ATMOSPHERIC ATTENUATION AT MILLIMETER WAVELENGTHS 



909 



The accuracy of the measurements Avill be affected, of course, by spuri- 

 ous refiections in the neighborhood of the corner-refiectors. The sites for 

 the experiment were chosen to minimize such refiections and checks were 

 made by observing the decrease in the return signals when the corner- 

 refiectors were covered by absorbent material. In all cases, the back- 

 ground reflections were at least 30 db below the signal from the corner- 

 reflector. 



The method of measuring the reflected signals is illustrated in Fig. 2. 

 The transmitted signal is frequency modulated in a saw tooth manner 

 with a small total frequency excursion, F. The signal reflected from the 

 near corner-reflector is delayed \ni\\ respect to the transmitted signal 

 by a time, n , equal to twice the distance to the reflector divided by the 

 velocity of light. During a portion, Ti — rx , of the sa^^i:ooth cycle, there 

 is a constant frequency difference, /, between the transmitted and re- 

 ceived signals, {f/F = ti/Ti). Power at this frequency is produced by 

 mixing the initial source signal with the delayed received signal and am- 

 plifying the difference frequency in a narrow-band amplifier centered at 

 frequency /. The output of this amplifier is, therefore, a pulse at fre- 

 quency /, of length Ti — n and repetition rate 1/Ti . 



To measure the signal returned from the far corner-reflector it is neces- 

 sary merely to increase the period of the sawtooth modulation propor- 

 tionate to the increase in distance. The frequency excursion, F, re- 

 mains the same; hence the average power output of the transmitter is 

 unchanged. As may be seen in Fig. 2, the freciuency difi"erence, /, between 

 the transmitted and received signals is unchanged; thus the same am- 

 plifier and output meter can be used for the two cases. Another advan- 

 tage in changing only the sawtooth repetition rate is that the delay is 

 the same fraction of a period in both cases; therefore the duty cycle is 

 unchanged and the intermediate frequency pulses can be detected by 

 either an average or a peak measuring device. 



Since the beat frequency, /, is not affected by slow changes in the fre- 



ANTENNA 

 EFFECTIVE AREA 



CORNER REFLECTOR 



R1 

 EFFECTIVE AREA At 



SOURCE 



\>- 



-d,- 



HI 



_i_ 



CORNER REFLECTOR 



R2 



EFFECTIVE AREA Ag 



Fig. 1 — Siting arrangement for the atmospheric absorption measurements. 



