334 BELL SYSTEM TECHNICAL JOURNAL 



frequency difference of the desired and interfering carriers is only a 

 few cycles per second. As this difference is increased the flutter is 

 transformed into a more sustained sound, rather harsh in character, and 

 as it is still further increased a low growl appears which becomes more 

 objectionable as it rises in frequency. The pitch of this growl cannot 

 exceed 100 cycles unless one or both stations are violating the 50 cycle 

 regulation. With the increasing use of very precise frequency control, 

 heterodyne frequencies of a few cycles have become very common, and 

 so, therefore, have flutter effects. 



It has been pointed out in an earlier paper ^ that the magnitude of 

 the flutter effect will depend upon the type of rectifier employed in the 

 receiving set, and that it will be very much more objectionable when a 

 square law detector is used than when a linear detector is employed. 

 This is to be expected, since in the former case the audio-frequency 

 output of the receiver will be proportional to the amplitude of the in- 

 coming carrier, while in the latter case the output will be essentially 

 independent of the carrier amplitude, provided over-modulation does 

 not occur. However, these statements refer to the case in which 

 automatic gain control is not used. When the receiver is equipped 

 with automatic control, as in most better grade modern receivers, the 

 superiority of the linear detector is nullified and a serious flutter may 

 occur. 



In addition to displaced sideband interference and flutter, trouble 

 may arise from distortion of the desired program by the action of the 

 interfering carrier. One or both of the first two types of interference 

 are likely to occur at lower field strength ratios than is the last, but at 

 higher levels of the undesired carrier all three types are of importance 

 and combine to degrade the quality of reception. In this paper, 

 studies of all these types will be reported. Audible beat interference 

 will not be discussed since it has been considered in other papers and, 

 as just mentioned, is much less important than it used to be. 



Theoretical Estimation of Flutter Effects 



As has already been stated, the flutter effect is due to the rise and 

 fall of the level of the noise background with variation in the effective 

 amplitude of the impressed carrier. In order to study this effect, let 

 us suppose that there are impressed upon the detector a component of 

 radio frequency noise which may be represented by iVcos (co -{- n)t, and 

 a desired carrier E cos w/. nj2-K is assumed to be an audio-frequency. 



If a square law, or quadratic, detector is employed, the audio- 



^" Theory of the Detection of Two Modulated Waves by a Linear Rectifier." 

 Proc. 1. R. E., Vol. 21, pp. 601-629, April, 1933. 



