Air Acoustics: Listening device model Van Soest (1929 – )
The negative experiences that Van Soest and his assistant Groot had with commercial listening devices led to their development of a Dutch listening device in 1929. The initial design was with paraboloids of plaster and paper. That version was soon replaced by a vertically balanced silumin version (5 mm thickness; reduced later to 3 mm). Earlier investigations (described above) indicated that the distance between the two ears of a listener was sufficient as the basis for a listening device. The directional hearing of the unarmed ears, after all, was much more accurate than the disturbance of this directional hearing by atmospheric noise. The listening device had to be light in weight so that a bearing for deriving the map angle and elevation could work silently.
Van Soest designed a device with two half parabolas which were sealed by flat plates separated at a mutual distance of 16 centimetres with 8-centimetre holes at the focal points. Bearings for the elevation setting and inflatable earpads were fitted around the holes in a way that the head placed between them could not cause any disturbing noise.
This construction was merged with a vertical column to which a chair for the listener operator was mounted. The adjustment of the map angle in the horizontal plane took place by rotation around the vertical column. This rotation was realised by the muscle strength of the arms and legs. Great care was devoted to the easy and silent operation of the bearing without slack or dead passage. The best listener thus achieved an accuracy of less than one degree in the map angle using the time difference in arrival between the two received signals.
After eliminating all disturbing mechanical noises when calibrating the listening device, only a single disturbing sound source remained. The almost always present wind caused noise due to turbulence around parts of the listening device. This noise was reduced by using a listening cage. The cage was built from a concrete-iron frame. A thin, more or less transparent, jute cloth was pulled over the frame. Without this cage, the device could not be used with a wind speed of 4 metres per second. With the use of the cage, listening was still possible at 6 metres per second.
A second advantage of the cage was that the listener was not distracted by movements in his field of view. Much to the disappointment of the Measurements Building, the military authorities did not support the further development of the cage, apparently in the belief that the cage was too impractical.
The directional hearing in the map angle was mainly determined by reducing the time difference of arrival of the sound in the left and right ear to zero by adjusting the angle of the ears of the hearing shells. The elevation was set so that a maximum intensity of the sound occurred. This difference was accentuated more sharply by the shadow effect of the partitioning divisions in the parabolic halves.
On February 18 and 20, and March 4 1930, various comparative measurements were made with the Van Soest listening device and the three foreign industrial devices: the Goerz listening device (half), the Doppelt Richtungshörer device, and the Barbier, Benard and Turenne device.
The Goerz device turned out to have a second systematic and not negligible error: an “acoustic parallax” between the visual and the acoustic direction [at an aircraft speed v m/s and angle alpha between the axis of view and flight direction, the error is v/330 *sin (alpha) radians; with an alpha of 45 degrees and 40 m/s the error is 85 milliradians. With the earlier described vertical error, this results in 1.6 times 85 milliradians= 8 degrees. With an alpha of 90 degrees and v= 50 m/s this even becomes 14 degrees].
A conclusion about the Barbier, Bénard and Turenne listening device was that it is hard to operate, is bulky and turns slowly. The Doppelt Richtinghörer has a low sound reception; listening with a stethoscope raised objections and the sound tubes had to be kept at a constant length.
At the same time, a Reisz microphone, amplifier and headphones were used to observe the aircraft. An optical distance meter was used to determine the distance within which a listening device worked properly.
Every listening post and the post at the microphone could press a switch as soon as there was an observation. The switch was connected to a light in a row of lights. Via pinholes, the lights could illuminate a steadily moving film (‘chronograph’). Every 10 seconds a time signal was turned on or off allowing for a 1/10 second time accuracy of the recorded signals on the film. The chronograph allowed for half an hour of recordings.
The Groot device (a.k.a. Van Soest device) had the best test results. Easily movable, this device detected an aircraft 5 km away (Barbier: 3 km; Doppelt Richtunghörer 1.5 km).
Very striking in these tests was that the listener in the Van Soest device had a clear impression that the aircraft noise moved from left to right or from right to left in the back of his head with an accent on the centre perception, while this effect was very weak using the other devices under test.
During one of the experiments, the direct sound of the aircraft was weaker than the low-hanging clouds reflected sound: a mirror sound image was observed.
The confidence in the Dutch listening device model Van Soest was great. The costs were also relatively low: listening device Dfl. 1,000, optional corrector Dfl 600, two-step system transmitters Dfl 300 and two-step system receivers Dfl 700 (estimate 12/1932 by the Commission for Physical Armourment). Therefore, the Sappers of the Army commissioned the industrial production of the listening device.