68 BELL SYSTEM TECHNICAL JOURNAL 



Changing loads also change the frequency, since they take the form of 

 variable resistance and reactance. 



Considerable work has been done by various individuals to make 

 the inductance and capacity of standard apparatus as free from the 

 effects of vibration and temperature changes as possible. Work of 

 that kind is not discussed here. Variable voltages occur in practically 

 all installations and changing load impedance in many. These two 

 things are the actual cause of the larger part of frequency variation in 

 many installations. 



Several vacuum tube circuits have been devised to surmount these 

 difficulties. They may be divided roughly into two groups: first, 

 those in which the attempt has been to minimize the change of fre- 

 quency with battery voltage, and second, those in which the attempt 

 has been to prevent the change. In the first group we have two types: 

 first, circuits in which the effective resistance has been reduced to as 

 low a figure as possible, and second, those in which a high impedance 

 has been inserted between the tube and the tuned circuit in order to 

 reduce the relative effect of changes in the tube. Considerable success 

 has attended the efforts of a number of engineers led by J. W. Horton 

 in these directions. More recently, circuits of the first type have been 

 applied to the production of relatively high frequencies.^ 



The second group in which the attempt has been to prevent the 

 frequency change developed from the work of Messrs. J. F. Farrington 

 and C. F. P. Rose. They found that a certain critical value of an 

 impedance between the vacuum tube and the tuned circuit apparently 

 produced a constant frequency over a limited range when the battery 

 voltages were varied. They experimented with various forms of net- 

 works for this stabilizing impedance and developed several in which 

 the output power was not reduced by stabilization. 



Theory 



The writer attacked the problem from a theoretical standpoint and 

 showed that in certain cases the mathematical procedure indicates 

 means of making the oscillator frequency independent not only of a 

 variable load resistance, but also of the battery voltages. The pur- 

 pose of this paper is to develop the general theory and application of 

 these circuits and to show how several circuits in particular may be 

 made to produce practically constant frequency with customary varia- 

 tions of voltage and load resistance. The relations necessary to main- 

 tain the frequency constant at any given setting when it is desirable 



^ Ross Gunn, "A new frequency stabilized oscillator system," Proc. I. R. E., 18, 

 September, 1930. 



