REGENERATION THEORY AND EXPERIMENT 687 



versed feed-back oscillator circuit, the three curves marked a, b, c, cor- 

 responding to progressively increasing gains around the loop. It will 

 be observed that after the maximum gain has reached and exceeded 

 unity, that the circuit is unstable, since the point (1, 0) is then enclosed. 

 This state of affairs may be contrasted with that existing in the par- 

 ticular form of feed-back circuit to which Fig. 4 applies. Again the 



Fig. 4 — Transfer factor diagram for a particular form of feed-back circuit, curves 

 a, 0, c, corresponding to increasing gains around the feed-back loop. Conditions a 

 and c are stable, b is unstable. 



three curves a, b, c, correspond to progressively increasing gains around 

 the feed-back loop. As the gain is increased the system is first stable 

 (a), then unstable (&), and finally stable (c), since it is only within 

 curve (b) that the point (1, 0) is enclosed. This striking example is the 

 one which was investigated experimentally. The methods used in de- 

 termining the transfer factor diagram form the subject of the next 

 section. 



Measuring Methods 



Application of the Nyquist stability criterion requires the de- 

 termination of the vector transfer factor around the feed-back loop 

 at all frequencies. This is usually effected by opening the circuit at any 

 point which provides convenient impedances looking in both directions 

 from the break. These points are then connected to an oscillator and 

 to suitable measuring circuits, which are to be described. Care must 

 be taken to ensure that the oscillator and measuring circuit impedances 

 are equal to the output and input impedances respectively of the 

 circuit under test. This precaution is necessary in order that the trans- 

 fer factor in the measuring condition may not differ significantly 

 from that existing in the operating condition. 



