NEGATIVE VOLTAGE FEEDBACK AND THE STABILIZED GAIN AMPLIFIER 



output voltage d Vji^ + d Vr^. Therefore the input to the valve is the difference 

 between the signal dVin and a fixed fraction of the output, so this is a negative 

 feedback system, the feedback being applied round one stage only. The A 



Figure 11.5 



is {^i{RI^ + RK))l{ra + {Rl + Rk)) and the B is RkI{Rl + Rr)- Typically 

 Rl might be 50,000 Q, /•„ = 10,000 Q, the bias 2 volts, the anode current 

 2 milliamps and ^ equal to 50. From these figures R^ must be 1,000^, 

 so the gain without feedback will be 



Our B 



50(50,000 + 1,000) 

 (10,000 + 50,000+ 1,000) 



1,000 



42 



50,000+ 1,000 



0-02 



AB is therefore only 0-84, which is not ^1, and therefore the amount of 

 feedback is small. If in course of time fj, falls to half its original value, the 

 gain with feedback falls from 42/(1 + 0-84) to 21/(1 + 0-42), i.e. from 23 to 

 15.* 



The gain can hardly be said to be stabilized. To improve matters we need 

 more Rj^. If we merely increase R^ in Figure 1 1.6 the valve tries to give 



HT + 



iVn 



SVf 



Rk 



Figure 11.6 



-18 volts 

 Figure 11.7 



itself more bias and the current falls. To increase R^ but preserve the status 

 quo in other respects we need a new circuit ; we return the bottom of Rj^ 

 to a fixed negative potential (Figure 11.7). Suppose we make R^ 10,000 O. 



* Assuming we take the output to be dVR^^ + SVr^. Usually only dVR^ is used, so the 

 gain is effectively lessened by the factor RlK^l + Rr)- 



167 



