Several attempts were made to waterproof telegraph cables, for instance to link parts of cities across rivers, but none succeeded (rubber was a contender but still an inadequately understood material)
Gutta Percha was introduced to London in 1842. It is derived from the resin of SE Asian trees, where the inhabitants used it as an easily worked material, hard below about 65OC and waterproof. Its properties were soon found ideal for many purposes, and a large import trade developed. Faraday suggested it could be used to cover underground telegraph wires in areas of heavy rainfall. Henry Bewley devised a machine to make gutta percha tubing and Charles Hancock modified the design to coat copper wire – to make “cable core”. They formed the Gutta Percha Company in 1845. Cable cores came to have three or four layers of gutta percha to prevent a defect in one from being catastrophic.
In 1849 an experimental (unarmoured) cable core was laid from a ship in Folkstone harbour and connected to the South Eastern Railway’s telegraph to London, and messages sent from the ship to London. In 1850 the Brett brothers, John & Jacob, commissioned a cable to cross between Dover and Calais; it had a single copper wire with gutta percha insulation – and worked for a few days - long enough for the Brett Brothers to retain their exclusive license to land cables in France.
In 1851 an improved cable was laid. It had four cores of the 1850 type helically wound together, bound in hemp saturated in pitch & tallow and 10 galvanised armour wires around the whole. But the end of the cable ran out a mile short of the coast. The extra cable needed to finish the system was made by W Kuper & Co. This firm was based in Camberwell on the Surrey Canal, but had recently expanded to Morden Wharf in Greenwich – it is possible that the extra cable was made there. The system was completed and worked for a number of years. More cables followed, between Britain and Europe, across the Irish Sea, and in the Mediterranean.
Americans were keen to get news from England – an example shows why. One of Canada’s main exports was of grain to Britain; Montreal set the price paid to farmers – London paid according to British market conditions: normally the prices kept in step but the Irish famine caused major problems. In 1850 the US, Canada, and Britain were all on the gold standard (£1 = $4.82), and coping with price fluctuations involved transhipments of gold.
Steamships had already reduced the transit time to two weeks instead of eight or so by sail. If first landfall was at Halifax then news could be rushed by carrier pigeon to Boston and thence by telegraph a couple of days before the ship reached New York. When the telegraph reached St John in 1849 a pony express across Nova Scotia was added to the mix – by the end of that year the telegraph reached Halifax.
The Superintendent of Lines in Nova Scotia was Frederick Newton Gisborne, an Englishman, and in 1852 he was the first to lay a (British) deep-sea cable in North America between Prince Edward Island and New Brunswick. In 1854 he met Cyrus West Field in New York, and suggested a New York to Newfoundland cable; when “and London” was added Field became interested. He approached the top experts in America, Matthew Maury an oceanography expert and head of the National Observatory and Professor Samuel Morse. Both thought the idea feasible. A survey across the North Atlantic had previously been made, showing a plateau at not too great a depth. Surveys at that time were at widely spaced locations, and taken from ships unsuited to maintaining station while a plumb bob was slowly lowered to the deep ocean bed - this was a hundred years before Sonar.
Cyrus Field then went to England looking for the best firms to undertake the job, and for financial backing. The Atlantic Telegraph Company was established as a business venture, with no in-house technical expertise. One of the investors was John Pender, a Manchester cotton merchant.
By then Kuper & Co had become Glass, Elliott & Co. In early 1857 they were awarded a contract for half the system (and expanded again onto Enderby Wharf just upstream from Morden Wharf); the other half went to R S Newall & Co. The cable core, made by the Gutta Percha Co, had a seven strand copper wire and three layers of gutta percha. The cable companies wound on 18, bought in, armour wires. Extra armour was added for the shore ends to guard against cable chafing from wave action or anchor damage. (Some modern fishing equipment can easily damage cables, but what was then in use was a minor concern).
The British and American governments provided the ships to lay the cable (as the ships were in commission there was no hiring charge): HMS Agamemnon loaded the Glass, Elliott cable at Greenwich, and USS Niagara the R S Newall cable at Birkenhead. The lay was commenced but the cable parted and was lost after only 300 nm had been laid. Cable still aboard was unloaded and more to replace what had been lost was made for another attempt in 1858. Glass, Elliott & Co had found that winding the armour wires with a right hand lay made it possible to coil the cable in a clockwise manner; R S Newall & Co used the traditional rope-making left hand lay. As cable is laid the tension rises as it takes the weight of what has just left the ship before it reaches the sea bottom, causing the armour wires to untwist – as the cable reaches the seabed the tension drops and they regain their original lay. Jointing cables with oppositely laid armour wires would break the joint during the descent; this was circumvented with a joint housing that rigidly held both cables, lowering it to the ocean bed in the middle of the Atlantic, then laying the two cables to the terminals.
This time the lay was successful, but the system failed after a few weeks. This was blamed on the excessive test voltages that had been applied to the cable, but another factor was that the cable kept since 1857 had been stored in the open and had been warmed by the sun sufficiently for the gutta percha to soften and allow the centre conductor to drop to one side. (After 1858 cable was always stored in covered tanks that were kept wet – by flooding them, or by a continuous water spray.
The demise of the 1858 cable was a serious setback and even led to stories that the whole enterprise had been a hoax. Nevertheless useful traffic had been carried, and a transatlantic telegraph shown to be practicable. The British government and the Atlantic Telegraph Co set up a commission in 1859 to study both the design and manufacturing methods used in cable manufacture. It published its report in 1861. This led to the merger of the Gutta Percha Co and Glass, Elliott & Co to form the Telegraph Construction and Maintenance Company – known as Telcon – to be able to control the quality at all stages of manufacture. The person to organise the merger and become the first Chairman was John Pender
Cyrus Field and the Atlantic Telegraph Co kept faith, and a new cable, to an improved design was ordered from Telcon. The SS Great Eastern, built by Brunel, had failed commercially as a liner and was being sold off – but being big, with screw, paddle and sail propulsion was seen as ideal for laying cable. With one funnel removed to make way for cable tanks, the ship was capable of taking the entire length of cable. Too big to come to Greenwich, hulks were used to ferry the cable to her at Sheerness. The lay started in July 1865. But again the cable was lost, this time only 600 miles from America.
The Atlantic Telegraph Co could not afford a fourth cable. The Anglo-American Telegraph company was set up and commissioned a new cable to be laid in 1866. This was successful; and after laying it the Great Eastern went back to the end of the 1865 cable, and spliced on sufficient cable to complete that system too. By the end of 1866 both systems were working. Messages were priced at £1 per 5 letter word for the first 20 words and 5s thereafter – leading to coded and abbreviated messages. As it turned out the 1865 cable had the longer life.
Dr Moyle described the well documented tale of the completion of the 1865 trans-Atlantic telegraph cable: the failed recovery of the cable when it was lost, and then its pioneering recovery in 1866 leading to the completion of the system. In 1865 the Great Eastern had sailed back behind the point they had reached, then sailed across the line of the cable with a grappling hook, had picked up the cable but the (steel) grappling rope broke. This was repeated twice more. In 1866, with a stronger grappling rope, they partially raised the cable and buoyed it off; they went further back and partially raised more of the cable. Gradually they (there was an escort ship which assisted) recovered the end of the 1865 cable.
The grapple design had a shaft with rows of hooks along its length, the inner diameter of the hooks matched to the cable diameter. When towed across the line of the cable it was hoped that one of the hooks would catch it. This would be felt by a man feeling the grappling rope on board ship as it bumped over the seabed, and detecting it becoming taught. The grappling rope would then be wound in and the strain on the winding engine monitored. If the tension stayed the same they had caught a piece of debris; but an increase showed they had the cable, adding to the tension as it lifted off the seabed.
All the major American and British contributors to this triumph of engineering: Cyrus West who had seen it through, scientists, manufactures, engineers, ships’ captains, etc were honoured or otherwise rewarded – except John Pender. He was out of favour in the British establishment having been embroiled in a bribery case while seeking election to a rotten borough. However later in life he was awarded KCMG then GCMG, becoming Sir John Pender. His London address was 18 Arlington Street (only a stone’s throw from the Royal Institution) - and the then Prime Minister, 3rd Marquis of Salisbury, lived at No.20.
Part 2. Where the Telegraph went next
Nigel Linge “Encircling the Globe”
The success of the Transatlantic Telegraph in 1866 was the trigger that unleashed a proliferation of submarine telegraph systems, which continued to 1903. On busy routes, particularly across the Atlantic, systems were duplicated and triplicated. The French put in a transatlantic system in 1869, made by Telcon, from Brest to Duxbury, then the longest system in the world.
On new routes segments were built by individual companies, so that difficulties encountered, and losses made, by one would not bring the whole scheme to a halt; once the whole system was established these companies were amalgamated. John Pender resigned as chairman of Telcon – in order to organise the majority of this activity and chair such companies. He naturally directed their cable purchases to Telcon.
The British were keen to have connections to the Empire, India being the first target. This started with the Falmouth – Gibralta – Malta system; except that Falmouth had too much shipping for the safety of the cable and they went from Porthcurno instead. Another cable reached the east of the Mediterranean, then there was an overland route through the Middle East to the Indian Ocean where a further cable went to Bombay. Completion was in 1870. The overland section in particular was slow, and it was bypassed by cables installed via Suez. In 1872 the companies that built the several system segments were amalgamated into Eastern Telegraph Company with Pender as chairman – a post he held until he died in 1896.
The cable systems went on to the Far East, eventually to China and Australia, again with separate companies amalgamated later.
South Africa was first reached by a system which looped down the west coast from one country to the next. This was slow and a cable via Zanzibar down the east coast was laid to speed things up. A more direct cable was also laid down the Atlantic via Ascension Island in 1899 at the time of the Boer War (the British Government spent £5500 on its messages while the press spent ten times as much).
The Americas were opened up by routes in and through the Caribbean from the USA, by the Central & South American Telegraph Company from 1881-3. The India-Rubber, Gutta-Percha & Telegraph Works Co Ltd at Silvertown were their main supplier. With a landline across Panama, the northern, eastern and western coasts of South America were served.
When Pender died he had responsibility for one third of the mileage of telegraph cable wordwide, handling over 2 million messages a year, with 1800 staff, and 10 cable ships – he was called the Cable King. One of his initiatives had been to set up the Global Trust Company – it had shares in telegraph cable systems worldwide, which overall made a consistent profit – and made a good investment for savers.
In 1902 the final link was made in an ‘all red line’, a cable system around the world where all the cable stations were in a part of the British Empire. This was when the Pacific was crossed (its longest segment from Bamfield, Vancouver to Fanning Island); and made use of the trans-Canadian landline of 1871, masterminded by Sit Stanford Flemming.
Queen Victoria came to the throne in 1837, the year of the first telegraph patent; and died in 1901, just a year short of this ultimate achievement in submarine telegraphy. Techniques, many devised at the outset, stood the test of time in cable manufacture, cable laying and system design.
The first half of the 20th century saw a considerable increase in the capability of cable, with corresponding increase in the equipment at the terminals.
One thinks of ‘earth’ as having zero potential. In one’s locality it does (except that an earth rod in resistive ground is unlikely to be perfect). Across an ocean earth potentials can fluctuate significantly.
A Mirror Galvanometer comprises a coil suspended by its input and output wires between the poles of a magnet, with a small mirror attached to the coil. The incoming current from the cable is passed through the coil, causing it to rotate in the magnetic field. A rising current will turn it one way, a negative current the opposite way. A light shone on the mirror is reflected onto screen marked with a scale. Before starting transmission the spot of light is centred. From then the spot is read as a dot for a rising current and a dash for a falling current – even though the mean position of the spot wanders across the scale (due to earth potential fluctuations). Two operators were needed – one to read the spot and the other to write down the message.
Syphon Recorder. The first improvement was to replace the mirror with a syphon, fed with ink, its end just above a paper ribbon on which it traced a line corresponding to the input signal. Only one operator was now needed.
Initially Simplex working was used, where a message was sent, and the cable given time to discharge before transmitting in the opposite direction. Duplexworking, enabling transmission in both directions simultaneously, was introduced in 1873. At both ends of the cable Send and Receive circuits were connected to two arms of a Bridge, the other connections being to the Cable and a balancing Artificial Line. John Muirhead developed the form of duplex working used on submarine telegraph cables, and held the patent.
Drum Relay. In 1899 S G Brown devised a relay to allow an incoming signal (after initial removal of earth potential fluctuations) to be retransmitted without operator interaction. It was a modification of the syphon recorder, the syphon replaced by a probe which just touched the edge of a drum. This was made of three discs with mica insulation between them, the probe running on the central disc with no signal, one side disc for a dot the other for a dash.
Cable Loading. Oliver Heaviside postulated in 1890s that for distortionless transmission the inductance (L) of a cable mattered as well as its capacitance (C), resistance (R) and leakance (G), ie RC = GL. The inductance of telegraph cables was quite low and had not been taken into account: adding inductance to a cable would enable it to transmit at a higher rate. Early experiments with added loading coils showed no improvement. However, Western Electric developed Permalloy , a nickel-iron alloy, which could be wound as a tape around the centre conductor and provide the necessary inductance. Shortly afterwards, in 1923, Telcon inventedMu‑metal, with small proportions of other metals, which was much easier to apply (as a wire) and just as effective. Rates of 1000 words a minute became possible. Duplex working came with corresponding improvements in artificial lines (in temperature controlled rooms). An early application was a new cable for the Vancouver to Fanning Island segment of the trans-Pacific Cable. Cable manufacture then had a new lease of life as faster cables replaced the old ones.
1000 words a minute was far faster than could be sent by one operator. Mechanisation came in. The Kleinschmidt Perforator enabled the operator to perforate a paper tape at a manageable rate, the tape then fed to a machine for transmission at cable rate (with other operators’ tapes).
Cable loading did not reduce cable losses. The Heurtley Hot Wire Magnifier was developed as a receiver, coming into service in 1919. The incoming signal was passed through a platinum resistance wire arranged to move a needle as it warmed or cooled – and the movement used to operate a relay. The magnifier was a very sophisticated piece of mechanical equipment and worked reliably (before the same could be said for valve electronics).
Multiplexers. Several, eg 5, channels were connected to a single cable by a rotary switch. Pulses from each channel were chopped to one fifth of their normal width and transmitted sequentially. At the receiving end a similar switch distributed the pulses to ongoing channels where they were restored to their normal length. Once synchronised the switches were kept in step by tuning forks. Other equipment included: Varioplex which divided its time between many channels, going faster if few were in use, or slower if the system was busier; and Translators which turned landline (Baudot) code to Telegraph code.
After 1922 staffing at Valencia dropped to 10% of its level in 1919.
There was a brief discussion about the non-use of valves (vacuum tubes) in Telegraph equipment. Valves steadily gained in reliability and longevity during the 20thcentury, but designers remained dubious. However, a feature of Telegraph was the very low frequency spectrum it used; to handle this very large capacitors would be needed – and these were notoriously unreliable until 1970s.
John Moyle “Looking for a Needle in a Haystack – Locating Causes of Malfunction”
Accounts of cable systems ignore reliability! But the number of repair ships in Victorian times was equal to the number of cable laying ships.
Malfunctions were of two types:
Records of malfunctions were rarely made public, and on routes with duplicate cables could escape notice. The Eastern Telegraph Co registered 2.5 malfunctions per 1000nm. However cables, no doubt with the odd repair, could attain a 50 year life.
There was a range of causes, some natural or due to outside agencies; others in-house:
- Earthquakes and Volcanoes: the mid-Atlantic Ridge and other sea bottom irregularities round the world were unknown. There were also mud slides off edges of continental shelves, particularly by river deltas.
- Lightning: if a cable hut was struck a large earth current could flow through the cable to the sea.
- Boring Animals: it was found that brass tape wound around the outside of the gutta percha was too smooth for borers to get a start.
- Anchor damage: choose cable landing places away from ports (even if they are the eventual destination).
- Inadequate Surveying: it usually was - even with well separated positions it was a slow process.
- Chafing: inadequate trenching at the landing site, or insufficient armouring.
- Inadequate Slack: no one wanted a length of cable to be suspended between high points on the seabed. Slack was usually set at 10% in shallow water and 20% in deep water in case of unexpected ruggedness of the sea bottom. Loops of cable on the sea bed were accepted.
- Storage: lengths of repair cable mattered as well as cable awaiting the arrival of a cable laying ship.
- Malevolence: damage inflicted by rival companies; cutting enemy cables in war time.
Locating a break could be done from one or both ends of a system by measuring the resistance of the centre conductor (assuming a short to earth at the break) and comparing the result to measurements made in the cable tanks before laying. Due allowance would be made for the sea bed temperature and the 0.3V effect of a copper/steel cell that would arise at the break. Copper cores generally had a resistance of 10 ohm/nm but could differ.
Defective Insulation did not necessarily stop the cable working; if near shore a counter current could mitigate the effect until a repair could be made.
Dr Moyle had studied many records - of the cable companies, and in the press (who only covered the particularly noticeable malfunctions). He had found that there was an average of one repair voyage per annum for every 500nm of cable.
Part 3. Telephony and Data
Jacob Ward “The Politics of Automating the Telephone Network in post-WWII Britain”
This was tangential to other papers in the conference, but fascinating. It described a traditional political process gradually evolving as it grappled with technical advances.
Britain had a series of Strowger (relay operated) telephone exchanges each specifically wired for its location. Updates to the system, such as adding a new exchange to serve more subscribers, meant physical changes to all the exchanges that would need to make connections to it.
A system using Group Routing and Charging Equipment (GRACE) was introduced in 1959 to allow subscriber trunk dialling, instead having operator assistance. This was a development of automation that London and other cities already had, and worked well.
Nevertheless General Purpose Electronic Exchanges were seen as the way forward - but they would have to work with existing exchanges until conversion was complete.
A major problem was ‘prestige’. The Post Office (BT had not then been hived off) had a fine, international, reputation. Problems were to be solved, however intractable, and the timetable met. A prototype exchange was built. It was soon found that the circuits (made with individual power hungry components) required far more power than a Strowger exchange – an extra fan floor was built. 1958 became 1963. Other problems loomed – different subcontractors had used different component types - high pulse rate interconnections across the exchange interfered with low pulse rate input and output lines. In 1965 they gave up and closed the exchange. However lessons had been learnt and the next attempt was better managed
In 1967 Mr J Merriman revised the philosophy of a general purpose exchange, though telephony was still at its heart. The Empress Exchange was to be the first System X exchange and use pulse code modulation, the design led by Roy Harris. It was based on a ‘switch’, which was ready in 1976, but development continued to 1979 before a general launch. The Carter Committee, 1975-7, made a major recommendation in its report to separate responsibility for procurement from development.
The idea for Integrated Services Digital Network (ISDN), first propounded by Kituhara of NTT in Japan in 1972, was taken up. A series of specifications was prepared and completed in 1978. This led to the use of packet switching where all types of data – telex, voice, video, etc - are coded into a standard digital form, put into packets with a header giving the destination and content details.
This is the traffic which optical fibre cables carry internationally.
At the time GRACE was introduced coaxial submarine cables carrying telephony (with some circuits given over to telegraphy) were superseding the telegraph cables - and experiencing a demand far greater than they could carry – so highly lucrative rates could be charged. An international control centre in London was planned for them. Construction of the massive Mondial House began in 1969 on a site just east of Cannon Street Station; it opened in 1978. It flourished in 1980s, but did not survive the advent of the World Wide Web and in 2005 was sold and soon demolished.
Derek Cassidy “Submarine Networks: the next stage of their evolution”
Overseas telephone connections have been wanted since the advent of the telephone – and some telegraph cables were able to provide a single channel. In 1884 it was done across San Fransico Bay to Oakland. In 1921 an inductively loaded cable from Key West to Havana was used successfully.
But what really made it practicable was coaxial cable, where a conductor of copper or aluminium, was put round the core to provide a controlled low resistance return path for the signal (as opposed to armour wires or brass tape). This was patented in 1923. In 1928 a coaxial system from Newfoundland to Ireland was proposed, but abandoned in 1930. The first was the Trans-Atlantic Telephone (TAT 1) system in 1956 for 36 channels, with one cable for each direction of transmission and repeaters to amplify the signal.
In 1947 the Irish Post Office and GPO had laid a system from Dollymount to Holyhead comprising a pair of cables with balanced earth, capable of carrying 50 channels.