Page:EB1922 - Volume 32.djvu/266

248

and No. 8948 of 1918: sec also Phil. Trans. Roy. Soc., A, vol. 221, p. 389). This depends on the principle utilized in the hot-wire anemometer, i.e. the change of electrical resistance consequent on the change of temperature of a heated wire which ensues when the air round it is set in motion. A very fine wire of platinum, whose resistance at atmospheric temperature approaches 100 ohms, is mounted in the form of a grid over a circular hole some 7 mm. in diameter (fig. 8). It is provided with terminals. It is included in one arm of a Wheatstone bridge, and sufficient current passed through the network to heat the wire to a dull red. The bridge is balanced so that when the air round the wire is undisturbed no current passes through the galvanometer. Motion of the air causes the resistance of the wire to decrease, upsets the balance, and so causes a current to pass through the galvanometer.

The microphone wire is mounted in front of an air container of some 16 litres capacity. Resonance effects in this container may be partly eliminated by small openings made in the wall. The instru- ment so completed is insensitive to all sounds of speech, musical sounds, traffic, or even rifle fire. It responds readily, however, to gun sounds (which arc low frequency disturbances), even when they are inaudible, and records also the shell-wave. Its reaction is very rapid, and the small lag which does occur appears to be the same for all similarly constructed instruments.

An essential part of the recording instrument is a galvanometer for each circuit which shall respond very rapidly to the current caused by a sound reaching the microphone. For rapid response it is necessary that the moving part of the galvanometer shall be very light indeed the movinjj coil or moving magnet type is out of the question. The condition is satisfied by the Eirithoven galvanometer, in which the moving part is a very fine wire (through which flows the current to be detected) mounted in a magnetic field. The wire moves in a direction normal to its length and to the direction of the field. Six wires, insulated from one another, and provided with separate terminals, can be mounted side by side in the field pro- duced by a single small electromagnet. This provides in a small space what is essentially six independent galvanometers, one of which is included in the bridge belonging to each microphone. As the sound reaches successively different microphone posts the corre- sponding galvanometer wires move in rapid response.

The instant at which each wire begins to move is registered on a moving photographic film. The camera in which the film runs vertically is furnished with a horizontal slit, a cylindrical lens in front of the slit reducing its effective breadth. Shadows of the per- pendicular galvanometer wires, cast by means of an electric lamp and an optical svstem mounted in the pierced poles of the galvanom- eter magnet, fall on the slit, and are focusscd on the film, appearing there as six points of shadow on a horizontal line of light. A_s long as the wires are still each point leaves on the running film a straight line; the movement of a wire registers itself as a break in this trace.

If the film ran at a uniform speed measurement on the developed film of the distance between the breaks would give the required time intervals. As this is not the case the following device is adopted: a wheel provided with ten flat spokes, one of which is somewhat wider than the others, is mounted in the case containing the lamp, so that, when it is rotated, the spokes successively interrupt the light which illuminates the galvanometer wires. The wheel is actuated by a synchronous motor controlled by a tuning fork, and rotates ten times a second. As a result of this arrangement there appear on the film lines perpendicular to the direction of the motion, the intervals between which correspond to hundredths of a second, every tenth of a second being marked by a wider line. This recording apparatus was devised by Dr. Lucien Bull, of the Institut Marey, near Paris.

Originally the film was cut off after the required record had been taken, and developed in a small dark room adjacent to the instru- ment. Later a method of automatic development was devised, by which the film passed successively through developer and fixer while running, and emerged ready for interpretation.

Fig. 9 shows some typical records, (a) and (b) are records of two differently situated 5'9-in. howitzers taken by six posts in each case. The burst of the shell was also registered on these films, but as it occurs several seconds later space docs not permit the inclusion of the part of the record in question, (c) is a record of a field gun, showing both shell-wave and gun-wave. Only five posts were used for this record. The varying interval between the two sounds at the different microphones is well shown: at the flank microphone, corresponding to the lowest trace, only one sound is heard, (d) is a record of the burst of a British shell on a German position.

Influence of Weather Conditions. The method in use demands that to every time interval shall correspond an exact distance, a standard velocity of sound being assumed, which corresponds to some standard temperature, and still air. (The velocity of sound does not, of course, vary with the pressure, and the effect of humidity is in general negligible.) Hence the time interval read off from the film has to be corrected for temperature and wind before it is used on the board prepared for location. For the temperature variations which occur in ordinary circumstances the increase of velocity of sound may be taken as proportional to the increase in temperature, so that the temperature correction is easily applied. Simple geometrical con- siderations show that the correction for wind depends only on the velocity and direction of the wind and the position of the microphone, and not at all on the position of the gun. With given microphone positions a diagram can be prepared which allows the rapid graphical determination of the correction for a known wind.

It has been found by experiment that the temperature and wind which are concerned in these corrections are not those prevailing at ground level, but at a height of between 250 and 500 ft. up.

Owing to the refraction of sound by wind the record of a given sound at ground level is greatly influenced by the variations of wind velocity at different heights above the ground. This wind gradient determines largely whether the conditions are favourable or un- favourable for the detection of sounds. In the case of a wind increas- ing in velocity with height, a following wind, besides increasing the velocity of the sound, tilts the wave front so that the sound con- verges on the listener or instrument on the ground, and is well heard. An opposing wind causes the sound to tend to pass up- wards and leave the ground. Hence a wind blowing from the instru- ments towards the hostile piece often renders sound-ranging almost impossible if it be of any strength. The temperature gradient also plays a part in the refraction of sound.

(C) -

(d)

FIG. 9. Typical records of Bull apparatus.

Location from Record. -Having seen how the intervals between the arrival of thr sound at different posts can be accurately obtained and corrected to standard conditions it remains to discuss how these intervals can be made to supply the position of the gun with as little delay as possible. A map board is prepared with an accurate " grid " (coordinate system of squares) covering the region in which loca- tions are expected. On tnis the microphone positions are accurately marked. The posts are usually taken consecutively in pairs; with each pair as foci a family of hyperbolae may be drawn giving the loci corresponding to various time intervals (at standard velocity of sound). In practice, however, to avoid the labour of preparing the hyperbolae it is usually preferred to use the asymptotes instead of the curves themselves: for these it is only necessary to have a thread attached to each mid-point between pairs of consecutive posts, and a scale plotted round the edge of the board for each base, graduated in time intervals, so that when the thread is placed to pass through a given graduation it is the asymptote to the hyperbola corresponding to the interval. To allow for the divergence of the asymptote from the hyperbola, which becomes serious as the base is approached, tables are prepared giving the corrections (always additive), in terms of the length of the base and the distance from the mid-point of the base, to be applied to the time intervals obtained from the record. The asymptote corrections having been applied to the various intervals, already corrected for temperature and wind, the asymptote corresponding is laid out for each base. The various lines should all intersect at a point : in general they do not, but form a small polygon from which the position of the gun can be estimated.

Estimation of Calibre. The position of the hostile burst may be obtained from the record of its sound in the same way as the position of the piece, and the interval between the departure and burst of the shell, i.e. the time of flight, can be computed from the record on one microphone. Thus the record gives the time of flight corresponding to a given range, which affords an indication of the calibre of the piece. In the case of guns, as distinct from howitzers, a further indi-