690 



NATURE 



192 



Radio Direction Finding by Reception. 



''I'llKKh lire in use to-day three prindpiil sy.stcniH 

 JL of direction finding by which the apparent 

 direction of arrival of a train of electromagnetic waves 

 can be observed and, under suitable conditions, the 

 direction of a radio transmitting station d ' ! 



These are usually described as the Bellini-'I' n, 



the single frame system, and the Robinson system. 

 The Bellini-Tosi system has been very fully developed 

 by the Marconi Company for use on land and on board 

 ship as an aid to navigation, and is the system most 

 usually employed in this country for that purpose. 

 The coil frame system has received most attention in 

 the United States, and has there been the subject of a 

 great amount of research work. The Robinson system, 

 for reasons which will appear later, is specially suitable 

 for use in connexion with direction finding in the air 

 and has mainly been developed with that end in 

 view. All three systems have been for the most 

 part the subject of independent development and 

 their several merits have been the subject of con- 

 siderable controversy. 



In general, however, all the systems operate upon 

 the principle that the magnitude of the electromotive 

 force induced in a vertical loop or coil of wire by an 

 electromagnetic wave depends upon the angle between 

 the plane of the loop and the wave front of the arriving 

 wave. An electromagnetic wave can be considered 

 as consisting of electric and magnetic forces which are 

 at right angles to each other and to the direction of 

 travel of the wave. These two force vectors are in 

 phase with each other and each varies rapidly in a 

 periodic manner. The effect on a wire placed in the 

 field due to such a wave can be deduced from con- 

 sideration of the effect of either the electric or the 

 magnetic fields in the wave front. In the case of a 

 single coil vertical loop of wire it can be shown that 

 the periodic magnetic field due to an electromagnetic 

 wave the wave front of which is plane, though not 

 necessarily vertical, introduces in the loop an alter- 

 nating electromotive force the maximum value of 

 which is given by the following expression : — 



E,„ = wH^A cos a 



where w is the periodicity of the wave, H,„ the maxi- 

 mum value of the horizontal component of the magnetic 

 jBeld in the wave front, A the area of the loop, and 

 90° - a the angle between the plane of the loop and the 

 horizontal component of the magnetic field. The 

 effect of the vertical component of the magnetic field 

 may be disregarded since the plane of the loop is 

 vertical and therefore cannot be linked by this com- 

 ponent. If the loop is rotated so that the electro- 

 motive force becomes zero the plane of the loop is then 

 parallel to the horizontal component of the magnetic 

 field, and the direction whence the waves are travelling 

 is thus at right angles to the plane of the loop. The 

 direction of arrival of the waves can therefore be 

 determined with an ambiguity of 180°. It can be 

 shown that, in general particulars, the underlying 

 principles of all the three systems in use to-day are the 

 same — and that the systems are in their action 

 essentially equivalent to the single rotating loop. 

 The single coil system most closely approximates to 



NO. 2819, VOL. I 12] 



the .simi)le thcortiif a! « ase. .\ 

 ever, usually inlroduced in 



which the amplifying and dtticujig upj>araiui ii con- 

 nected ; but it can Ik? shown that the potential 

 difference across this condenser depends on the 

 tion of the frame in the same manner as the . 

 motive force induced in the frame. The single coil. ;is 

 used in practice, consists, as a rule, of several turns u\ 

 wire instead of a single turn. These turns are usually 

 spaced in a series of equally dimensioned loops in 

 nearly parallel planes (box type coil), or are wound 

 spirally in the same plane (pancake type coil). In the 

 case of a pancake coil the quantity A in the expression 

 given above is replaced by the effective or mean area 

 of the coil. In the case of box type coil, sinre the 

 winding of the coil is, of necessity, slightly a 

 the axis of the coil, there is the possibility of an « ; 

 turn of wire being introduced the plane of which 1- 

 practically at right angles to the main turns of tl « 

 coil ; the effect due to such a turn, however, introdu< t - 

 an error not exceeding o'i°, which is p< 'li'ii.i, f,.r 

 ordinar}' purposes. 



The connexion of the tuning condenM r ai.u i« « ri\ n - 

 apparatus to the coil introduces certain disturbing 

 effects. First the electromotive ftirce picked up by 

 the leads and the whole circuit, although small com- 

 pared with the maximum value of the potential 

 difference across the tuning condenser, may be suffi- 

 ciently strong to give an audible signal when the coil 

 is orientated so as to make this potential difference 

 zero {i.e. cos a = o). In this way an ill-defined minimum 

 may be produced, and accurate determination of the 

 bearing made more difficult. Secondly, a bad zero 

 may be produced by what is known as " vertical " or 

 " antenna " effect. One side of the tuning condenser 

 is connected to the grid of the first receiving valve, 

 while the other side is connected to the filaments of 

 the valves, to which in turn are connected the filament 

 and anode batteries. The capacity to earth of the 

 two sides of the coil joined to the tuning condenser 

 will, therefore, in general be unequal, and a potential 

 difference will be produced across this condenser even 

 when there is no circulating current in the coil. One 

 result is a blurring of the minimum, and another is 

 that the two minimum positions are found on rotating 

 the coil not to be exactly 180° apart, owing to the fact 

 that the superimposed potentials due to the antenna 

 action of the coil are nondirectional. These effects 

 can be eliminated, however, by the introduction of a 

 small variable balancing condenser between the grid 

 of the first valve and the tuning condenser. 



In the Bellini-Tosi system two large rectangular or 

 triangular loops each of a single turn are erected with 

 their planes at right angles. To each of these a small 

 field coil is connected in series. These small coils are 

 again mounted with their planes at right angles and 

 between them is pivoted a small search coil, attached 

 to which is a pointer which moves across a horizontal 

 circular scale divided into degrees from which the 

 observed bearings are read off. 



The two field coils reproduce in miniature the field 

 in which the main coils are placed, and the search coil 

 turning within the small field coils is equivalent to 



