English Summary of the Danish Ph.D. dissertation:

The Arc Transmitter - a Comparative Study of the Invention, Development and Innovation of the Poulsen System in Denmark, England and the United States

by Hans Buhl, 1995

THE dissertation is a comparative study of the development of the so-called Poulsen arc transmitter in Denmark, Great Britain and the United States. The arc transmitter which - contrary to all previous types of radio transmitters - generated continuous radio waves was invented by the Danish engineer Valdemar Poulsen in 1902. After a few years of development the arc technology was transferred to Germany and Great Britain in 1906 by Poulsen, his collaborator P.O. Pedersen and their financial backers. In 1909 the American patents as well as a few arcs were bought by a young Californian engineer, Cyril F. Elwell. The subsequent development in Europe and the United States was rather different, since i Europe there were for many years severe difficulties in innovating the Poulsen technology, whereas in the States an extended commercial telegraph system was soon established. Later the US Navy also adopted the Poulsen system and made it the most important radio system for about a decade until it was superseded by radio tube systems. So in the sense that it was transformed into a widespread and important technology the Poulsen system became a much bigger success in the United States than in Denmark.

The aim of the dissertation has been to uncover possible reasons to the very divergent development in the various countries through a comparative study of the many technical as well as non-technical factors that might have influenced the development of the Poulsen system in these countries. Besides pointing out important factors for a successful development of a new technology, this also shows how technical designs are shaped by a continual interaction with the socio-economic context in the widest sense.

Sparks, waves and the invention of the Poulsen Arc

Chapter 2 gives the background for the invention of the Poulsen arc and discusses the possible inspiration and motivation for Poulsen's work. It also contains a short outline of the early development of Poulsen's invention.

Until the advent of the Poulsen arc the only means of wireless telegraphy had been spark transmitters, which despite their popularity had some serious and inevitable flaws. One problem was that they were "dirty" transmitters, since they emitted strongly damped radio waves, which by their nature had a broad frequency spectrum. This hindered sharp tuning, and consequently nearby radio stations disturbed each other heavily. Another problem was that the damped waves could not be modulated to carry audio signals, i.e. wireless telephony.

There were many attempts to solve the problems within the spark concept. But a number of scientists and engineers early realized that the only satisfactory solution was to use continuous waves. These, however, turned out to be very difficult to produce. Just after the turn of the century the Canadian engineer Reginald Fessenden tried to generate high frequency oscillations using a very rapidly rotating alternating current generator, an alternator. But the very high rotating speeds caused severe mechanical difficulties.

Contrary to this Valdemar Poulsen managed to invent the first high frequency generator without moving parts. The technological ancestor of the Poulsen arc was William Duddell's "musical arc", an electric oscillator made from an arc lamp, shunted by a condenser and a coil to form a tuned resonant circuit. Duddell had proposed that it might be used as a generator for continuous radio waves, but in 1903 experiments forced him to conclude that the arc could not oscillate at radio frequencies. He had found an upper limit of much less than 100 kHz, too low for radio work.

At this time, however, Valdemar Poulsen had already turned an arc lamp into the first efficient generator for continuous radio waves. Poulsen's point of departure is not known. Having been employed in the Copenhagen Telephone Company (KTAS) and being the inventor of the telegraphone, the world's first functioning magnetic recorder, he had worked with telecommunication before, but apparently he had never shown interest in wireless communication. Having heard about Duddell's oscillating arc Poulsen, however, started experiments with electrical arcs. Being a very practically oriented engineer, Poulsen performed intuitive experiments in which he tried out many different modifications of singing arcs, and one day he observed that the frequency of the oscillations could be increased by letting the arc burn in the vapor of a spirit lamp. Joking, he later said that he had been inspired by the stimulating effect of alcohol. Systematic experiments, however, confirmed his assumption that the important thing was to let the arc burn in an atmosphere of hydrocarbon vapor or even pure hydrogen. Later he also discovered that a magnetic field perpendicular to the arc had a useful influence on the frequency and efficiency of the oscillations.

During the two years following the invention in 1902 Valdemar Poulsen patented the arc generator in Denmark and 13 other countries.

The Poulsen transmitter was a radical invention in the sense that it did not fit into the existing radio technological system. For instance Poulsen and Pedersen had to develop a new type of detector, the so-called "tikker", to utilize the continuous waves. When this was done in the summer of 1904 they established communication between the laboratory in Copenhagen and an experimental radio station in Lyngby, 15 kilometers outside of Copenhagen, and less than two years later they crossed Denmark by wireless, connecting Lyngby to Esbjerg, about 270 kilometers away.

To give an impression of the possibilities to innovate the Poulsen system in Denmark chapter 3 is formed as a survey of the electrotechnical environment in Denmark around the turn of the century. This shows that the electrotechnical activities only accounted for a small fraction of the industrial activities. On the other hand it was the fastest growing field of all. Denmark experienced a very rapid electrification in the beginning of this century and especially within the telecommunication field there were a very fast development. Even before the turn of the century Denmark had an extended telegraph network and from around 1910 it had the highest telephone density in Europe. There were, however, only few activities involving wireless communication and especially there was no Danish production of wireless equipment.

Parallel to the electrotechnical development professional societies were founded, and in 1903 the Polytechnical College established an electrotechnical line. So in general the technical support network in Denmark was quite good.

Most of the telephone and telegraph equipment was supplied by domestic factories. But they were rather small and the production was mainly based on foreign knowledge. Thus it is doubtful that they could have been very supportive to the development of a new radio system. The Great Northern Telegraph Company (Det Store Nordiske Telegrafselskab) probably was the only exception. It had been founded in 1869 and operated telegraph lines between Denmark, England, France, Norway, Sweden, Finland, Russia, China and Japan. The company also ran the largest telegraph workshop i Scandinavia. But when it was first approached by the Poulsen group it showed no interest in the new competitive technology.

The first endeavors to innovate the Poulsen arc

In 1906 when the basic development of the Poulsen system had been carried out Poulsen and Pedersen tried to innovate the system, which is the main subject of chapter 4. To utilize the transmitting range of the Poulsen system, it was almost necessary to take an international approach. Since Great Northern showed no interest, Poulsen and Pedersen turned to foreign investors in order to finance the commercial exploitation of the Poulsen arc. After fruitless negotiations with among others Telefunken in Germany they turned to England and in July 1906 they eventually signed a contract with Lord Armstrong and others who owned the British De Forest Wireless Telegraph Syndicate. The agreement joined the Danish and British interests into a new company, the Amalgamated Radio Telegraph Company, Limited.

The superior properties of the arc transmitter were soon displayed since Poulsen and Pedersen within a year managed to establish experimental wireless connections from Denmark to Cullercoats, Newcastle, England, (about 900 kilometers) and to Knockroe, Ireland, (about 1500 kilometers). The latter station was intended for transatlantic transmissions.

The positive results achieved in the Amalgamated, however, masked a smoldering conflict between the Danish engineers and the managers in London. To the latter the arc was just a supplement to their other activities whereas Poulsen and Pedersen wanted to innovate the arc transmitter as much and as fast as possible. But before this conflict burst into flames all activities in the Amalgamated Radio Telegraph Company were brought to a sudden halt at the end of 1907 when lord Armstrong went bankrupt. As a result, the Amalgamated was liquidated before any commercial service had been implemented.

The Poulsen system in Denmark from 1908

Before the Amalgamated was liquidated Danish investors established a new, all-Danish company, the Continental Syndicate (Det Kontinentale Syndikat for Poulsen Radio-Telegrafi A/S), which became the organizational basis for Poulsen and Pedersen's work with the Poulsen-system during the following years (chapter 5). They personally got a substantial income from their patents, but for several reasons they never managed to innovate the Poulsen system into a prominent communication technology in Denmark.

One of the major obstacles was the legislation. Since the end of the eighteenth century Denmark had had a state monopoly in the telephone and telegraph field. In 1907 the first Danish radio law also gave the state the exclusive rights to establish and run wireless stations, except for experimental purposes with special permission. The law not even contained the possibility to grant concessions to private companies. This of course limited the possible applications of the Poulsen system in Denmark and forced the Continental Syndicate to try to sell Poulsen equipment to the authorities. They, however, showed very little interest.

The principal user of wireless equipment in those days was the Navy and hence there were several approaches between the Continental Syndicate and the Danish Navy, which is the subject of chapter 6. The officers of the Sea Mine Corps (Søminekorpset), which were responsible for the Navy's electrical equipment originally had a positive attitude towards the Poulsen system but in test after test it turned out that Poulsen's transmitters and receivers did not work as well as expected. Since the Poulsen system did not show the expected superiority there was no desire to let Poulsen's equipment replace well-known, old spark transmitters. And in reality it had to be a complete replacement since the arc and spark systems were incompatible. This means that system incompatibility was one of the obstacles for the introduction of the Poulsen system in the Danish Navy.

The lack of orders created a vicious circle, since without a market there was no regular production of arcs in Denmark. And without such a production it was difficult to develop the system and compete with other systems when a transmitter was ordered from time to time. As a result only few arc transmitters were ever in use in Denmark.

At last, however, the Danish state got a Poulsen station. But that was mainly due to World War I. Immediately after the outbreak of the war in the summer of 1914 the military authorities seized the Lyngby station, and before the end of the war the state had bought it. The intention was not to establish regular wireless connections but primarily to have a wireless back up in case the submarine cables should be cut or otherwise damaged. During the war the Poulsen system had achieved international recognition, mainly due to a rapid development in the United States and therefore the old transmitter in Lyngby was replaced twice by bigger transmitters, which for instance were used in the first Danish radio broadcasts (chapter 8). But this late Danish acceptance of the Poulsen system, which was mainly due to the foreign breakthrough, came far too late to get any practical importance for the development of the Poulsen system in Denmark.

The Poulsen system in Great Britain

Probably as a consequence of the Danish-English collaboration in the Amalgamated there were for many year quite close ties between the development of the Poulsen system in the two countries (chapter 7). Among the advantages of the development in Great Britain compared to Denmark was the possibility to get licenses to run commercial radio stations and the favorable position for transatlantic connections. Therefore such connections were the primary goal for several short lived British Poulsen companies.

Engineering science

Inspired by the work with the transatlantic transmitters and because the war made transmission experiments impossible, P.O. Pedersen in 1914 initiated careful scientific investigations of the Poulsen arc (chapter 10). His goal was to develop a theory, since no deep understanding of the physics behind the arc generator existed. In the first decade of the development of the Poulsen arc, intuition had played a major role, but Pedersen knew that further improvements in design and efficiency required a much better knowledge of the interaction between the different design parameters of the arc.

Pedersen's investigations can be divided into two types. The first type consisted of pure scientific and idealized experiments concerning e.g. electron theory and ionization by collision in the arc. Even though the results gave a deeper knowledge of the fundamental physics of the arc, they were far too general for practical use. The other type of investigations used a much more direct technical approach. In these experiments Pedersen used real Poulsen arcs and examined the relation between magnetic field strength, wave form of oscillations, arc efficiency, etc.

These investigations resulted in several articles and a few patents concerning arc generator improvements, but only few of Pedersen's ideas were implemented in practical use.

The Poulsen arc in the USA

The first time the Poulsen arc principle was used in the United States was in 1906-1907 when Lee de Forest without having the rights to it constructed a wireless telephone which was based on a Poulsen arc (chapter 4 and 9). It was demonstrated at a regatta on Lake Eire, and later he managed to sell about 30 sets of the telephone to the US Navy. When it was operated carefully it did actually work, but it was met with so much opposition from many Navy officers that it was not generally accepted. So the successful transfer of the Poulsen arc did not happen until the summer of 1909 when a young American engineer, Cyril F. Elwell, brought the American patent rights back to California from Copenhagen (chapter 9).

Before his journey to Denmark Elwell had made experiments with a wireless telephone based on a spark transmitter. But like other engineers who had tried that, he realized that wireless telephony would only be possible if it was based on continuous waves. Therefore, having heard about Poulsen's invention Elwell immediately went to Copenhagen to examine the system and eventually get the American patent rights. This starting point had the important consequence that the Poulsen arc never became a goal in itself to Elwell.

When Elwell returned to Palo Alto from Copenhagen in 1909 he had not only acquired the patent rights to the Poulsen arc in the United States. He had also bought a small 100 W transmitter built by Poulsen himself, which very soon made it possible to set up public demonstrations of wireless telegraphy. They helped to raise enough money to pay for two 5 kW transmitters Elwell had ordered in Copenhagen.

Thanks to the earnings from the early demonstrations Elwell was able to form the Poulsen Wireless Telephone and Telegraph Company with himself as president and chief engineer, and thanks to the assistance from Danish engineers arc stations were established in Stockton, Sacramento and San Francisco during 1910. At the same time Elwell founded an arc generator factory in Palo Alto to avoid the high cost of importing Poulsen arcs from Denmark.

Despite their preliminary success, Elwell and his company constantly lacked capital. But this problem found a solution when at the end of 1910 Elwell formed a new company, the Federal Telegraph Company, together with Beach Thompson, a San Francisco financier who had been involved in the construction of electrical power lines. Beach Thompson had the necessary capital as well as a market strategy like Elwell's. Their idea was to construct a network of radio telegraph stations connecting the major cities on the Pacific coast and extending eastward into the rest of the United States. The service was intended to compete with the wired telegraph systems by undercutting their rates. Press traffic at discount rates was expected to secure a sufficient volume of traffic, but the service would also be available to the public.

During the next two years many new stations were constructed and by the end of 1912 Federal operated 14 stations on a commercial basis.

Beside the building of a commercial telegraph service Elwell wanted to convince the US Navy about the superiority of the arc transmitter. Even though the Navy was rather reluctant - perhaps because they had just installed a new 100 kW rotary spark transmitter at the headquarters at Arlington - Elwell in the beginning of 1913 got permission to install a temporary 30 kW transmitter on the condition that he did not put any nail holes in the walls. But the demonstrations showed that the arc transmitter with its continuous waves had a better performance than the three times more powerful spark generator. The Navy responded by ordering a 100 kW arc transmitter from Federal.

Elwell took the order knowing very well that for the time being he didn't know how to build such a powerful transmitter. In fact he had recently tested a 60 kW arc, which for unknown reasons supplied less power to the antenna than the 30 kW transmitter. A rather tricky and unpleasant problem!

The solution was found by Leonard Fuller, a new, competent engineer in the company. He had studied arc generators for his thesis, and realized that the curious high-power behavior was caused by the fact that Elwell had simply scaled all the dimensions of a smaller arc to get the bigger one. To obtain the necessary knowledge of the various arc parameters, Fuller conducted systematic experiments from which he derived empirical formulas and curves, that made it possible to predict how the efficiency of an arc changes as a function of the important design parameters (chapter 10). This introduction of scientific knowledge and engineering theory was crucial in making it possible to deliver the 100 kW transmitter to the US Navy, and later to construct transmitters of thousands of kilowatts.

During the war the US Navy established a world wide communication system based on Federal's high power arc transmitters. In the same period most Navy Yards and bigger battle ships were equipped with smaller Poulsen-stations. The US Navy's total commitment to the Poulsen system probably was the single most important reason for the international breakthrough of the Poulsen system.

Copyright, Hans Buhl, 1996.


Revised July 18 1996 by stenohb@aau.dk