584 BELL SYSTEM TECHNICAL JOURNAL 



after the spark ceases and is called the re-ignition voltage of the gap. It 

 will be shown that this re-ignition voltage determines to a large extent 

 the starting voltage of the fixed gaps. 



Before describing the sequence of events required for starting and operat- 

 ing, it is desirable to define our terms more precisely than we have defined 

 them up to this point. The minimum operating voltage is the lowest switch 

 voltage at which the tubes will continue to break down 100% of the time 

 under the action of the trigger pulse, and the maximum operating voltage 

 is that higher switch voltage at which spontaneous breakdown of the series 

 of gaps never occurs. Thus the operating range of voltage is that which 

 includes those voltages existing across the series of gaps, at the time of 

 application of trigger pulse, for which the tubes always break down under 

 the action of the trigger pulse but never before . Starting voltage is defined as 

 the minimum value of d-c voltage at which a series of gaps can be made to 

 break dowii under the action of the trigger pulse. Starting thus differs 

 fundamentally from operating in that while operating demands that the 

 series of gaps always breaks down under the application of the trigger pulse, 

 starting requires only that the gaps break dowTi once in many trigger pulses 

 occurring in a fraction of a minute. Thus, a starting voltage is always lower 

 than the minimum operating voltage. However, due to the doubling of the 

 switch voltage when starting occurs, the d-c power supply voltage required 

 to start may be higher than the d-c power supply voltage at the minimum. 



The results of a quantitative oscillographic analysis of starting and oper- 

 ating characteristics of a pair of preproduction W. E. 1B22 tubes^ are now 

 presented in detail, for they are qualitatively representative of all spark gaps. 

 These tubes operate in a two-gap circuit, a schematic of which is shown in 

 Fig. 5 (a). The analysis is carried out by an examination of the voltage- 

 time wave which occurs at the point of application of the trigger pulse, the 

 midpoint of the two gaps. It will help in understanding the oscillograms^ 

 which follow if it is borne in mind that the voltage across gap 1 is the voltage 

 shown on the oscillogram with respect to ground or "O" voltage, while the 

 voltage across gap 2 is the voltage shown on the oscillogram with respect 

 to the switch voltage. 



The sequence required for starting is shown in Fig. 16 (a). Just before 

 the application of the trigger pulse the voltage at the midpoint of the two 

 gaps is half that of the applied d-c by virtue of the resistance divider. When 

 the trigger pulse is applied, the voltage rises to A (3.8 kv) which is the mini- 

 mum breakdown voltage for these tubes with voltage rates of rise 

 encountered in the trigger pulse. Gap 1, therefore, may break down at A 



^ These tubes contain both corona points and radium to reduce spark delay time (see 



n-(e)). 



' The time scales of these oscOlograms are expanded in regions of very rapidly reversing 

 voltage in order to make clear the sequence of events. 



