These devices served as the transmitters for most wireless telegraphy systems for the first three decades of radio (1887–1916), and the first demonstrations of practical radio were carried out using them. In later years somewhat more efficient transmitters were developed based on rotary machines like the high-speed Alexanderson alternators and the static Poulsen Arc generators, but spark transmitters were still preferred by most operators. This was because of their uncomplicated design and because the carrier stopped when the telegraph key was released, which allowed the operator to "listen through" for a reply. With other types of transmitter, the carrier could not be controlled so easily, and they required elaborate measures to modulate the carrier and to prevent transmitter leakage from de-sensitizing the receiver. After WWI, greatly improved transmitters based on vacuum tubes became available, which overcame these problems, and by the late 1920s the only spark transmitters still in regular operation were "legacy" installations on naval vessels. Even when vacuum tube based transmitters had been installed, many vessels retained their crude but reliable spark transmitters as an emergency backup. However, by 1940, the technology was no longer used for communication. Use of the spark-gap transmitter led to many radio operators being nicknamed "Sparks" long after spark transmitters ceased to be used. Even today, the German verb "funken", literally, "to spark", also means "to send a radio message/signal".
The history of radio shows that the spark gap transmitter was the product of many people, often working in competition. In 1862 James Clerk Maxwell predicted the propagation of electromagnetic waves through a vacuum.
In 1878, David E. Hughes used a spark gap to generate radio signals, achieving a detectable range of approximately 500 metres.
In 1888 physicist Heinrich Hertz set out to scientifically verify Maxwell's predictions. Hertz used a tuned spark gap transmitter and a tuned spark gap detector (consisting of a loop of wire connected to a small spark gap) located a few meters away. In a series of UHF experiments, Hertz verified that electromagnetic waves were being produced by the transmitter. When the transmitter sparked, small sparks also appeared across the receiver's spark gap, which could be seen under a microscope.
During his visit to the Paris Exposition Universelle in 1889 the Serbian American engineer Nikola Tesla learned of Hertz’s experiments with electromagnetic waves using coils and spark gaps and went on to duplicate Hertz's experiments. Tesla proposed methods to sychronise sparks with the peak output of an alternator, which he patented in 1896. Tesla never developed any practical communication system, and his primary interest in the wireless phenomenon lay in his belief that it could be used for wireless lighting and power distribution.
Marconi began experimenting with wireless telegraphy in the early 1890s. In 1895 he succeeded in transmitting over a distance of 1 1/4 miles. His first transmitter consisted of an induction coil connected between a wire antenna and ground, with a spark gap across it. Every time the induction coil pulsed, the antenna would be momentarily charged up to tens (sometimes hundreds) of thousands of volts until the spark gap started to arc over. This acted as a switch, essentially connecting the charged antenna to ground, producing a very brief burst of electromagnetic radiation.
While the various early systems of spark transmitters worked well enough to prove the concept of wireless telegraphy, the primitive spark gap assemblies used had some severe shortcomings. The biggest problem was that the maximum power that could be transmitted was directly determined by how much electrical charge the antenna could hold. Because the capacitance of practical antennas is quite small, the only way to get a reasonable power output was to charge it up to very high voltages. However, this made transmission impossible in rainy or even damp conditions. Also, it necessitated a quite wide spark gap, with a very high electrical resistance, with the result that most of the electrical energy was used simply to heat up the air in the spark gap.
Another problem with the spark transmitter was a result of the shape of the waveform produced by each burst of electromagnetic radiation. These transmitters radiated an extremely "dirty" wide band signal which could greatly interfere with the reception of other transmissions on nearby frequencies. Receiving sets located relatively close to such a transmitter would have entire sections of a band masked by this wide band noise.
Despite these flaws, Marconi was able to generate sufficient interest from the British Admiralty in these originally crude systems to eventually finance the development of a commercial wireless telegraph service between United States and Europe using vastly improved equipment.
Reginald Fessenden's first attempts to transmit voice employed a spark transmitter operating at approximately 10,000 sparks/second. To modulate this transmitter he inserted a carbon microphone in series with the supply lead. He experienced great difficulty in achieving intelligible sound. At least one high-powered audio transmitter used water cooling for the microphone.
In 1905 a "state of the art" spark gap transmitter generated a signal having a wavelength between 250 meters (1.2 MHz) and 550 meters (545 kHz). 600 meters (500 kHz) became the International distress frequency. The receivers were simple unamplified magnetic detectors or electrolytic detectors. This later gave way to the famous and more sensitive galena crystal sets. Tuners were primitive or nonexistent. Early amateur radio operators built low power spark gap transmitters using the spark coil from Ford Model T automobiles. But a typical commercial station in 1916 might include a 1/2 kW transformer that supplied 14,000 volts, an eight section capacitor, and a rotary gap capable of handling a peak current of several hundred amperes.
Shipboard installations usually used a DC motor (usually run off the ship's DC lighting supply) to drive an alternator whose AC output was then stepped up to 10,000–14,000 volts by a transformer. This was a very convenient arrangement, since the signal could be easily modulated by simply connecting a relay between the relatively low voltage alternator output and the transformer's primary winding, and activating it with the morse key. (Lower-powered units sometimes used the morse key to directly switch the AC, but this required a heavier key making it more difficult to operate).
Spark gap transmitters generate fairly broad-band signals. As the more efficient transmission mode of continuous waves (CW) became easier to produce and band crowding and interference worsened, spark-gap transmitters and damped waves were legislated off the new shorter wavelengths by international treaty, and replaced by Poulsen arc converters and high frequency alternators which developed a sharply defined transmitter frequency. These approaches later yielded to vacuum tube technology and the 'electric age' of radio would end. Long after they stopped being used for communications, spark gap transmitters were employed for radio jamming. As late as 1955, a Japanese radio-controlled toy bus used a spark transmitter and coherer receiver; the spark was visible behind a sheet of blue transparent plastic. Spark gap oscillators are still used to generate high frequency high voltage to initiate welding arcs in gas tungsten arc welding. Powerful spark gap pulse generators are still used to simulate EMPs. Most high power gas-discharge street lamps (mercury and sodium vapor) still use modified spark transmitters as switch-on ignitors.
The function of the spark gap is to initially present a high resistance to the circuit so that the C1 capacitor is allowed to charge. When the breakdown voltage of the gap is reached, the air in the gap ionizes, the resistance across the gap is dramatically lower and a current pulse flows across the arc to the rest of the circuit. The gap is set so that the discharge coincides with a maximum or near maximum of charge in C1 and it is as if a high speed switch is turned on at just the right moment to allow the C1 capacitor to discharge its stored energy into the other circuit elements. This pulse of energy is rapidly exchanged back and forth between the C and L elements and takes the form of a damped oscillation at a radio frequency. The back and forth exchange of energy is in the form of an alternating current and voltage wave with much of the energy flowing out to the antenna. These waves are called "damped waves" because the wave tends to "die out" or "dampen out" between discharges of the spark gap as opposed to modern continuous waves (CW), which don't die out. Because the "damped waves" are a train of regularly spaced radio frequency triangle waves that occur at an audio rate, early crystal, magnetic and Fleming Valve detectors would hear them as a musical note rich in harmonics, making it easy for the human ear to "copy" messages and identify stations by their unique sound, even under adverse conditions.
The exchange of energy in this kind of oscillator occurs at a rate or frequency determined by the resonant frequency of its "tank circuit" which is composed of the combined capacitance of C1 and C2 and the inductance of L, famously known as an LC circuit. Capacitance of C2 was generally small and generally not shown in most diagrams. C2 represents the stray circuit capacitance, but C1 was relatively huge both in size and capacity so that it could store the large amount of high voltage energy necessary for high power transmission (P=EI). Some installations had whole buildings devoted to the C1 capacitor (as in the Cape Breton Transmitter). The inductance coils (L) were relatively small so that the entire circuit could resonate at a reasonably "high" frequency, given the large value of C1. Frequencies much above 1 MHz were impractical because L could not get electrically smaller and not enough energy could be stored in a small C1, although a small C1 would have been necessary because of the resonate characteristics of "shortwave" frequencies.
In addition to the size and robustness of the oscillator components, the lower frequency components were likewise robust. This is because a very large induced EMF occurs when the spark is struck, causing a strain on the insulation in the primary transformer. To overcome this, the construction of even low-power sets was very solid and a radio frequency choke coil or a resistor (R shown in this diagram) was necessary to protect the transformer or induction coil. The telegraph key (essentially an easy to operate on/off switch) many times had to carry large currents and high voltages and so it also was generally quite robust too.
Although ubiquitous in the early days of radio, the technical features that finally doomed the spark gap transmitter was its extremely broad frequency spectrum and damped wave output. Damped waves are excellent for radiotelegraph use when used with the early radio detectors of the period, but they are very wasteful of bandwidth, and this factor limited the number of stations that could effectively use a band because of all the interference. In addition, the wide bandwidth meant that useful intelligence was spread out over a large spectrum and only a fraction of the transmit power was useful for communications. Finally, the damped wave is already a form of AM and it can not be further modulated for voice with any intelligibility. Only the continuous wave oscillators, made possible by vacuum tube technology, could be used at high frequency (HF) and beyond and only then were efficient radiotelegraph and voice/data transmissions made possible.
A simple spark gap consists of two conducting electrodes separated by a gap immersed within a gas (typically air). When a sufficiently high voltage is applied, a spark will bridge the gap, ionizing the gas and drastically reducing its electrical resistance. An electric current then flows until the path of ionized gas is broken or the current is reduced below a minimum value called the 'holding current'. This usually occurs when the voltage across the gap drops sufficiently, but the process may also be assisted by cooling the spark channel or by physically separating the electrodes. This breaks the conductive filament of ionized gas, allowing the capacitor to recharge, and permitting the recharging/discharging cycle to repeat. The action of ionizing the gas is quite sudden and violent (disruptive), and it creates a sharp sound (ranging from a snap for a spark plug, to a loud bang for a wider gap). The noise from the spark mechanism, especially from the higher powered transmitters, was so loud that it could seriously interfere with the operator's ability to receive messages after transmitting. Higher powered spark gap mechanisms were isolated from the operator's station in an insulated space called a "Silent Room" which, when the radio was transmitting, was anything but silent inside. The spark gap also liberates light and heat.
Quenching the arc
Quenching refers to the act of extinguishing the previously established arc within the spark gap. This is considerably more difficult than initiating spark breakdown in the gap. As transmitter power was increased, the problem of quenching arose.
A cold, non-firing spark gap contains no ionized gases. Once the voltage across the gap reaches its breakdown voltage, gas molecules in the gap are very quickly ionized along a path, creating a hot electric arc, or plasma, that consists of large numbers of ions and free electrons between the electrodes. The arc also heats part of the electrodes to incandescence. The incandescent regions contribute free electrons via thermionic emission, and (easily ionized) metal vapor. The mixture of ions and free electrons in the plasma is highly conductive, resulting in a sharp drop in the gap's electrical resistance. This highly conductive arc supports efficient tank circuit oscillations. However, the oscillating current also sustains the arc and, until it can be extinguished, the tank capacitor cannot be recharged for the next pulse.
Several methods were applied to quench the arc.
- Jets of air that cool, stretch, and literally 'blow out' the plasma,
- multi-plate discharger of Max Wien to cool the arcs in medium power spark sets, known as the "whistling spark" for its distinctive signal,
- using a different gas, such as hydrogen, that quenches more efficiently by providing more effective electrode cooling,
- a magnetic field (from a pair of permanent magnets or poles of an electromagnets) oriented at right angles to the gap to stretch and cool the arc.
Spark gaps used in early radio transmitters varied in construction, depending on the power to be handled. Some were fairly simple, consisting of one or more fixed (static) gaps connected in series, while others were significantly more complex. Because sparks were quite hot and erosive, electrode wear and cooling were constant problems.
The need to extinguish arcs in increasingly higher power transmitters led to the development of the rotating spark gap. These devices were used with an alternating current power supply, produced a more regular spark, and could handle more power than conventional static spark gaps. The inner rotating metal disc typically had a number of studs on its outer edge. A discharge would take place when two of the studs lined up with the two outer contacts which carried the high voltage. The resulting arcs were rapidly stretched, cooled, and broken as the disk rotated.
Rotary gaps were operated in two modes, synchronous and asynchronous. A synchronous gap was driven by a synchronous AC motor so that it ran at a fixed speed, and the gap fired in direct relation to the waveform of the A.C. supply that recharged the tank capacitor. The point in the waveform where the gaps were closest was changed by adjusting the rotor position on the motor shaft relative to the stator's studs. By properly adjusting the synchronous gap, it was possible to have the gap fire only at the voltage peaks of the input current. This technique allowed the tank circuit to fire only at successive voltage peaks, thereby delivering maximum energy from the fully charged tank capacitor each time the gap fired. The break rate was thus fixed at twice the incoming power frequency (typically, 100 or 120 breaks/second, corresponding to 50 Hz or 60 Hz supply). When properly engineered and adjusted, synchronous spark gap systems delivered the largest amount of power to the antenna. However, electrode wear would progressively change the gap's firing point, so synchronous gaps were somewhat temperamental and difficult to maintain.
Asynchronous gaps were considerably more common. In an asynchronous gap, the rotation of the motor had no fixed relationship relative to the incoming AC waveform. Asynchronous gaps worked quite well and were much easier to maintain. By using a larger number of rotating studs or a higher rotational speed, many asynchronous gaps operated at break rates in excess of 400 breaks/second. Since the gap could be fired more often than the input waveform switched polarity, the tank capacitor was charged and discharged more rapidly than a synchronous gap. However, each discharge would occur at a varying voltage that was almost always lower than the consistent peak voltage obtained from a synchronous gap.
Rotary gaps also served to alter the tone of the transmitter, since changing either the number of studs or the rotational speed changed the spark discharge frequency which was audible in receivers with detectors that could detect the modulation on the spark signal. This enabled listeners to distinguish between different transmitters that were nominally tuned to the same frequency. A typical high-power multiple spark system (as it was also called) used a 9-to-24-inch-diameter (230 to 610 mm) rotating commutator with six to twelve studs per wheel, typically switching several thousand volts.
The output of a rotary spark gap transmitter was turned on and off by the operator using a special kind of telegraph key that switched power going to the high voltage power supply. The key was designed with large contacts to carry the heavy current that flowed into the low voltage (primary) side of the high voltage transformer (often in excess of 20 amps). Alternatively a relay was used to do the actual switching.
- Crystal Radio
- Radio receiver
- Antique radio
- Television interference (electrical interference)
- James O'Neill, Prodigal Genius: The Life of Nikola Tesla, page 86
- Marc Seifer, Wizard: The Life and Times of Nikola Tesla - page 1721
- Ken Beauchamp, History of Telegraphy, page 193
- Lewis Coe, Wireless Radio: A Brief History, page 112
- Radio: Brian Regal, The Life Story of a Technology, page 22
- W. Bernard Carlson, Tesla: Inventor of the Electrical Age, page 132
- Brian Regal, Radio: The Life Story of a Technology, page 23
- A. B. Rolfe-Martin (1914). "IX Spark Gaps and Dischargers". wireless Telegraphy. London: Adam and Charles Black. p. 103. efficiency is 25%
- [dead link]
|Wikimedia Commons has media related to Spark-gap transmitters.|
- Alternator, Arc and Spark
- Fessenden and the Early History of Radio Science
- Brief history of spark
- Massie Spark Transmitter The new England Wireless and Steam Museum
- "The Sounds of a Spark Transmitter with audio". Archived from the original on July 18, 2011.
- The Sparks Telegraph Key Review
- Radio Technology in common use circa 1914