BELL SYSTEM TECHNICAL JOURNAL 

 I / (inH! (uMm me^\ io^V^ \ / , ^, /-x 



In which 



flo = 1 + -^ + — g^ . ai = Mg + -^ + — g^ 



flo = ^ + ^, 6o = 1 + 3.1/2g2^ ^ (7a) 



Comparison Between Detectors 



It is now possible to make a comparison between the performance 

 of the straight Hne and the square law detectors. In Figs. 1 to 4 are 

 shown the relative amplitudes of the interfering frequencies in the two 

 cases for various degrees of modulation. The data for the square law 

 case are indicated by dashed lines and for the straight line case by 

 solid lines, and where the two coincide this is noted on the figures. It 

 is to be noted that the expression for the amplitude of the desired 

 frequency P/2t is a complicated function. However, computation 

 shows that over the range in which we are interested, the value of this 

 expression does not differ from ME by more than 1 per cent and, 

 therefore, this value has been assumed in computing the relative 

 amplitudes of the other frequencies. 



Probably the most striking feature to be noted in comparing the two 

 cases is the similarity of the results. This is particularly evidenced by 

 the carrier beat note of frequency u/2t the amplitude of which differs 

 in the two cases by an inappreciable amount. The spurious fre- 

 quencies (q ± n)/2Tr also are practically identical for both detectors. 

 There are, howe\er, several important differences as follows: 



The group of spurious frequencies of angular velocity p ± q ± u, 

 which is of appreciable importance in the square law case, is entirely 

 absent from the range of magnitude considered when a straight line 

 detector is employed. The frequencies {p ± u)l2Tr are greater in the 

 square law case over the range which we have considered, but the 

 curve which represents them has a smaller slope than in the straight 

 line case and for larger values of the interfering signal the intensities of 

 these frequencies would be relatively less with the square law detector. 

 The intensity of the undesired speech q is definitely less in the straight 

 line case than in the square law case but the slope of the q curves is 



