Non-linear optical amplifiers on operators’ cable communication lines: theory and practice

Non-linear optical amplifiers on operators’ cable communication lines: theory and practice
Vladimir Myagkiy
head of telecommunications networks

Russia, along with other countries, has made a breakthrough in the development of its telecommunications sector within a very short period of time. It seems as though just yesterday, the very concept of electric communication was consistently associated in our minds with a regular dial telephone and ever-present short beeps. In the past few years, the main trend of development for the telecommunication market has been the continuous growth of the throughput capacity of communication channels. The dialup Internet access has been forgotten already; solutions based on DSL technology are outliving their usefulness.

Currently, the growth of throughput capacity has snowballed because the telecommunication market has been flooded with the latest broadband access solutions offering each subscriber access to various services at the speed that would have been considered a real breakthrough by communication operators even a few years ago.

Even a couple of years ago, analysts predicted the following growth trend of penetration of broadband access services (see Fig. 1).

The Russian telecommunications market has outpaced those rather optimistic forecasts, and even now the penetration of broadband access services to households is assessed at 50%. The increase of throughput capacity of subscriber channels has caused the need to expand the bands of trunk and local channels.

Even at the beginning of the past decade, the low-capacity systems based on fibre-optic communication lines (PDH, SDH) were replaced with multichannel systems with wavelength-division multiplexing, allowing to arrange dozens of channels with the throughput of 10, 40, 100 Gbps each in a single fibre.

In the past few years, manufacturers of transmission system equipment have aimed to increase the number of channels and the throughput capacity of each minuscule portion. Even now many operators are expanding their existing lines to reach the maximum possible density of all available fibres.

A major criterion for equipment applicability in the trunk lines of operators is section length of the fibre-optic communication line without intermediate reinforcement. Even a few years, ago this kind of problem was typical only for transcontinental communication lines and was explained by the difficulty of signal regeneration in an extended underwater section of a FOCL. These days, market conditions require that operators minimise the cost of equipment acquisition and costs of communication node operation for every section. A simple pattern occurs: the fewer nodes for communication with equipment there are in a line, the more effective are the investments in the building of a modern telecommunication infrastructure. The described relation is typical not only for Russian telecommunication market; thus, equipment manufacturers have long been looking for a method to improve the energy characteristics of a FOCL line. The improvement of characteristics of optical amplifiers based on EDFA technology can no longer ensure the required breakthrough in both increasing the length and expanding the width of the bank, because these amplifiers are characterised by insufficient signal/noise ratio in case of a large amplifying coefficient.

A way out was found: the problem could be solved by using non-linear effects in the optic fibre to build optical amplifiers.

The effect that allowed taking telecommunication equipment to a new level was discovered as early as in 1928 by a group of Indian physicists led by Chandrasekhara Venkata Raman and became known as Raman scattering; the researcher received the Nobel Prize for his discovery.

The purpose of this article is not to describe the theory of building non-linear amplifiers, though we believe it necessary to provide some theoretical grounds before we move on to the analysis of practical experience in using non-linear amplifiers in Russian operators’ communication networks.

The Raman effect is essentially a stimulated inelastic scattering, where a photon of an incident beam (in modern optical systems, this is the laser pumping beam installed in the amplifier) is divided into a photon of a smaller frequency and a phonon. At that, the so-called Stokes wave occurs, having a frequency equal to the difference of the pumping frequency and the main signal frequency.

Achievements of other outstanding scientists must also be mentioned here. For example, in case of different pumping powers, one of them arising if the power is equal 10 mW became known as the stimulated Brillouin scattering (SBS), and the second one, arising in case of a pumping frequency of about 1 W, became known as stimulated Raman scattering or combined scattering (SRS).

A feature of these two phenomena is that the intensity of the two scatterings may increase by a hundred million times as compared to the principal signal, which allows using non-linear effects in the development of fibre-optic amplifiers. If the frequency of the pumping signal is selected correctly for each effect and working range, this will result in a materially amplified principal signal at the output of the device's amplifier.

At the current stage of development of science and technology, the SBS phenomenon has not yet found practical application in the shape of a completed product. The obstacle for using this discovery at the current stage of development of optical science and technology is insufficient band width, making it possible to amplify only one wave length so far, as well as various side phenomena (for example, the influence of the backscattered wave). It is likely that in a few years there will be amplifiers utilising the SBS effect in optical systems, and detailed descriptions of serial amplifiers based on SBS will be given in research publications.

The Raman effect is free of the above inconvenient features, so Raman amplifiers appeared before SBS amplifiers and were implemented in modern telecommunication equipment. The first serial amplifier was presented in 2012. These days, practically all manufacturers of telecommunication equipment offer solutions based on Raman amplifiers.

The pass bandwidth of Raman amplifiers amounts to 5-10 THz, which allows successfully using these amplifiers in dense wave-division multiplexing (DWDM). Amplification coefficients of industrial-grade Raman amplifiers reach 20-30 dB.

However, EDFA amplifiers should not be considered a thing of the past. Firstly, the amplifiers built based on the Raman effect are not inexpensive, and using them in short sections of a fibre-optic network is at least unreasonable. Moreover, the experience of implementing these solutions even in longer sections has demonstrated certain characteristics of the devices, making their application more difficult.

If we look at the history of the matter, we will see that communications officers have always aimed to create branched networks with a large number of intermediate points to simplify traffic output in any direction. This approach was considered competent and flexible planning. Many engineers were justifiably proud that they were able to plan a network so that in a single intermediate cross they could switch to practically any direction with a simple patch cord.

All these solutions that were generated for years and carefully included in the cross list became the worst enemy of the tracks using Raman amplifiers.

The reason for this is simple and has to do with reflections of the pumping signal from optical connections, dangerous for the receiving paths.

Let us examine this phenomenon for a system including a Raman amplifier with back pumping or counter-pumping.

The pumping wave spreads to meet the weak principal signal and after interacting with it creates a reverse Stokes wave, amplified to the working level of receiving paths of the equipment. At the same time, the powerful pumping signal reaches the nearby connection of optic fibres at the cross. The signal reflected from the connection returns to the receiving paths of the equipment. In case of material irregularity as well as a short distance to the connection, the reflected signal has material power. Such signal contacting the pass band of DWDM equipment may in time damage the sensitive input paths of the equipment.

For protection from this damage, equipment manufacturers include special reflected-signal level analysis circuits in Raman amplifiers to automatically lower the power of the pumping signal and prevent the amplifier from going beyond the estimated/maximum amplification coefficients, as a result of which the required power characteristics of paths are not achieved at the fibre-optic line section.

The only way to counteract this phenomenon is to fight the harmful reflections. As the experience of our company specialists has demonstrated, this struggle takes place at practically all sections involving Raman amplification. After a successful launch of a large number of lines using these amplifiers, a conclusion was made that the difficulties with launching were not due to lack of qualified engineers, but to characteristics of existing cable lines and the culture of their maintenance. Consistent implementation of measures recommended below enabled our company to achieve positive results.

The first and most widely spread cause of incomplete operability of amplifiers is the impact of regular contaminations on the switched connections. Unfortunately, maintenance services have not been paying due attention to regular procedures of maintaining perfect cleanliness of optical connections. All switched optical connections need to be thoroughly cleaned from contamination, though a common blowing out of a contact will bring no results, nor will the use of cleaning tapes and other methods of cleaning. Without the use of a microscope, all of the above methods are useless. It is necessary to thoroughly clean all switched connections in the crosses and patch cords at the optical distance of at least 10 km from the place of installation of a counter-pumping Raman amplifier. The operation of our amplifier will be influenced by a single fibre connected to the Raman amplifier, but we recommend bringing both paths to the state of operational readiness and correctly cleaning all connections in the cable section, from receiver to receiver.

As a rule, after a thorough performance of all of the above actions, the number of inoperable or partially operable paths is approximately halved. Do not be surprised if in the same section one of the direction works and the return one, having an absolutely identical length and route, refuses to work - this is a widespread phenomenon for paths with non-linear amplifiers, and it means only that the problem needs to be found along the route, close to the malfunctioning Raman amplifier.

The following activities for launching amplifiers are the most complex and uncomfortable for the cable infrastructure.

Determine the optical connections closest to the amplifier that fails to work at the estimated power. One of the methods of lowering the level of harmful reflections is the replacement of optical connections with connections having the “angular end polish”. This type of pigtails and patch cords are not used in networks frequently, but our experience confirms that Step Logic specialists were able to launch a number of sections using this method. I am confident that after performing the above activities the number of problems along the route will be reduced by another 20-30%. The physical nature of this phenomenon is simple and will be clear to practically any technologically trained specialist: the intensity of reflection from the joint of two surfaces placed at an angle other than 90 degrees will be materially lower, which ensures the lowering of power of the reflected signal. After replacing the pigtails and patch cords do not forget to once again clean the connections from contamination with a mandatory use of a microscope.

Further measures to improve the line’s characteristics have to do with material changes in cable infrastructure.

At sections where, despite all efforts, the amplifiers cannot be brought to maximum power, you will have to consistently “straighten” the fibres in the crosses. You will have to begin with the nearest cross. A direct connection of optical cables completely eliminates the influence of reflection of the pumping wave; therefore, after removing all optical connections at the distance of about 10 km along the optical fibre from the place of installation of a Raman amplifier, you will be able to launch all intervals of your line using Raman amplifiers.

A lot of specialists will probably argue that the activities described above are based on trial and error, and if a set of reflectometric measurements is performed on a fibre-optic line section in advance, no one will have to search for the problem because it will be clear where to straighten and where to clean.

It is true, but we also recommend taking reflectograms of each elementary section before connecting amplifiers so as to determine every unevenness in the optical fibre. Besides, you need to bear in mind that in our case reflectometric measurements need to be performed on two sides of an elementary cable section to be able to receive reliable data on the characteristics of reflections in each of the directions. Measurements results will primarily allow to reliably determine every unevenness of the optical line, which might include low-quality welds in joints, low-quality welds in crosses, low-quality patch cords, and fibre damage.

Moreover, it needs to be mentioned that in a large number of sections, after receiving satisfactory ORL results (about -45…-50 dB), Step Logic specialists were forced to perform the above set of activities to successfully launch the paths. At the same time, in some sections, where ORL results were completely unacceptable (about -25…-30 dB) the very same Raman amplifiers were launched and worked without interruptions.

This is another mystery we have been unable to solve so far. The optical cross closest to the equipment with a Raman amplifier, as a rule, has little to no influence over the launch of paths, though the reflection from the closest connection was the most intensive. At that, at several sections straightening of the fibre located within 10 optical kilometres from the equipment caused immediate reach of full power by the amplifier.

It is possible that in future our colleagues or we will be able to demonstrate the nature of the above mysteries.

We hope that the described experience of Step Logic employees will help other specialists in our sector to build communication networks for telecommunication companies.

Back to all opinions