APPLICATION OF PRINTING TELEGRAPH 603 



TABLE I 



** N is assumed to be measured in a constant band width of about 2200 cycles 

 using the "warbler method" ^ of noise measurement. 



*** 5 is assumed to be about 5 db above 1 milliwatt where speech is at reference 

 volume. 



tion is rather to be expected inasmuch as ultimately in the system 

 there must be a relay mechanism operated by the signals. This relay 

 must with a certain degree of accuracy reproduce the length of the 

 signal impulse. It is desirable to have the relay remain unbiased over a 

 considerable range of variation in signal level. If a signal impulse is 

 transmitted only for marking, the spacing signal becomes an interval 

 of no current and the restoring force on the relay must be applied lo- 

 cally by either electrical or mechanical means. Then with signals of 

 the usual rounded wave shape, if the relay operating force varies while 

 the restoring force remains constant, the signals are either "heavy" or 

 "light," that is, the marking intervals are either lengthened or short- 

 ened and the system becomes biased.^ * 



The most obvious way of avoiding these difficulties is some arrange- 

 ment in which the restoring force on the relay is varied in a manner 

 similar to the operating force resulting from the received signal. One 

 method of accomplishing this result which has been found quite effec- 

 tive is the two-tone method of transmission. As far as the radio circuit 

 is concerned the signals consist of a marking and a spacing signal trans- 

 mitted on slightly different frequencies. Since these two signals 

 traverse the same transmission medium, they are, at least when there 

 is no selective fading, subjected to similar variations in the equivalent 

 of the transmission path. Therefore, if a polar receiving relay is 

 operated by using one of these frequencies to produce the operating 

 force and the other frequency to produce the restoring force, no bias 

 results. The increase in magnitude of variations of the transmission 

 * Details are given on this effect and methods for its measurement in reference 7. 



