202 BELL SYSTEM TECHNICAL JOURNAL 



showed that the variations in angle of arrival of the distant signals would be 

 small compared to the beam width of 10-foot antennas. First experiments 

 were therefore made with 10-foot diameter parabolic "dish" type antennas. 

 The experimental models were made of wood with a metallized reflecting 

 surface consisting of silver conducting paint. Fairly satisfactory tolerances 

 were met in these first models, but it was anticipated that trouble would be 

 experienced in constructing a permanent metal paraboloid of that size to the 

 required tolerances without the use of a heavy and costly supporting means 

 for the parabolic sheet. It was also found that an ice coating a quarter 

 wavelength thick on the reflecting surface, when wet, acted as an effective 

 absorber of power,t since the sheet of water is resistive and is backed up by 

 the reflector. Such a condition could produce an intolerable drop in re- 

 ceived signal and would have to be prevented by providing the dish with a 

 plastic cover. As this cover should preferably house the feed also, it would 

 have presented a difficult supporting problem. 



Two electrical shortcomings of the paraboloid antenna also presented 

 themselves. First, it was found extremely difficult to obtain a satisfactory 

 impedance match between the antenna and the feed line. This was true 

 partly because of energy reflected from the dish re-entering the feed horn, 

 (this produced a constant 0.6 db standing wave ratio in the feed line), and 

 partly because of the problem of matching the feed horn itself over the de- 

 sired 400 megacycle band. Secondly, the mutual interference or "crosstalk" 

 between two paraboloids was found to be only 50 to 60 db down when placed 

 back to back** (Fig. III-l). 



A type of reflector antenna was later ijivestigated, which, although larger 

 physically than a dish having the same aperture area, overcomes the above 

 two objections.^^ It is shown sketched in Fig. III-2. The photograph of 

 Fig. III-3 shows the antenna lying on its side. It can be seen that the feed is 

 effectively "offset" and reflection back toward the feed is eliminated; the 

 experimental model of Fig. III-3 showed only 0.1 db standing wave ratio in 

 the feed line over a 10% band of frequencies. Furthermore, a horn or 

 "shielded" type feed is used which confines the energy and minimizes stray 

 radiation, and measurements indicate that the back-to-back crosstalk sup- 

 pression of two such antennas will be high. This long horn is also partly 

 responsible for the excellent impedance match. A horn having a large 

 aperture "matches" free space quite well and the slight mismatch at the 

 throat can be tuned out over a wide band of frequencies. This is not true 



t A waveguide termination in common use today employs a resistive sheet placed one- 

 quarter wavelength in front of a conducting plate; this device absorbs practically all the 

 power falling on it. 



** Back to back crosstalk suppression in the order of 125 db would be desirable for 

 repeaters receiving and transmitting on the same frequency. 



»' U, S. Patent * 2,236,393, H. T. Friis and A. C. Beck. 



