(/.(/.V OF IliOir.iy TEI.EI'IIOM- KIU'F.AIERS 25 



(if piilMS li.i\ini; .1 fr»«nuiuy im|ii.i1 Id tli.il nf tlu- lonlrollinj; currrnl 

 and \.ir\iiii; in in.i^nitiKli- .urordinj; to llic shii|H- of tin- voicr w.ixi- 

 litinj; iransmiltid. riii-sc piilsi-s will W pailialK rrtlrctiil at tin- 

 irrt-nularity B^ and pari of •heir t>nerg\- will ritiirn to tlu- rt(KMtiT. 

 Dur to ihc fad that a tinilt' time is ritiuirud for ihu pulses to pass 

 from R to Bi and bark, tlun' an- likily to arri\i' at tliu riKlU in<jmi'nt 

 to fnul the anipliliLT set for ainplit'iration in the opposite dirertiort, 

 in which case they will pass through towards .-1. l"or a single irrcg- 

 ularit>-, it wonkl !«? possible to select a freciuency such that the pulse 

 would return when the repeater is set against it, but this would re- 

 quire a dilTcrent fre(iuency for each irri'i;ul,iril\ which is obviously 

 impossible. 



In case the line cannot transmit the high fretiuency pulses, their 

 energy would be stored in the inductance or capacit>- of the first ele- 

 ments of the line L^ and returned to the amplifier when it is in condi- 

 tion to transmit from Li to L\. To a\oid the latter objection it has 

 lx!en proposed to employ low pass filters on the output side of each 

 one-way amplifier to convert the high frecjuency pulses back into 

 ordinary voice waves before passing thcni into the line, but this 

 obviously defeats the object sought in using the high frequency 

 control of the amplification because each amplifier now receives 

 ordinary voice waves and gives out enlarged copies of them which 

 are subject to the same reflections as if plain one-way amplifiers 

 without the high frequency control had been used. 



From these considerations it will readily be seen that repeater 

 systems depending upon high frequency variation of the gain to 

 avoid singing and the necessitN- for impedance balances are inherently 

 unworkable. 



