Although the office fax machine has only become popular in the past decade, this year sees its 150th anniversary. It was patented on 27 May 1843, 30 years even before the telephone. But whereas the telephone quickly established itself as an essential tool for business, commercial success has been much longer coming for a machine that could transmit pictures and documents, within seconds, from one office to another.
The inventor of the idea was Alexander Bain, who was born in 1810. Bain, a Scotsman from a remote croft in Caithness, is reputed to have performed his early experiments using cattle jawbones for hinges, heather for springs and metal plates buried in the earth for batteries. He was apprenticed to a clockmaker in Wick and invented the first electric clock, which used electromagnetism to pull a pendulum from side to side. He then moved to London and patented his fax machine. The basic principle is very simple. The image to be sent is divided into fine lines. Each line then consists of white segments and black segments, which can be sent by telegraph like the dots and dashes of Morse code, and reassembled at the receiving end (see Figure 1).FIG-mg18604601.jpg
Advertisement
For the TV programme The Secret Life of the Fax Machine we managed to send a fax by semaphore using this principle. I walked backwards and forwards over a giant sheet of paper with a message written on it, timing my steps with a metronome, and signalling whenever I trod on a black bit of the message. Half a mile away, my collaborator Rex Garrod walked over another sheet of paper, keeping in step using another metronome and applying a black paint roller whenever he saw my signal.
Bain used printer’s type to assemble his message for faxing. A stylus wiped over the metal type, making contact with the raised parts, while a stylus at the receiving end wiped over a sheet of paper soaked in potassium ferrocyanide. Whenever the sending end made contact and completed the circuit, the electrolytic chemical reaction turned the paper at the receiving end black. To move each stylus from side to side, he used his electromagnetic pendulums, together with a separate clockwork mechanism to lower the paper a line at a time.
Brilliantly ingenious though his machine was, Bain never developed it. He realised that sending written messages by fax was much slower than sending the Morse code for the letters that make up the message. Compared to the long stream of black and white segments needed to transmit the word ‘fax’, the Morse code is simply:
In 1846, Bain combined a paper tape punched with Morse code at the sending end with his soggy electrolytic paper at the receiving end to perfect a ‘chemical’ telegraph which held the world speed record of 253 words per minute for many years. He developed various other printing telegraphs, the predecessors of the teleprinter, but got into fierce patent battles with Charles Wheatstone in Britain and Samuel Morse in America, who had both patented telegraphs. He eventually lost all his money, but in 1872 the British government finally recognised his achievements and granted him a small pension. He died five years later, in a home for incurables in his native Scotland.
MESSAGES IN METAL
When the first commercial fax service opened in 1865, between Paris and Lyon, it used fax machines working on Bain’s principles and perfected by a French engineer called L ‘Abbe Caselli. I found one of his machines, called a Pantelegraphe, in the Musee des Techniques in Paris. The message to be faxed was written on a thin sheet of metal in ‘insulating’ ink, so that the stylus made contact only where there was no ink. The metal sheet was wrapped round one of the curved plates on the side of the machine. A long pendulum wiped the stylus backwards and forwards over the plate, and there was a mechanism to move it across by one line at the end of each swing. At the receiving end was an identical machine with Bain’s soggy paper on the curved plate. To keep the two styluses moving exactly in time with each other, Caselli fitted a clockwork chronometer at each end to release the pendulums at the start of every line. The quality of the image the Pantelegraphe achieved is extraordinary. I came back from Paris so inspired by it that I built a simplified version myself. It works well at close range, and greatly increases my admiration for Caselli, whose machines worked 400 kilometres apart.
Despite the technical achievement of the Pantelegraphe, it was not a commercial success. The telegraph already provided a quick, simple way of sending written messages, and the pace of business life was still slow enough for there to have been no real demand for sending drawings or pictures instantaneously. The service ceased in 1870 with the outbreak of the Franco-Prussian War.
In 1878, the British Post Office adopted a facsimile device called the Telewriter which worked in an entirely different way. A pen at the sending end was connected via a pantograph to two variable resistors. These were connected to two electromagnets at the receiving end which moved another pantograph, copying the movements of the pen at the sending end. Office machines like this, also known as electrowriters, continued to be made until the 1960s.
None of these devices, however, could fax a paper and ink image. This limitation was overcome by the discovery in 1878 that the resistance of the element selenium falls when light shines onto it. A fine beam of light is focused onto the image, and light reflected from it illuminates the selenium cell. White parts of an image reflect more light than black parts, changing the cell’s resistance. This could be used to scan any black-and-white image, even photographs. In 1906, a German physicist, Arthur Korn, developed the first photographic fax machine. At the receiving end he had a sensitive galvanometer which moved a shutter over a light, varying the amount reaching a moving piece of photographic paper. This created lines of light of varying width that reproduced the image. By this time synchronous motors had been developed whose speed could be accurately controlled. Korn was able to do away with Caselli’s pendulums and build machines that resembled small lathes.
For the TV programme, Garrod and I converted a pair of metalworking lathes into fax machines based on Korn’s design. I wrote a message on a piece of paper and wrapped it round a drum fixed in my lathe. The tool post held a reflective optosensor (the modern equivalent of the selenium cell) which moved slowly along the drum. The sensor powered a buzzer, taped to my telephone handset, which squeaked whenever it passed over a black area of the paper. At the receiving end, Garrod had a piece of Bain’s electrolytic paper wrapped round an identical drum in his lathe. Resting against it was a stylus, mounted in the tool post in the same way as the sensor at the sending end. A microphone taped to his phone handset sent a current to the stylus via an amplifier whenever it picked up the squeak, reproducing the black bits of the image. Our main problem was getting the lathes to run at the same speed – a difference of 1 in 10 000 revolutions was enough to make the fax illegible. Even so, we did manage to send several quite legible faxes.
In the mid-1920s, after various inventors had improved the fax process sufficiently for photographs to be sent down telephone lines, newspapers started to make regular use of faxed photos, using lathe-like machines. But there were still serious limitations. Ordinary long-distance telephone lines included devices to improve the quality of voice transmission, but these devices distorted fax messages, so specially ‘conditioned’ telephone lines had to be used for fax transmissions. Also, high speeds and high resolution made speed regulation of the drums even more critical. As a result, a complete transmitter and receiver required a whole room full of equipment.
FAXES BRING THE NEWS
Newspapers were the first major users of the fax. Agencies such as Associated Press still fax newspaper photos every day, transmitting them worldwide by satellite. Newspapers preview the daily selection of 250 photos on a computer monitor, and then print out any they are interested in using. Many complete newspapers, such as the International Herald Tribune, are also transmitted by fax and satellite, from their editorial offices to local printing plants all over the world.
Meteorological offices are another well-established group of users of fax technology. They have been transmitting weather maps this way since the 1950s. The British Met Office still uses antique fax machines, which need the special telephone lines, to send weather maps and satellite cloud pictures to its regional weather centres. They even print out their images on Bain’s soggy electrolytic paper.
While facsimile was invaluable for these specialist uses, little effort went into developing an office fax machine because teleprinter technology, pioneered by Bain and Wheatstone, could already send written messages instantly all over the world. As electronics improved in the 1950s, however, a few companies, particularly Magnavox in Germany, started to develop office fax machines. Xerox, after its success with the office photocopier, developed a fax system called Long Distance Xerography in 1964, and then bought the rights to Magnavox’s machines and started marketing them as Telecopiers. A telephone handset had to be placed in a holder on the side of the machine because, until the 1970s, telephone companies did not permit the direct connection of any equipment to their telephone lines. The machines were also very expensive, took six minutes to send a single page, and could communicate only with other Telecopiers; different manufacturers’ machines were incompatible.
Attempts began in 1960 to reach an international standard. Each manufacturer wanted its own machine to become the standard, but the national telephone companies did not consider any of them good enough. After much discussion, the first American standard appeared in 1966. Although heralded as a triumph, problems with compatibility could still leave the received copy stretched or with parts near the edge of the page missing. Two years later, the International Telegraph and Telephone Consultative Committee, CCITT, produced its first attempt at an international standard, Group I. Faxes could now be sent from Europe to America, but not the other way round, and no one could send faxes into or out of France. After another eight years, in 1976, the committee produced the first truly international standard, Group II, though the machines were so complicated and expensive that most users stuck with Group I. Finally, in 1980, the committee produced today’s digital standard, Group III.
The principle of a digital fax machine is simple. The image being faxed is not only divided into lines as before, but each line is also divided to create tiny squares called pels, each of which can be black or white (see Figure 1). Group III machines have a maximum resolution of 203 pels to the inch (and 196 lines per inch), small enough not to be too noticeable. The advantage, as with all digital systems, is that there is no ambiguity: the image is received exactly as sent even if there is a certain amount of interference on the line.
CODING FOR SPEED
The speed of transmission can be increased by ingenious coding to compress the pels. Instead of sending each individual pel, the machine sends information about how many pels of the same colour follow each other in an unbroken run: for example, six black, three white, one black, five white, one black, five white and one black. An American mathematician, David Huffman, worked out a code for all the possible run lengths of black and white. He gave the most common ones on an ordinary typewritten letter the shortest codes. For instance, two black squares together are very common so its code is the brief 11; 19 squares is much less common and so its code is much longer: 00001100111. The Huffman coding explains why faxes do not print out at a constant speed, and why pages of type are usually faster than detailed drawings or photos.
The digital codes are then compressed further by comparing the runs of pels on a line with the line above, and sending an abbreviated code if they are the same. In practice, a large proportion of any line is identical to the previous one, so this is an effective form of compression. However, if an error does slip through (usually caused by noise on the telephone line) it would be propagated down the page. To prevent this, complete lines of Huffman code are sent for every other line. Even doing this, fax machines achieve a compression rate of about 20 to 1.
Group III machines also incorporate a form of error correction. There are 1728 pels to a line on A4 paper, and at the end of each line the receiving machine counts up the decompressed pels to check they add up to 1728. If they do not, most fax machines simply reprint the previous line, making the error less noticeable. Many of the latest machines incorporate memories and store the whole page digitally before sending it. The receiving machine then stores the page digitally and, before printing it out, asks for any lines that do not add up to 1728 pels to be sent again.
The first digital fax machine, the Dacom Rapidfax, was made in the US in 1974. But the real catalyst in creating practical, affordable fax machines was the Japanese language. Japanese uses more than 2000 characters, while the international teleprinter code allowed a maximum of only 56 characters. The convenience of sending handwritten messages spurred the Japanese into action. First, they developed large-scale integrated circuits to do the complicated digital coding and compression. Then they adopted another American invention, the thermal printer. This works with the familiar thin, shiny fax paper that turns black when heated (try holding a piece in front of an electric fire). The fax machine has a row of tiny electric heating elements which can heat up and cool down again more than three hundred times a second. The only moving part of the printer is a stepper motor to move the paper over the row of elements, a line at a time. This simplicity not only makes the printer very cheap, but also very reliable.
One of the most complicated parts of sending a fax is making the machines at each end ‘talk’ to each other. This ‘handshake’ procedure (which makes the characteristic fax warbling sound) is similar to starting a telephone conversation (see Figure 2). The handshake procedure takes place at a relatively low frequency of 300 bits per second, but the fax itself is sent much faster. During the ‘training check format’, the machines first try the highest possible speed, which is 9600 bits per second. If this is not received perfectly, they try progressively slower speeds, down to 2400. The actual fax transmission sounds like a hiss, because the digital bits are not sent one at a time, but in groups of four, achieved by varying both the phase and the amplitude of the signal. Telephone lines are designed to cope with frequencies of less than 4000 hertz but, by using phase and amplitude in this way, the signal has only to change 2400 times a second to transmit 9600 bits.FIG-mg18604602.jpg
The popularity of the fax has led to large sums being invested in its development. The whole-page memories being fitted in some fax machines not only give error-free reception, but also allow a confidential fax to be stored inside the receiving machine until the recipient keys in a code. Machines that work with plain paper are also becoming more common. Most current models rely on expensive laser printer mechanisms, but manufacturers are developing simple ink-jet printers with enough jets to print a whole line of pels simultaneously.
DITHER OVER STANDARDS
Machines which transmit faster and at higher resolution are also being manufactured. The CCITT set an international standard for advanced Group IV machines as early as 1984. The standard is designed to work with the digital telephone networks gradually being introduced all over the world, and increases the number of bits that can be sent per second from 9600 to 64 000. Group IV machines use this increased speed to send about six pages a minute with an increased resolution of 300 pels per inch. They also have sophisticated algorithms for converting shades of grey to patterns of dots, called dither patterns, which greatly improves the reproduction of photographs.
Although the results look impressive, the Group IV standard does have its disadvantages. The handshake procedure is extremely complicated because it has facilities for telex and electronic mail, which were included because the enormous success of Group III fax was not anticipated at the time. Satellite links were also less common in 1984, and it has since been found that Group IV signals do not travel well over satellite links – a big drawback for international communication. Some manufacturers are committed to promoting Group IV, while others are promoting an enhanced Group III standard. By changing the method of defining each group of four bits during transmission, engineers have achieved rates of 14 400 bits per second with Group III machines, and believe 24 000 will be possible.
Even without further improvements in the technology, the effects of the fax have been far-reaching. The speed and ease of sending engineering drawings allows companies to undertake contracts all over the world that would previously have been impractical. Fax enables individuals anywhere to send and receive written information instantly, which can give it a political dimension: the Chinese students in Tiananmen Square and Boris Yeltsin, besieged in the Russian parliament building, used fax to stay in contact with the outside world.
The success of fax has also set back the coming of the electronic, paperless office. A faxed note, often handwritten, creates paper and produces information in a form that is incompatible with computerisation. Electronic mail, which sends digital information directly from one computer to another, has been available for longer than the digital fax machine but has never achieved the same general acceptance.
The fax machine is not only internationally compatible, it is also very simple to use. Despite the enormous complexity of the digital compression and handshake circuitry inside, the press of a single button is all it takes to make it work. Very little computer technology has achieved quite this elegant and user-friendly simplicity. Perhaps the fax machine is more ‘intelligent’ than all the expensive computer technology it is incompatible with.
Tim Hunkin has researched and written a TV series for Channel 4 about technology in the office. The first programme, The Secret Life of the Fax Machine, is to be shown next Thursday, 18 February, at 8.30 pm. ‘The secret life of the fax’ exhibition at the Science Museum will contain the Hunkin version of the Pantelegraphe described in this article. The exhibition runs from 16 February to 26 April (Tel: 071-938 8000 for details).
