The Evolution Of Submarine Fibre Optic Cable Technology - Direct Optical Detection And Chromatic Dispersion

The era of subsea fibre optic technology began with a very simple, elegant approach. In the digital world we rely on binary code as the language for networks, computer processing, and storage. One advantage of the binary format is that any data can be represented by only two symbols, zero and one. This means faster and cheaper processing since there are only two fundamental building blocks to which technology must develop physical counterparts. Binary makes this particularly easy. The absence of  an electric charge, magnetic orientation or pulse of light is interpreted as a zero. Their presence is interpreted as a one. This simple schema leads to relatively infrequent errors or mistakes. And just as importantly, errors are easily detected and fixed. In contrast, an analog system relies on a continuous range of values and hence can have many fundamental building blocks. While this might seem more efficient due to the infinite richness of potential values rather than relying on longer strings of zeros and ones, it is not. The more potential values, the more complicated the computer and laser technologies must become. Furthermore, analog signals are notoriously susceptible to errors and it is more difficult to detect and correct them. In general, analog networking and computers are much more expensive and less reliable than their digital counterparts. They also do not enjoy Moore's law. 

So fibre optic networks communicated from the 1980s through 2008 simply by varying light intensity between two values, on and off or one and zero. However, every laser light pulse consists of different frequencies. A pulse of light includes a range of frequencies or colors with the modal frequency having the greatest optical power or electromagnetic energy. It is the peak of a symmetric distribution of frequencies plotted against power. See the chart below. 

This graph shows optical power or light intensity of a laser pulse graphed against frequency or wavelength. The modal frequency (half of the spectrum lies below and above it) is 1549.5 and it shines the most brightly of all the colors comprising the laser flash. Unfortunately, electromagnetic frequencies do not travel at the same speed through the glass of a fibre optic strand. Frequencies reach the far end at different times. This is called chromatic dispersion. This link provides some intuition into the phenomenon: https://www.youtube.com/watch?v=PWaNMjimtP0. Chromatic dispersion confuses the optical receivers that must interpret light pulses as binary code. Over a long enough distance on a long haul link the following binary signal can become scrambled: 1 0 0.  The one pulse of light travels a long distance and becomes spread out over time as some frequencies race ahead and others lag behind. Hence while the laser may have emitted a light pulse during the first interval, but nothing during the second and third, the optical receiver sees the same pulse spread out over the three intervals. Hence the receiver will interpret it incorrectly as 1 1 1 when the laser actually sent 1 0 0. 

Once lasers reached the 10 billion per second bit transmission rate, the bit intervals became so small that dispersion blocked further throughput improvements. Receivers simply could not distinguish at acceptable error rates pulses of light from their absence. As we will see in the next post, this led to digital signal processing as the next step in optical communication evolution.