"Demonstration of bandwidth capabilities of beyond-horizon tropospheric radio propagation"

Abstract

Tropospheric transmission beyond the horizon has opened up many new radio possibilities. The principal application seems to be for point to point communication services at distances of 150-200 miles or more. This type of transmission is characterized by rapid and selective fading, which suggests a possible limitation on the useful bandwidth. Radar pulse tests made about 5 years ago by Bell Telephone Laboratories at 3700 mc showed that one microsecond pulses were not substantially widened after transmission over distances up to 285 miles. From these tests it was postulated that bandwidths of several megacycles might be available. A more convincing test required large antennas and a high power CW transmitter that could be modulated. Suitable equipment became available in late 1953 and wideband tests have been carried out by Bell Telephone Laboratories in cooperation with Lincoln Laboratory of Massachusetts Institute of Technology. These tests were made on a 188 statute mile path between the Bell Telephone Laboratories' site at Holmdel, New Jersey and the MIT Round Hill field station at South Dartmouth, Massachusetts. The transmitter used an experimental klystron developed by Sperry to provide a power output of about 300 watts at 5050 mc. Outside of the power amplifier, the basic radio components were from the conventional TD-2 radio relay equipment which is used in the Bell System nation-wide microwave network. 28 ft. parabolic antennas were used at both ends of the path. The receiver had a bandwidth capability of up to 30 megacycles and a noise figure of 9 db. Two types of tests have been made to determine the distortion caused by selective fading. First, a 12 voice channel multiplex system was tested for inter-channel crosstalk. In this case a 1.3 megacycle filter was inserted in the receiver IF to improve its performance. when this test showed that the distortion caused by selective fading was negligible for this bandwidth, a second test using a much wider band was proposed. Since television transmission is known to require a wide band, television tests were made using both a window pattern and television pictures. In this case also no significant impairment in system quality could be attributed to distortion in the transmission medium. This test circuit does not have adequate signal to noise margin to provide commercial reliability with the present transmitter power and antennas. Provision of a satisfactory signal to noise ratio seems to be the major problem with this type of transmission, since there appears to be no inherent bandwidth limitation up to several megacycles.