426 



REGULATORY CIRCUITS 



at o) = 20 in order to provide sufficient attenuation of amplitude noise 

 fluctuations. This requirement, in connection with the static gain require- 

 ment, fixes the location of the low-frequency corner of the AGC filter at 

 4 rad/sec. The AGC loop transfer characteristic thus developed is shown 

 in Fig. 8-20. 



In order to maintain the fidelity of transmitted modulation, a network 

 will be introduced into the transfer function to provide a null at the lobing 

 frequency as described in Paragraph 8-19. It will be supposed that the 

 phase shift must be kept less than 1.5° and the lobing frequency, its 

 regulation, and the angle tracking bandwidth have the values assumed 

 in the example in Paragraph 8-18. 



Fig. 8-21 shows a parallel-T net- 

 work which can be used to provide 

 the required null.^' The transfer 

 function of this network is 



mR 



Fig. 8-21 Parallel-T Null Network with 

 Symmetry Pattern m. 



Voltage transfer function of 

 network 



u'^ 1 



H- + // 



(-^) 



+ 1 



(8-18) 



where u = jcoRC = jco /coc 



CO = angular frequency, rad/sec 

 Wc = null location, rad/sec 



m = symmetry parameter determining null sharpness. 



The effect of this network with a null of 50 cps = 314 rad /sec and a value 

 of m = 0.54 is shown in Fig. 8-20 in combination with the single time 

 constant RC filter. The value of m (Fig. 8-21) was selected to provide a 

 total attenuation of 31.9 db (as required in Paragraph 8-19) at 292 rad/sec, 

 the lowest possible modulation frequency. The high-frequency gain margin 

 should be checked, particularly at one-half the PRF. An inspection of 

 Fig, 8-20 shows that a minimum gain margin of 20 db, in comparison with 

 the 6-db requirement, is maintained at all high frequencies. 



The closed-loop response of the AGC loop indicates most directly the 

 dynamic AGC action in attenuating low-frequency amplitude noise and 

 transmitting modulation frequencies. Fig. 8-22 shows the closed-loop 

 response corresponding to the trial design illustrated in Fig. 8-20. 



"C. F. White, Tmns/er Characteristics of a Bridged Paratlel-T Network, NRL Report R-3I67, 

 27 August 1947. 



