below the positive gain represented by the gain curve 44 . Thus the gain is greater than the loss throughout the S-band 42, as clearly indicated by the shaded area 50. A more suitable roll-off loss curve 38 is that the gain is at least 5 dB greater than the loss in the short-wave band 42 . For more information on how to select an appropriate roll-off loss curve, the reader is referred to US Patent Application 10/095,303, filed March 8, 2002.

The sub-band amplifying device of the present invention utilizes the advantages of the W-shaped optical fiber designed according to the above rules. In fact, the sub-band amplifying device of the present invention preferably uses a W-shaped optical fiber designed according to the above-mentioned rules, its active substance 18 is erbium, the short-wave band is the S-band or a part of the S-band, and the long-wave band covers the entire C-band and/or L-band or parts of both bands. Usually, the main material of the optical fiber 10 is silicate glass, such as alumino-germanium silicate glass, lanthanum-doped germanium silicate glass, aluminum/lanthanum-doped germanium silicate glass, or phosphorus-doped germanium silicate glass. For example, an erbium-doped fiber amplifier 68 (EDFA) made of germanium aluminosilicate glass as shown in FIG. 5 can be used as the sub-band amplification device. In this example, an erbium-doped fiber amplifier 68 has a core 70 of refractive index _n0 doped with erbium at a concentration of 0.1% by weight. The core 70 is surrounded by a depressive cladding 72 with a refractive index _n1 and a second cladding 74 with a refractive index _n2 . Erbium-doped fiber amplifier 68 has a protective sheath 76 wrapped around second cladding 74 to provide mechanical stability and protect Erbium-doped fiber amplifier 68 against external influences.

Optical signal 78 having a first wavelength _λ1 contained within the S-band is transmitted from optical fiber 80 to erbium-doped fiber amplifier 68 for amplification. Optical signal 78 may be an optical signal carrying useful information that needs to be amplified.

Optical fiber 80 and optical fiber 82 are coupled in a combiner 84 . Optical fiber 82 is used to couple pump light 88 from pump source 86 into erbium-doped fiber amplifier 68 . A pump source 86, preferably a laser diode, provides a pump light 88 with a pump wavelength λp of about 980 nm for pumping erbium ions in the core 70 to obtain a high level of relative population inversion D. The variation range of the parameter D is from D=-1 to D=1, the former means no inversion of particle number, and the latter means complete inversion of particle number. When D=0, it means that exactly half of the erbium ions are in the excited energy state or excited state clusters, and the other half are in the ground energy state. In this case, the erbium-doped fiber amplifier 68 is almost transparent (3-level transitions with a wavelength near 1530 nm). For an erbium-doped fiber with unbalanced switching, the parameter D represents the mean value of the population inversion. In this embodiment, the intensity of the pump light 88 must ensure that the population inversion D≥0.7 of the erbium ions.

After the pump light 88 and the signal light 78 are mixed in the multiplexer 84 , both paths of light are sent to the erbium-doped fiber amplifier 68 by the optical fiber 80 . In particular, both signal light 78 and pump light 88 are coupled from fiber 80 into core 70 .

Both the core 70 and the claddings 72, 74 in this embodiment have circular cross-sections. According to the method of the present invention, the cross section and the refractive indices n ₀ , n ₁ , n ₂ are selected so that the cut-off wavelength λc is located near 1525 nm. That is, the cut-off wavelength λc is selected between the S-band 42 and the long-band 46 or between the C-band and the L-band.

It is very important that the refractive index n0 of the core 70 is chosen to provide a large negative slope of the effective refractive index n _eff corresponding to wavelength, typically about .008/1,000 nm, near the cutoff wavelength λc. As a result, the roll-off loss curve exhibits a rapid drop in the region of wavelengths below the cut-off wavelength λc, thereby ensuring that the loss is lower than the positive gain in the S-band 42 . The loss of the roll-off loss curve increases rapidly in the region where the wavelength is greater than the cut-off wavelength λc. Thus the loss in the C-band and L-band 46 is at least comparable to the high gain.

As shown in the figure, the erbium-doped fiber amplifier 68 designed according to the present invention will be able to ensure that the optical signal 78 with a wavelength of λ ₁ is amplified, and the wavelength of any C-band and L-band is λ ₂ , especially at λ ₂ The amplified spontaneous emission = 1530 nm is suppressed into the cladding 74 . Positive gain for S-band 42 is typically on the order of 1dB per meter (or 0.2-5dB/meter, depending on fiber design or to obtain maximum benefit at S-band wavelength) above loss, so that for effective To amplify the optical signal 78, the erbium-doped fiber amplifier 68 needs to have a certain length L. The smaller the difference between the positive gain and loss in the S-band, the longer the length L of the erbium-doped fiber amplifier 68 that provides effective amplification of the optical signal 78 . For a useful 25dB of S-band amplification, the required length is about 5 to 100 meters.

FIG. 6 shows a sub-band amplifying device 100 of the present invention. The sub-band amplification device 100 is provided with an input optical fiber 102 for receiving an optical signal 104 to be amplified. As shown, the optical signal 104 spans a long-band signal 106 and a short-band signal 108 separated by a crossover wavelength λcross. The sub-band amplification device 100 also has a pump input fiber 112 for receiving pump radiation 116 from a pump source 114 . In the present embodiment, the pump source 114 is a laser diode emitting pump radiation 116 at a pump wavelength λ _p =980 nm.

The sub-band amplification device 100 has an overlapping portion 118 shared by the first portion 130 and the second portion 132 . That is, the overlapping portion 118 is where the first portion 130 and the second portion 132 overlap. The second part 132 is provided with a fiber amplifier 138 behind the overlapping part 118 for amplifying the short-wavelength signal 108 in the optical signal 104 . Fiber amplifier 138 is preferably a first erbium-doped fiber amplifier. The overlapping portion 118 is provided with an amplifier 122 for amplifying the long-wavelength signal 106 of the optical signal 104 . It should be noted that amplifier 122 does not necessarily have to be an erbium-doped fiber amplifier, or even a fiber amplifier.

Amplifier 122 should provide some amplification (or at least zero amplification, but no loss) to short-band signal 108 and a useful gain to long-band signal 106 . In a preferred sub-band amplifying device 100, the amplifier 122 should provide sufficient amplification for the short-band signal 108 to compensate for the loss of the short-band signal 108 in the components preceding the fiber amplifier 138. A preferred sub-band amplification device 100 scheme is that the amplifier 122 should provide low noise amplification of the short-band signal 108 to minimize the noise effect of the loss of the short-band signal 108 in the components before the fiber amplifier 138 . The overlapping portion 118 is also provided with a coupler 120 and a coupler 124 . The coupler 120 is any suitable optical coupling device, such as a wavelength division multiplexing coupler capable of combining the optical signal 104 and the pumping radiation 116 and then passing through the amplifier 122 to obtain optical amplification of the optical signal 104 . In this embodiment, amplifier 122 is a second erbium-doped fiber amplifier and long-band signal 106 spans the C-band. The coupler 124 is any applicable optical splitter, such as a wavelength division multiplexing coupler capable of splitting the optical signal 104 into two optical signals 104A and 104B. In this example, the coupler 124 is a 50/50 wavelength division multiplexing coupler, which splits the optical signal 104 into two signals 104A, 104B of equal intensity and spectral band. Alternatively, couplers with different splitting ratios or band combiners capable of separating the short-band signal 108 from the long-band signal 106 may be used.

Coupler 124 is connected to first portion 130 and second portion 132 . In particular, coupler 124 sends signal 104A to first section 130 and signal 104B to second section 132 . The first part 130 is provided with a delay device 126 , and the delay device may be an optical fiber with a certain length for compensating the time delay caused by the transmission between the first part 130 and the second part 132 . The first part 130 and the second part 132 are combined together through the coupler 128. The coupler 128 is preferably a band synthesizer, which can synthesize the two-way signals 104A, 104B of the short-wave band 108 and the long-wave band 106 and reject signals of other wavelengths. radiation, such as residual pump radiation 116'.

The second part 132 is provided with a coupler 134 for coupling the residual pump radiation 116 ′ out of the second part 132 . After the coupler 134, an isolator 136 and a first erbium-doped fiber amplifier 138 are arranged in sequence. Isolator 136 is designed to pass signal 104B to first erbium-doped fiber amplifier 138 , blocking any radiation propagating back to overlapping portion 118 .

A first erbium-doped fiber amplifier 138 constructed according to the principles discussed above is used to amplify the signal 104B in the short wave band 108, the S-band of this embodiment. A pump source 140 is provided to generate pump radiation 142 for the first erbium-doped fiber amplifier 138 . The pump source 140 is a diode laser emitting pump radiation 142 with a pump wavelength λ _p =980 nm. Input fiber 144 and coupler 146 are used to deliver pump radiation 142 to first erbium-doped fiber amplifier 138 . The coupler 146 is connected such that the pump radiation 142 passes from the opposite direction to the preceding optical signal 104B. That is, within the first erbium-doped fiber amplifier 138 the pump radiation 142 propagates counter-propagated with the optical signal 104B. It should be noted that any residual pump radiation 142' is intercepted by coupler 134 and isolator 136 to prevent it from passing back into overlapping portion 118. The counterpropagation of the pump radiation within amplifier 138 provides high pump-to-signal conversion efficiency.

The coupler 146 is configured to transmit the optical signal 104B amplified by the first erbium-doped fiber amplifier 138 to the coupler 128 in the short-wave band, ie, the S-band. The coupler 128 combines the amplified signal 104A at the long wavelength band 106 with the amplified signal 104B at the short wavelength band 108 to form an amplified optical signal 104'. A preferred filter 148, a gain flattening filter (GFF), can be used to provide a uniformly amplified optical signal 104'.

During operation, the optical signal 104 enters the sub-band amplifying device 100 through the input fiber 102, and then transmits to the second erbium-doped fiber amplifier 122 of the overlapping part 118 along the same direction (same geometric propagation direction) with the pump signal 116. The second erbium-doped fiber amplifier 122 amplifies the portion of the signal 104 within the long-wavelength band 106 and, in a preferred embodiment, also provides a small amount of amplification for the portion of the signal 104 within the short-wavelength band 108 .

The coupler 124 decomposes the optical signal 104 amplified in the long-wavelength band into a signal 104A and a signal 104B, and then sends them to the non-overlapping parts of the first part 130 and the second part 132 respectively. Signal 104A is not further amplified within first portion 130 . Those signals in the short wavelength band 108 of the signal 104B are amplified by the first erbium-doped fiber amplifier 138 .

Most of the pumping radiation 116 is preferably used up at the second erbium-doped fiber amplifier 122, and any residual pumping radiation 116' that is potentially harmful to the operation of the pump source 140 of the first erbium-doped fiber amplifier 138 is exhausted. Coupler 134 splits out. Any remaining pump radiation 142 ′ from backpropagating that affects the operation of pump source 114 is also blocked by coupler 134 and isolator 136 .

Delay 126 ensures that amplified signals 104A and 104B are combined within coupler 128 into one signal 104' with close group delay. In this case, it is preferred that the filter 148 adjusts the level of amplification of the signal 104' over the entire band, ie the short band 108 and the long band 106 or the S-band and the C-band.

The sub-band amplification device 100 is very efficient for amplifying the signal 104 with a minimum of components. It should also be pointed out that the short-wave band 108 and the long-wave band 106 can be selected based on application conditions. Although in telecommunications, the typical short-wave band 108 includes at least a portion of the S-band, other short-wave bands are also applicable by varying the parameters of the first erbium-doped fiber amplifier 138 according to the previously described rules. In telecommunication applications, the long wave band 106 includes at least part of the C-band or the L-band or both. Of course, the long wavelength band 106 can be any desired long wavelength band by selecting an appropriate amplifier 122 or even replacing the amplifier 122 with multiple amplifiers connected in series or in parallel.

Another advantage of the sub-band amplification device 100 is its good noise figure. Especially in the disposition of each component in the sub-band amplification device 100, since the input signal is directly passed into a co-pumped erbium-doped fiber amplifier without distributed loss, a good noise figure is produced. Generally, the noise figure is best if the gain occurs before all the losses (100% of the loss occurring before the gain adds to the noise figure). It is also beneficial to use common pumping for the shared amplifier 122 as this results in high population inversion. It is also beneficial when the length of the shared amplifier 122 is chosen such that the population inversion of the amplifier is close to 100%, since then the amplified spontaneous emission across the band is minimal. With nearly complete population inversion, the amplification stage contributes minimally to the noise figure. Using standard erbium-doped fiber, this approach provides only a small amount of short-band gain (approximately 5dB) without sacrificing population inversion. If the fiber length is increased beyond the optimal length, the increase in population inversion will negatively affect the noise figure and limit the S-band gain (note that the C-band gain will continue to increase in this case). If the fiber length is increased beyond the optimal length and further increased, it will eventually lead to a reduction in the S-band gain and even a net loss in the S-band. This latter case (S-band loss) is the most common case for the best erbium-doped fiber amplifiers operating preferentially at C-band.

It is also beneficial for the noise figure that the shared amplifier 122 is a conventional erbium-doped fiber amplifier without a fundamental mode cutoff. For this reason, it is unlikely that there will be additional distribution losses in addition to erbium absorption losses (ie inversion < 100%). The absence of this distributed loss further improves the good noise figure.

The basic design of the sub-band amplification device 100 is to split the low noise and amplified signal into two paths. This splitting (and its associated losses) is done after the signal goes through the primary amplifier, and therefore has less impact on the noise figure than if the split was done before the primary amplifier. In addition, the inherently poor noise figure of the S-band of the Erbium-doped fiber amplifier 138 is effectively de-amplified in the S-band branch due to the low gain of the initial pole that amplifies the good noise figure.

The sub-band amplifying means can be configured in many different ways. FIG. 7 depicts an alternative sub-band amplification arrangement 150 having separate first and second sections 152,154 and using a common pump source 156. In this embodiment, the common pump source 156 is also a diode laser emitting pump radiation 116 at the pump wavelength λ _p =980 nm. In fact, elements of device 150 are referenced by the same parameters relative to elements of device 100 .

Optical signal 104 is received by coupler 158 . The coupler 158 is arranged to split the signal 104 into a first signal 160 spanning the short band 108 and a second signal 162 spanning the long band 106 . A first signal 160 is sent by an optical fiber 164 belonging to the first part 152 to a coupler 166 . Coupler 166 also receives pump radiation 116 from laser diode 156 . The output of the coupler 166 mixes the pump radiation 116 and the first signal 160 and sends them to the fiber amplifier 168 of the first part 152 .

The fiber amplifier is set up to amplify the signal 160 according to the above rules. On the other hand, the fiber amplifier 168 is arranged to amplify the short wave band 108 . Fiber amplifier 168 is preferably a first erbium-doped fiber amplifier. The first erbium-doped fiber amplifier 168 is followed by a coupler 170 arranged to transmit the amplified signal 160 ′ to an optical fiber 172 belonging to the first part 152 and to transmit the pump radiation 116 to a coupler 174 belonging to the second part 154 .

At the same time, the second signal 162 is transmitted to the coupler 174 by the optical fiber 176 belonging to the second part 154 . The coupler 174 mixes the undepleted pump radiation 116 of the first erbium-doped fiber amplifier 168 with the second signal 162 and sends them to an amplifier 178 which is designed to amplify the long-wave band 106 . In this embodiment, amplifier 178 is a second fiber amplifier, particularly a second Erbium-doped fiber amplifier. A second erbium-doped fiber amplifier 178 amplifies signal 162 to produce an amplified signal 162'.

The multiplexer 180 combines the amplified signal 160' from the first branch 152 and the amplified signal 162' from the second branch 154 into one signal. The combiner 180 is followed by an isolator 182 which passes only the newly synthesized signal 104' and eliminates any remaining pump radiation 116.

The sub-band amplifying device 150 utilizes the pumping source efficiently due to the advantage of using a single common pumping source 156. In particular, the S-band fiber amplifier 168 can be operated at full pump power in a co-pumped configuration, thus making it possible to obtain the highest population inversion required for the best amplification performance in the S-band. When an S-band erbium-doped fiber amplifier is operated with very high population inversion D, it naturally absorbs the pump radiation ineffectively, thus producing a large amount of "garbage" pump light. The sub-band amplification device 150 uses these so-called "junk" pump lights to pump the amplifier 178. Of course, those skilled in the art will appreciate that the sub-band amplification device 150 can be reconfigured to use two pump radiation sources and/or to provide geometrically co-pumped and counter-pumped. Furthermore, the optical fiber amplifier can be several optical fiber amplifier units separated by isolators as required. Using a "mid-span" isolator can improve the noise figure and increase the net gain of the amplifier due to the elimination of backpropagating amplified spontaneous emissions that would otherwise rob the amplifier of pump power, reducing population inversion and thus cause the amplifier to saturate.

FIG. 8 shows an optical communication system 200 using a signal source 202A providing a signal 204A in the S-band. Signal source 202A is a diode laser belonging to signal source arrays 202A-202H. Actually, the optical communication system 200 is a wavelength division multiplexing system (WDM), even a dense wavelength division multiplexing system (DWDM) or a coarse wavelength division multiplexing system (CWDM), in which all 8 signal sources 202A-202H generate signals at 8 different wavelengths or 8 wave channels in the S-band. It should be pointed out that in this embodiment, we only use 8 channels to describe the wavelength division multiplexing system (WDM). In fact, it can be known from the prior art that the number of channels used by the wavelength division multiplexing system (WDM) can be Any number from 2-1000. Signal source arrays 202A-202H are controlled by a laser array controller 210 . Laser array controller 210 may be any suitable electronic or optical pumping mechanism for driving diode lasers 202A-202H.

Groups of optics 206A-206H, in this embodiment lenses, are used to focus and couple the signals 204A-204H into the multiplexer 208 for wavelength division multiplexing. Multiplexer 208 multiplexes and passes signals 204A-204H through first section 212A of common communication fiber 212 in accordance with well-known WDM protocols. The first section 212A of the communication optical fiber 212 has a very long span, such as tens of kilometers. Communication fibers 212 are selected from those that exhibit a low and sufficient amount of dispersion in the S-band to prevent nonlinear effects affecting signals 204A-204H. For example, the chromatic dispersion of the communication fiber 212 can be 3-10 ps/(nm.km), but cannot be lower than 1 ps/(nm.km) in the S-band. Also, it is very beneficial that the communication fiber 212 has a low loss such as 0.2 to 0.3 dB/km in the S-band.

Communication fiber 212 is connected between WDM multiplexer 208 and WDM demultiplexer 228 . After the first section 212A, the communication fiber 212 is coupled with the fiber 214 in a combiner 216 . An optical fiber 214 is used to couple pump radiation 218 from a pump source 220 into an erbium-doped fiber amplifier 222 arranged after the first section 212A. The pump source 220 is a laser diode emitting pump radiation having a wavelength λ of about 980 nm for pumping the erbium ions in the core of the erbium-doped fiber amplifier 222 to the relative population inversion D of the high energy state . The pump source 220 is controlled by a pump laser controller 224, which may be any suitable electronic or optical pumping device. An optics group 226 , in the exemplary embodiment a lens, is used to couple the pump radiation 218 into the optical fiber 214 .

Erbium-doped fiber amplifier 222 is coupled to second section 212B of optical fiber 212 . In practice the system 200 may have multiple sections of fiber and multiple erbium-doped fiber amplifiers spaced at regular intervals between these sections. In this embodiment, the terminal of section 212B is connected to the demultiplexer 228 of WDM. The WDM demultiplexer 228 separates the signals 204A-204H by wavelength and transmits them to respective receivers 230A-230H using sets of optics 232A-232H. In this embodiment, optics group 232A-232H is a group of lenses. Receive controller 236 is coupled to receivers 230A-230H. Receive controller 236 may include any signal resizing and other functions, like signal processing functions required to detect and process signals 204A-204B.

As shown in this embodiment, the erbium-doped fiber amplifier 222 is used by the communication system 200 as an in-line amplifier. In particular, Erbium-doped fiber amplifier 222 has a tailorable length L that can provide sufficient S-band gain to amplify signals 204A-204H. For example the length L may range from 5 to 50 meters. At the same time, the erbium-doped fiber amplifier 222 is tuned to exhibit a low noise figure, eg, 6 dB or less, and a flat gain spectrum. A suitable gain flattening filter (not shown) may be used in the erbium-doped fiber amplifier 222 to ensure a flat gain spectrum, as will be appreciated by those skilled in the art.

Alternatively, Erbium-doped fiber amplifier 222 may be used as a power amplifier. Therefore, the power of the pumping source 220 should be increased to increase the saturation power of the Erbium-doped fiber amplifier 222 . Other design changes and optimizations can be made at the same time. For example, to increase the absorption efficiency of the pump radiation 218, the concentration and the length L of the erbium dopant can be increased at the expense of the noise figure. Erbium-doped fiber amplifier 222 can also be designed to be cladding pumped in this configuration. Cladding pumping is a well known technique in the field of C-band Erbium-doped fiber amplifier fabrication. When the erbium-doped fiber amplifier 222 is designed to be cladding pumped, a low-cost, high-power, broadband 980nm diode-pumped laser is used as the pumping source 220 . By canceling the gain flattening filter, reducing the gain flattening amount or moving the gain flattening filter to the front part of this embodiment, where the erbium-doped fiber amplifier is made up of multiple amplifiers (multi-stage erbium-doped fiber amplifier), the S-band Erbium-doped fiber amplifiers are also optimized for high power. Since the internal loss of the Erbium-doped fiber amplifier is reduced, the overall efficiency of the Erbium-doped fiber amplifier will be increased and thus the saturation power will be increased. Typically, a power amplifier usually works in an unsaturated state or sometimes has only one signal path, so when it is used as an in-line amplifier, its gain flatness is not as important as when it is used as a power amplifier.

Communications system 200 may use erbium-doped fiber amplifier 222 as a preamplifier for signals 204A-204H. In this case, the length L is reduced to less than 50 m or even less than 10 m, and the noise figure is carefully controlled by a combination of fiber and gain-flattening filter parameters. Preamplifiers are typically used for single signal channels and low optical power levels, but moderate gain is required. So gain flattening is usually not needed, and amplification efficiency is not very important. A key specification for preamplifier design is noise figure.

Figure 9 shows a portion of another optical communication system 250 of the present invention. Among them are several erbium-doped fiber amplifiers 252A-252H used as preamplifiers. System 250 has a WDM demultiplexer 254 for demultiplexing a plurality of signals 256A-256H carried by communication fiber 260 . A set of lenses 258A-258H is provided for focusing the signals 256A-256H demultiplexed by the WDM demultiplexer 254 to a corresponding set of receivers 260A-260H. Acceptors 260A-260H are controlled by acceptor controller 262 as in the previous embodiments.

In this embodiment, the erbium-doped fiber amplifier 252 is arranged after the demultiplexer 254 of the WDM. This is the best setting when the erbium-doped fiber amplifier 252 is designed as a preamplifier. This is due to the fact that each erbium-doped fiber amplifier 252A-252H processes only one of the signals 256A-256H at this location, which is better tuned to obtain low noise preamplification.

FIG. 10 shows another optical communication system 300 of the present invention employing S-band erbium-doped fiber amplifiers 302A and 302B for different portions or sub-bands of the S-band. Erbium-doped fiber amplifiers 302A, 302B are installed between two segments of optical fibers 304A, 304B. A WDM demultiplexer or demultiplexer 306 and WDM multiplexer or combiner 308 are used to deliver the appropriate portion of the S-band wavelengths to each erbium-doped fiber amplifier 302A, 302B. In this embodiment, erbium-doped fiber amplifier 302A is optimized to amplify the sub-band of wavelengths between 1460 and 1490 nm, while amplifier 302B is optimized to amplify the sub-band of wavelengths between 1490 and 1520 nm. Of course more than two erbium-doped fiber amplifiers can be used to amplify smaller portions of the S-band if desired. The communication system 300 is capable of using the wider entire range of the S-band.

It should be noted that the S-band erbium-doped fiber amplifiers 302A, 302B can be combined with the C-band and/or L-band erbium-doped fiber amplifiers if the optical communication system 300 is designed to transmit signals that lie in the entire band or in certain bands. Usually those skilled in the art will recognize that S-band erbium-doped fiber amplifiers and C-band or L-band erbium-doped fiber amplifiers and/or erbium-doped fiber amplifiers with two or more sub-bands of S-bands can be used Anywhere in the communication system 300, not limited to use as in-line amplifiers, preamplifiers, and power amplifiers.

It is clear to those skilled in the art that the above-described embodiments can be replaced in various ways outside the scope of the present invention. Accordingly, the scope of the present invention is to be determined by the following claims and their legal equivalents.

Claims (34)

1, a kind of subrane amplifying device comprises

A) be used for the first of the long-wave band of amplifying optical signals; With

B) be used to amplify the second portion that is provided with fiber amplifier of the short-wave band of described optical signal, described fiber amplifier comprises:

1) one has fiber core cross section and refractive index n ₀Fibre core;

2) a kind of active substance that mixes in the described fibre core;

3) covering that forces down that surrounds described fibre core, the described covering that forces down forces down covering cross section and refractive index n ₁

4) one is surrounded described second covering that forces down covering, and described second covering has second covering cross section and the refractive index n ₂With

5) one is used for the described active substance of pumping to the high pumping source of population inversion D relatively, so that described active substance presents postiive gain and presents high-gain in described long-wave band at described short-wave band;

It is characterized in that described fiber core cross section, described covering cross section and the described refractive index n of forcing down ₀, n ₁, and n ₂Being selected to and producing one is λ about cutoff wavelength _cThe damage curve that roll-offs, the described curve that roll-offs produces comparable with described high-gain at least loss in described long-wave band, and produces the loss of muching littler than described postiive gain at described short-wave band.

2, subrane amplifying device according to claim 1, it is characterized in that: described active substance is an erbium, promptly described fiber amplifier is first erbium-doped fiber amplifier.

3, subrane amplifying device according to claim 2 is characterized in that described short-wave band comprises part S-wave band at least and described long-wave band comprises portion C-wave band at least, described cutoff wavelength λ _cBe that wavelength is got in friendship between described S-wave band and C-wave band.

4, subrane amplifying device according to claim 3 is characterized in that described long-wave band at least also comprises partial L-wave band.

5, subrane amplifying device according to claim 3 is characterized in that described cutoff wavelength is about 1530nm.

6, subrane amplifying device according to claim 1 is characterized in that described first comprises one second erbium-doped fiber amplifier.

7, subrane amplifying device according to claim 6 is characterized in that described active substance is an erbium, and promptly described fiber amplifier is first erbium-doped fiber amplifier.

8, subrane amplifying device according to claim 7 further comprises a public pumping source, and being used for provides the pumping width of cloth to penetrate to described first erbium-doped fiber amplifier and described second erbium-doped fiber amplifier.

9, subrane amplifying device according to claim 8 is characterized in that described public pumping source comprises a laser diode, and being used for provides described pumping radiation at about 980nm place.

10, subrane amplifying device according to claim 1 is characterized in that described first and described second portion share a lap.

11, subrane amplifying device according to claim 10 is characterized in that described lap comprises second erbium-doped fiber amplifier that is used to amplify described long-wave band.

12, a kind of subrane amplifying device that uses carries out the method that subrane amplifies, and described method comprises:

A) provide a first that is used for the long-wave band of amplifying optical signals; With

B) provide a second portion that is provided with fiber amplifier that is used to amplify the short-wave band of described optical signal, described fiber amplifier constitutes like this:

1) provide one to have fiber core cross section and refractive index n ₀Fibre core;

2) described active substance is incorporated in the described fibre core;

3) provide one to force down the described fibre core of encompasses, the described covering that forces down forces down covering cross section and refractive index n ₁

4) provide one the second described covering that forces down of encompasses, described second covering has second covering cross section and the refractive index n ₂With

5) select described fiber core cross section, described covering cross section and the described refractive index n of forcing down ₀, n ₁, and n ₂To produce one about cutoff wavelength λ _cThe damage curve that roll-offs, the described damage curve that roll-offs produces comparable with described high-gain at least loss in described long-wave band, and produces the loss of muching littler than described postiive gain at described short-wave band.

13, method according to claim 12 comprises further and selects erbium as described active substance that promptly described fiber amplifier is first erbium-doped fiber amplifier.

14, method according to claim 13 further comprises described first erbium-doped fiber amplifier of backward pumping.

15, method according to claim 12 further comprises providing the described first that is provided with second erbium-doped fiber amplifier to be used to amplify described long-wave band.

16, method according to claim 15 comprises further and selects erbium as described active substance that promptly described fiber amplifier is first erbium-doped fiber amplifier.

17, method according to claim 16 further comprises with a public pumping source providing public pumping to described first erbium-doped fiber amplifier and described second erbium-doped fiber amplifier.

18, method according to claim 17 is characterized in that described public pumping source provides wavelength to be about the pumping radiation of 980nm.

19, method according to claim 12 is characterized in that described short-wave band comprises part S-wave band at least and described long-wave band comprises portion C-wave band at least, and described cutoff wavelength λ _cBe that wavelength is got in friendship between described S-wave band and described C-wave band.

20, method according to claim 19 is characterized in that described long-wave band at least also comprises partial L-wave band.

21, method according to claim 19 is characterized in that described cutoff wavelength λ _cBe positioned at about 1530nm.

22, a kind of optical communication system comprises:

A) signal source that is used to provide S-wave band wavelength signals;

B) fiber amplifier that is used to amplify described signal, described fiber amplifier has:

1) one has fiber core cross section and refractive index n ₀The er-doped fibre core;

2) covering that forces down that surrounds described fibre core, the described covering that forces down forces down covering cross section and refractive index n ₁

3) one is surrounded described second covering that forces down covering, and described second covering has second covering cross section and the refractive index n ₂

4) erbium that is used for the described fibre core of pumping is to the high pumping source of population inversion D relatively, and promptly described active substance presents postiive gain and presents high-gain in the long-wave band of being longer than described S-wave band at described S-wave band;

It is characterized in that described fiber core cross section, described covering cross section and the described refractive index n of forcing down ₀, n ₁, and n ₂Be selected in described long-wave band and produce comparable with described high-gain at least loss, and produce the loss of muching littler than described postiive gain at described S-wave band.

23, optical communication system according to claim 22 is characterized in that described fiber amplifier is from by preamplifier, an amplifier of selecting among the amplifier group that power amplifier and inline amplifier are formed.

24, optical communication system according to claim 22 is characterized in that described optical communication system is a wavelength division multiplexing communications systems.

25, optical communication system according to claim 24 is characterized in that described fiber amplifier is from by preamplifier, an amplifier of selecting among the amplifier group that power amplifier and inline amplifier are formed.

26, optical communication system according to claim 24 further comprises:

A) telecommunication optical fiber;

B) wavelength division multiplexer is used for carrying out multiplexing to the described signal of described telecommunication optical fiber; With

C) Wave decomposing multiplexer is used for described signal demultiplexing.

27, optical communication system according to claim 26 is characterized in that the preamplifier after described fiber amplifier is arranged on described Wave decomposing multiplexer.

28, optical communication system according to claim 26 is characterized in that described fiber amplifier is arranged between described wavelength division multiplexer and the described Wave decomposing multiplexer.

29, optical communication system according to claim 28 is characterized in that described fiber amplifier is from by preamplifier, an amplifier of selecting among the amplifier group that power amplifier and inline amplifier are formed.

30, a kind of optical communication system according to claim 24 is characterized in that described signal source comprises a laser instrument that belongs to laser array.

31, optical communication system according to claim 22 is characterized in that described long-wave band comprises portion C-wave band or L-wave band at least.

32, optical communication system according to claim 22 is characterized in that described pumping source is the laser diode that pumping radiation is provided at about 980nm.

33, optical communication system according to claim 22 is characterized in that described signal is the signal of carrying information.

34, optical communication system according to claim 22 is characterized in that described fiber amplifier is by the tuning sub-band that is used to amplify described S-wave band.

Images