SEI.ECTIIE CIRCUITS .1X1) STATIC IXTERI-ERENCE 279 



band filter of a fnv sections can be designed to have a fisurr of merit 



iloselv approximating the ideal limiting value, , 



(wj — Ul) 



This prop<isition is merely a special case of the general principle 

 that, as regards static interference, it is useless to employ extremely 

 hijih select i\ity. The gain obtainable, as compared with only a 

 mcKlerate amount of selectivity is slight and is inherently accom- 

 panied by an increased sluggishness of the circuit. That is to say, as 

 the selectivity is increased, the time required for the signals to build up 

 is increasetl, with a reduction in (|iiality and possible signaling speed. 



Another circuit of practical interest, which has been proposed as a 

 solution of the "static" problem in radio-communication consists of a 

 .series of sharply tuned oscillation circuits, unilaterally coupled through 

 amplifiers.* This circuit is designed to receive only a single frequency 

 to which all the indi\'idual oscillation circuits are tuiu'd. Tlu' figure 

 of merit of this circuit is approximately 



^-^-^ (2«-2)! 



where n denotes the number of sections or stages, and L and K are 

 the inductance and resistance of the indi\idual oscillation circuits. 

 The outstanding fact in this formula is the slow rate of increase of 5 

 with the number of stages. For example, if the number of stages is 

 increased from 1 to 5, the figure of merit increases only by the factor 

 ."i.titi, while for a further increase in « the gain is very slow.^ This gain, 

 furthermore, is accompanied by a serious increase in the "sluggish- 

 ness" of the circuit: That is, in the partciular example cited, by an 

 increase of 5 to 1 in the time required for signals to build up to their 

 steady state. 



The analysis of a number of representative schemes, such as the 

 introduction of resistance to damp out disturbances, balancing 

 schemes designed to neutralize static without affecting the signal, 

 detuning to change the natural oscillation frequency of the circuit, 

 demodulation through several frequency stages, etc., has shown that 

 they are one and all without value in increasing the ratio of mean 

 square signal to interference current. In the light of the general 

 theory", the reason for this is clear and the limitation imposed on the 

 solution of the static problem by means of selective circuits is seen 

 to be inherent in the nature of the interference it.self. 



'See r. S. Patent No. 1173079 to .Xlexanderson. 



' When the number of stages n is fairly large, the selective figure of merit liecomes 

 proportional to /n and the building-up time to n. 



