Page:EB1922 - Volume 30.djvu/74

SUN

Position 2

Plain Glass Plate, capable ot being rotated through a measured angle

Spherical bubble

Piism

i. ' A > .' "v y

RAE SEXTANT FIG. 24.

carefully chosen to be equal also to this distance, the bubble will remain in focus and will appear to move with the sun or star if the instrument should rock in the hand.

Gyrostatic Horizons. When sextant observations are made at a ground station it is best to employ an artificial horizon, usually in the form of a bath of mercury. The sextant is then used to measure the angle between the heavenly body itself and its image seen in the reflecting surface of the mercury; half this angle is the angle of elevation of the body above the horizontal. Such a method is inapplicable to an aircraft for two reasons: first, that the vibration would cover the mercury surface with ripples and cause it to reflect a shimmer instead of a definite image ; and secondly, that the accelera- tion forces would act on the mercury and cause its surface to tilt in one direction or another. For this reason use has sometimes been made of a little gyrostat spinning on a pivot and carrying a small circular mirror fixed at right angles to the axis of rotation. If this gyrostat accurately kept its axis vertical the little mirror would form a convenient substitute for the mercury bath. But it also is subject to the disturbing effect of acceleration forces, and is thereby deflected more or less from the desired position. Its behaviour in this respect is, however, much in advance of that of a simple pendulum or bubble; although since it is a rotating body it has the double disadvantage of requiring power to drive it, and of being adversely affected in its performance by the inevitable wear of its pivot. It is still uncertain whether a sextant using a bubble or a little gyrostat will in the long run prove the more suitable for air purposes. Gyrostatic means of measurement are, however, of much importance for air navigation, and the first application on a wide scale is that of the gyrostatic " turning indicator." In this device a gyrostat is spun in bearings so that its axis lies normally in a horizontal plane. If then the frame- work containing the bearings is turned about a vertical axis due to the aircraft carrying it turning to port or starboard the gyrostat will tend to turn itself about an axis perpendicular alike to that about which the forced turn occurs, and that about which the gyrostat is itself rotating. This effect is called " precession " and the couple brought into play is called the " precessional couple"; this couple is caused either to compress or to wind up a spring and in so doing to move a pointer, the indications of which give a measure of the degree of rapidity of the turn, and whether the direction is to port or starboard. Such turning indicators are invaluable when flying in cloud, mist or fog. Without them a pilot tends to lose all sense of direction, and the indications of the compass, which might be thought a sufficient safeguard against such uncertainty, are in some cases so affected by the large and sudden acceleration forces brought into play as to be quite misleading in their indications. The reason for this will be dealt with at greater length in what follows. The gyro turning indicator was first employed for measuring the rate of roll of ships (apparatus for this purpose was made both by J. B. Henderson and H. E. Wimperis prior to the World War) and its use on aircraft came in the later stages of the war. In the meantime an aircraft turning indicator due to H. Darwin had been employed; this depended on the static air pressure at the two wing tips being communicated to a differential manometer (air-speed indicator type) and a reading being given whenever the aircraft turned, since in so doing it introduced centrifugal forces which disturbed the balance of the two pressures and so gave a plus or minus deflection of the manometer needle. The instrument works well, but needs more attention than the gyro device.

Gyrostats are also used in aircraft as azimuth indicators for experimental or test purposes; they may some day be used as part of a gyrostatic compass, but the necessary weight limit will make their introduction for this purpose a matter of some difficulty.

Magnetic Compass. The design of the magnetic compass as applied to aircraft has in late years undergone a marked improve- ment. Quite early tests showed that the compass should be a liquid one, and that to avoid the effect of engine vibrations the pivot should be above the cup. But most of the early compasses had

periodic times of oscillation about equal to those of the airplanes on which they were carried, and resonance in vibration took place, so that when the airplane rolled even a little, the compasses oscillated through considerable angles. Moreover, such short compasses gave false readings of a turn when flying on any course between N.E. and N.W. The simplest explanation of this phenomenon (first given by Keith Lucas at the Royal Aircraft Factory in 1915) is that since in these latitudes the north-seeking end of a balanced magnetic needle tends to dip downwards it is customary to add a weight to the south end in order to keep the compass card horizontal. When an airplane flying N. begins to turn to starboard this little weight is acted upon by a centrifugal force acting from E. to W. and hence tends to turn the compass card also to starboard. An ideal compass would remain pointing exactly N., and the turn of the aircraft to starboard would be noticed by the apparent motion of the lubber mark from N. towards E. around the compass card ; but if the card is also rotating in the same direction, and at perhaps a greater angular speed than the airplane, the lubber mark may appear to move towards the W., giving the false impression of a turn to port. Hence a flier unable to see the ground may infer quite wrongly that he is turning to port when he is really turning to starboard. In order, as he thinks to correct his turn, he tends still more to starboard whereas he really should have turned to port. The compass therefore fails to keep him on a straight course. Many of the earlier types of compass had this defect, but by making the compass period very much longer (as suggested by Keith Lucas), or by making the damping friction very much greater (as suggested later by Campbell & Bennett), the northerly turning error was either eliminated or greatly reduced. There is, however, a practical limit to the length of the periodic time, since if this be too great it becomes difficult to use the compass for ordinary navigation : it is too sluggish in giving its indications. This limit also concerns the highly damped or aperiodic compass, but not in the same degree. It is easier to construct a good compass by making the degree of damping approach the aperiodic than in any other way. Theory indicates that the performance of compasses is governed more by the product of undamped periodic time and the damping coefficient than by any other equally simple factor. In the early types of compass both elements entering into the product were too low; this was remedied by Keith Lucas in the one direction and by Campbell & Bennett in the other. Actually it is best to use both means subject always to the limit of not making the compass too slow in its movements.

Air Speed and Height_ Measurements. The measurements of air speed and height are linked together, since both depend on the temperature, pressure and density of the air. The usual form of air- speed indicator, first made by M. O'Gorman in 1911, makes use of the difference in the air pressure in two tubes, one of which has an open end facing the direction of motion, and the other a closed end, but with a hole in the side. In the latter the static pressure is read, and in the former the larger pressure due to the addition to the static of the kinetic effect of the air speed. A simple instance of a similar effect is seen when a plank is dipped vertically into a flowing stream; the surface facing up-stream will be wetted higher up than the one facing down-stream. The difference in height is a measure of the velocity or rather of the square of the velocity of the stream. In the case of a compressible fluid like air it also depends on its density. In fact, the reading of the air-speed indicator is proportional to the product of the density of the air, by the square of the velocity through the air. Since such instruments are always calibrated so as to read correctly at sea level, it follows that the " indicated " air speed will always be less than the true air speed at altitude. Thus an aeroplane travelling at 140 m. an hour at a height of, say, 21,000 ft. will only be credited with 100 m.p.h. on the air-speed indicator. Such indicators are therefore sometimes provided with circular cal- culators around their circumferences to enable the true air speed to be read for navigational purposes. For aerodynamic purposes such corrections are quite unnecessary since the forces due to air pressure acting on the wings, the fins, the tail and all other surfaces will also be proportional to the product of air density by the square of the speed, and an instrument like the air-speed indicator which gives a reading proportional to this product is, for this purpose, ideal and needs no correction. So that, although for purely navigational requirements it might be thought advisable to introduce a type of air-speed indicator giving true air speed, such action would be disadvantageous from the purely flying point of view. Hence it is best to retain the present instrument and to add for navigational purposes a circular calculator to effect the conversion. The case of the aneroid is not entirely parallel, but it also needs a supplementary device if the true height is to be read. Almost all altimeters in use are based on the pre-flight aneroid in which the trade convention was to assume everywhere an atmospheric temperature of 10 C. Although this is not widely out for the average surface temperature it is manifestly most incorrect at a height, since on the average the temperature falls by about 6 C. for every km. (3,281 ft.) of ascent. Thus at 7 km. (23,000 ft.) the mean temperature of the atmosphere would be about 21 below the assumed steady level of IOC.; a difference of about 7 %, leading to an over-estimate of height by the same amount. This is corrected by reading the temperature at height on a strut thermometer and using a circular calculator (the A.M.L. height computer) as in the case of the air-speed indicator