One notable mistake in this article is that it consistently misspells inventor Valdemar Poulsen's last name as "Poulson".
Wireless Telegraphy and High Frequency Electricity, H. LaVerne Twining, 1909, pages 180-196:
Plate I

William  Dubilier
Chief  Electrician  of  the  Collins  Wireless  Telephone  Company


    Up to the present time every known wireless telegraph system has been utilizing damped electric oscillations. It was not until lately that some of our greatest scientists and inventors have expressed their belief that by such means the greatest drawback that we have to contend with, that is, imperfect tuning will never be eliminated and perfect tuning will only be possible by using undamped or persistent electric oscillations.
    Since the possibility of the wireless telephone depends entirely upon the production of such oscillations and suitable means for varying them, we may predict that in the near future the wireless telephone will not only progress far ahead of the wireless telegraph, but take its place. For it can be used either for wireless telegraphy or wireless telephony. It also does away with the spark.
WM.  DUBILIER.        

    There are a number of different ways of producing electric oscillations, the best known being an induction coil or transformer, and the one that is about the least known being an ordinary arc lamp, energized by a direct current. The difference between the two being that in the former they are damped and in the latter they are undamped or continuous.
    Undamped or persistent oscillations are high frequency alternating currents, just as are the alternating currents used for electric lighting and the transmission of power, the only difference being that the frequency of one is anywhere from 1,000 to 100,000 times greater than the other.
    There are several methods used for producing such currents. One is by the use of the high frequency alternator, the invention of which dates back to 1889, when arc lighting by alternating currents became popular, the sound of which they tried to eliminate by increasing the frequency.
    Nikola Tesla constructed a machine which consisted of a fixed ring-shaped field magnet with magnetic poles inwards, and a rotating armature in the form of a fly wheel. The magnet had 400 radial poles in the circumference and 400 coils on the armature. When driven at a speed of 3,000 revolutions per minute or 50 per second, it produced an alternating current of 10,000 cycles. The output of this was limited to a small amount of energy, probably not more than ½ kilowatt. It was dangerous, however, to run such a machine.
    The Westinghouse Co. has built for Mr. G. B. Famme an alternator having a 2-kilowatt capacity at a frequency of 10,000. It is of an induction type and has 200 polar projections.
    In all these machines it is customary to make the field magnet the revolving part, the armature being stationary.
    Duddell succeeded in building a machine of the induction type which, at a speed of 30,000 revolutions per minute, gave a current of one ampere and a frequency of 15,000 per second at 40 volts.
    There is claimed to be built a machine possible to create an alternating current having a frequency of 100,000, when the disc is driven at a speed of 600 revolutions per second, the output being only 0.1 of an ampere at 2 volts.
    R. A. Fessenden claims to have, constructed a machine with a frequency of 60,000, with an output of not more than 200 watts, at a speed of 10,000 revolutions per minute. Although this machine was sufficient for experimental purposes, it was far from being practical for wireless, the output being too small and the machine being too dangerous to run.
    It is said, on one occasion, while one of these machines was going at its normal speed, that a magnet flew from the field, clear through a two-foot brick wall and 250 feet out into the field. The efficiency being very low, the machine dangerous, and the output small, tend to make the high frequency alternator very impractical and useless at its present stage.
    From the experience so far received by the scientific world, we may conclude: First, that an attempt to run alternators at high speeds, say above 5,000 revolutions per minute, involves the loss of considerable energy due to air friction and churning, hence it cannot have a high efficiency; second, that the size of the armature and its peripheral velocity has its practical limitations; third, in using an induction type of motor, it labors under the disadvantage that an attempt to take a current out of the machine generally results in a large drop in the terminal potential difference. It is, therefore, exceedingly hard to combine in one alternator the properties of high frequency, high, power and a large power output. Such machines are not as yet commercial articles, hence the alternating method of producing undamped oscillations has up to the present only come into limited use, although there is a possibility of it being improved.

The  Arc  Method  of  Producing  Undamped  Oscillations

    Up to the time Duddell described his singing arc, many inventors struggled to combine an arc lamp with a capacity and inductance for producing oscillations, but met with little success. As early back as 1840, Grove describes an arc lamp burning in hydrogen and its effects. In 1875 de La Rue and Hugo Miller used an arc in hydrogen in experimenting on some vacuum tubes, and in 1892 Elihu Thompson patented the following method for transforming an alternating current in an alternator: Fig. 1
    In Fig. 1, G is a direct current generator in the same circuit with a very high inductance R, a spark gap, and two metal balls S. These balls are connected in another circuit, consisting of a condenser C and an inductance L in series. When the spark balls are brought in contact, a current is drawn through the inductance L. If the balls are separated, the condenser will become charged by the difference of potential created, and when fully charged it discharges across the gap. Thompson claimed to obtain oscillations of 30,000 per second, but no proof was given in his specifications that these were not intermittent. Although this was quite theoretical, it shows that he was trying to find means for producing undamped oscillations.
    About the same time Firth and Rodgers gave out the statement that the current through an arc was oscillating and that they had succeeded in converting 3% of the continuous current into an oscillating one.
    It was not until Duddell made his discovery in 1900 that the matter was seriously thought of. He described some of his observations made before the London Institute of Electric Engineers on the solid carbon arc lamp, having a capacity and inductance shunted across it, showing its oscillating nature. In his circuit he used a direct current generator of 3.5 amperes and a potential difference of 42 volts. Around the arc he shunted a capacity of about 3 microfarads in series with an inductance of 5 millihenries. Under such conditions the arc gave out a musical tone, the pitch of which depended upon the capacity and inductance. Since the musical tone is due to the rapid change in the arc, a very important factor arises which may open the way later for a new method of producing undamped oscillation, and the author, working on this theory, has been quite successful. An important factor must be taken into consideration in the Duddell arc. The inductance in the direct current circuit must have a high resistance and inductance as compared with the resistance or inductance in the oscillating circuit. If we draw the characteristic curve of a D.C. arc, we will find that it does not quite agree with Ohm's law; that is, it is not a straight line as the case would be with a metallic conducting circuit.
    If we take observations with a voltmeter and ammeter on a solid carbon direct current arc, for various constants of the arc, using the potential difference in volts as the ordinate, and the current in amperes as abscissa, we will find a curve that is concave upward and as the current increases it slopes downward; it is therefore a curve that slopes in the opposite direction to the curves that obey Ohm's law. All this phenomena has been investigated by Messrs. Ayrton, Upson and others, and the conclusion is that in all cases, whether between carbon and carbon, or carbon and metal, or these with gases, the curves slope downward, showing that as we increase the current through the arc the potential difference decreases.
    The action of the capacity and inductance on the arc may be as follows: Fig. 2
    In shunting the capacity C and inductance L across an arc (see Fig. 2) that is burning steadily, the capacity instantly takes upon itself a charge and the current through the arcs is at the same time diminished, the potential difference therefore increases across the arc and this tends further to charge the condenser. This reacts on the arc and still further increases its current, which in turn lowers the potential difference.
    Since it discharges through an inductance L, it not only fully discharges but becomes charged in the opposite direction, just as a pendulum, when pulled to one side and let go, will not only go back to its original position, but go far beyond it in the opposite direction.
    When in this condition, it is ready to repeat the operation with more vigor than before, and so, persistent and undamped oscillations are set up by the condenser charging and discharging.
    Suppose in swinging the pendulum, we apply enough force on each swing to make up for the friction and other losses and make it come back to the same position all the time. This can be accomplished only when we apply the force just about the time it starts to swing in the opposite direction, since it has its own time period of oscillation, depending upon the length. Now if we should strike it before it starts to swing back, we will have two forces in the opposite direction applied to the same points and they will have a tendency to neutralize each other.
    The same applies to the oscillating circuit. If the capacity and inductance, each having its own natural time period of oscillation (into which part of the direct current is converted), are not in resonance, that is, if the capacity does not fit the inductance, we will have very weak oscillations, one counteracting the other.

Poulson's  Improvements

    In 1903 the Danish physicist, Poulson, formed an arc between a water cooled metallic electrode S and a solid carbon S1 (Fig. 3), the chief improvement being, however, the fact that he burned his arc in a medium of coal gas and later used alcohol. With this arrangement he succeeded in obtaining much more forceful oscillations than were heretofore known. The frequency varied from 500,000 to 1,000,000 cycles. When the machine was operated, a great amount of heat was evolved, and although the water cooled the copper rod to some extent, one may readily understand how inconvenient such an arrangement is. However, the advantages gained by the fact that undamped oscillations were obtained, which as stated in the beginning makes tuning possible, induced him to proceed at once and apply his machine to the practice of wireless telegraphy. Fig. 3
    Now in summing up the work done with the arc, H. Simon and Fleming came to the conclusion, that in order to obtain strong undamped oscillations one must have an artificially cooled electrode for positive process, and this, I think, has been solved by Mr. A. Frederick Collins. Since 1900 he has been working on the combination of an arc lamp and transmitter for wireless telephonic work, thus being practically ahead of all other physicists.
    At the time Duddell conceived of the musical arc, he had no idea of its being used in connection with a transmitter for wireless telephony. A description of this was given in the Scientific American of July 19, 1902, showing that he was the first scientist to apply an arc lamp for wireless telephony. The publication of Poulson's experiments, showing that the cooling of the arc lamp electrodes was the cause of powerful oscillations, led Mr. Collins to deeply investigate and evolve a perfect system of wireless telephony.
    In the Poulson method of producing oscillations, if the arc was left burning for some time, the machine and its parts would gradually heat up, and the water in the tank would become warm. It was not safe to connect the water cooled electrode to a water pipe, since this would ground the machine and interfere with its operation. The question of cooling was therefore an important one, as pointed out before. Mr. Collins then produced his revolving arc lamp, in which the electrodes were revolved by a small motor or clockwork. This at once eliminated all troubles due to heating, also to getting rid of a large amount of energy dissipated as heat.
    The first application of the direct current arc to wireless telephony was made by Collins in 1902, and since that time he has devised many a form of arc lamp for the production of in sustained oscillations, one of which is shown photographically in Fig. 4, top view Fig. 5 and in cross section in Fig. 6. Figure 4
    In 1903, when experimenting with the musical arc, Poulson found that more intense oscillations were obtained, if the arc is formed between a cool metallic electrode and a solid carbon.
    Collins has ascertained that a greater percentage of direct current is converted into high frequency oscillations, providing carbons are used, and one or both are kept at a low temperature. In order to accomplish this in practice, he employs a pair of carbon or graphite disks as the anode and the cathode. These disks are mounted on parallel spindles so that they are in the same plane and are connected by means of beveled gears to an insulated shaft.
    The disks are insulated from each other by fiber brushings inserted in the gearings, the casing forming one of the connections, while the insulated bearings in the bottom of the casing forms the other. The gearing is so arranged that carbon disks are rotated in opposite directions, the power being furnished by a 1/8 horse-power motor. One of the bearings in the shaft is mounted to a keyed sleeve which permits the spindle carrying one of the disks to be moved toward or away from the opposite disk so that the length of the arc can be placed in a metal casing while the rotating mechanism is attached to the bottom casing.
    The casing is supported between the poles of an electromagnet, and through the ends of the poles and at right angles to them, are polar rods of soft iron which are threaded. These are screwed through the extremities of the magnet and at right angles to the arc. The ends terminating in the casing are pointed, while those projecting outside have disks of hard rubber so that they may be adjusted in positions to the arc. The magnet coils are placed in each of the leads of the supply circuit, and serve as well to choke back the oscillations from reaching the generator. The casing is supported between the poles of the magnet and the magnet in turn is held in position by an iron base. Figure 5
    The magnets provide a strong magnetic field in which the arc burns and so increases the resistance between the carbons and hence raises the voltage. The adjustable poles of the magnet are used primarily to blow back and keep the arc between the carbons where the distance is shortest. Were this not done, the arc would follow the revolving carbons until broken. The arc has been burned in different gases, under pressure and in vacuum.
    In experimenting with this arc with different gases, the author has discovered that certain conditions existed in the arc chamber heretofore unknown, one being that certain gases under certain conditions do not burn continuously but explode with a very great rapidity. It was on one occasion when using this gas in connection with the arc that undamped oscillations were obtained in the aerial system which indicated two times as much current on a hot wire ammeter than was previously obtained.
    Upon further investigation more detailed specifications will be made public in the near future.
    The rotating oscillation arc eliminates the disadvantageous features of the stationary arc in that a constantly fresh and cool surface is presented to the arc, and in that it prevents the burning away of the electrodes which gives rise to untoward variations in the frequency of the oscillations, and finally in that the optimum length of the arc, namely, at the length when the frequency of the oscillations is the greatest, may be maintained for long periods of time, which is quite impossible when the carbons are. stationary. Figure 6
    Across this arc is connected an oscillation circuit having a variable condenser (see Fig. 7) consisting of metal plates placed above one another in a large tank of insulating white paraffine oil. One set of plates is fixed on a shaft so that it can be revolved and brought between the other set, so that any variation of capacity can be obtained. It is upon this condenser that free oscillations of considerable force, so to speak, depend. The variable inductance included in this circuit is a single helix of bare wire, which can also be varied so that any combination of inductance and capacity can be obtained. There is, however, one important point to bear in mind, and that is the capacity must be of a small value as compared with the inductance and adjusted so that a frequency may be obtained anywhere from 100,000 to 1,000,000 cycles per second.
    In one test made by Mr. Collins between Newark and Philadelphia, a distance of ninety miles, described in the Scientific American of September 19, 1908, a revolving arc lamp energized by a current of 8 amperes at 500 volts was set in operation in connection with a resonance tube used for tuning. This consists of an exhausted glass tube 13 inches in length and 13/4 inches in diameter. Sealed in the ends are platinum wires 1/16 inch in diameter, and these extend longitudinally through the center of the tube until the ends almost touch each other. The outside terminals are connected in shunt with the induction coil. Now, when the first feeble oscillations begin to surge in the closed circuit, one or the other will glow, or both of the free ends of the enclosed wires will glow, depending upon the oscillatory nature of the current. As the current strength of the oscillations increases, the glow light extends farther and farther toward the ends of the tube, always keeping close to the oppositely disposed wires. Figure 7
    The length of the glow on the wires is proportional to the current strength, and thus the tube may also be used as a measuring apparatus instead of the milliammeter usually employed. The characteristics of the oscillations can also be easily observed; for if they are positive the light will appear almost entirely on the end of one of the wires, and if the current is reversed, on the opposite end; while if the current is oscillating with equal electromotive forces, the light will have the same degree of intensity on both wires. By means of a revolving mirror the oscillations may be segregated, and it is then easy to see whether they are periodic or continuous, and if the latter, to analyze the wave form of the spoken words.
    Upon the Land Title Building, Philadelphia, Pa., were raised three kites in tandem to which the aerial was connected. The aerial at Newark consisted of 1,500 feet of phosphor bronze wire. By means of a reel at Philadelphia, about the same length of aluminum wire was let out, which made the attuning of both instruments quite easy. Plate I shows Mr. Collins at the time talking to Philadelphia, where the speech was received quite audibly and clearly.
    Although very good results were obtained by him a short time previous between his Newark laboratory and the Singer Building, New York, a distance of 9 miles, and between Newark and Rockland Lake, a distance of about 40 miles, the Philadelphia test was the greatest distance ever made on this side of the Atlantic Ocean. Fig. 8 shows a wiring diagram of the apparatus.
Figure 8

Controlling  the  Waves  by  Means  of  a  Telephone  Transmitter

    Many different combinations and arrangements have been tried in connecting up the transmitter with the oscillating circuit, but in all my experiments with the wireless telephone, I have found it most practical, in fact the only possible way to get good results, to work the transmitter on an independent circuit of its own and connect that inductively to the arc, or superimpose it upon the direct current supplied to the arc.
    Many experimenters claim results with the transmitter connected to the ground circuit. Upon experiment, this will be found to be almost impossible, as the high frequency oscillations of three or four amperes would arc the carbon and burn it out.
    In the last distance tests made, the terminals of a small transformer coil were shunted across the arc, but a condenser of a large capacity is interposed to check the high voltage direct current from flowing through it. The primary of the transformer was connected in series with a 25-volt generator and a telephone transmitter, as shown in the wiring diagram. Figure 9
    Now when the arc is set in operation, a slight change in its resistance would vary the oscillating circuit and hence change the amplitude of the waves sent out. Upon speaking into the transmitter, the current through the primary of the transformer produces an alternating current at the ends of the secondary circuit on the direct current of the arc, and changes its resistance, which in turn varies the oscillating circuit.
    The amplitude of the electric waves changes in the same manner and is proportional to the change of air pressure against the diaphragm and the current through the transmitter. The transmitter may also be inductively connected to the inductance or to some plates of the condenser.

Marjorana's  Liquid  Transmitter

    Marjorana has been using the intermittent discharge of a condenser by increasing its rapidity and he has produced discharges at the rate of 10,000 per second: these discharges in turn consist of a train of oscillations. This he has done by the use of a very short spark gap, a high inductance in series with the electromotive force and large impressed voltage. In his transmitter he utilizes the action of a liquid flowing from a tube, which is sensitive to sound vibrations.
    A fine stream of liquid flows out at one end, and, when there is no sound, a straight and unbroken column of water passes between two conductors to which the instruments are connected. When a sound is made, the water column is found to contract in certain places which forms a wavy column. Contact is made by the liquid between the two terminals, and when the liquid flows unevenly, we have a varying resistance between the two terminals. Figure 10

Receiving  Instruments

    The receiving instruments used for wireless telephony contain certain forms of detectors, as all wireless telegraph receivers are not suitable for wireless telephony. For example, detectors of the coherer or imperfect contact type will only detect oscillations, but do not indicate changes in their amplitude.
    Three forms of detectors have been used with much success, viz.: The thermo-electric, electrolytic, and the ionized gas detector. Of these the first seemed to work about the best, as a form has been devised by Mr. Collins which eliminates all troubles of adjusting after once placed in position. It is different from all other detectors previously invented, and the principle upon which it works is as follows: Two exceeding fine wires of different metals, crossing at right angles, are made into a thermo-couple and so constructed that the conduction losses are far greater than the radiation losses. Another wire made of a very high special resistance material and which is heated by the received oscillation surging in it, is mounted on a movable block just underneath the couple and its distance from it can be regulated (see Fig. 10). When the received oscillations pass through this wire of a very high specific resistance, it heats up, which in turn acts upon the thermo-couple, the resulting electromotive force effecting a very sensitive receiver and producing the voice.
    An improvement upon this detector was recently made by the author by making use of the wire which is heated up by oscillations, as one of the metals of the thermo-couple. This detector is shown in a photograph of the receiving set used by Mr. Collins in telephoning 81 miles between Newark and Philadelphia. Fig. 9 is a photograph of the complete receiving outfit.