FREQUENCY SHIFT TELEGRAPHY 



299 



the figure for each curve. The signal- to-noi^e level scales refer to the highest 

 level portion of the diversity signal. Since the amplitude modulator which 

 was used to simulate the selective fading did not produce phase shifts or 



■20 -16 



-12 -8 -4 4 8 )2 



LEVEL RATIO (A/b) "^ DECIBELS 



20 



Fig. 36. — Diversity combination characteristic obtained by series addition of discrimi- 

 nator outputs — levels measured at output of 400 kc I.F. amplifier. 



30 



20 



■20 



-30 



-16 -12 -8-4 4 8 12 16 



LEVEL RATIO (^/q) IN DECIBELS 



Fig. 37. — Diversity combination characteristic obtained by parallel addition of dis- 

 criminator outputs — levels measured at output of 400 kc I.F. amplifier. 



transmission delays as would actually exist over H.F. radio circuits the actual 

 distortions shown by the curves are optimistic. 



The ideal diversity selection circuit should theoretically give a signal- 

 to-noise characteristic identical to that of a single channel under the signal- 

 to-noise condition corresponding to the signal of the best momentary re- 

 ception. It will be seen that the test results of Fig. 38 approach this limit 

 within 2 or 3 db at a peak distortion of 20%. Part of this difference is 



