48 



BELL SYSTEM TECUMCAL JOURNAL 



In Figs. 13 and 14 are shcnvn the probable frecjuencies of Iicnvling 

 for these two transmitters as the tube length of the coupler is in- 

 creased. The shaded areas are the so-called breaking points where 

 the howling may be at either of the fre(iuencies shown. 



With these facts in mind let us review the conclusions reached by 

 Kennell>' and Up.son given in the beginning of this paper. It is seen 

 that conclusion (1) is not warranted. The transmitter and circuit 

 conditions as well as the receiver diaphragm influence the mean 

 frequency of humming. The second conclusion regarding the branches 



PROBABLE FREQUENCIES OF HOWLING 

 NORMAL TRANSMITTER 



30 40 50 60 



LENGTH OF TUBE f CMS") 



Fig. 13 



PROBABLE FREQUENCIES OF HOWLING 

 "hollow' TRANSMITTER 



30 40 50 feO 



LENGTH OF TOBE (CMS) 



Fig. 14 



30 



of the curves representing the relation between frequency and tube 

 length is correct and the explanation has just been given. This 

 periodic relation is not only true of the mean frcquenc>- line l)ut for 

 every constant frequency line. 



iZ 



V 



The terms corresponding to 4 ~r and 4 77 depend upm a number of 



factors including the circuit and end conditions. Conclusion (3) is 

 partially correct, the range of the howling frequencies dei)ending 

 upon the efficiencies of the transmitter, recei\'er, and circuit is e\ident 

 from equation (54). Calculati:)ns sh;)w that conclusion (4) is generally 

 correct although not necessarily so. 



