§ 4. Referring to D. Many model builders make this mistake, i.e., the mistake of getting as low a centre of gravity as possible under the quite erroneous idea that they are thereby increasing the stability of the machine. Theoretically the centre of gravity should be the centre of head resistance, as also the centre of pressure.

In practice some prefer to put the centre of gravity in models slightly above the centre of head resistance, the reason being that, generally speaking, wind gusts have a "lifting" action on the machine. It must be carefully borne in mind, however, that if the centre of wind pressure on the aerofoil surface and the centre of gravity do not coincide, no matter at what point propulsive action be applied, it can be proved by quite elementary mechanics that such an arrangement, known as "acentric," produces a couple tending to upset the machine.

This action is the probable cause of many failures.

§ 5. Referring to E. If the propulsive action does not pass through the centre of gravity the system again becomes "acentric." Even supposing condition D fulfilled, and we arrive at the following most important result, viz., that for stability:—

The centres of gravity, of pressure, of head resistance, should be coincident, and the propulsive action of the propeller pass through this same point.

§ 6. Referring to F and N—the problem of longitudinal stability. There is one absolutely essential feature not mentioned in F or N, and that is for automatic longitudinal stability the two surfaces, the aerofoil proper and the balancer (elevator or tail, or both), must be separated by some considerable distance, a distance not less than four times the width of the main aerofoil.[9] More is better.

§ 7. With one exception (Pénaud) early experimenters with model aeroplanes had not grasped this all-important fact, and their models would not fly, only make a series of jumps, because they failed to balance longitudinally. In Stringfellow's and Tatin's models the main aerofoil and balancer (tail) are practically contiguous.

Pénaud in his rubber-motored models appears to have fully realised this (vide Fig. 7), and also the necessity for using long strands of rubber. Some of his models flew 150 ft., and showed considerable stability.

With three surfaces one would set the elevator at a slight plus angle, main aerofoil horizontal (neither positive nor negative), and the tail at a corresponding negative angle to the positive one of the elevator.

Referring to O.[10] One would naturally be inclined to put a keel surface—or, in other words, vertical fins—beneath the centre of gravity, but D shows us this may have the opposite effect to what we might expect.

In full-sized machines, those in which the distance between the main aerofoil and balancers is considerable (like the Farman) show considerable automatic longitudinal stability, and those in which it is short (like the Wright) are purposely made so with the idea of doing away with it, and rendering the machine quicker and more sensitive to personal control. In the case of the Stringfellow and Tatin models we have the extreme case—practically the bird entirely volitional and personal—which is the opposite in every way to what we desire on a model under no personal or volitional control at all.

The theoretical conditions stated in F and N are fully borne out in practice.

And since a curved aerofoil even when set at a slight negative angle has still considerable powers of sustentation, it is possible to give the main aerofoil a slight negative angle and the elevator a slight positive one. This fact is of the greatest importance, since it enables us to counteract the effect of the travel of the "centre of pressure."[11]

§ 8. Referring to I. This, again, is of primary importance in longitudinal stability. The Farman machine has three such planes—elevator, main aerofoil, tail the Wright originally had not, but is now being fitted with a tail, and experiments on the Short-Wright biplane have quite proved its stabilising efficiency.

The three plane (triple monoplane) in the case of models has been tried, but possesses no advantage so far over the double monoplane type. The writer has made many experiments with vertical fins, and has found the machine very stable, even when the fin or vertical keel is placed some distance above the centre of gravity.

§ 9. The question of transverse (side to side) stability at once brings us to the question of the dihedral angle, practically similar in its action to a flat plane with vertical fins.

§ 10. The setting up of the front surface at an angle to the rear, or the setting of these at corresponding compensatory angles already dealt with, is nothing more nor less than the principle of the dihedral angle for longitudinal stability.

As early as the commencement of last century Sir George Cayley (a man more than a hundred years ahead of his times) was the first to point out that two planes at a dihedral angle constitute a basis of stability. For, on the machine heeling over, the side which is required to rise gains resistance by its new position, and that which is required to sink loses it.

§ 11. The dihedral angle principle may take many forms.

As in Fig. 12 a is a monoplane, the rest biplanes. The angles and curves are somewhat exaggerated. It is quite a mistake to make the angle excessive, the "lift" being thereby diminished. A few degrees should suffice.

Whilst it is evident enough that transverse stability is promoted by making the sustaining surface trough-shaped, it is not so evident what form of cross section is the most efficient for sustentation and equilibrium combined.

It is evident that the righting moment of a unit of surface of an aeroplane is greater at the outer edge than elsewhere, owing to the greater lever arm.

§ 12. The "upturned tip" dihedral certainly appears to have the advantage.

The outer edges of the aerofoil then should be turned upward for the purpose of transverse stability, while the inner surface should remain flat or concave for greater support.

§ 13. The exact most favourable outline of transverse section for stability, steadiness and buoyancy has not yet been found; but the writer has found the section given in Fig. 13, a very efficient one.

CHAPTER IV.

Section I.—Rubber Motors.

§ 2. Let us take a piece of elastic (rubber) cord, and stretch it with known weights and observe carefully what happens. We shall find that, first of all: the extension is proportional to the weight suspended—but soon we have an increasing increase of extension. In one experiment made by the writer, when the weights were removed the rubber cord remained 1/8 of an inch longer, and at the end of an hour recovered itself to the extent of 1/16, remaining finally permanently 1/16 of an inch longer. Length of elastic cord used in this experiment 8-1/8 inches, 3/16 of an inch thick. Suspended weights, 1 oz. up to 64 oz. Extension from ¼ inch up to 24-5/8 inches. Graph drawn in Fig. 14, No. B abscissæ extension in eighths of an inch, ordinates weights in ounces. So long as the graph is a straight line it shows the extension is proportional to the suspended weight; afterwards in excess.

In this experiment we have been able to stretch (distort) a piece of rubber to more than three times its original length, and afterwards it finally returns to almost its original length: not only so, a piece of rubber cord can be stretched to eight or nine times its original length without fracture. Herein lies its supreme advantage over steel or other springs. Weight for weight more energy can be got or more work be done by stretched (or twisted, or, to speak more correctly, by stretched-twisted) rubber cord than from any form of steel spring.[12] It is true it is stretched—twisted—far beyond what is called the "elastic limit," and its efficiency falls off, but with care not nearly so quickly as is commonly supposed, but in spite of this and other drawbacks its advantages far more than counterbalance these.

§ 3. Experimenting with cords of varying thickness we find that: the extension is inversely proportional to the thickness. If we leave a weight hanging on a piece of rubber cord (stretched, of course, beyond its "elastic limit") we find that: the cord continues to elongate as long as the weight is left on. For example: a 1 lb. weight hung on a piece of rubber cord, 8-1/8 inches long and 1/8 of an inch thick, stretched it—at first—6¼ inches; after two minutes this had increased to 6-5/8 (3/8 of an inch more). One hour later 1/8 of an inch more, and sixteen hours later 1/8 of an inch more, i.e. a sixteen hours' hang produced an additional extension of ¾ of an inch. On a thinner cord (half the thickness) same weight produced an additional extension (after 14 hours) of 10-3/8 in.

N.B.—An elastic cord or spring balance should never have a weight left permanently on it—or be subjected to a distorting force for a longer time than necessary, or it will take a "permanent set," and not return to even approximately its original length or form.

In a rubber cord the extension is directly proportional to the length as well as inversely proportional to the thickness and to the weight suspended—true only within the limits of elasticity.

§ 4. When a Rubber Cord is stretched there is an Increase of Volume.—On stretching a piece of rubber cord to twice its original (natural) length, we should perhaps expect to find that the string would only be half as thick, as would be the case if the volume remained the same. Performing the experiment, and measuring the cord as accurately as possible with a micrometer, measuring to the one-thousandth of an inch, we at once perceive that this is not the case, being about two-thirds of its former volume.

§ 5. In the case of rubber cord used for a motive power on model aeroplanes, the rubber is both twisted and stretched, but chiefly the latter.

Thirty-six strands of rubber, weight about 56 grammes, at 150 turns give a torque of 4 oz. on a 5-in. arm, but an end thrust, or end pull, of about 3½ lb. (Ball bearings, or some such device, can be used to obviate this end thrust when desirable.) A series of experiments undertaken by the writer on the torque produced by twisted rubber strands, varying in number, length, etc., and afterwards carefully plotted out in graph form, have led to some very interesting and instructive results. Ball bearings were used, and the torque, measured in eighths of an ounce, was taken (in each case) from an arm 5 in. in length.

The following are the principal results arrived at. For graphs, see Fig. 16.

§ 6. A. Increasing the number of (rubber) strands by one-half (length and thickness of rubber remaining constant) increases the torque (unwinding tendency) twofold, i.e., doubles the motive power.

B. Doubling the number of strands increases the torque more than three times—about 3-1/3 times, 3 times up to 100 turns, 3½ times from 100 to 250 turns.

C. Trebling the number of strands increases the torque at least seven times.

The increased size of the coils, and thereby increased extension, explains this result. As we increase the number of strands, the number of twists or turns that can be given it becomes less.

D. Doubling the number of strands (length, etc., remaining constant) diminishes the number of turns by one-third to one-half. (In few strands one-third, in 30 and over one-half.)

E. If we halve the length of the rubber strands, keeping the number of strands the same, the torque is but slightly increased for the first 100 turns; at 240 turns it is double. But the greater number of turns—in ratio of about 2:1—that can be given the longer strand much more than compensates for this.

F. No arrangement of the strands, per se, gets more energy (more motive power) out of them than any other, but there are special reasons for making the strands—

G. As long and as few in number as possible.

1. More turns can be given it.

2. It gives a far more even torque. Twelve strands 2 ft. 6 in. long give practically a line of small constant angle. Thirty-six strands same length a much steeper angle, with considerable variations.

A very good result, which the writer has verified in practice, paying due regard to both propeller and motor, is to make—

H. The length of the rubber strands twice[13] in feet the number of the strands in inches,[14] e.g., if the number of strands is 12 their length should be 2 ft., if 18, 3 ft., and so on.

§ 7. Experiments with 32 to 38 strands 2 ft. 6 in. long give a torque curve almost precisely similar to that obtained from experiments made with flat spiral steel springs, similar to those used in watches and clocks; and, as we know, the torque given by such springs is very uneven, and has to be equalised by use of a fusee, or some such device. In the case of such springs it must not be forgotten that the turning moment (unwinding tendency) is NOT proportional to the amount of winding up, this being true only in the "balance" springs of watches, etc., where both ends of the spring are rigidly fastened.

In the case of Spring Motors.[15]

I. The turning moment (unwinding tendency) is proportional to the difference between the angle of winding and yielding, proportional to the moment of inertia of its section, i.e., to the breadth and the cube of its thickness, also proportional to the modulus of elasticity of the substance used, and inversely proportional to the length of the strip.

§ 8. Referring back to A, B, C, there are one or two practical deductions which should be carefully noted.

Supposing we have a model with one propeller and 36 strands of elastic. If we decide to fit it with twin screws, then, other reasons apart, we shall require two sets of strands of more than 18 in number each to have the same motive power (27 if the same torque be required).[16] This is an important point, and one not to be lost sight of when thinking of using two propellers.

Experiments on—

§9. The Number of Revolutions (turns) that can be given to Rubber Motors led to interesting results, e.g., the number of turns to produce a double knot in the cord from end to end were, in the case of rubber, one yard long:—

It will be at once noticed that the greater the number of rubber strands used in a given length, the fewer turns will it stand in proportion. For instance, 8 strands double knot at 310, and 4 at 440 (and not at 620), 16 at 200, and 8 at 310 (and not 400), and so on. The reason, of course, is the more the strands the greater the distance they have to travel round themselves.

§ 10. The Maximum Number of Turns.—As to the maximum number of permissible turns, rubber has rupture stress of 330 lb. per sq. in., but a very high permissible stress, as much as 80 per cent. The resilience (power of recovery after distortion) in tension of rubber is in considerable excess of any other substance, silk being the only other substance which at all approaches it in this respect, the ratio being about 11 : 9. The resilience of steel spiral spring is very slight in comparison.

A rubber motor in which the double knot is not exceeded by more than 100 turns (rubber one yard in length) should last a good time. When trying for a record flight, using new elastic, as many as even 500 or 600 or even more turns have been given in the case of 32-36 strands a yard in length; but such a severe strain soon spoils the rubber.

§ 11. On the Use of "Lubricants."—One of the drawbacks to rubber is that if it be excessively strained it soon begins to break up. One of the chief causes of this is that the strands stick together—they should always be carefully separated, if necessary, after a flight—and an undue strain is thereby cast on certain parts. Apart also from this the various strands are not subject to the same tension. It has been suggested that if some means could be devised to prevent this, and allow the strands to slip over one another, a considerable increase of power might result. It must, however, be carefully borne in mind that anything of an oily or greasy nature has an injurious effect on the rubber, and must be avoided at all costs. Benzol, petroleum, ether, volatile oils, turpentine, chloroform, naphtha, vaseline, soap, and all kinds of oil must be carefully avoided, as they soften the rubber, and reduce it more or less to the consistence of a sticky mass. The only oil which is said to have no action on rubber, or practically none, is castor oil; all the same, I do not advise its use as a lubricant.

There are three only which we need consider:—

• 1. Soda and water.
• 2. French chalk.
• 3. Pure redistilled glycerine.

The first is perfectly satisfactory when freshly applied, but soon dries up and evaporates.

The second falls off; and unless the chalk be of the softest kind, free from all grit and hard particles, it will soon do more harm than good.

The third, glycerine, is for ordinary purposes by far the best, and has a beneficial rather than a deleterious effect on the rubber; but it must be pure. The redistilled kind, free from all traces of arsenic, grease, etc., is the only kind permissible. It does not evaporate, and a few drops, comparatively speaking, will lubricate fifty or sixty yards of rubber.

Being of a sticky or tacky nature it naturally gathers up dust and particles of dirt in course of time. To prevent these grinding into the rubber, wash it from time to time in warm soda, and warm and apply fresh glycerine when required.

Glycerine, unlike vaseline (a product of petroleum), is not a grease; it is formed from fats by a process known as saponification, or treatment of the oil with caustic alkali, which decomposes the compound, forming an alkaline stearate (soap), and liberating the glycerine which remains in solution when the soap is separated by throwing in common salt. In order to obtain pure glycerine, the fat can be decomposed by lead oxide, the glycerine remaining in solution, and the lead soap or plaster being precipitated.

By using glycerine as a lubricant the number of turns that can be given a rubber motor is greatly increased, and the coils slip over one another freely and easily, and prevent the throwing of undue strain on some particular portion, and absolutely prevent the strands from sticking together.

§ 12. The Action of Copper upon Rubber.—Copper, whether in the form of the metal, the oxides, or the soluble salts, has a marked injurious action upon rubber.

In the case of metallic copper this action has been attributed to oxidation induced by the dissolved oxygen in the copper. In working drawings for model aeroplanes I have noticed designs in which the hooks on which the rubber strands were to be stretched were made of copper. In no case should the strands be placed upon bare metal. I always cover mine with a piece of valve tubing, which can easily be renewed from time to time.

§ 12A. The Action of Water, etc., on Rubber.—Rubber is quite insoluble in water; but it must not be forgotten that it will absorb about 25 per cent. into its pores after soaking for some time.

Ether, chloroform, carbon-tetrachloride, turpentine, carbon bi-sulphide, petroleum spirit, benzene and its homologues found in coal-tar naphtha, dissolve rubber readily. Alcohol is absorbed by rubber, but is not a solvent of it.

§ 12B. How to Preserve Rubber.—In the first place, in order that it shall be possible to preserve and keep rubber in the best condition of efficiency, it is absolutely essential that the rubber shall be, when obtained, fresh and of the best kind. Only the best Para rubber should be bought; to obtain it fresh it should be got in as large quantities as possible direct from a manufacturer or reliable rubber shop. The composition of the best Para rubber is as follows:—Carbon, 87·46 per cent.; hydrogen, 12·00 per cent.; oxygen and ash, 0·54 per cent.

In order to increase its elasticity the pure rubber has to be vulcanised before being made into the sheet some sixty or eighty yards in length, from which the rubber threads are cut; after vulcanization the substance consists of rubber plus about 3 per cent. of sulphur. Now, unfortunately, the presence of the sulphur makes the rubber more prone to atmospheric oxidation. Vulcanized rubber, compared to pure rubber, has then but a limited life. It is to this process of oxidation that the more or less rapid deterioration of rubber is due.

To preserve rubber it should be kept from the sun's rays, or, indeed, any actinic rays, in a cool, airy place, and subjected to as even a temperature as possible. Great extremes of temperature have a very injurious effect on rubber, and it should be washed from time to time in warm soda water. It should be subjected to no tension or compression.

Deteriorated rubber is absolutely useless for model aeroplanes.

§ 13. To Test Rubber.—Good elastic thread composed of pure Para rubber and sulphur should, if properly made, stretch to seven times its length, and then return to its original length. It should also possess a stretching limit at least ten times its original length.

As already stated, the threads or strands are cut from sheets; these threads can now be cut fifty to the inch. For rubber motors a very great deal so far as length of life depends on the accuracy and skill with which the strands are cut. When examined under a microscope (not too powerful) the strands having the least ragged edge, i.e., the best cut, are to be preferred.

§ 14. The Section—Strip or Ribbon versus Square.—In section the square and not the ribbon or strip should be used. The edge of the strip I have always found more ragged under the microscope than the square. I have also found it less efficient. Theoretically no doubt a round section would be best, but none such (in small sizes) is on the market. Models have been fitted with a tubular section, but such should on no account be used.

§ 15. Size of the Section.—One-sixteenth or one-twelfth is the best size for ordinary models; personally, I prefer the thinner. If more than a certain number of strands are required to provide the necessary power, a larger size should be used. It is not easy to say what this number is, but fifty may probably be taken as an outside limit. Remember the size increases by area section; twice the sectional height and breadth means four times the rubber.

§ 16. Geared Rubber Motors.—It is quite a mistake to suppose that any advantage can be obtained by using a four to one gearing, say; all that you do obtain is one-fourth of the power minus the increased friction, minus the added weight. This presumes, of course, you make no alteration in your rubber strands.

Gearing such as this means short rubber strands, and such are not to be desired; in any case, there is the difficulty of increased friction and added weight to overcome. It is true by splitting up your rubber motor into two sets of strands instead of one you can obtain more turns, but, as we have seen, you must increase the number of strands to get the same thrust, and you have this to counteract any advantage you gain as well as added weight and friction.

§ 17. The writer has tried endless experiments with all kinds of geared rubber motors, and the only one worth a moment's consideration is the following, viz., one in which two gear wheels—same size, weight, and number of teeth—are made use of, the propeller being attached to the axle of one of them, and the same number of strands are used on each axle. The success or non-success of this motor depends entirely on the method used in its construction. At first sight it may appear that no great skill is required in the construction of such a simple piece of apparatus. No greater mistake could be made. It is absolutely necessary that the friction and weight be reduced to a minimum, and the strength be a maximum. The torque of the rubber strands on so short an arm is very great.

Ordinary light brass cogwheels will not stand the strain.

A. The cogwheels should be of steel[17] and accurately cut of diameter sufficient to separate the two strands the requisite distance, but no more.

B. The weight must be a minimum. This is best attained by using solid wheels, and lightening by drilling and turning.

C. The friction must be a minimum. Use the lightest ball bearings obtainable (these weigh only 0·3 gramme), adjust the wheels so that they run with the greatest freedom, but see that the teeth overlap sufficiently to stand the strain and slight variations in direction without fear of slipping. Shallow teeth are useless.

D. Use vaseline on the cogs to make them run as easily as possible.

E. The material of the containing framework must be of maximum strength and minimum lightness. Construct it of minimum size, box shaped, use the thinnest tin (really tinned sheet-iron) procurable, and lighten by drilling holes, not too large, all over it. Do not use aluminium or magnalium. Steel, could it be procured thin enough, would be better still.

F. Use steel pianoforte wire for the spindles, and hooks for the rubber strands, using as thin wire as will stand the strain.

Unless these directions are carefully carried out no advantage will be gained—the writer speaks from experience. The requisite number of rubber strands to give the best result must be determined by experiment.

§ 18. One advantage in using such a motor as this is that the two equal strands untwisting in opposite directions have a decided steadying effect on the model, similar almost to the case in which two propellers are used.

The "best" model flights that the writer has achieved have been obtained with a motor of this description.[18]

In the case of twin screws two such gearings can be used, and the rubber split up into four strands. The containing framework in this case can be simply light pieces of tubing let into the wooden framework, or very light iron pieces fastened thereto.

Do not attempt to split up the rubber into more than two strands to each propeller.

Section II.—Other Forms of Motors.

§ 18A. Spring Motors.—This question has already been dealt with more or less whilst dealing with rubber motors, and the superiority of the latter over the former pointed out. Rubber has a much greater superiority over steel or other springs, because in stretch-twisted rubber far more energy can be stored up weight for weight. One pound weight of elastic can be made to store up some 320 ft.-lb. of energy, and steel only some 65 lb. And in addition to this there is the question of gearing, involving extra weight and friction; that is, if flat steel springs similar to those used in clockwork mechanism be made use of, as is generally the case. The only instance in which such springs are of use is for the purpose of studying the effects of different distributions of weight on the model, and its effect on the balance of the machine; but effects such as this can be brought about without a change of motor.

§ 18B. A more efficient form of spring motor, doing away with gearing troubles, is to use a long spiral spring (as long as the rubber strands) made of medium-sized piano wire, similar in principle to those used in some roller-blinds, but longer and of thinner steel.

The writer has experimented with such, as well as scores of other forms of spring motors, but none can compare with rubber.

The long spiral form of steel spring is, however, much the best.

§ 18C. Compressed Air Motors.—This is a very fascinating form of motor, on paper, and appears at first sight the ideal form. It is so easy to write: "Its weight is negligible, and it can be provided free of cost; all that is necessary is to work a bicycle pump for as many minutes as the motor is desired to run. This stored-up energy can be contained in a mere tube, of aluminium or magnalium, forming the central rib of the machine, and the engine mechanism necessary for conveying this stored-up energy to the revolving propeller need weigh only a few ounces." Another writer recommends "a pressure of 300 lb."

§ 18D. A pneumatic drill generally works at about 80 lb. pressure, and when developing 1 horse-power, uses about 55 cubic ft. of free air per minute. Now if we apply this to a model aeroplane of average size, taking a reservoir 3 ft. long by 1½ in. internal diameter, made of magnalium, say—steel would, of course, be much better—the weight of which would certainly not be less than 4 oz., we find that at 80 lb. pressure such a motor would use

55 / Horse Power (H.P.)

cub. ft. per minute.

Now 80 lb. is about 5½ atmospheres, and the cubical contents of the above motor some 63 cub. in. The time during which such a model would fly depends on the H.P. necessary for flight; but a fair allowance gives a flight of from 10 to 30 sec. I take 80 lb. pressure as a fair practical limit.

§ 18E. The pressure in a motor-car tyre runs from 40 to 80 lb., usually about 70 lb. Now 260 strokes are required with an ordinary inflator to obtain so low a pressure as 70 lb., and it is no easy job, as those who have done it know.

§ 19. Prior to 1893 Mr. Hargraves (of cellular kite fame) studied the question of compressed-air motors for model flying machines. His motor was described as a marvel of simplicity and lightness, its cylinder was made like a common tin can, the cylinder covers cut from sheet tin and pressed to shape, the piston and junk rings of ebonite.

One of his receivers was 23-3/8 in. long, and 5·5 in. diameter, of aluminium plate 0·2 in. thick, 3/8 in. by 1/8 in. riveting strips were insufficient to make tight joints; it weighed 26 oz., and at 80 lb. water pressure one of the ends blew out, the fracture occurring at the bend of the flange, and not along the line of rivets. The receiver which was successful being apparently a tin-iron one; steel tubing was not to be had at that date in Sydney. With a receiver of this character, and the engine referred to above, a flight of 343 ft. was obtained, this flight being the best. (The models constructed by him were not on the aeroplane, but ornithoptere, or wing-flapping principle.) The time of flight was 23 seconds, with 54½ double vibrations of the engines. The efficiency of this motor was estimated to be 29 per cent.

§ 20. By using compressed air, and heating it in its passage to the cylinder, far greater efficiency can be obtained. Steel cylinders can be obtained containing air under the enormous pressure of 120 atmospheres.[19] This is practically liquid air. A 20-ft. cylinder weighs empty 23 lb. The smaller the cylinder the less the proportionate pressure that it will stand; and supposing a small steel cylinder, produced of suitable form and weight, and capable of withstanding with safety a pressure of from 300 to 600 lb. per sq. in., or from 20 to 40 atmospheres. The most economical way of working would be to admit the air from the reservoir directly to the motor cylinders; but this would mean a very great range in the initial working pressure, entailing not-to-be-thought-of weight in the form of multi-cylinder compound engines, variable expansion gear, etc.

§ 21. This means relinquishing the advantages of the high initial pressure, and the passing of the air through a reducing valve, whereby a constant pressure, say, of 90 to 150, according to circumstances, could be maintained. By a variation in the ratio of expansion the air could be worked down to, say, 30 lb.

The initial loss entailed by the use of a reducing valve may be in a great measure restored by heating the air before using it in the motor cylinders; by heating it to a temperature of only 320°F., by means of a suitable burner, the volume of air is increased by one half, the consumption being reduced in the same proportion; the consumption of air used in this way being 24 lb. per indicated horse-power per hour. But this means extra weight in the form of fuel and burners, and what we gain in one way we lose in another. It is, of course, desirable that the motor should work at as low a pressure as possible, since as the store of air is used up the pressure in the reservoir falls, until it reaches a limit below which it cannot usefully be employed. The air then remaining is dead and useless, adding only to the weight of the aeroplane.

§ 22. From calculations made by the writer the entire weight of a compressed-air model motor plant would be at least one-third the weight of the aeroplane, and on a small scale probably one-half, and cannot therefore hold comparison with the steam engine discussed in the next paragraph. In concluding these remarks on compressed-air motors, I do not wish to dissuade anyone from trying this form of motor; but they must not embark on experiments with the idea that anything useful or anything superior to results obtained with infinitely less expense by means of rubber can be brought to pass with a bicycle pump, a bit of magnalium tube, and 60 lb. pressure.

§ 22A. In Tatin's air-compressed motor the reservoir weighed 700 grammes, and had a capacity of 8 litres. It was tested to withstand a pressure of 20 atmospheres, but was worked only up to seven. The little engine attached thereto weighed 300 grammes, and developed a motive power of 2 kilogram-metres per second (see ch. iii.).

§ 23. Steam-Driven Motors.—Several successful steam-engined model aeroplanes have been constructed, the most famous being those of Professor Langley.

Having constructed over 30 modifications of rubber-driven models, and experimented with compressed air, carbonic-acid gas, electricity, and other methods of obtaining energy, he finally settled upon the steam engine (the petrol motor was not available at that time, 1893). After many months' work it was found that the weight could not be reduced below 40 lb., whilst the engine would only develop ½ H.P., and finally the model was condemned. A second apparatus to be worked by compressed air was tried, but the power proved insufficient. Then came another with a carbonic-acid gas engine. Then others with various applications of electricity and gas, etc., but the steam engine was found most suitable; yet it seemed to become more and more doubtful whether it could ever be made sufficiently light, and whether the desired end could be attained at all. The chief obstacle proved not to be with the engines, which were made surprisingly light after sufficient experiment. The great difficulty was to make a boiler of almost no weight which would give steam enough.

§ 24. At last a satisfactory boiler and engine were produced.

The engine was of 1 to 1½ H.P., total weight (including moving parts) 26 oz. The cylinders, two in number, had each a diameter of 1¼ in., and piston stroke 2 in.

The boiler, with its firegrate, weighed a little over 5 lb. It consisted of a continuous helix of copper tubing, 3/8 in. external diameter, the diameter of the coil being 3 in. altogether. Through the centre of this was driven the blast from an "Ælopile," a modification of the naphtha blow-torch used by plumbers, the flame of which is about 2000° F.[20] The pressure of steam issuing into the engines varied from 100 to 150 lb. per sq. in.; 4 lb. weight of water and about 10 oz. of naphtha could be carried. The boiler evaporated 1 lb. of water per minute.

The twin propellers, 39 in. in diam., pitch 1¼, revolved from 800 to 1000 a minute. The entire aeroplane was 15 ft. in length, the aerofoils from tip to tip about 14 ft., and the total weight slightly less than 30 lb., of which one-fourth was contained in the machinery. Its flight was a little over half a mile in length, and of 1½ minutes' duration. Another model flew for about three-quarters of a mile, at a rate of about 30 miles an hour.

It will be noted that engine, generator, etc., work out at about 7 lb. per H.P. Considerable advance has been made in the construction of light and powerful model steam engines since Langley's time, chiefly in connexion with model hydroplanes, and a pressure of from 500 to 600 lb. per sq. in. has been employed; the steam turbine has been brought to a high state of perfection, and it is now possible to make a model De Laval turbine of considerable power weighing almost next to nothing,[21] the real trouble, in fact the only one, being the steam generator. An economization of weight means a waste of steam, of which models can easily spend their only weight in five minutes.

§ 25. One way to economize without increased weight in the shape of a condenser is to use spirit (methylated spirit, for instance) for both fuel and boiler, and cause the exhaust from the engines to be ejected on to the burning spirit, where it itself serves as fuel. By using spirit, or some very volatile hydrocarbon, instead of water, we have a further advantage from the fact that such vaporize at a much lower temperature than water.

§ 26. When experimenting with an engine of the turbine type we must use a propeller of small diameter and pitch, owing to the very high velocity at which such engines run.

Anyone, however, who is not an expert on such matters would do well to leave such motors alone, as the very highest technical skill, combined with many preliminary disappointments and trials, are sure to be encountered before success is attained.

§ 27. And the smaller the model the more difficult the problem—halve your aeroplane, and your difficulties increase anything from fourfold to tenfold.

The boiler would in any case be of the flash type of either copper or steel tubing (the former for safety), with a magnalium container for the spirit, and a working pressure of from 150 to 200 lb. per sq. in. Anything less than this would not be worth consideration.

§ 28. Some ten months after Professor Langley's successful model flights (1896), experiments were made in France at Carquenez, near Toulon. The total weight of the model aeroplane in this case was 70 lb.; the engine power a little more than 1 H.P. Twin screws were used—one in front and one behind. The maximum velocity obtained was 40 miles per hour; but the length of run only 154 yards, and duration of flight only a few seconds. This result compares very poorly with Langley's distance (of best flight), nearly one mile, duration 1 min. 45 sec. The maximum velocity was greater—30 to 40 miles per hour. The total breadth of this large model was rather more than 6 metres, and the surface a little more than 8 sq. metres.

§ 29. Petrol Motors.—Here it would appear at first thought is the true solution of the problem of the model aeroplane motor. Such a motor has solved the problem of aerial locomotion, as the steam engine solved that of terrestrial and marine travel, both full sized and model; and if in the case of full sized machines, then why not models.