DURING THE past two years in the Hydrographic Office, it has been the writer’s privilege and duty to work with the Instrument Section of the Bureau of Aeronautics in the development and testing of aircraft drift sights. Within this period various aircraft navigational instruments have been flight-tested, including six types of drift sights, and one ground-speed meter. Since drift sights, ground-speed meters, and drift determination in general is a highly controversial subject, the results of these tests are offered as personal opinions.
Surface navigators have been trying for centuries to develop some simplified means of measuring ground speed and drift to use in dead-reckoning navigation. Even today, the navigator of a modern battleship with thousands of dollars of modern navigational equipment may find an error in dead reckoning comparable with that of his contemporary in the air. No one denies the desirability of a ground-speed and drift meter as it would certainly simplify dead-reckoning navigation both for the surface and air navigator.
Several attempts have been made in recent years to develop such a ground-speed meter for aircraft, notably by the Bureau of Standards and by Mr. Harold Gatty. There is no doubt but that such indicators are or can be made mechanically perfect, but unfortunately they are predicated to true altitude. These instruments solve ground speed by developing a series of similar triangles between the observer’s eye-—the instrument’s reference plane— and the surface of the earth. This is accomplished by the observer sighting through a moving grid, and varying the height of eye or the speed of the grid until its movement is seen to synchronize with objects on the surface. A study of similar triangles will soon show that an error of 10 per cent in the altitude will be an error of 10 per cent in the ground speed.
Such ground-speed meters will offer great possibilities in air navigation as soon as some practicable method is developed for measuring true altitude. We know the present-day altimeter is nothing more than a barometer indicating height above a given datum plane in accordance with a computed “standard atmosphere,” which is seldom if ever encountered under actual conditions. It is safe to say that the altimeter, with all of its errors of lag, temperature, seasoning, hysteresis, etc., is more accurate than the calibrated yardstick (standard atmosphere) accepted as the measure of height above sea level. The fact that the altimeter is a sensitive type does not make it more accurate, as such an altimeter has only an open scale in comparison with the non-sensitive type.
Disregarding ground-speed meters, we find two means remaining for the air navigator to determine the plane’s track and speed over the earth’s surface. One is that of taking the wind direction and velocity as given by the aerologist and applying this wind vector to the plane’s course and air speed, which will give plane’s heading and ground speed. This method gives excellent results, but unfortunately is unreliable as the plane leaves the vicinity of the aerologist’s observation point. The other means of wind determination is accomplished by observing two drift angles, on courses approximately 90 degrees apart at a known constant air speed and solving the wind star problem for the force and direction of the wind. This is the only reliable method that can be used under all circumstances where the surface can be seen and is the standard method of the Navy.
The question is now introduced: What is the best type of drift sight? There are two general classifications of drift sights. One is the rotatable grid type, such as the Army type D-I-A, and the other, the relative bearing type, such as the Navy Mark II. Each type possesses distinct advantages under certain conditions, which will be outlined in the following paragraphs.
The principle of the first type is a rotatable grid of parallel lines set to read zero when parallel to the center line of the plane. The grid is rotated until objects on the surface, seen through this grid, run parallel to the reference lines. The movement of the grid to the right or left is read off an angular scale as the drift angle. The grid-drift sight can be used over land continuously with excellent results, except in very rough air when the plane is yawing and rolling about its athwartships axis. This type has a distinct advantage over the relative bearing sight, in that there is only a slight error in the drift angle if the plane is off course at the instant of drift observation. The above fact will be apparent to the reader if he will analyze the wind triangle and note that being off course causes an error in the wind angle instead of an error in the drift angle reading, as is the case with the relative drift sight.
It is not such a simple matter to use this same grid drift sight over the water as there are no fixed objects on the surface. Those navigators who use whitecaps for reference points are likely sooner or later to get some surprisingly inaccurate results.
The fact that waves move over the surface of the sea at a velocity that sometimes exceeds that of the wind has been definitely proved by various authorities on wave motion. Whitecaps form on the crest of waves and travel with that wave until they break and form sea foam. The length of time that any one whitecap travels with a wave depends upon the existing state of the sea. The estimated average time of travel would be 5 seconds for a moderate sea condition. There are so many factors involved in the speed of waves, such as (a) duration of the wind, (b) force of the wind, (c) extent of open water, that it is not accurate to use whitecaps as a reference point for observing aerial drift with any type of drift sight.
There is considerable doubt if even sea foam should be used as a reference point for a drift sight. Various experiments were conducted to determine if there was any movement of sea surfaces.
An analine dye bomb and an aluminum powder bomb were lashed together and dropped from an altitude of 500 feet. The analine dye colored the water for a depth of several feet and served as a point of reference while the aluminum powder floated on the surface. The column of colored water remains fixed while the aluminum powder floated down wind approximately 100 yards in 9 minutes. Under the existing conditions of light winds, the surface movement was estimated at less than one knot. Other experiments with anchored smoke bombs and aluminum powder gave similar results. However, it is believed that surface movement can be neglected for the purpose of measuring aerial drift, except in high surface winds and when flying at high altitudes where it is difficult to distinguish sea foam from whitecaps.
There is yet another factor to consider if accurate drift angles are to be measured and that is the sea movement due to tides and currents. In certain areas the velocity and direction of currents can be predicted and a correction factor made for the drift angle depending upon the plane’s speed and track over the water.
Summing up the advantages for the grid-type drift sight, we find that it is probably the best type for over-land navigation, but considering that it cannot be used over water at night, at high altitudes, over glassy water, or over rough seas, it is doubtful if this type has any value for over-sea navigation.
The relative bearing type is simply a pelorus used to measure the relative angle between the center line of the plane and an object astern. This type of drift sight can be used with fair success over land, for the drift angle will be as accurate as the angle measured between the plane’s heading and any object ahead, or astern of the plane. It does require close coordination between the pilot and the drift observer as an off-course error of 2 degrees introduces the same 2-degree error in the drift angle.
A serious fault with the present type of relative drift sight is that it is mounted on the side of the plane, requiring the observer to stick his head out into the wind blast to take drift readings. This feat is rapidly becoming impossible with the increased speed of planes. The solution to this problem will probably be a wind shield around the sight or an optical drift sight inside the cockpit. Many types of optical sights have been developed and tested, but to date it has been impracticable to use them with float lights due to the extreme difficulty of locating the drift float in the water.
In observing drift angles over the water, it is necessary to drop a drift float that will produce smoke by day and light by night, to serve as a reference point on the water. Then the relative bearing can best be determined by taking at least three and preferably five different readings, averaging the results.
In addition to the two above methods of observing drift there is yet a third way. This method, it is believed, was first used by Captain A. C. Read, U. S. Navy, on his transatlantic flight in the NC-4 and is one method now used by the Pan-American Airways on their over-seas flights. Each Pan-American Airway boat has a Navy Mark II sight mounted in the cockpit. The sight is inverted with the base piece fixed to the top of the cockpit, above and forward of the pilot’s head. The sight is pointed forward and downward at an angle of approximately 70 degrees. The observer looks through the barrel of the sight and rotates it until the “speed lines” over the water are parallel to the fixed marker inside the barrel. The drift angle is then read off the base scale in degrees right or left.
What the explanation of these “speed lines” is, the writer has never been able to determine. It is not whitecaps, sea foam, wind streaks, or surface movement of water. Perhaps the eye integrates the relative movement of the plane to the right or left, and when there is no drift the speed lines are parallel to the plane’s heading. At any rate the speed lines are definitely there on the surface.
Discussing this method of obtaining drift with Pan-American Airways’ pilots, the writer found that they all agree that best results are obtained at 1,000 feet altitude or less, but accurate drift could be obtained up to 3,000 feet. Furthermore, that some training with the sight was necessary before being able to use it in this manner.
This method of observing drift certainly deserves a fair service test to determine its possibilities. Otherwise, we shall be forced to use some type of relative drift sight to determine the drift factor in aerial navigation. It is hoped that an entirely new approach to this project will be developed.