Fiber Bragg Grating Dispersion Compensation Enables Cost-Efficient Submarine Optical Transport By Fredrik Sjostrom, Proximion Fiber Systems Undersea optical transport is an important part of the infrastructure making up the world-spanning digital transport network. Due to extreme link length and stringent reliability requirements, this important part of the optical backbone poses quite a challenge compared to the more conventional terrestrial systems hence specific technical solutions are required. Almost every submarine network system needs to be tailor made in order to maximize link performance while minimizing cost. Achieving cost effectiveness without jeopardizing the reliability of these long term investments in infrastructure poses a real challenge for system designers. This article discusses how the unique properties and design possibilities relating to Fiber Bragg Grating (FBG)-based dispersion compensation technology can facilitate desirable cost savings by supporting new terminal designs and link optimized compensations. Chromatic Dispersion Chromatic dispersion is basically a phenomena caused by the frequency-wavelength) dependent group velocity of an optical signal, i.e. different colors of light travels at slightly different speeds in an optical fiber. Chromatic dispersion is measured in picoseconds per nanometer (ps/nm) and is defined as the derivative of the group delay with respect to the wavelength (Equation 1). Equation 1: The effect of chromatic dispersion can quite simply be described as a smearing or spreading of light pulses as they traverse an optical fiber. If left uncompensated, this pulse distortion will lead to overall
signal degradation and inter symbol interference, hence leading to erroneous transmissions at quite moderate distances, way below those addressed by Ultra Long Haul, (ULH), undersea systems. Submarine Optical Transport Overview There are basically two main types of submarine transport systems: repeatered and un-repeatered. Repeatered submarine systems are typically ULH systems relying on submerged distributed amplifiers to boost the signal as it is attenuated along the transmission fiber. The word repeaters may actually be somewhat misguiding since today s systems mainly rely on optical amplifiers and not Optical-Electrical- Optical (OEO) repeaters, as was the case in early systems. For some unknown reason, this classification name seems to have survived this change and remained within the submarine community. Un-repeatered systems are systems with no submerged amplification hence these links are normally shorter than the repeatered ones. These systems are typically coast or island bound, so called festoon networks, relying on shore-based landing stations for amplification when needed hence making them quite similar to terrestrial transport systems. Due to the great variety of link distances, transport capacity (bit rate; channel count) and overall system configuration the choice of fiber types used varies a lot from one undersea project to another. For shorter distances a single fiber type is typically used, but when the distance increases, various types of hybrid solutions (fiber mixing) can be used. Two examples of hybrid schemes can be seen in Figure 1. Figure 1: In-line dispersion compensation approach utilizing multiple fiber types
In systems not utilizing a fiber mix, i.e. single fiber systems, the choice of fiber is typically decided by the number of channels required. High channel count automatically means increased non-linearity penalties. This is normally mitigated by the use of Large Effective Area fiber (LEAF). However, LEAF is normally associated with a quite high dispersion slope, which in turn leads to high length dependent residual slope build-up, i.e. different channels will experience very different amounts of dispersion throughout the link. For longer links where different fibers are mixed, the situation is quite similar with the exception of bulk dispersion elimination. By mixing fibers with positive and negative dispersion the residual dispersion for the center channel can be reduced to zero, but due to the difference of the slope parameter between the fibers there will still be a residual dispersion present as described in Figure 2. Figure 2: Typical accumulated residual dispersion in a submarine optical link This residual dispersion can be quite substantial for the edge channels and values of several thousands of ps/nm are quite common. This poses huge problems in terms of link design because in order to optimize transmission performance, the desired residual dispersion value for each channel may be as low as 50 ps/nm for certain systems.
It is not only the direct effect of high, channel specific, dispersion values directly requiring per-channel dispersion compensation in order to function properly that causes problem when designing these long links, but also a number of additional and potentially harmful and cost driving issues need to be taken into account as well. For instance, high and especially channel dependent dispersion, give rise to optimization problems in today s Dense Wavelength Division Multiplexing (DWDM) networks. DWDM is a technology in which a single fiber carries numerous closely spaced optical signals at different wavelengths simultaneously. In DWDM networks adjacent signals tend to affect each other negatively, causing non-linear effects such as four-wave mixing (FWM) and cross phase modulation (XPM) etc. The fact that penalties caused by these effects are coupled with the individual dispersion induced pulse distortion, makes these links difficult to design hence access to suitable dispersion management components are vital in order to lessen the impact and presence of non-linearities throughout the link. It is also a known fact that the residual dispersion historically has been one of the main hurdles in deploying wide-band (full C-band) optical transport supporting dense channel plans (> 80 channels). In recent years so called dispersion managed or dispersion flattened fibers have come into play for the especially challenging links in order to address this problem. These dispersion managed hybrids typically use two specialty fibers with opposite dispersion as well as opposite dispersion slope that when used together at a specific length ratio minimizes the residual dispersion. These fiber cables are based on careful selection of specialty fibers making them quite expensive and these links do currently not yet constitute any greater part of the installed fiber base. Also worth mentioning is that the larger portion of the installed submarine fiber base is today unlit. Since a large number of these fiber cables were deployed around the turn of the century, hence predatingthe deployment of dispersion flattened cables, the majority of these demonstrates substantial residual dispersion. As discussed above, the amount of residual dispersion poses a lot of problems and in some case even limits the range of usable wavelength to a quite narrow span centered at 1550 nm. In recent years, as the demand for bandwidth steadily increases, the interest in utilizing this dark fiber capacity has gained much interest and is emerging as a very cost competitive alternative to new fiber link deployment. Also, upgrading older 2.5 G lines to 10 G is an economical viable approach to increasing capacity. However, since the sensitivity to dispersion basically scales with the square of the bit rate, adding adequate dispersion compensation is a necessity, hence providing an application space in which FGB- DCMs really can provide great value.
Fiber Bragg Gratings vs. Dispersion Compensating Fibers Chromatic dispersion compensation using highly efficient reflective FBGs is fundamentally different from the incumbent technology used for dispersion compensation, namely Dispersion Compensating Fiber (DCF). Dispersion management utilizing FBGs is based on the introduction of wavelength specific time delays through the use of a precisely chirped fiber grating which can be tailored to mimic the dispersion characteristics of the fiber or span intended for compensation. A graphical illustration of an FBG is shown in Figure 3 while the optical compensation principle is shown in Figure 4. Figure 3: Reflective Fiber Bragg Grating Figure 4: The FBG-based dispersion compensation principle
The traditional means of overcoming the issue of residual dispersion has been to incorporate bundles of DCF into the submarine terminals. DCF comes in different types, but can generally be described as a specialty fiber having a dispersion coefficient with an opposite sign than that of the dispersion for which to compensate. Typically, in the case of negative DCFs, the high level of dispersion necessary is achieved by reducing the diameter of the fiber core and utilizing depressed cladding designs. Unfortunately these DCFs suffer from two major problems. First, because a substantial part of the supported fiber mode resides in the cladding, these fibers exhibit quite high loss, leading to additional amplification needs, hence increased Amplified Spontaneous Emission (ASE). Second, the small effective area leads to an inherently large nonlinearity parameter which in combination with the high relative optical intensity, given a fixed power in the larger transport fiber, leads to severe non-linearity penalties. Due to the immense fiber length required to compensate the accumulated dispersion in long links these components also tends to be extremely bulky thereby requiring much expensive terminal space. On the other hand, FBG-DCMs are small, low loss devices and since they just consists of a couple of decimeters or meters of fiber, not tens or hundreds of kilometers as with fiber based dispersion compensation, the non-linear effects are negligible even at extreme optical power levels. FBG-DCMs also have the ability to provide both negative and positive dispersion by simply changing the grating chirp compared with fiber based dispersion compensation, which requires the use of totally different fiber types. Two main types of FBG-based dispersion compensators are commercially available today channelized and continuous. The channelized version provides channel spacing specific, or grid specific compensation whereas the continuous type provides, in much the same manner as a DCF, continuous compensation throughout the C or L band, hence providing total channel plan independence, an important feature especially for ensuring future upgradeability. Furthermore, an increasingly popular method to solve time varying dispersion and save costs on commissioning and part number reduction is to utilize colorless tunable FBGs units in the system architecture. The favorable performance regarding insertion loss, latency and the lack of non-linear effects at high optical power in combination with a significantly smaller form factor is a key property not only allowing for cost-efficient network designs in general, but proving especially suitable for submarine applications.
Channelized Compensation Solutions FBG technology has proven itself to be a very suitable and cost-effective technology for channel-specific dispersion compensation, i.e. channelized devices, having set dispersion for a specific channel or channel plan. The technique of cascading channelized FBG units in a per-channel approach, with suitable and differentiated dispersion values, to counteract channel dependent residual dispersion has gained in popularity lately due to the low cost and immense savings on costly terminal space that can be achieved. A typical channelized DCM basically operates over the full C-band and provides a set dispersion for any defined ITU channel. The most common channel spacing for these components is 100 GHz but even though modern submarine systems normally utilize very dense channel spacing schemes, down to 33 GHz or even 25 GHz, this poses little or no problems since the channels can, by design, be offset with an arbitrary frequency, hence any channel can be compensated. A total dispersion compensation solution can be achieved using a suitable menu of compensators with suitable offsets, while maintaining a low part number count hence keeping down inventory and OPEX. An illustration of the basic principle can be seen in Figure 5. Figure 5: Residual dispersion per-channel compensation utilizing channelized FBG-DCMs Figure 5 depicts a system with only post compensation but both pre and post compensation is normally utilized to achieve the desired dispersion map. One of the immense advantages favoring this solution is the small form factor that can be achieved. A channelized FBG for negative dispersion compensation is about 10 to 20 times smaller than a DCF equivalent and for positive compensations the space saving is even greater since ordinary SMF is normally used for this, hence requiring spools with hundreds of kilometers of fiber in the terminals.
The form factor is especially important when upgrading systems. Normally, the terminal space is limited and since the dark channels normally are situated far from the C-band center, high residual dispersion, i.e. large compensations, are required. In some cases using DCF and especially ordinary SMF (for the positive compensation) is not even an option for these upgrades because the space requirement is just to great and/or the extra amplification needs cannot be accommodated. Furthermore, the extreme low insertion loss, resilience to non-linearities and the ability to provide dispersion values of thousands of ps/nm in a single module makes this an ideal component submarine terminal use. Continuous Compensation Solutions Another method of dealing with residual dispersion is to utilize continuous FBG-DCMs. Unique properties and ability to tailor make these gratings to fit virtually any residual dispersion slope, in theory eliminating the need for channel specific compensation, has gained a lot of interest in the submarine community. To tailor a DCF for a specific dispersion slope other than a standard fiber like SMF or LEAF, is not really economical viable. FBG based solutions, on the other hand, are very different in nature since the dispersion properties are solely decided by the grating characteristics and not the fiber design itself, i.e. no special pre-form, drawing equipment or index profile design has to be developed and manufactured. To manufacture tailored FBG-DCM compensations, with specific dispersions and dispersion slopes, basically just involves exposing gratings with different chirp properties, which provides a cost efficient and fast design-to-product alternative to the channel or banded DCF compensation approach. Furthermore, in comparison to channelized dispersion compensators, these components have the ability to continuously compensate every channel over the full C-band or over specific sub-bands. The continuous nature of the compensation also makes the solution totally channel plan independent ensuring future upgradeability. The possibility of combining different FBGs enables some quite remarkable designs as well. For instance it is possible to achieve zero dispersion at a specific wavelength, while maintaining a controlled slope, by combining two gratings. The optical principle, group delay and dispersion properties of such a component, a.k.a. Slope Compensator, is presented in Figures 6a, 6b, and 6c.
Figure 6a: Double grating FBG-based dispersion compensation principle Figure 6b: Group delay versus Wavelength measurement graph
Figure 6c: Dispersion versus Wavelength measurement graph In Figure 7, a number of different approaches to continuous compensation are described, ranging from pure slope compensation to flat compensations. The choice of dispersion characteristics to utilize will naturally vary based on terminal design and specific system requirements. Figure 7: Different types viable of continuous dispersion compensation designs
Much in the same manner as for channelized FBG-DCMs, the use of continuous FBG-DCMs instead of bulky spools of DCF, can massively reduce the space needed for compensation. Also, low loss is important in these applications since less amplification is needed, even further addressing the cost and space requirements. FBGs, A Technology Addressing Every Aspect Of Terminal Design Figure 8 depicts a generic terminal design and aims to exemplify where in the terminal topology different types of FBG based dispersion compensations can be used. Figure 8: Terminal example including different types of FBG-based DCMs The great variety and immense design freedom in combination with unique properties such as low loss, small form factor, tunability and resilience to non-linearities provided by FBG based DCMs has in recent years made this technology the technology of choice for numerous designers of undersea systems. It has provided the undersea telecommunication industry with unparalleled possibilities when it comes to cost and performance network optimization.
About The Author Fredrik Sjostrom (fredrik.sjostrom@proximion.com) is Technical Sales Manager at Proximion Fiber System AB (www.proximion.com).