The design of EDC is dedicated to solve the dispersion interference, in order to reduce the optical signal damage caused by the dispersion. There are three types of dispersion including: chromatic dispersion, modal dispersion, and polarization mode dispersion. When the fiber transmission signal rate reaches 10Gbps, the chromatic dispersion interference problem appears, and as the data transmission symbol rate increases, the problem becomes more serious and becomes the biggest obstacle to increasing the rate. Using EDC technology, the 10Gbps SONET single-mode fiber connection distance can reach 145 kilometers (the poor fiber can reach 120km), and OC-48 (2488.32Mbit/s) can be seamlessly upgraded.
color Degree dispersion
After optical transmission for a certain distance, the optical pulse is stretched, causing inter-symbol interference. This is because different wavelengths of light have different transmission speeds in the same medium (the propagation speed of light of different wavelengths in vacuum is the same, but it is different in other media).
Mode dispersion
Mode dispersion mainly refers to the multimode fiber used in short-distance data center and building backbone links. Multimode optical fiber transmits lasers with multiple modes. Different modes of light have different path lengths, and finally form dispersion when it reaches the receiving end.
Polarization mode dispersion
The problem considered in single-mode fiber is that when the single pulse light at the transmitting end reaches the receiving end, it becomes multi-pulse light. If the fiber is round enough and not bent, etc., then this will not happen, because the light of different polarizations arrives at the receiving end at the same time.
Electronic Dispersion Compensation Algorithm Selection
There are many equalization algorithms based on them to realize effective electronic dispersion compensation. The continuous time filter (CTF) is the simplest one to implement on the chip and has the advantage of low power. The continuous-time filter adjusts the analog bandwidth of the optical front end by expanding and limiting the relevant frequency band.
Continuous-time filters can bring benefits to optical applications where the optical signal-to-noise ratio (optical noise) is limited by band-limited channels, and can also compensate for multi-chromatic dispersion through wave shaping. The continuous-time filter has limited benefits for noise-loaded channels that require high frequency boost because it affects the signal-to-noise ratio.
In terms of the implementation of electronic dispersion compensation, the most common architecture is based on a combination of feedforward equalizer (FFE) and/or decision feedback equalizer (DFE), which uses a continuous-time filter More complex signal conditioning methods used in the. FFE and DFE are usually multi-tap architectures and are effective methods to compensate for inter-symbol interference. When there is only a single unit interval interference, FFE/DFE only needs to determine whether a symbol has been extended to adjacent symbols, and then increase or subtract the symbol accordingly. When there is more than one unit interval of interference, not only does a single symbol extend to adjacent symbols, each symbol can be distorted by several adjacent symbols. The FFE part of the design focuses more on eliminating the distortion before the main energy point of the symbol (also called the precursor region), while the DFE part aims to compensate for the interference behind the main energy point of the symbol (also called the back body region).
The most common FFE (feedforward equalizer) implementation method is based on an analog distributed amplifier, in which the delay element is implemented by various delay lines on the chip. Implementing DFE (Decision Feedback Equalizer) requires a bit rate clock and uses sampled data to determine signal quality. The design of DFE can be analog or digital, depending on the selected architecture. For analog design, its energy consumption is generally lower than digital design, because the analog signal does not have to be converted to the digital domain, which eliminates the need for high-speed analog-to-digital converters and digital signal processing (DSP). When comparing analog and digital FFE/DFE implementation methods, performance stability at the operating angle is another trade-off factor that must be considered. There are also some more complex equalizer architectures, their implementation takes the form of a maximum likelihood series estimator (mlse), using the viterbi decoder algorithm. MLSes are generally digital designs, and more complex digital signal processing methods are required to perform filtering. The maximum likelihood series estimator can achieve better performance than the decision feedback equalizer, but the DSP implementation of the filter is generally more complicated and often consumes 2 to 4 times the energy. In this way, the maximum likelihood series estimator is often reserved for applications where the performance value provided is really worthwhile, such as when serious nonlinear problems are encountered in the application, or for ultra-long-distance optical fibers.
Problems in the realization of electronic dispersion compensation
The ideal situation is that an implementation of electronic dispersion compensation can dynamically adapt to any link. However, each optical link has different characteristics, including its length, quality, fiber condition, and other distinguishing factors. Currently, long-distance optical links use a dispersion compensation filter or some other fixed means to manually adjust the distance and wavelength. If the electronic dispersion compensation algorithm is self-adaptive, network technicians can simply insert a new line card instead of adjusting the settings based on the single link connected by the line card, making the installation process move toward a true "plug and play" A step away. In addition, since the characteristics of the optical fiber degrade over time, that is, more kinks will appear in the optical fiber, the line card can re-adjust the connection frequently without human involvement. The adaptive electronic dispersion compensation algorithm also facilitates the use of a single circuit board design to cope with multiple applications.
In order to achieve self-adaptability, the realization of electronic dispersion compensation algorithms often uses a very mature least squares (LMS) algorithm, while applying a feedback mechanism and a method of estimating signal quality. In the implementation process, a closed loop is adopted, and the line card can be self-adjusted to make small adjustments to the gain and filter to obtain the best signal response. When the electronic dispersion compensation equalizer is directly integrated on the transceiver device, dynamic adaptation is easier to achieve.
On multimode fibers, modal dispersion is generally more prominent, and it can extend to several unit intervals instead of just one or two unit intervals. Due to these factors, the electronic dispersion compensation algorithm must provide more complex equalization for short-distance multimode fibers than for single-mode fibers with a distance of 145km.
Another important component of the electronic dispersion compensation design is the variable gain amplifier (VGA). By the time the optical signal reaches the receiver, its amplitude has been significantly reduced. The variable gain amplifier gives gain according to the input signal to provide the filter with the maximum dynamic range. The variable gain amplifier keeps the output stable regardless of changes in the input signal flow within a given dynamic range.
Standardization of electronic dispersion compensation
Electronic dispersion compensation is a very critical technology, so OIF and ITU are developing application codes for SONET long-distance applications , IEEE developed the standard 802.3aq based on electronic dispersion compensation for 10GBASE-LRM.
The 802.3aq standard is aimed at applications with longer spans, and its minimum chromatic dispersion must be at least 2400ps/nm, which is equivalent to a nominal optical fiber of approximately 140km. The goal of the standard is to enable the existing OC-48 link to be upgraded to 10Gbps/OC-192 without having to replace the existing fiber or use dispersion compensation fiber. This will allow telecom operators to replace transponder modules (and back-end devices such as framing). The end result is the ability to upgrade equipment without having to upgrade the link itself.