SOFT VALVES 



milliamps. Whilst they are therefore similar in principle to reference tubes, 

 they are much larger and more robust affairs. 



To ensure that the tubes will strike we arrange them in order, with high 

 resistances across the less refractory ones, and apply the same rule as 

 for reference tubes, except that the voltage seen by the tubes is now not V^^ but 



K{RlKR + Rl)}- 



I R 



o — ^-^vvw- 



^i ^. 



R, 



Figure 7.5 



Figure 7.6 



If the string of stabihzer tubes has an incremental resistance Rf, then so 

 far as the reduction of the effect of fluctuations is concerned the equivalent 

 circuit is Figure 7.6. We have now to consider two aspects of stabilizer 

 performance. 'Forward' stabilization is the tendency of the arrangement to 

 maintain V^ despite fluctuations in K„. 'Backward' stabilization is the ten- 

 dency of the arrangement to maintain Vg despite fluctuations in i?^, that is, 

 despite variations in load current. Clearly the forward stabilization is good 

 if R^ or Rj^ are much smaller than R. Also the backward stabihzation is 

 good if the impedance looking back towards the stabilizer from the load is 

 low compared with the load resistance. Therefore for good all-round 

 stabilization we need R and Ri^ large and R^ smafl. On putting in some 

 figures it turns out that the design of stabihzer circuits is surprisingly com- 

 plicated and may even be impossible. For, caUing the striking potential of 

 the tube or string of tubes V^, we have: that they may just strike, F^ = 

 K(Rl)I(Rl + R)- Also that once they have struck F„ = /i? + V,. 

 Eliminating R between these equations 



K- 



V. 



+ ^i = 



VuRl 



or 



IRr 



V = V- — — 



" ' IRr 



F. 



Looking at the denominator of this it is evident that if the striking voltage 

 Vi is so high that it approaches IRj^ then the expression goes to infinity, 

 which means that the tubes will not strike however high we make F„. The 

 arrangement is probably intolerably inefficient if F„ is greater than twice F^, 

 that is, one should aim for 



2> 



IR 

 IR, 



■ F. 



or 



/^i>2F, 



122 



