Abstract :
A brief description of the nature of polarization error, and the historical background of earlier work and measuring techniques, introduces the instrumental problems associated with improving the performances of U-type Adcock direction-finders, particularly on sites of poor conductivity. An Army transportable U-type Adcock direction-finder, having surface feeders and an earthing system designed for speedy installation, was tested on a site of good conductivity, and compared well with similar but more elaborate systems, indicating a standard-wave error of only a few degrees. A series of tests on ground of abnormally low conductivity, over the frequency range 2?10 Mc/s, showed that a simple counterpoise of eight radial elements could have an effectiveness comparable with that of elaborate earthing arrangements of a more permanent nature. Quantitative comparisons of the effects of various earthing arrangements (buried feeders and extended earthing plates associated with Marconi direction-finders Types DFG24/2 and DFG25) are necessarily restricted to the small frequency range 2?5 Mc/s, on account of limitations which prevent the local source from simulating plane (or natural) wave conditions and also to enable a simple workable analysis, applicable only to electrically small systems, to be used. However, this lower region of the high-frequency band is a part where reduction of polarization error is particularly required. The computed performance of a simple unscreened U-Adcock direction-finder reveals that adding a feeder screen without adequate earth return at its ends, can have a deleterious effect, but with care a fair improvement over a closed loop (reduction in standard-wave error from 35 deg to 20 deg) can be achieved on an exceptionally poor site. On a good site the performance of a U-Adcock direction-finder, even without feeder screens, is far superior to a loop. Tests on an average type of unfavourable site to be found in this country, show that the use of an e- ight-radial counterpoise, which, on average reduces polarization error by 4 to 1, results in a performance comparable with that for a very good site. Data available from polarization-error measurements on a directionfinder with a large circular earth mat refer only to limited measuring conditions and do not allow a strict comparison, but from general considerations a greater efficacy in reducing polarization error as compared with an 8-wire radial system can be expected. The present tests, though limited, reveal fundamental points which would suggest that a layer of crushed coke might confer similar benefits to an earth mat. The simplicity of the counterpoise of eight long radials is such, however, that its use is suggested for all but the best sites, so as to make performance less dependent on weather, and because it facilitates the approach of an external cable with a minimum of disturbance. Even with the suggested coke layer its use should still be beneficial.
