Fig. 15.—The manner of building up the large screws.
About the time that I was making these experiments, my duties called me to Paris, and while there I called on my old friend Gaston Tissandier. Through his influence I was permitted to see some models of the screws that were alleged to have been used by Captain Renard in his experiments for the French Government, and I was somewhat surprised to find the form of the blades, the same as shown at h, [Fig. 14], and completely without any twist. On my return to England, I made a screw of this description. It is also shown in the photographic illustration, [Fig. 9]. Upon testing this screw, I found that its efficiency was only 40 per cent. of that of a—that is, the energy or acceleration imparted to the air was only 40 per cent. of the readings of the dynamometer. It then occurred to me that this particular form of screw was probably the one that the French had for exhibition purposes, but not the one they intended to use. Having tried all the various forms of screws and other objects shown in [Fig. 9], I made some sheet metal screws; also a screw which consisted of a steel frame covered with woven fabric, and which was identical with screws that I had seen described in various works relating to aerial navigation. It was found quite impossible to keep the fabric taut and smooth, and the results were very bad indeed, it being only 40 per cent. as efficient as a well-made wooden screw.
Fig. 16.—A fabric-covered screw with a very low efficiency.
Having thus ascertained the best form of a screw, I built up my first large screws, which were 17 feet 10 inches in diameter, after the well-known manner of making wooden patterns for casting steamship propellers. [Fig. 15] shows the form of the end of the blade, the middle of the blade, and the hub. My first pair of large screws had a pitch of 24 feet, but these were too great a drag on the engine. I therefore made another pair with 16 feet pitch which greatly increased the piston speed, and permitted the engines to develop much more power; the screw thrust was also increased just in an inverse ratio to the pitch of the screws. Another pair of screws was tried with 14 feet pitch and 12 feet in diameter, but these did not do so well. My large screws were made with a great degree of accuracy; they were perfectly smooth and even on both sides, the blades being thin and held in position by a strip of rigid wood on the back of the blade. In order to prevent the thrust from collapsing the blades, wires were extended backwards and attached to a prolongation of the shaft. Like the small screws, they were made of the very best kind of seasoned American white pine, and when finished were varnished on both sides with hot glue. When this was thoroughly dry, they were sand-papered again and made perfectly smooth and even. The blades were then covered with strong Irish linen fabric of the smoothest and best make. Glue was used for attaching the fabric, and when dry another coat of glue was applied, the surface rubbed down again and then painted with zinc white in the ordinary way and varnished. These screws worked exceedingly well. I had means of ascertaining, with a great degree of accuracy, the thrust of the screw, the number of turns per minute, the speed of the machine, and, in fact, all the events that were taking place on the machine. It was found that when the screw thrust in pounds was multiplied by the pitch in feet, and by the number of revolutions made in a minute of time, it exactly corresponded to the power that the engines were developing, and that the amount of loss in skin friction was so small as to be practically negligible.
Fig. 17.—The hub and one of the blades of the screw on the Farman machine. The blade c, is a sheet of metal riveted to the rod b, and forms a projection on the back of the blade which greatly reduces its efficiency. The peculiar form of hub employed makes it possible to change the diameter and pitch of this screw at will.
In connection with this subject I would say that many experimenters claim to have shown that the skin friction on screws is considerable, in fact, so great as to be a very important factor in the equation of flight. I am, however, of the opinion that these experimenters have not had well-made screws. If the surface of the screw is uneven, irregular, or rough, a considerable amount of energy is lost, as shown in the French screw and the fabric covered screw. It is simply a question of having a screw well-made. In those recently employed in France (see [Fig. 17]), the blades are of hammered sheet metal, the twist is not uniform or true, and what is worst of all, the arm b projects on the back of the blade and offers a good deal of resistance to the air. This form of screw, however, is very ingenious; as will be seen by the drawing, the pitch and diameter can be changed at will. It is, however, heavy, wasteful of power, and altogether too small for the work it has to do. The skin friction of screws in a steamship has led inventors to suppose that the same laws relate to screws running in air, but such is by no means the case. In designing a steamship, we have to make a compromise in regard to the size of the screw. If the screw is too small, an increase in diameter is, of course, an advantage, and it may also be an advantage, not only to increase the diameter, but also to reduce the pitch; however, a point is soon reached where the skin friction will more than neutralise the advantages of engaging a larger volume of water. This is because the water adheres to the surface; in fact, the skin friction of a ship and its screw consumes fully 80 per cent. of the total power of the engines, but with an air propeller its surface is not wetted and the air does not stick to its surface. If made of polished wood, the friction is so extremely small as to be almost unmeasurable. The diameter of a well-made screw running in air is therefore not limited in any degree by skin friction, as is the case with a screw running in water; in fact, it is rather a question of its weight, and its efficiency ought to increase in direct ratio with its diameter, because the area of the disc increases with the square of the diameter. The screw slip is therefore reduced by one-half by simply doubling the diameter of the screw. It will be understood that by doubling the diameter of the screw, four times as much air will be engaged. If we push this back at half the speed, we shall have the same screw thrust, because the resistance of the air is in proportion to the square of the velocity that we impart to it, so that one just balances the other, and the diminution of wasteful slip is just in proportion to the increase in diameter. In all cases, the screw should be made as large as possible.
Fig. 18.—Section of screw blades having radial edges. With screws of this form, the blades, of course, become narrower as the hub is approached, and if it is a true screw and the edges radial, the sine of the angle will be the same at all points. It is 2 inches in this case.