A D V E N T U R E S in C Y B E R S O U N DExperiments with a "Singing Arc" and Tesla Coil by W. Moeller
Since radio discovery began, the methods used to generate high frequencies have often changed. There was first Heinrich Hertz's resonant circuit excited by sparks; this method, while applicable to telegraphy, creates only damped waves unsuitable for modulation. In 1900 while seeking a method of created undamped waves, William Duddell discovered the "singing arc". He demonstrated that an arc light is capable of exciting continuous oscillation in a parallel resonant circuit. In 1906 Valdemar Poulsen constructed the first practical arc transmitter for transmission work; but it was soon superseded by the vacuum-tube transmitter. However, students of radio development will find experiments with the arc-light generator quite interesting; and they can be performed with comparitively simple means, such as a small laboratory affords. The first is Duddell's classic experiment, showing the generation of alternating current in any circuit containing capacitance and inductance as soon as it is connected across an electric arc.
 The layout of the equipment used in these high-voltage, high-frequency experiments The condenser and coil determine the tuning of the circuit in which the arc sets up oscillations On the right, several experiments in the field of the coil and the method of keeping arc in an atmosphere of hydrocarbons
 Fig. 1: The fundamental circuit of the arc-light generator
The direct-current supply should be about 200 volts. The current flows through the regulating Resistance R and through the audio Choke L1 to the carbons; returning through the Ammeter A and the second audio Choke L2. The chokes and resistor should be capable of passing the necessary current without becoming too hot. The carbons are a quarter of an inch thick, or so. The part of the circuit at the right, including the Condenser C and Coil L, with the hot-wire ammeter HA - is switched in later. The capacity of C is 4 mf. and its rating should be at least 700-volt D.C. test; the Coil L is 2 inches in diameter and has 60 turns of No. 18 wire, suitably insulated. (The frequency of this tuned circuit falls within the audible range.) The experiment begins by bringing the two carbons in contact and separating them a small fraction of an inch; thus starting the arc. The Resistor, R, is set to give a suitable current flow. As much as 2 amperes may be readily drawn from the light-lines. If we now switch in the resonant ciruit C-L, the arc gives out a whistling note. At the same time, we see, the hot-wire ammeter HA gives a kick. Since this circuit is blocked to direct current, by the blocking Condenser, C, it is evident that the cause must be a flow of alternating current. Therefore we know that the circuit is oscillating, and the current in it is relatively great. In the writer's experiments it has amounted to 4 amperes, registered by the hot-wire ammeter.
The cause of the sound deserves explanation. When the resonant circuit is shocked into oscillation, an alternating current is set up, and flows over the leads into the carbons; they cannot pass out into the feed-lines, because the Chokes L1 and L2 stop their flow. Consequently, the alternating current is superimposed on the direct current in the carbons, causing alternate strengthening and weakening. This also causes a fluctuation in the emission of ionized gas from the arc; and the temperature of the air about it fluctuates also, setting up successive compressions and rarefactions, which we term "sound waves", and which our ears recognize.
 Fig. 2a, 2b
Our second experiment is to vary the capacity of the Condenser C; if it is replaced by one of 2-mf., the note becomes higher, with an increase in the frequency of the oscillating circuit. We may, as In Fig. 2a, substitute a group of condensers so that we can switch from one to another; the pitch will change. Similarly, as in Fig. 2b, we may substitute coils with different numbers of turns: the altered inductance also affecting the pitch of the note emitted by the arc. More inductance (that is, more turns on the same diameter with same spacing) the deeper the note. If we substitute for the 4-mf. condenser with which we started, a radio tuning condenser, with a value of .0005-mf., for instance, there is no oscillation in the radio-frequency resonating circuit thus produced, Not only will there be no audible note, but the hot-wire ammeter will register no current. We will deal smith this phenomenon further on.
All alternating currents have the property of causing magnetic coupling between circuits. If we make two turns of heavier copper wire, say No. 12, two inches in diameter., and connect the ends across a small flashlight bulb, (Detail 'A'), attach an insulating handle; and approach the resonating Coil L. with the exploring coil held parallel to the windings of L, we shall find that the lamp lights up brilliantly. We can also increase the inductance of the Coil L by introducing into its center a piece of iron rod (Detail 'B'). This will at once produce a decided change in the pitch of the tone emitted by the arc. Alternating currents produce, by induction, local currents in metal masses of any shape or form. The presence of these "eddy currents" may be demonstrated by putting a few small nails or screws in a test tube. covering them with water, and introducing the tube into the field of the coil L (Detail 'C'). Not only will this have its effect on the tone of the arc: but after a few minutes we shall find the glass warmed by the heat generated by the "eddy currents" in the metal, and transmitted by the water, (This is the principle of the high-frequency furnace.)
As said before, we failed to generate radio frequency current by the coil when we reduced the capacity of the condenser to a proper value. Even a .0001-mf., the writer failed to obtain a reading on the hot-wire ammeter. However, we can easily obtain a positive result here; we have only to let the carbons burn in an atmosphere of hydrocarbons. This is easily obtained by placing a Bunsen burner under the arc, with the air holes at its base closed. If we now replace the big condensers by a variable condenser, of .0005-mf., or one of the older 001-mf. instruments, then our hot-wire ammeter at once registers. Also, our exploring coil will show that high-frequency current is being generated in the oscillating circuit. There is however no whistle from the arc even in the range of radio-frequency vibrations, above the audio range. We have in fact, a miniature "arc transmitter".
 Fig. 3
The resonance of radio circuits similarly tuned is at the basis of radio transmission and reception. We can illustrate the principle (Fig. 3) by setting up a similar oscillating circuit with a Coil LE, wound like L, and with a similar variable condenser across its ends in series with a small neon lamp. When the circuits are tuned the bulb lights up. It makes no difference whether we tune the receiver to the transmitter, or the transmitter to the receiver. If we take the receiving circuit further away, we thus reduce the coupling and increase the sharpness of response. If we come closer, the sharpness is reduced, but the current received is much greater.
 Fig. 4
We may place the test coil some distance away, at right angles to the oscillating Coil L, and make a direct coupling with a piece of wire (Fig. 4) the lamp will light.The effect of tuning may be again demonstrated.
 The small Tesla coil used by the author with its coronal discharge from the high-voltage end The large primary is for connection to a high-voltage generator
Beginning of first sentence lost in original scan and then continues... of a condenser, to introduce capacity. The distributed capacity of a winding may serve, as in the well-known Tesla coils. For these experiments, it is easy to make a Tesla coil by winding on any suitable insulating tube 14 inches long, and 1.25 inches in diameter a full layer of No. 32 D.S.C. wire. This is close wound, and the spool is covered with a suitable lacquer, or collodion. to hold the windings in place. The beginning of the winding is grounded through a binding-post attached to its base and the other end is connected to a brass ball, at the top of the tube - as illustrated in the photograph reproduced at the head of this section. The base holds the coil upright. Such a coil has considerable self-capacity between its respective windings, one with another which reaches a very respectable total. This capacity makes it possible for the whole coil to have a natural fundamental frequency. In response to excitation. this is capable of oscillating: as we shall proceed to demonstrate by coupling it to the arc generator. The arc has been introduced into an atmosphere of gas (Detail D) as explained above. The capacity C is now represented by a .001-mf. Variable condenser: where the Coil L has now been replaced by five turns of No. 12 heavy bare copper wire. 8 inches in diameter, and almost self-supporting. The Tesla coil is placed in the center with its "foot" grounded. The current through the Tesla coil will be too small to light a filament-type incandescent lamp but the alternating voltage at the terminal of the coil is very high, and we can utilize this fact to light a neon lamp, one of those for instance, used to test a radio receiver. We apply the test prod of such a neon lamp to the ball at the top of our Tesla coil.
 Fig. 5
As soon as the oscillating circuit of our arc-generator is tuned to the fundamental frequency of the Tesla coil, the lamp applied to the top of the latter will light brilliantly with the rushes of current back and forth, into and from this capacity (Fig. 5).
 Fig. 6
We may also if we wish test direct instead of inductive coupling to the Tesla coil as we did before with our external tuned (receiving) circuit. This is shown in Fig. 6, where the Coil L is common to both the Tesla coil's circuit and that of the arc-generator circuit. Grounding is effected now through a 1 mf. fixed Condenser C. As before, the neon lamp will light up, to denote the oscillating condition in the Tesla coil.
1. Special thanks to Bill Noble for his assistance in tracking down the original version of this article. 2. The original article (in print form) needed to be deconstructed then reconstructed for web publication. Aside from the one error introduced in the scanning process (hoped to be fixed at some time), the rest of the article is essentially true to the original except that the diagrams have been placed within the text rather than as 'sidebars'. Colour has also been introduced in the 'breadboard' diagram. Comments 1 From the person who inspired my searching for the 'lost arcs', David Quinlan.
"...Of course there is still one step left. Moller's article describes a generator with one foot on Duddell's "Singing Arc" and the other on the Poulsen Arc (because it uses a hydrocarbon atmosphere) The real Poulsen arcs used a magnetic field transverse to the arc itself. Back to the Top | Essays Index | Quit | eMail: Dr Russell Naughton |