REFLEX OSCILLATORS 



615 



Investigation of the pulling figure of early models of the 2K55 led, how- 

 ever, to the disclosure of one unforeseen pitfall arising from the existence of 

 electronic hysteresis. It had at first been considered that electronic hystere- 

 sis would not be of importance in the transmitter tube, where the feature of 

 electronic tuning was of no importance. This might be true in a CW oscilla- 



FREQUENCY 



POWER 



Fig. 104. — Rieke diagram for the 2K54 oscillator at a nominal frequency of 4350 

 megacycles. The unity iswr point was obtained as described in Fig. 103. 



tor, but in a pulsed oscillator the existence of hysteresis resulted in an un- 

 foreseen reduction of the sink margin. Since the oscillator is being pulsed, 

 for each pulse the oscillating conditions are being re-established. Although 

 the cathode-repeller voltage need not vary during the pulsing, the fact that 

 the cathode-resonator voltage is being changed means that for each pulse 

 the drift angle in the repeller space varies on the rise and fall of the pulse. 

 The effect during the rise of the pulse is the same as though the repeller 



