DESIGN PROCEDURE 



master oscillator; the thyratron HT should not be applied until its heater 

 has warmed up. It seems a pity to use a thermal delay switch just for one 

 valve. We might use an indirectly heated thermionic rectifier, which also 

 takes some time to warm up. Or we might provide a 'Thyratron HT switch' 

 to be operated by the user, say, 30 seconds after closing the main switch. 

 Perhaps it would be safer not to use a thyratron at all ; what about a blocking 

 oscillator? 



Spotting latent difficulties such as these can save trouble later on, but it 

 seems that, however hard one tries, a few always manage to get by unnoticed. 



The next phase is to ponder upon the work so far, because it is at this 

 juncture that ideas come which can lead to important simplifications, and 

 may even be the germ of an original piece of circuitry. For example, in the 

 time-base generator, if we applied the trigger signal to the sanatron control 

 valve at the control grid, instead of at the suppressor, we could use a triode 

 instead of a pentode : and if a triode were to prove a satisfactory alternative 

 to a pentode for the paraphase amplifier, then we could use one double 

 triode instead of two pentodes. Net profit — one valve saved. Again, in 

 the stimulator, suppose that instead of having — in each channel — a double 

 triode for the delay flip-flop and a power pentode for the output, we use 

 the screen and control grid of the power pentode as one half of the flip-flop. 

 Then one triode-pentode valve would perform the duty both of delay stage 

 and output stage. 



Having arrived at a few notions of this kind one is naturally anxious to 

 try them out, and the time is now ripe to begin practical work. For circuit 

 development one needs a good stock of reliable old components — perhaps 

 stripped from obsolete equipment; it is a pity to use new parts here. In 

 addition, certain items of electronic test equipment are required. In the 

 author's opinion the following are essential : 



(1) A good measuring-oscilloscope (i.e., one with the Y shift calibrated 

 in volts and the X shift in time) provided with direct-coupled amplifiers. 



(2) An audio-frequency oscillator, range about 10-100,000 c/s, with 

 calibrated attenuator, and output switchable to either sine wave or square 

 wave. 



(3) Two multi-range universal meters (i.e., combined voltmeter, ammeter, 

 and ohmmeter, a.c. and d.c.) of which at least one should be of high sensi- 

 tivity — say 20,000 ohms/volt. 



Since the design of power packs for apparatus is best left till last, when 

 the output and degree of stability required are known, a wide-range variable 

 HT supply having at least 3 separate outputs will be found invaluable. 

 Failing this, keep a stock of half a dozen HT batteries of the type well 

 endowed with tappings. Specimens discarded as too old for reliable 

 amplifier operation are commonly quite satisfactory for powering develop- 

 mental pulse circuits. A 6V car battery and charger are also required. 



A valuable optional piece of apparatus which can reduce unnecessary 

 wastage of doubtful components, and establish their values accurately when 

 these have to be known, is a resistance-and-capacitance bridge. These may 

 be bought commercially, but they are quite simple affairs and may be home 

 made. M. G. Scroggie {Radio Laboratory Handbook. London; Iliffe) 

 has described an excellent instrument which would be very suitable. 



673 



