FEW ENOUNTERS WITH nature leave the observer more profoundly impressed than a close brush with lightning. Those who escape from the experience with a whole skin never forget the blinding flash, fearful concussion and accompanying odor of "brimstone."

The power of nature's big spark to rip trees and other objects apart has long intrigued experimenters. Yet relatively few have undertaken even a small-scale study of the instantaneous discharge of electricity through gas at atmospheric and higher pressures. This is difficult to .explain in view of the success which followed the study of comparable discharges through gases at low pressure. Among other accomplishments, this latter work opened the field of electronics.

A host of volumes record this work. In contrast, the word spark appears in the index of only about one physics text in four, and these usually dismiss the subject with a few paragraphs on the venerable Rhumkortf, coil. Thus far only two noteworthy applications have been found for the spark. Pioneers of the wireless telegraph used it as an automatic switch in the so-called spark transmitter', and modern engineers find it a handy means for igniting gasoline in engines and cigar lighters.

We may soon hear more about sparks, if the work of two engineers at the University of Michigan lives up to its promise. H. C. Early, research engineer, and E. A. Martin, professor of electrical engineering, have been looking into the nature and applications of high-energy sparks under the sponsorship of the U. S. Army Office of Ordnance Research. They have succeeded in concentrating about half the peak current of a major lightning stroke into a hair-thin path a half-inch long that offers appreciable electrical resistance to the discharge. The results should fascinate amateurs who like sound and fury in their experiments. The sparks made by Early and Martin punch holes through metal plates, emit light which pales the sun and, in one proposed application, emboss paper with printed characters at speeds impressively higher than those of any process now in use. Numerous other applications are implied. Apart from its potential uses, the high-energy spark turns out to be a simple and relatively inexpensive device for investigating the behavior of matter under extremes of temperature and pressure that should open a whole new field of interest to amateurs.

Little energy is represented in ordinary sparks such as those developed across the electrodes of induction coils or capacitors charged to high voltage: the resistance of air normally drops so low during the discharge that the spark gap acts as little more than a short-circuit, The electrical power which appears in any load, according to Ohm's law, is equal to the square of the current multiplied by the resistance of the load. Hence the problem of developing a high-energy spark cannot be solved effectively by the brute-force method of merely substituting sources of higher voltage and current. Attempts to deliver more power to a load which takes he form of a short-circuit simply end in more power being dissipated by the associated apparatus and wiring. The solution of the problem of creating energetic sparks lies in increasing the resistance of the gap.

Early and Martin have devised two ingenious ways of doing this. In their first arrangement the gap is immersed in water, the spark being initiated through a fine wire which literally explodes when power is applied. The power source is designed for a rate of rise in current of tens of thousands of amperes per microsecond. Accordingly, an instant after the switch is closed the spark gap is transformed into a column of plasma expanding against water, the inertia of which causes a build-up in pressure on the order of 10,000 atmospheres. In the second method recessed electrodes are sandwiched between two sheets of insulating material. This forces the spark into the form of a wide, restrained ribbon under high pressure. In each arrangement the electrical resistance of the path increases with pressure. The source of power is a capacitor charged to high voltage by a transformer-and-rectifier combination t: identical in principle with those found in radio receiving sets.

Most of the underwater experiments have been made with the capacitor charged to 25,000 volts, Measurements show that at this initial potential the spark absorbs some 400 million watts from the capacitor. The current reaches a peak value of 85,000 amperes. For a few microseconds the radiated energy, much of it within the visible portion of the spectrum, equals that of a black body heated to a temperature of 54,000 degrees Fahrenheit, some six times hotter than the surface of the sun. The resulting shock wave propagated through the water is capable of exerting pressures up to 70 tons per square inch.

Amateurs should not find the construction of the apparatus difficult. The circuit for underwater discharges assembled by Early and Martin employs five General Electric Pyranol capacitors connected in parallel for a total capacity of 5.8 microfarads. These units are relatively costly if purchased new, but some are available on the surplus market. It is also possible to build a capacitor by stacking sheets of aluminum foil between sheets of insulating material such as Mylar film alternate sheets of foil being connected to common leads. The unit is then potted in a wooden tank filled with transformer oil.

A number of 25,000-volt transformers from the early days of amateur radio are still around; one of these will provide the necessary voltage. If one cannot be located, pay a visit to the local power company. Pole transformers used on 13,000-volt transmission lines are occasionally discarded. Some are of the three-phase type, and can easily be converted to step 110 volts up-to 26,000 volts. Ignition transformers used in home oil-burners, or a neon-light transformer with the magnetic shunt removed, may also be employed. When they are connected in a voltage doubling circuit, these make an inexpensive voltage source.

Either of two types of rectifiers may be used. Vacuum-tube rectifiers designed for high-voltage applications are manufactured by the Westinghouse Electric Corporation. They are known as Kenotron diodes and are capable of handling high power. Another suitable rectifier is the Radio Corporation of America 5825 tube, which has an inverse peak-voltage rating of 60,000 volts and is relatively small and inexpensive. The RCA 1B3-GT8016,'with an inverse peak rating of 30,000 volts, is also small and inexpensive.

The high current-handling ability of Kenotron tubes permits the capacitors to be charged quickly, a convenience but not a necessity. The tubes are costly, and require relatively large, expensive transformers. The current available from voltage-doubling circuits employing the RCA types is small. When they are used, precautions must be taken to avoid corona leakage encouraged by sharp edges and corners on the high-voltage conductors, or the charge will leak off through the air as 'fast as it is delivered by the power supply.

The entire power supply, including the rectifier, must be housed in a well-shielded cabinet, not only to prevent stray radiation, which is capable of creating radio interference, but also as a safety measure. The voltages are lethal. All doors to the cabinet should be fitted with interlock switches for breaking the line voltage automatically and short-circuiting the capacitors with a copper bar when the doors are opened. A manually operated switch for closing the circuit between the capacitor and spark gap should be mounted inside the cabinet. A type of switch suitable for the purpose uses two metal' spheres about an inch in diameter, one of which is movable. The arrangement may be thought of as a knife-switch with spheres substituting for the conventional switch blades. The spheres reduce corona discharge. The device is operated by means 'of a long pull-cord of Fiberglas. When the movable sphere is moved close to the stationary one, a spark jumps and initiates the discharge.

Because of the intense pressure wave generated by the underwater spark, the gap assembly is housed in a two-gallon welded steel tank with 8/32-inch walls. The tank is fitted with a splash-proof cover, as shown in Roger Hayward's drawing in Figure 9. Leads for the underwater portion of the circuit may be made of RG-8-U coaxial cable, which is stocked by most dealers in amateur radio supplies. The outer conductor of braided wire is stripped off. The spark is observed through a Plexiglas window of convenient size set in rubber gaskets in the front wall of the tank.

A serviceable gap may be formed by cutting back the insulation of the coaxial cable to expose an eighth of an inch or so of the central conductor, and bending the ends of the cables toward one another. The cables are stiff enough to retain their position prior to the discharge, so the gap need not be supported on an insulating column. Early and Martin found it desirable, however, to control the position and shape of the spark by connecting a wire of .001 inch diameter between the electrodes. This "initiating" wire prevents the spark from taking an erratic path and hence permits experimental results to be reproduced closely. The tank may be filled with tap water.

Short leads of strap copper about an inch wide should be used between the capacitor and the underwater portion of the circuit. They may be cut from copper flashing or other thin sheet. By minimizing self-inductance, leads of this shape permit the current to discharge at maximum rate.

The high-energy spark is an excellent light source for high-speed photography. Radiation from the underwater discharge persists for about 30 microseconds. Even if the window of the tank is masked to pinhole size, the minute though brilliant flash will blacken a sheet of Super XX film at a distance of 50 feet! The exposure interval may be shortened to a microsecond or less by means of a "shatter shutter" devised by Early and Martin. This essentially consists of a thin sheet of glass painted black and supported within a few thousandths of an inch of the initiating wire. The opposite side of the wire is backed by a block of porcelain, as shown in the drawing Figure 6. Light is transmitted through a slit or pinhole scraped in the paint. When the assembly is placed in the tank, care must be taken to assure that no bubbles remain between the glass and block. Within about a twentieth of a microsecond following-the explosion of the initiating wire, the shock wave starts shattering the glass into very fine particles. The glass is transformed into a translucent body which scatters the light within the block until repeated reflections with the block walls absorb it. The shatter shutter is attractive for applications where high intrinsic brilliance and ultra-short exposures are desired, as is often the case in schlieren and shadowgraph systems. Oscillograms of light pulses emitted by the spark, both with and without the shatter shutter, are reproduced in Figures 4 and 5. The sweep speed (horizontal scale) in both oscillograms is four microseconds per division. The oscillograms were made by feeding the output of a photoelectric cell into a cathode-ray oscilloscope, the sweep along the time axis being triggered by a pulse from the high-voltage circuit. Some idea of the intensity of the light source can be gained from the shadowgraph of the compass tips in Figure 10. The exposure, limited to two microseconds by a shatter shutter, was made on Super XX film. In this experiment the size of the source was limited to a pinhole aperture measuring 1/16 inch in diameter, yet more light was transmitted than required for proper exposure: The spark was located 29 feet from the film and the compass was supported at a distance of 18 inches in front of the film. Although no lens was used, the knurling on the heads of the clamping screws was recorded clearly.

Another accessory which not only enhances the spark's versatility as a light source but enables the experimenter to investigate some of the microsecond events associated with the growth and decay of the discharge is the Kerr cell. This device, invented in the last century by the Scottish physicist John Kerr, is based on the property of some transparent substances, such as nitrobenzene, to alter the polarization of transmitted light when subjected to a strong electrical field. The cell essentially consists of a pair of electrodes, resembling the plates of an air capacitor, spaced a few millimeters apart and immersed in nitrobenzene contained in a liquid-tight cell fitted with windows on opposite sides. A filter of high-extinction Polaroid film is mounted in front of each window, one filter fixed rigidly and the other arranged so that it can be rotated about the optical axis of the cell A beam of light is polarized by the first filter and proceeds to the second, where it is absorbed if the plano of the second filter's polarization is crossed with respect to the first. If a pulse of voltage is now impressed on the electrodes, the optical property of the nitrobenzene is altered so that the plane of the beam's polarization is rotated to correspond with that of the second filter. Accordingly the device, normally opaque, becomes transparent whenever a voltage pulse of proper magnitude is applied to its electrodes. Thus the Kerr cell is a high-speed shutter, one capable of opening and closing in a small fraction of a microsecond. Combined with the underwater spark and a schlieren optical system, the Kerr cell makes it possible to shoot knife-sharp pictures of high-velocity events such as the passage of a rifle bullet. Clear pictures of the spark channel at any stage of development can also be made with the Kerr cell. The size of the channel, its rate of growth, the relative intensity of the radiation and other characteristics of the spark may thus be investigated in considerable detail. A series of pictures made by Early and Martin of typical spark-channel growth appears in Figures 1 though 3.

In constructing the Kerr cell it is desirable to restrict the light so that the full beam passes between the electrodes This can be accomplished by mounting a pair of apertures just beyond each filter. It is well to remember that nitrobenzene, although the best liquid for most Kerr cell experiments, is highly volatile, poisonous, inflammable and a wonderful solvent for the cements one would most like to use in putting the glass parts of the cell together. The cell is triggered by a 7,000-volt direct-current pulse which may be derived from any of the conventional pulse-generator circuits. A number are described in Electronics: Experimental Techniques by William C. Elmore and Matthew Sands.

In addition to the energy radiated by the spark in the form of light, heat and ultraviolet rays, much is dissipated by the shock wave which develops during the early stages of the discharge. After about five microseconds both the pressure and the electrical resistance of the channel drop appreciably, halting the absorption of energy from the capacitor. Even so the shock wave packs enough wallop in the form of mechanical energy to suggest a number of interesting experiments. In one Early and Martin filled a small, thick-walled pressure vessel with polluted swamp water containing a variety of microorganisms. The water was subjected to a single discharge to learn whether the steep-front pressure wave could sterilize the water. Subsequent microscopic examination failed to show a trace of living organisms. Later, however, it was determined by means of culture techniques that some spores survived and were capable of growth.

The underwater spark's capacity for doing mechanical work can be demonstrated by exploding an underwater spark about 3/8 inch from the bottom of a sheet-metal tank. If a perforated steel die is placed against the underside of the tank, a clean hole will be punched in the bottom of the tank. In one such experiment the metal slug was ejected with such force that it ricocheted off a block of metal and buried itself in a plank of wood.

Similar, though less intense, forces can be developed by sparks in air. The essential increase in the resistance of the spark channel is achieved by recessing a pair of wire electrodes in the face of a ceramic block and covering them with a second block. This confines the discharge to the thin space between the blocks, as shown in the drawing in Figure 11. If the surfaces are relatively flat, so that the film of air between them is on the order of a thousandth of an inch thick, the spark will be accompanied by a shock wave capable of blowing the top block l0 feet into the air. Impressive effects can be observed even when the spark is powered by a relatively small source. A .02-microfarad capacitor charged to 10,000 volts, for example, is sufficient to produce a shock wave in air with a peak pressure of 100 atmospheres.

One of many possible applications suggested for an "air gun" of this type is that of high-speed printing. The paper to be printed is placed on top of the block containing the recessed electrodes and covered with a die or stencil of the character to be printed. The shock wave punches a clean replica of the stencil out of the paper. Raised characters can similarly be embossed in paper if an engraving is substituted for the stencil die. The embossed paper may be passed under an inked roller which coats the tops of the raised characters and thus improves their legibility. Sufficient energy for embossing 100-pound, highly calendered paper can be derived from a .01-microfarad capacitor charged to 8,000 volts, corresponding to about .3 watt-seconds.

Early and Martin devised a simple experiment for investigating the upper speed-limit of the embossing process. A paper disk is clamped in a sanding flange and chucked into a variable-speed drill-press. The rim of the paper runs between the face of the recessed block containing the electrodes and that of the die. The gap is fired after the disk has reached a predetermined speed. At a rim speed of 150 feet per second, EarIy and Martin found that the printed character, whether embossed by a die or punched from a stencil, was sharp and clean. At 260 feet per second perceptible blurring occurred, but the printing was still legible. Assuming characters of the size commonly used in typewriters, this corresponds to a printing rate of 374,400 words per minute! A practical machine based on the principle would necessarily reproduce text at a much lower rate, of course, because time would be consumed in moving the various dies into position. This could be accomplished by engraving the alphabet and other characters on the rim of a thick disk, somewhat as in the conventional stock ticker. The desired character would be printed by firing the gap at the appropriate instant.

Aside from its interest as a printing mechanism, an apparatus for embossing paper by this method also affords a means for investigating certain properties of the shock wave. The gas pressure generated by the spark can be determined roughly by tm analysis of the shape of the embossed character and the time required to make the impression. From the known speed of the Martin-Early spinning disk, for example, the rim velocity was calculated to be .007 centimeters per microsecond. The travel of the paper during the embossing operation (as determined from the blurring) was approximately .0025 centimeter. The embossing time was therefore on the order of three microseconds. The punched part of the paper was displaced .02 centimeter in three microseconds. The mass of the paper, determined by weighing a sample, was .006 grams per square centimeter.

The acceleration of the paper is equal to twice the distance traveled, 2 X .02 centimeters, divided by the square of the time, (3 X 10^-6)² seconds, or about 4.5 X 10⁹ centimeters per second per second. Force is defined as mass times acceleration. The force generated by the spark is therefore equal to the mass of the paper, 6 X 1O^-3 grams, multiplied by the acceleration, 4.5 X 10⁹ centimeters per second per second, or 2.7 X 10⁷ dynes per square centimeter.; This is equivalent to about 27 atmospheres or 375 pounds per square inch.

It has been proposed that the underwater spark could be used for measuring such properties as the elasticity of materials under high stress. If a small piece of wire screening is placed behind the underwater spark, for example, Kerr-cell photographs show a discontinuity in the screen's image caused by optical refraction at the cylindrical shock front. A prism of the material to be studied could be placed in the tank so that the difference in sonic velocity through the prism and through the water would show up in the photograph as an optical displacement of the image of the screening caused by refraction. The difference in displacement between the refracted and unrefracted shock front would give the angle of refraction. From this information the velocity of the shock wave through the' material can be calculated. The square of the velocity, when multiplied by the density, gives the elasticity, the quantity desired, and suggests how the amateur may investigate numerous other physical properties of materials by means of shock-wave effects generated by sparks both in air and water.

Richard J. Blume of the Watson Scientific Computing Laboratory at Columbia University writes: "Quartz crystal clocks are a professional interest of mine, and I should therefore like to offer comment on the article by W. W. Withrow, Jr., which appeared in 'The Amateur Scientist' for September. Withrow has demonstrated that the construction of a quartz clock is within the means and capability of the amateur. The task can be made considerably easier, however, than the text leads one to suppose. Specifically, the article contains several terrifying references to the trickiness of multivibrator circuits, especially the one which is required to divide the 120-kilocycle signal by precisely 20 to produce the 6-kilocycle signal. Contrary to the impression created by the article, a multivibrator is a tame and dependable object–if it is not required to divide by too large a number. Twenty is too large. If the division by 20 were carried out in three successive steps of 2, 2 and 5 (by the substitution of three successive multivibrators for Withrow's one) the amateur would experience much less difficulty during the initial adjustment of the circuits and the long-term stability of the multivibrator chain would be vastly increased."

In commenting on the same article Robert Kruse, of the Robert Kruse Laboratory in Madison, Conn., writes: "It is only fair to warn those intending to duplicate Withrow's 60-cycle quartz clock that these devices do not attain their final stability for several months, during which time frequent adjustments may be necessary. The change in operating characteristics with age occurs partly in the tubes, partly in other components, and sometimes in the crystal' itself. In one extreme case observed about 12 years ago a crystal bar drifted almost linearly for over a year. A happier 't choice of tubes might also have been recommended. Since replacement of a multivibrator tube may require readjustment of the resistors, it is desirable to use the 6SN7 tube rather than the 6SL7, the emission of which is somewhat limited for this service.

"Incidentally, multivibrators are rich in harmonics up to the 50th or higher. Accordingly a crystal clock may also be used as a source of stable frequencies, both radio and audio. These, may be tapped from the multivibrator through vacuum-tube amplifier, if the amplifier is not too tightly coupled to thc source. A suitable amplifier for this service can be built around the 6J5 tube, being similar in circuitry to Withrow's preamplifiers. Such a system has been made commercially ill highly refined form for about 20 years; with the aid. of a thermostatically controlled-.enclosure for the crystal the daily variation of the frequency is about 1 part in 200 million. An even older application of the idea in a unit now in service at the University of Minnesota was used for some years in a factory where intermittent motor loads caused very bad line-volt. age fluctuation. Yet we were able to keep the daily variation to 1 part in 5 million as judged by the beat-note of the 5-megacycle harmonic against the carrier frequency of WWV."

Bibliography

ELECTRICAL BREAKDOWN OF GASES. J. M. Meek and J. D. Craggs. Clarendon Press, 1953.

SUPERSONIC FLOW AND SHOCK WAVES. R. Courant and K. O. Friedricks. Interscience Publishers, Inc., 1948.

TEMPERATURE OF THE UNDER WATER SPARK AS COMPUTED FROM DISTRIBUTION OF INTENSITY IN OH ABSORPTION BANDS. Earl D. Wilson in Journal of the Optical Society of America, Vol. y 17, pages 37-46; July, 1928.

The Society for Amateur Scientists (SAS) is a nonprofit research and educational organization dedicated to helping people enrich their lives by following their passion to take part in scientific adventures of all kinds.