Fullscreen HTML5Med Res (??MB)HTML4

UNDERSEA COMMUNICATIONS DATA TRANSMISSION & POWER CABLES We don’t hear much about undersea communications cables, even though they carry around 99% of international internet traffic. Without undersea cables, the internet as we know it would not exist. By Dr David Maddison, VK3DSM U ndersea communications cables are the most invisible and yet one of the most important parts of the internet. Compared to alternatives such as satellites, cables are much cheaper and offer much lower latency (delays due to the time the signal takes). As of June 2024, there were 600 active or planned submarine communications cables and 1.4 million kilometres of cables in service (see Fig.1). The lengths range from 131km for the CeltixConnect-1 (CC-1) from Dublin, Ireland to Holyhead, United Kingdom to 20,000km for the Asia-­America Gateway (AAG) Cable System from the United States to various places in Asia and the Pacific. There are usually multiple cables connecting each country to provide 14 Silicon Chip redundancy in case of accidental or deliberate damage. Communications cables carry not only internet traffic including video but also telephone calls and private computer networks. Undersea communications cables originate with the first undersea telegraph cables. There are also undersea power-­ c arrying cables. Undersea telegraph cables Before there was significant (or any) radio traffic, there was an extensive network of undersea telegraph cables. Fig.2 shows the Eastern Telegraph Company’s international telegraph network in 1901. On the 12th of December, 1901, Marconi conducted the first transatlantic Australia's electronics magazine radio transmission from Cornwall (UK) to Newfoundland (Canada), using a 150m-long kite-supported antenna for reception. Marconi established a commercial service for ships at sea in 1904 and a transatlantic radio-telegraph service in 1907. However, that service was not reliable for many years. Thus, there was still a demand for cabled telegraph services in the early 1900s. Today, there is still competition for communications between optical fibre and wireless services, including via satellites. Land-based cables were uninsulated and suspended between poles, but subsea cables must be insulated. Few suitable materials were known in the early 1800s. siliconchip.com.au In 1843, Michael Faraday sent samples of the natural rubber-like material gutta-percha from a tree of the same name from Singapore to London for testing. In 1845, Sir Charles Wheatstone suggested it be used to insulate a cable between Dover and Calais. The cable was laid in 1850 and was successful. The first attempt at laying a telegraph cable across the Atlantic was in 1858; it was ultimately unsuccessful. It was laid between Ireland and Newfoundland and worked extremely slowly for a few weeks before being destroyed by applying too high a voltage (2000V) to it in an attempt to speed it up. The problem was that signals were ‘smeared out’ at the receiving end, significantly reducing the transmission rate, as subsequent signals would interfere with prior signals. This was due to cable capacitance. The cable acts as a long, thin capacitor, with one electrode being the conductive seawater on the outside and the other the central conductor. This meant the transmission rate had to be dramatically reduced to receive intelligible signals. The speed was so slow that a 99-word transmission between Queen Victoria and President James Buchanan took 16.5 hours, or ten minutes per word. Incidentally, in a classic engineering error, two cables were ordered from two suppliers and were provided with cable twists running in opposite directions. This would have made splicing them impossible, so a special bracket was improvised to hold the wires. Fig.1: just some of the current submarine cables worldwide. Source: www. submarinecablemap.com Transmission line theory In those early years, transmission line theory, shown in Fig.3, was poorly understood. In 1855, the future Lord Kelvin (William Thomson) made some theoretical progress and developed a model that predicted the poor performance of the 1858 cable. However, that did not lead to a complete understanding because he only considered capacitance and resistance but not inductance in the cable. Although Lord Kelvin was involved in that cable project, his concerns were not heeded due to internal company squabbles. He wanted a thicker cable. Nevertheless, he developed a highly sensitive mirror galvanometer to detect signals on the cable (see Fig.4). Morse dots and dashes were siliconchip.com.au Fig.2: undersea and overland telegraph cables of the Eastern Telegraph Company, the largest cable company in the world in 1901. Source: www. zmescience.com/other/great-pics/map-undersea-cables-18112010 Fig.3: an electrical model of a transmission line, such as an undersea telegraph cable, with resistive (R), conductive (G), capacitive (C) & inductive (L) components Australia's electronics magazine December 2024  15 represented by negative or positive pulses rather than pulses of differing duration. In 1876, Oliver Heaviside revolutionised the understanding of transmission lines and published the first of his papers on analysing the propagation of signals in cables. They included the ‘telegrapher’s equations’: δ/δx V(x, t) = -L δ/δt I(x, t) − RI(x, t) δ/δx I(x, t) = -C δ/δt V(x, t) − GV(x, t) These use resistance, conductance, inductance and current to predict voltage and current distributions in transmission lines as a function of distance and time. They are derived from Maxwell’s equations. More transatlantic cables A second cable was laid in 1865, but it broke over halfway across and could not be recovered after numerous attempts. A third cable was laid in 1866, which was successful. The 1865 cable was also retrieved and repaired, so there were two cables in service. Remarkably, although it took several attempts, the 1865 cable was recovered with a grappling hook at a depth of 4km. The line speed was decent at seven words per minute, much faster than the 1858 cable. Like the 1858 cable, these were laid between Ireland and Newfoundland. The effective line speed was further improved with Julius Wilhelm Gintl’s development of duplex transmission Fig.4: Thomson’s mirror galvanometer could detect extremely small currents. Source: https://w.wiki/AoEY 16 Silicon Chip in 1872, which allowed two messages to be sent simultaneously in different directions. In 1874, Thomas Edison invented quadruplex transmission, which allowed four messages to be sent simultaneously on one cable, two in each direction. Fig.5 shows manufacturer samples of the 1858, 1865 and 1866 cables. Each cable had a thicker version for the shore and continental shelf sections and a thinner version for the deep ocean. Note that there was no repeater technology in this period (as used routinely today), so a signal needed to travel the entire length of the cable without being amplified or having its waveform conditioned in any way. That made the feat of transoceanic communications even more formidable. In 1866, a transatlantic telegraph message cost US$10 per word with a ten-word minimum. Back then, $100 was 10 weeks’ pay for a skilled worker. That is equivalent to US$2000 or $3000 today – for a ten-word message! The first undersea telegraph cable connecting Australia ● The first undersea telegraph cable connecting the Australian mainland to Tasmania was built in 1859. It had numerous problems and was abandoned in 1861. Another cable was installed in 1869, running from Cape Shanck, Vic to Low Head, Tas. ● In 1871, the first cable connecting Australia to the rest of the world was installed from Darwin to Singapore via Java (see Figs.6 & 7). It was described at the time thus (siliconchip. au/link/abyz): The cable consists of seven small copper wires—a central one, with the six twisted round it. It is insulated by gutta-percha, over this is a coating of tarred hemp, then a sheathing of galvanised iron wire, with an outside covering of tarred hemp. The deep sea portion is three-quarters of an inch in diameter, the intermediate one inch, and the shore ends (twenty miles in length) three inches in diameter. There is much information about this cable at siliconchip.au/link/abz0 ● In 1876, the first undersea cable was laid between Australia and New Zealand. ● In 1889, a third international link was laid from Broome, WA to Batavia (Jakarta). ● In 1891, a cable was laid from Bundaberg, Qld to Gomen (New Caledonia). ● In 1901, another cable was run from the Cocos-Keeling Islands to Perth, part of the global “Red Route” cable through British territories. ● In 1902, a cable was added from Southport, Qld to Canada via Fiji and Norfolk Island. For more information on the Southport cable, see the telegraph display in The Gold Coast Historical Museum Fig.5: a manufacturer’s sample case of products for the 1858, 1865 and 1866 Atlantic Cables manufactured by Glass, Elliot, and Co. They merged into Telegraph Construction and Maintenance Co. Source: https://atlantic-cable. com/Article/AtlanticCables Australia's electronics magazine siliconchip.com.au (www.gcmuseum.com.au) at 8 Elliot St, Surfers Paradise. You can see the remains of the cable hut of the Pacific Cable Station at Cable Park, Main Beach Parade, Main Beach, Gold Coast City. The Cable Station operated from 1902 to 1962. The All Red Line The All Red Line was a system of telegraph lines and undersea cables that linked most countries of the British Empire (Fig.8). The colour red was the traditional colour used on maps to indicate British Empire countries and colonies. It was built because the UK had security concerns about a vital cable network with landfalls that were not on territory they controlled. The first successful part of the cable was from Ireland to Newfoundland, Canada in 1866. The network was completed in 1902 with a final trans-Pacific cable from British Columbia, Canada to Fanning Island (then part of the UK and roughly in the middle of the Pacific Ocean). That section of the cable was funded by the UK, Canada, New Zealand, New South Wales, Victoria and Queensland. Australia’s first connection to the cable was from Darwin to Singapore via Java in 1871. Fig.6: a portion of the original Darwin to Java cable recovered from the Timor Sea in 2016. Source: https://digital-classroom.nma.gov.au/images/section-portdarwin-java-underwater-telegraph-cable-1871-72 Fig.7: bringing the cable to shore at Darwin in 1871. Source: www.pastmasters. org.au/overland-telegraph-amp-undersea-cables.html Cable circuits Telegraph cables generally had one central conductor. The return current path of single-core telegraph cables was through the sea; although sea water is not nearly as conductive as copper, the cross-section is high, so the resistance is low. At the low frequency of Morse transmission, such an arrangement worked satisfactorily. The currents involved in transoceanic telegraphy were extremely small and susceptible to many forms of landbased electrical interference. Therefore, the Earth electrodes for cables were run many kilometres out to sea to minimise such interference (see siliconchip.au/link/abz1 for further information). Fig.8: the All Red Line of telegraph cables connecting the British Empire, built between 1866 and 1902. Source: https://w.wiki/AoEZ Increasing telegraph speed One way of increasing the speed of a telegraph cable was to wrap the inner conductor with mu-metal, which is typically used today for magnetic shielding. Mu-metal was invented in 1923 and was used to provide inductive loading of subsea telegraph cables (see Fig.9) to compensate for the siliconchip.com.au Fig.9: “Loaded cable” as used on part of the Pacific cable route to increase transmission speed between England and Australia: (a) conductor made of copper; (b) continuous winding of “mumetal” wire; (c) gutta-percha insulation; (d) inner wrapping of jute; (e) sheathing of steel wires; (f) coating of composition; (g) outer wrapping of jute with external coating. Source: https://atlantic-cable.com/Cables/1902PacificGB Australia's electronics magazine December 2024  17 plastic jacket dielectric insulator metallic shield centre core Fig.10: the structure of a typical coaxial cable. A subsea cable has many more layers of insulation, reinforcement and armour. Source: https://w.wiki/AoEa Fig.12: how the repeaters were powered for the first transatlantic communications cable, TAT-1. cable’s capacitance. This enabled a much greater transmission rate. For example, in 1926, the busiest part of the Pacific cable from Fiji to Vancouver was duplicated with this ‘loaded cable’, increasing the transmission from 200 to 1000 letters per minute. Telephony through subsea cables Fig.11: a cross-section of TAT-1 coaxial cable. Source: https://w.wiki/ AoEb 18 Silicon Chip Single-wire subsea telegraph cables with Earth returns are unsuitable for voice because the attenuation is too great at higher frequencies due to cable inductance and capacitance. The signal was distorted and the cables were also too susceptible to interference. In 1877, Alexander Graham Bell attempted to make a telephone call over the Atlantic telegraph cable but the experiment failed. One attempt to resolve such problems was to ‘pupinise’ (named after Michael Pupin) a subsea cable. This involved adding inductors (loading coils) at regular intervals along it with balanced pairs of wires to increase its inductance, thus offsetting its capacitance. This method also allowed the use of thinner, cheaper wires. This technique was independently discovered by George Campbell at AT&T and Michael Pupin at Columbia University, based on Oliver Heaviside’s theory. Still, there were limits to the distance over which this technique was effective. A pupinised cable was laid across Lake Constance in Switzerland in 1906, and in 1910, such a cable was laid across Chesapeake Bay with 17 pairs of conductors. Pupinised cables had problems; the waterproofing materials available at the time were inadequate, and bulges in the cable where the inductors were installed mechanically weakened it. Continuous loading, with no cutoff frequency, was a superior method of Australia's electronics magazine solving the same problems as pupinisation. A project to install a continuously loaded transatlantic cable was underway in the 1930s, but it was abandoned during the Great Depression. By the late 1930s, repeaters and multiplexing provided more capacity on the same number of circuits at a lower cost, so cable loading was no longer necessary. Transatlantic radiotelephony A transatlantic radiotelephony service was also established in 1927. It charged US$45 for three minutes, equivalent to about US$800 or $1200 today. Thus, plenty of financial incentives existed to develop a cheaper service, but certain technological advances were required. Such advances included synthetic polyethylene insulation to replace rubber and gutta-percha from 1947 and reliable vacuum tubes for repeaters and coaxial cable. Modern coaxial cable was patented in 1929, although Nikola Tesla obtained a similar patent in 1894. Coaxial subsea telephone cables Coaxial cables have an inner conductor plus a shield around the outside (see Fig.10). They can carry high-­frequency signals with low losses and are therefore suitable for many telephone circuits and/or data/video. Coaxial cables are superior to single or multiple conductors in subsea cables. The first transatlantic telegraph cables (from 1858) were coaxial, but transmission line theory was not fully developed at the time, so they could not operate at high speeds. The first modern subsea coaxial cable was laid in 1936 and ran 300km between Apollo Bay near Melbourne and Stanley, Tas. It carried six siliconchip.com.au Perspex Bar Supervisory Directional Filter Unit (removed) Power Bridge & Separating Equaliser Filter Amplifier Valves Directional Brass Resistor Filter Cylinder Box Housing Cable Centre Gland Cover Armour Wires Sea Cable Conductor Bridge & Power Armour Equaliser Separating Watertight Cable Diaphragm Bulkhead Seal Wire Clamp (removed) Filter Gland Fig.13: a cutaway of the repeaters used for TAT-1. Source: https://collection.sciencemuseumgroup.org.uk/objects/co33321/ submerged-repeater-for-tat-1-1956-amplifier telephone circuits, at least a dozen telegraph circuits and an 8.5kHz broadcast channel. For further information, see siliconchip.au/link/abz2 In 1956, the first intercontinental transatlantic coaxial cable, TAT-1 (Transatlantic No. 1), was installed (see Fig.11). It carried 35 telephone channels, with a 36th channel carrying 22 telegraph lines (used by Telex). There were two separate cables, one for each direction, each 41mm in diameter. TAT-1 used valve (vacuum tube) repeaters to boost and condition the signals. Each repeater had three valves. Valves were specially developed for this: the 6P12 for the shallow water portion and the 175HQ for the deepsea portion. The repeaters were at 69km intervals and were 2.74m long, 73mm in diameter and flexible so they could be wound over the cable drum – see Fig.13. Power was supplied via the cable (see Fig.12). Each repeater unit was unidirectional to minimise size, so it was compatible with cable-laying equipment while also minimising the effect of stray capacitance and inductance. For more details, see siliconchip.au/link/ abz3 and siliconchip.au/link/abz4 From 1963, TAT-1 carried the original primary circuit for the famous “Moscow–Washington hotline”. The original bandwidth of TAT-1 was 4kHz per phone channel, but it was reduced to 3kHz to allow for a total of 48 channels. Three additional channels were added using a carrier-­ suppressed ‘Type C’ modulation scheme (siliconchip.au/link/abz5). In 1960, a Time-Assignment Speech Interpolation (TASI) system was implemented on the cable, increasing the number of speech circuits to 72. TASI uses the idle time on calls to carry additional calls. For more information on TASI, see siliconchip. au/link/abz6 TAT-1 was in operation until 1978. siliconchip.com.au The valve repeaters proved extremely reliable, and the cable might still be in use had it not become obsolete due to its low bandwidth. Australia’s first submarine telephone cable The first subsea coaxial cable for telephony connecting Australia to the world was the COMPAC cable, which began service in 1963. It connected to Canada via New Zealand, Fiji and Hawaii, as shown in Fig.14. A microwave link across Canada and the transatlantic CANTAT cable connected it to the UK. It provided 80 two-way telephone channels or 1760 teleprinter circuits, including leased lines. The cable was 32mm in diameter in the offshore sections. A video from 1963 about the project, “80 Channels Under The Sea”, can be viewed at is at https://youtu. be/m1sfMjTyjPo Before the COMPAC cable, Australia had operated an international radio telephone service since the 30th of April 1930. People had to rely on booking a radiotelephone call, which was transmitted by HF radio and could only be made at particular times of day, depending upon atmospheric conditions. Optical fibre cables The next major development beyond submarine coaxial cables was optical fibre cables. Optical fibres for communications are made of high-­ purity glass that can transmit data via pulses of laser light at one or more frequencies. Light stays within the fibre due to total internal reflection. Optical fibres offer many advantages. The data rate achievable is many times faster than over coaxial cable, and the signal loss is lower. Fibre is immune to electrical interference and harder to intercept by hostile actors. More optical fibres can be inserted into an undersea cable (or anywhere) than coaxial cables, as they are much smaller in diameter and weigh less. Fig.14: a COMPAC cable map from Voices Through The Deep (1963), NZ Post Office. Source: https://heritageetal.blogspot.com/2020/09/the-many-lives-of-emervyn-taylors.html Australia's electronics magazine December 2024  19 Fig.15: a cross-section of a submarine optical fibre communications cable. The copper or aluminium tube is both for protection and to carry power, while the petroleum jelly provides lubrication. Original source: https://w.wiki/7ojk Fig.16 shows the basic elements of an individual ‘single-mode’ optical fibre for communications cables, while Fig.15 shows a bundle of optical fibres incorporated into an undersea communications cable. Single-mode fibre is typically used for long-distance communications cables as it can support a longer distance (up to 50 times more than multimode) and a higher data rate. However, it is more expensive and requires a light source with a narrow spectral width. Multi-mode fibre is cheaper but more suitable for short-to-­ mediumrange applications. The first undersea optical fibre was TAT-8, a transatlantic cable that opened in 1988 and retired in 2002. It had a capacity of 280Mb/s, equivalent to 4000 voice circuits. It contained two working fibres plus a spare. TAT-8 had repeaters every 67km. Wavelength division multiplexing (WDM) is used in modern cables to increase the bandwidth by utilising multiple laser wavelengths (colours), up to 30, over a single fibre instead of a single wavelength (see Fig.17). An older optical fibre cable may be able to be retrofitted with WDM terminal equipment to increase its capacity. Optical fibre repeaters (Fig.18) contain optical amplifiers and circuitry to condition and reform the signal. DC power to repeaters is provided via the cable, usually between 3kV and 15kV. The current for a 10kV supply might be 1.65A, meaning an incredible 16.5kW of power is running through the cable. One end of the cable is typically supplied with a positive voltage, the other with a negative voltage, resulting in a virtual Earth in the middle of the cable. The return current is through the seawater. A recent development (2021) is NEC’s multicore fibre. This refers to individual fibres that have four instead of just one optical pathway (see Figs.20 & 19). This quadruples the number of channels through an individual cable compared to a conventional cable of the same diameter. Fig.17: the principle of wavelength division multiplexing (WDM), as used on modern optical fibre communications cables. A ‘mux’ is a multiplexer, while a ‘demux’ is a demultiplexer. 20 Silicon Chip Fig.16: the structure of a typical single-mode optical fibre. This is an individual fibre with protection, not a complete communications cable. Original source: https://w.wiki/33S5 Information on the bandwidth of modern optic fibre cables is hard to come by. Still, the 6605km transatlantic MAREA cable with eight fibre pairs (owned by Microsoft, Meta and Telxius) is said to be rated at 224 terabits per second (224Tb/s). Google’s 15,000km West African Equiano cable with 12 fibre pairs is said to carry 150Tb/s. Modern fibre optic cables are 17-21mm in diameter, except on the continental shelf (typically to a depth of 1500m), where they are 40-50mm due to additional armouring against sea life and abrasion from storms etc. Different cable configurations are possible depending on the level of protection needed; see Fig.21. Additional protection may be provided by burying the cable in shallower areas. Reliability and redundancy Communications cables and repeaters have to be very tough and strong to withstand the bending of the cable as it is loaded, then unloaded and installed. Fig.18: an NEC repeater for the 9400km-long Trans-Asia cable as it goes into the sea. Source: www.nec.com/en/case/ asia_direct_cable Australia's electronics magazine siliconchip.com.au while Amazon is a major capacity buyer or part owner of 4 cables. Many of these cables are shown in Fig.1 (see siliconchip.au/link/abzf). Undersea cable manufacturers Fig.19: an LW-series optical fibre cable from OCC Corporation using 32 of NEC’s multicore optical fibres. It is 17mm in diameter, designed for depths up to 8km and can carry 15kV DC to power repeaters. Source: www.occjp. com/en/products/seabed/sc500.html Consider the tensile loading from the weight of several kilometres of cable as it hangs from the ship (possibly during rough seas) during laying and possible retrieval for cable repairs. The cable may be laid as deep as 8000m, such as in the Japan Trench, where the pressure is 800 atmospheres or 826kg/cm2. The temperature at the bottom of the ocean is around 4°C. Also, the cables have to be armoured to protect against certain marine life. Cables also have to be 100% reliable; no one wants to have to retrieve a cable that has a fault due to a quality control failure. Cables typically have redundant components in the repeaters that can be switched on if required, along with one or more redundant fibres. Who owns undersea cables? Apart from telecommunications companies and investors, about 1% of cables are owned by government entities. The Big Tech giants, Amazon, Alphabet (Google), Meta (Facebook) and Microsoft, own or have interests in many cables. After all, these companies are responsible for about 70% of internet traffic combined. Their business models rely on ample internet capacity. Google owns 17 cables outright and is part owner of an additional 16. Meta (Facebook) is a part owner or major capacity buyer in 15 cables and owns one outright. Microsoft is a part owner or major capacity buyer of 6 cables, siliconchip.com.au Companies that manufacture undersea cables include: ● SubCom LLC (www.subcom.com) ● Alcatel Submarine Networks (www.asn.com) ● HMN Technologies Co Ltd (www. hmntech.com) ● NEC (www.nec.com/en/global/ prod/nw/submarine) Components are made by Corning, General Cable and Norddeutsche Seekabelwerke. Fig.20: regular optical fibre (left) and NEC multicore optical fibre (right). 1000µm = 1mm. Original source: NEC – siliconchip.au/link/abzd Protection of cables by international law An international convention protects undersea cables: the Convention for the Protection of Submarine Telegraph Cables. This was brought into effect in 1884 and remains in force. It makes it an offence to damage submarine cables and outlines who is responsible in the event of accidental damage. The Australian colonies signed in 1885 (SA, Vic), 1886 (Qld) and 1888 (NSW, Tas & WA). Capacity metrics Two capacity metrics are used for optical communications cables. The potential capacity is the theoretical maximum capacity of a cable and is what is usually cited in promotions. There is also lit capacity, the capacity for which terminal equipment is installed at either end. When a cable is first put into service, the full capacity is not usually utilised as demand does not yet exist. Cable owners only install the amount of expensive transmission equipment needed at a given time. More is added as demand increases until the potential capacity is reached. Espionage In December 2016 (siliconchip.au/ Article/10459), we mentioned Operation Ivy Bells, a US operation to tap into a Soviet copper communications cable during the Cold War. There were undoubtedly many other such instances from all parties. Modern optical fibres are much harder to tap into, and end-to-end encryption makes intercepting and decoding communications very difficult. Australia's electronics magazine Fig.21: various possible configurations of optical subsea communications cables. Original source: ICPC – siliconchip.au/link/abze December 2024  21 Australia’s connections to the world Many cables connect Australia to the world (and other parts of Australia). We compiled the following list showing the name of each cable, its length and the year it was or will be put into service: 1995 Bass Strait-1 241km 1999 SeaMeWe-3 39,000km 2000 Southern Cross Cable Network (SCCN) 30,500km 2001 Australia-Japan Cable (AJC) 12,700km 2003 Bass Strait-2 239km 2005 Basslink 298km 2008 Gondwana-1 2151km 2008 Telstra Endeavour 9125km 2009 PIPE Pacific Cable-1 (PPC-1) 6900km 2016 North-West Cable System 2100km 2017 Tasman Global Access (TGA) Cable 2288km 2018 Australia-Singapore Cable (ASC) 4600km 2018 Hawaiki 14,000km 2019 INDIGO-Central 4850km 2019 INDIGO-West 4600km 2020 Coral Sea Cable System (CS2) 4700km 2020 Japan-Guam-Australia South (JGA-S) 7081km 2022 Oman Australia Cable (OAC) 11,000km 2022 Southern Cross NEXT 13,700km 2023 Darwin-Jakarta-Singapore Cable (DJSC) 1000km 2026 Honomoana unknown length 2026 Tabua unknown length 2026 Sydney-MelbourneAdelaide-Perth (SMAP) 5000km 2027 Asia Connect Cable-1 (ACC-1) 19,000km 2027 Hawaiki Nui 1 10,000km 2027 Te Waipounamu 3000km TBD Umoja unknown length 22 Silicon Chip It is possible to tap into optical fibres by bending them and then examining the light leakage at the bend. Depending on the cable, this may result in a detectable reduction in light levels. While encryption makes this less of a concern, protections have been proposed to prevent it, such as using ‘bend-insensitive cable’ or a ‘quantum alarm’ to detect it. Deliberate damage – a major vulnerability With 99% of internet traffic travelling through undersea communications cables, and significant amounts of electrical power for certain communities, nations are vulnerable to being ‘shut down’ very quickly by terrorist or enemy military action. There is no obvious practical way to adequately protect such infrastructure; damage to one cable can take weeks to repair under the best conditions. It would be virtually impossible to repair multiple points of damage on one or multiple cables in any reasonable time. Hazardous areas might include volcanic locations, hot water seeps, areas prone to landslides and ecologically sensitive areas with deepwater coral etc. The location of where cables come ashore is also carefully considered. Cables are carried by special ships on giant spools. One example is the Isaac Newton, shown in Fig.22. It can carry a total of 11,900 tonnes of cable on two spools, and can perform a variety of other functions. A sea plough is used to bury the cable to prevent damage in areas close to shore – see Fig.23. There are about 60 cable installation and repair ships in service worldwide. Damage or faults Undersea cables are periodically damaged. Causes include underwater landslides, earthquakes, volcanoes, marine life, fishing trawlers (38%), anchors (25%) and, closer to shore, extreme storms, strong currents and tsunamis. Around 70% of optical cable damage occurs at depths under 200m. Communications cable life Cable faults were only responsible Most cables have a design life of for about 6% of failures from 1959 to about 25 years. However, many are 2006. Worldwide, about 100 incidents retired early because their bandwidth of cable damage or faults are recorded becomes inadequate and higher-­ per year. capacity cables are more profitable to Sharks have been known to attack install. On occasion, unused cables unburied cables for unknown reasons, might be raised and relocated to as shown in Fig.24. Because of this, another location. This might be worth- cables have been provided with extra while for countries or companies with armour. However, the International limited budgets. Cable Protection Committee stated Sometimes cables are recovered for there was no damage from the incithe valuable materials in them such dent shown in Fig.24. as copper, aluminium, lead and steel. They also wrote that sharks and Collectors may go on diving expe- other fish were responsible for only ditions to retrieve samples of cables 1% of cable faults until 2006 and none of historic interest; for example, see since then (siliconchip.au/link/abz7). http://w1tp.com/mcable.htm In 1929, transatlantic telegraph cables were cut within 100km of an Cable costs & laying the cable earthquake epicentre due to landCables cost upwards of US$25,000 slides. ($38,000) per kilometre, and recent On the 30th of March 2016, 10 Africables have been in the price range of can countries were entirely off the US$250-$300 million ($380-450 mil- internet for two days when a fishing lion) for transatlantic and US$300- trawler inadvertently cut one cable. $400 million ($450-600 million) for In 2019, Tonga’s cable was cut by a trans-Pacific cables. ship’s anchor. During the planned routing of the Then, in 2022, the cable connecting cable, hazardous zones and ecologi- Tonga was cut for over a month due to cally sensitive zones are avoided using the Hunga Tonga-Hunga Ha’apai volseabed mapping systems, such as mul- canic eruption. An earthquake on the tibeam side-scan sonar (we covered 29th of June 2024 damaged it again. sonar in June 2019; siliconchip.au/ Tonga has only limited satellite conArticle/11664). nectivity and no backup cable. Australia's electronics magazine siliconchip.com.au Fig.22: a cutaway model of the cablelaying ship Isaac Newton. Source: https://w.wiki/AoEd Sometimes, ‘accidental’ cable damage is deliberate. In 1959, a Soviet fishing trawler cut five US cables in 12 locations. And in 2021 a research cable was severed off the coast of Norway by a fishing vessel, see https://youtu.be/ pw2lO4sxZn8 Repairing faults The location of cable breaks can be determined by time-domain reflectometry (TDR). With TDR, pulses are sent down the cable and reflections from a cable break are timed. The location of the break is determined by the time taken as a fraction of the speed of light in the cable. We published a DIY TDR design in December 2014 (siliconchip. au/Article/8121). Once a fault is located, a cable repair ship is dispatched to that location and the cable is retrieved with a grapnel (Fig.26) that hooks and locks onto it, a process that sounds much easier than it really is. A damaged cable is normally cut on the sea floor (if it already isn’t cut), both ends retrieved, and a new section added. Rejoining a broken cable is a delicate process, as shown in Fig.25. What about Starlink? Figures are hard to come by, but one estimate by the US FCC suggests that only 0.37% of their international internet traffic goes via satellite. The rest is by cable. Starlink is a wonderful technology that gives internet access to users and devices anywhere in the world, but it is unlikely to significantly relieve the demand for undersea cable bandwidth. The cost for a 60,000Gbps 9000km-long undersea cable with a service life of 25 years is around US$300 million (~$450 million) or US$12 million (~$18 million) per year. That gives a cost per Gbps per year of around US$200 (~$300). The cost for 10 Starlink v3 satellites to cover roughly the same distance is US$17 million (~$25 million), with approximately 50Gbps bandwidth and a service life of five years. That gives a cost per year of just US$1.7 million (~$2.6 million) but a cost per Gbps per year of US$34,000 (~$52,000)! So Starlink cannot compete with undersea cables in terms of cost, but that is not its purpose. Its purpose is to offer internet service everywhere, provide an alternative to land-based ISPs, siliconchip.com.au Fig.23: a Soil Machine Dynamics sea cable plough on Normandy Beach, used to bury cable. Source: https://x.com/MachinePix/status/623603135404187648 Fig.24: a shark attacking an undersea cable as seen from a remotely operated vehicle (ROV) – a “megabite”? Source: https://youtu. be/1ex7uTQf4bQ Fig.25: the delicate process of repairing a cable break (or making a new join). Source: KIS-ORCA – siliconchip.au/link/abzg Fig.26: an ETA-brand ‘cut and hold’ grapnel to cut and retrieve cables from the deepest parts of the ocean. There are many different designs of this type of device. Source: https://eta-ltd.com/cut-hold-grapnel Australia's electronics magazine December 2024  23 How much does internet infrastructure weigh? On the 21st of July 2024, ABC RN (Australia) rebroadcasted a BBC program in which they tried to estimate the weight of all internet infrastructure, including cables (siliconchip.au/link/abza). They concluded that subsea cables weighed two million tonnes, while the total weight of all infrastructure was 92.5 million tonnes. Naturally, that is a rough estimate. and provide internet access in places where free speech is compromised. Other uses of optical fibre cables Active fibre optic cables can be used for seismic measurements, as vibrations in the cable alter the scattering of light in the fibre. Such measurements generate 1Gb of data per minute (see siliconchip.au/link/abz8). The future As more devices and consumers (especially in developing countries) are connected to the internet and existing consumers demand more bandwidth, it is expected that more and more cable capacity will be required. The demand for cable capacity will only be slightly offset by increased satellite capacity, so demand for undersea cables will be strong. Undersea power cables While undersea communications cables are the most prevalent, there are also numerous undersea power cables (see Fig.27). They typically traverse much shorter distances than data cables. Numerous references mention the installation of the first underwater power cable in 1811 across the Isar River in Bavaria. However, we could not find an original source for this. We did find evidence that in 1811, Baron Pavel Lvovitch Schilling devised a water-resistant electrical wire that could be laid in wet earth or rivers for the remote control of mines or for telegraphy. It was coated with natural rubber and varnish. His first use of the wire in a river was for “operations with a subaqueous galvanic conducting cord through the river Neva, at St Petersburg, in the year 1812” – see https://w.wiki/Akii and https://w.wiki/Akij AC/DC Undersea power cables carry either alternating or direct current. AC is simpler because a transformer can easily change voltages at either end of the cable. DC transmission generally requires rectification at one end to convert AC to DC to send through the cable, then an inverter at the other end to convert the DC to AC. If the cable is used bi-directionally, then inverter and rectifier equipment is required at each end. DC transmission is considerably more complicated and expensive than simply having a transformer because it requires high-power, high-voltage rectifiers and inverters. However, DC transmission has the advantage of lower energy losses for longer cable runs. That is because DC has no losses from capacitance between conductors; with AC, this capacitance must be charged and discharged twice per cycle. For DC, that means less energy is wasted as heat, and less conductor material is needed. Also, there is no skin effect with DC transmission, so all of the conductor material is used to carry current, not just the outer layer. There is a maximum theoretical length for AC power transmission because, at some point, the entire current capacity of the cable is used to charge the remaining capacitance. Of course, there are other cable length limitations for both AC and DC cables. For both AC and DC undersea cables, there are greater losses and usually greater expense than for overhead power lines. So undersea cables are only used if there is no good alternative. AC transmission is generally used for shorter cable runs, while DC is used for longer runs where the extra cost is worthwhile due to reduced power losses. However, DC systems are considered less reliable due to the complicated (and therefore failure-prone) conversion equipment at either end. Other sources of energy loss in cables include: Fig.27: a cross-section view of a 150kV 3-phase undersea power (submarine) cable. Source: https://w.wiki/ ApBx 24 Silicon Chip Australia's electronics magazine siliconchip.com.au ● Ohmic power losses due to the resistance of the conductor material, which are proportional to the square of the current and can be reduced by using higher voltages (and thus lower currents for the same power). ● Reactive power losses due to capacitance between the conductors. ● Skin effect losses due to the concentration of alternating current near the surface of a conductor, which can be reduced with separately insulated, stranded conductors. ● Power losses due to proximity with other cables, avoided by spacing cables widely apart. ● Sheath losses due to the generation of eddy currents in the protective metal sheath (armour) around conductors within a cable. ● Leakage losses due to current flowing through the dielectric (insulation) material. DC cables can be configured as monopolar or bipolar, as shown in Fig.28, or another configuration, such as series-connected. Monopolar configurations, with just one conductor (either positive or negative) at a high voltage, are the simplest and cheapest, but bipolar configurations provide more flexibility and reliability. For monopolar configurations, return circuits can be through the Earth, sea or a metallic return cable. For bipolar configurations, one cable is positive and the other negative, both at high potential, with negligible return current under normal circumstances. If a fault occurs in one cable of a Fig.28: two possible configurations for HVDC cable systems, (a) monopolar and (b) bipolar. siliconchip.com.au Fig.30: a simplified electrical model of HVAC undersea power cables. bipolar system, the other cable can still be used but at 50% of the normal current, with a return path through the Earth, sea or a metallic return cable. Electrically, an AC undersea power cable can be considered as consisting of resistance, capacitance and an inductive load, as shown in Fig.30. Terminal stations provide additional resistive and inductive loads. The first high-capacity submarine electrical cable, Gotland 1, was laid in 1954. It was 98km long and went from Gotland Island off Sweden to the mainland, with a capacity of 20MW. It carried 100kV DC and used mercury arc rectifiers to turn AC to DC, then an inverter to convert the DC into AC again. In 1970, the service was upgraded to 150kV and 30MW using thyristors for rectification. The longest undersea power cables in the world are North Sea Link (720km, 515kV DC, 1.4GW), NorNed (580km, 450kV DC, 700MW) and SAPEI (420km, 500kV DC, 1000MW), all in Europe, with Australia’s Basslink the fourth-longest. Basslink was featured in the September 2008 issue (siliconchip. au/Article/1943). It is a 290km (undersea section) 400kV DC 500MW cable between Victoria and Tasmania. The cable weighs 60kg/m. It is of monopole configuration; Fig.29 shows a cross section. It actually consists of three separate cables bundled together with polypropylene rope. The bundle comprises the HVDC cable, a return cable and a 12-core fibre-optic cable for communications. Since the return cable is at low potential, it has much less insulation (and cost) than the power cable. The proposed SingaporeLink cable is 4300km long, has a 1.75GW power rating at 525-640kV DC between Darwin and Singapore to connect intermittent solar and wind electricity generation in Australia with Singapore (siliconchip.au/link/abz9). If it goes ahead, it will be by far the world’s longest undersea electricity cable. The cable would be made in 20km lengths spliced into 200km lengths. Some questions have been raised over its technical and economic SC feasibility. Fig.29: the configuration of the Basslink cable between Victoria and Tasmania. Original source: https://tasmaniantimes.com/2016/11/what-is-your-view-onwhat-caused-the-basslink-failure Australia's electronics magazine December 2024  25