SOME OF the world's most useful work is performed by oscillators, which are devices that cause a flow of energy to alternate between two directions. An oscillator is an essential element of apparatus as diverse as pendulum clocks, jackhammers, radios, lasers, hydraulic rams and electronic heart pacemakers. For every kind of oscillator that has been put to work at least a dozen other kinds repose on laboratory shelves as interesting but useless gadgets. Three new examples of such gadgetry recently came to the attention of this department. The first of the three, which is known as a salt oscillator, is the creation of Seelye Martin of the University of Washington. Martin writes:

"I discovered this fascinating device while setting up a demonstration of a buoyant jet for a class in meteorology. The apparatus consisted of a hypodermic syringe, with the plunger removed, and a beaker of fresh water. The syringe was partly immersed in the water and held in position by an apparatus stand [see illustration at right]. Into the syringe I poured a saturated solution of table salt. (I made the solution by stirring salt into fresh water until no more would dissolve.) The level of the brine in the syringe was well above the level of the fresh water in the beaker.

"As I expected, the dense brine streamed out of the hypodermic needle in the form of a turbulent jet and sank to the bottom of the beaker. Although both fluids were clear, the stream of brine could be seen almost as easily as an icicle in air, because the difference in density of brine and fresh water refracts light. As the surface level of the salt water fell substantially below that of the fresh water the velocity of the jet decreased and the flow eventually stopped.

"At this point an astonishing effect was observed. The flow reversed! A jet of fresh water spurted into the syringe from the top of the needle and rose to the surface of the brine. The upward flow continued for a time and then reversed. Saltwater again flowed out of the needle and sank to the bottom. The sequence of alternating jets continued for many cycles.

"To explain the cause of these oscillations one must examine the dynamics of the system. The fluid level inside the syringe oscillates back and forth between two equilibrium positions where the forces that generate flow in each direction are exactly balanced. The first position is reached when salt water fills the needle. In this position the system is in equilibrium because the weight of the column of salt water that extends from the tip of the needle to the surface of the salt water just equals the weight of an equivalent column of fresh water of the same cross section that extends from the tip of the needle to the relatively higher surface of fresh water outside the syringe. The column of salt water is shorter than the equivalent column of fresh water, but the salt water is proportionately denser; hence the weights of the two columns are equal.

"The second position of equilibrium occurs when fresh water fills the needle. This column now consists of two parts: fresh water in the needle and, adjoining it above, a column of salt water of equal cross section. Accordingly the compound column is less dense than the column of pure brine, but the surface of the brine in the barrel of the syringe is now proportionately higher than it was when brine filled the needle-precisely enough higher to restore the balance of the system. "Both positions of equilibrium are extremely sensitive to disturbances because heavy brine overlies light fresh water. Equilibrium can exist only if the interface between the brine and the fresh water is parallel to the free surface of the solutions and is motionless. Just as water flows out of an inverted vessel and air flows in, so does the interface overturn between brine and fresh water. Salt water tends to flow down one side of the needle and fresh water to flow up the other side.

"In the case of the salt oscillator, however, flow in one direction tends to choke off flow in the other direction if the difference in density and velocity of the two fluids is sufficient. To visualize the action assume that salt water fills the needle. Any disturbance of the interface between salt water and fresh water at the bottom of the needle causes fresh water, however little in amount, to enter the needle and displace an equal amount salt water. The column now consists of both salt water and fresh water. Accordingly it is lighter than the equivalent column of fresh water outside the syringe. The columns are unbalanced. The resulting force causes additional fresh water to enter the needle, further increasing the imbalance.

"Because the volume of the needle is much smaller than the volume of the barrel, this intrusion of fresh water into the needle barely affects the surface level of the brine in the syringe. As a result the higher fresh water rises in the needle, the larger becomes the force that accelerates its upward flow. Eventually the forces that accompany the upward acceleration of fresh water become large enough to choke off the downward flow of brine. One-way flow is established in the needle. Immediately thereafter a jet of fresh water erupts in the barrel of the syringe. The level of the diluted brine rises [see illustration at right].

"Upward flow continues until the free surface of salt water in the syringe approaches the level of the second point of equilibrium. There the flow stops. The system is again in equilibrium, with the needle full of fresh water. Instability then initiates a local overturning of the interface and starts salt water down the needle. Thereafter the flow of brine continues until the first equilibrium point is reached once again and a new cycle begins.

"The salt oscillator runs for many cycles. The flow of fresh water into the syringe during each cycle increases the dilution of the salt water and thereby reduces after each cycle the pressure caused by the infusion of brine or fresh water into the needle at the beginning of the downward or upward flow. After many cycles this force becomes so small that it can no longer choke off the downward-upward motion induced by the instability. At this point steady, two-directional flow forms in the needle: fresh water streams up one side and salt water moves down the other [see illustration at left]. Finally the salt solution in the syringe becomes so dilute that it cannot maintain convective mixing. All visible motion ceases.

"The period of oscillation depends primarily on the geometry of the syringe. It varies roughly in proportion to the length of the needle. Doubling the length of the needle approximately doubles the period. An increase in the radius of the barrel or a decrease in the radius of the needle also increases the period.

"In oscillators with periods of more than 10 seconds the frequency of oscillation is independent of differences in the density of the fluids and appears from experiments to be influenced only slightly by their viscosity. The viscosity of the fresh water is constant, of course, so that the duration of the upward flow is almost constant throughout the period of operation. On the other hand, the salt concentration decreases with each successive cycle, and so therefore does the viscosity of the brine. The duration of the downward flow is initially from 5 to 10 percent greater than that of the upward flow, but the two intervals approach equality as the cycles continue.

"I was particularly fascinated by an incidental feature of the oscillator. It can be made to generate sequences of small vortex rings by mounting the syringe in a container of deep water such as a 1,000-milliliter graduated cylinder or an equivalent transparent vessel with a depth of at least 20 centimeters. If the period of oscillation is three seconds or more, the downward flow of salt water will usually break up into a chain of vortex rings that are between five and 20 millimeters apart on a common vertical axis.

"The vortex rings interact by overtaking and passing through one another, a phenomenon that has been described as 11 follows by the British physicist George K. Batchelor: 'The velocity field associated with the rear vortex ring has a radially outward component at the position of the front ring, and so the radius of the front ring gradually increases. This leads to a decrease in its speed of travel, and there is a corresponding increase in the speed of travel of the rear vortex, which ultimately passes through the larger vortex and in turn becomes the -front vortex. This maneuver is then repeated. It is possible to demonstrate in the laboratory one or two such passages of one vortex through another before they decay [see illustration at right].

"Both the jets and the vortexes can be made strikingly colorful by adding a few grains of fluorescein dye to the salt solution. I use the water-soluble form of fluorescein sodium salt known as uranine. Very small concentrations of this nontoxic dye turn a bright fluorescent green when the fluid is illuminated by a strong lamp. Uranine is available from dealers in chemical supplies. One can also stain the fresh water lightly with a dye. A drop or two of eosin will turn the water pink. (Most red inks are eosin solutions.)

"The salt oscillator can be improvised from various materials and in a number of different arrangements. I used Yale Luer-lok hypodermic syringes and needles primarily because needles of uniform bore and accurately known radius could be easily interchanged. This feature enabled me conveniently to make a variety of oscillators for investigating the influence of the geometry of the needle on the period of oscillation.

"The sale of hypodermic syringes to laymen is restricted in some communities. A simple substitute can be improvised by softening a glass pipette in the flame of a gas burner and drawing the end to a length of from three to 25 millimeters and a bore of from one to three millimeters. The period of oscillators made from pipettes tends to vary because the radius of the 'needle' is not constant. A glass capillary tube of reasonably constant radius can be made by softening the center of a length of 10millimeter tubing and quickly stretching the glass until it shrinks to a diameter of about two millimeters. Scratch the glass lightly with a file at two points about a centimeter apart near the center of the stretched portion. The one-centimeter length can be removed by grasping the ends of the capillary and pulling them to crack the glass.

"To generate oscillations of uniform period the ends of the capillary must be cut at right angles with respect to the axis of the tube. Glass tubing will usually break at a right angle if one pulls the capillary apart without exerting a lateral, or bending, force. The broken end can be ground true on fine emery cloth. (An experimenter using a hypodermic needle must also square off the pointed end.) The syringe-like structure can be completed by fitting a stopper into a length of large-diameter tubing and pushing the capillary into a hole made in the center of the stopper [see illustration at left.]

"Thin stoppers can be made by fitting a conventional stopper into the large diameter tubing. Mark the protruding portion at the end of the glass with a sharp pencil. Remove the stopper and make transverse cuts on both sides of the mark with a razor blade. For use in the salt oscillator the capillary must be vertical, that is, parallel to the gravitational field. An oscillator equipped with a capillary of one-millimeter bore and a length of four millimeters will have a period of about eight seconds. An extremely simple oscillator of the salt type can be built by making a pinhole in the center of the bottom of a soup can. One that I built this way operated for four days."

ANOTHER recently developed oscillator of more than passing interest has been made by Patrick Peebles of London. The active element is a flame of burning gas. In some respects the device is similar to the "sensitive flames" that were objects of popular experimentation during the 19th century. The burner used in that era consisted of a simple nozzle about a millimeter in diameter. When the nozzle was supplied with methane or a similar gas at a pressure of one ounce per square inch, the resulting jet was characterized by laminar flow: the ignited gas burned as a steady, quiet flame. The length of the flame could be increased by raising the pressure of the gas.

At a certain critical pressure that depended in part on the diameter of the nozzle the jet abruptly became turbulent: the flame danced and emitted a low roar that could be heard at a distance of several meters. When the pressure was adjusted almost to the point of turbulent flow, the flame became remarkably sensitive to external sounds. For example, at the critical adjustment the flame could be momentarily shortened to half its length by a finger snap at distances of up to 10 meters from the apparatus.

The sensitive flame would also oscillate if it was put inside a resonant cavity. A typical cavity consisted of a vertical tube open at both ends. When the burner was placed inside the tube at a position where the pressure of the standing sound wave was maximum, the sensitive flame would flicker at the resonant frequency of the tube and would emit sound waves continuously. So, at least, one can read in accounts published at the time. The editor of this department has never succeeded in making the arrangement work as an oscillator. Convection currents in the vertical tube blow out the flame.

Peebles' oscillator, however, works nicely. It consists essentially of two nozzles of one-millimeter bore mounted at an acute angle of approximately 30 degrees One nozzle makes a vertical jet and the other deflects the jet from the perpendicular [see illustration at right]. The resulting fan-shaped flame is made sensitive both by adjusting the pressure of the gas almost to the point of instability and by inserting the tip of an open copper tube into the flame about a centimeter from the top. The copper tube is two centimeters long and about eight millimeters in diameter. The end that penetrates the flame to a depth of about three millimeters is cut at an angle of 45 degrees. The edges of the cut should be smooth and sharp.

The opposite end of the tube makes a tight fit with a hole in the center of a copper plate that is some 10 centimeters square. The plate functions as a heat sink for cooling the copper tube. When one end of a second copper tube of similar diameter and a length of three centimeters or more is placed close to the base of the flame, the system goes into oscillation, emitting continuous sound at a pitch determined by the geometry of the second tube.

Essentially the oscillator functions by amplifying a small initial disturbance. A portion of the amplified output serves as a succeeding disturbance that initiates the next cycle, and so on. When the system is properly arranged, it can also be made to function as a conventional amplifier. Indeed, Peebles modified the device in two ways for use as both an amplifier and a loudspeaker.

In one scheme he coupled the acoustic output of an earphone into the tubing that supplies gas to the vertical jet. Sound waves from the earphone alter the pressure of the vertical jet and reappear as amplified sound that is emitted by the flame. He estimates that the output is amplified about 2,500 times. Speech sounds can be understood five meters from the flame although the quality of the reproduction is not good. Peebles operates the earphone with a small transistor radio. The amplifier can also be driven electrically.

In Peebles' second scheme the output of a radio that develops about 10 watts is coupled through a step-up transformer to a pair of electrodes placed on opposite sides of the flame near the base. The electrodes are two squares of copper screening about 15 millimeters wide spaced 10 millimeters apart. An automobile ignition coil can be used for the transformer, although the quality of the amplified sound will be higher if the experimenter uses a grid transformer of the kind designed for coupling low-impedance input circuits to vacuum-tube amplifiers. Peebles makes nozzles for the system by softening the middle of a six millimeter length of glass tubing in a gas flame and drawing the softened portion down to a bore of about a millimeter. If the device refuses to work, try adjusting the depth to which the upper copper the flame and also the height at which it penetrates the flame.

The simplest of the three oscillators was brought to my attention by A. D. Moore, professor emeritus of electrical engineering at the University of Michigan. It consists of a U-shaped loop of wire suspended at the top by an electrically insulating clamp [see illustration at left]. When the wire is heated to a temperature of about 400 degrees Celsius by either an alternating current or a direct current, the loop vibrates like a child's swing. I have made a number of loops with Nichrome wire salvaged from replacement units for a 1,000-watt radiant heater.

I clamp the ends of the loop between two strips of quarter-inch Transite, a building material composed largely of asbestos. I make the loops about 10 times longer than their width. The lengths of my oscillators have ranged from 10 to 36 inches. The frequency of vibration varies inversely with the length of the loop. Loops made of iron and other metals oscillate but not as vigorously as those made of Nichrome. The amplitude of vibration may exceed 15 angular degrees. For sources of power I have used both storage batteries and step-down transformers. Power requirements depend on the size of the loop but in general will be on the order of 10 amperes at 12 volts.

Why does the loop oscillate? Readers may enjoy discovering the answer at the workbench and validating it with at least two different experiments before Moore presents the explanation in this department next month.

AN interesting new concept in the ancient art of making devices for transmitting power mechanically is submitted by Rick Freund of 273 Lawton Avenue, Cliffside Park, N.J. 07010. Essentially the mechanism consists of a pair of universal joints made from business cards, plastic soda straws and wooden dowels. Like the oscillators, Freund's gadget appears to have no practical application, but it does show that a constant-velocity, right-angle drive can be made from flat stampings.

Each of Freund's universal joints includes a hollow shaft that supports at one end a fork with two tines that engage slots in opposite edges of a disk. A second pair of slots in the disk are similar but are positioned at right angles with respect to the first set. The second pair of slots engage a cardboard link that transmits motion between identical disks. Links of two kinds are used [see illustration at right]. One is a flat, rectangular card notched at opposite ends for engaging the disks. The other link consists of a pair of notched cards joined rigidly at right angles. The hollow shaft of each universal joint turns on wooden dowels supported at an angle of 45 degrees by a common base of Styrofoam.

With the flat link in place the assembly operates like a pair of crown gears: the motion of one shaft is transmitted uniformly to the other. According to Freund, when the mechanism is made of precision stainless-steel stampings with disks the size of a dime, it transmits eight inch-ounces of torque with little more play than a pair of mating gears. The behavior of the device can be altered significantly by substituting the right angled link for the flat link. The mechanism is then analogous to a mating pair of elliptical gears. When one shaft is driven at constant speed, the other accelerates from maximum to minimum twice during each revolution. The ratio of maximum to minimum speed at the output shaft depends on the angle at which the shafts are supported by the base. The ratio varies from one, when the shafts are in alignment, to 16 to one at an angle of 60 degrees.

Roger Hayward, who illustrates this department, delved into the history of the universal joint and found that its invention is generally ascribed to the English physicist Robert Hooke. Hayward writes: "Volume 3 of American Mechanical Dictionary (Houghton, Mifflin & Co., 1876, page 2683) shows a pair of Hooke's joints. The article points out the need for symmetry with respect to the link and notes that the two shafts have to lie in the same plane. All of this is credited to Hooke.

"Because of Hooke's work in physics and mechanics I am inclined to believe he first recognized that accelerated motion induced in a system by a single universal joint can be canceled by the opposing acceleration of a second joint. That Hooke had no use for cyclically varying speed is to me rather trivial. You cannot figure out a way to eliminate an error without first understanding what the error is.

"Freund refers to his device as a toy. This makes me wonder what a toy is. Are my tools the toys that I really love or are toys the things that I make with the tools? There must be a wide spectrum of items that are toys to some people, including Freund's interesting contraption.

Bibliography

EXPERIMENTAL SCIENCE: ELEMENTARY, PRACTICAL AND EXPERIMENTAL PHYSICS. George M. Hopkins. Munn & co., 1890.

PERIODIC OSCILLATIONS IN A MODEL OF THERMAL CONVECTION. Joseph B. Keller in Journal of Fluid Mechanics, Vol. 26, Part 3, pages 599-606; November, 1966.

ELEMENTARY MECHANICS OF FLUIDS, Hunter Rouse. John Wiley & Sons, Inc., 1946.

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.