DESIGN PROCEDURE 



Assuming the necessary components, test gear and tools are to hand, 

 decide on the first part of the apparatus to be developed and jot down the 

 proposed circuit. Insert some hkely component values, bearing in mind 

 the design equations and limits given in earlier parts of this book. Happily 

 electronic design is not like solving a simultaneous equation. There is no 

 single correct combination of component values which is the only one which 

 will work: this may be true of parts of a circuit, e.g. filters, where certain 

 values are uniquely determined if others are known, but for the most part 

 considerable latitude is possible. Often ratios are of more importance than 

 absolute values, as in, for example, potential dividers. It follows that 

 electronic design is not a direct process in which every step is inflexibly 

 determined by its predecessors. For most of us there is no question of 

 sitting down with pen, paper, manufacturers' literature and slide rule, 

 doing some calculations, then handing one's technician a completed circuit 

 diagram in the confident knowledge that the circuit will work, exactly as 

 predicted, straight away. A more empirical approach is both easier and 

 safer. 



Having arrived at some values for the new apparatus, build it, one stage 

 at a time, and do not move on to the next stage until the first is giving the 

 required performance. The great thing is to get the stage going, if only 

 after a fashion. At least one then has a performance of some sort to measure, 

 waveforms to look at, and voltages to check. Thereafter, a proper under- 

 standing of the circuit action, coupled with the apphcation of common 

 sense, should suggest what component values to modify in order that the 

 required behaviour is obtained. The only point to watch is that, in one's 

 enthusiasm, some simple rule about component loading is not overlooked. 

 Over-run resistors and chokes generally draw attention to the fact by dis- 

 coloration and the emission of peculiar smells, or even smoke, but capacitors 

 can break down without warning, and valves may suffer silently for a time, 

 then fail prematurely. 



When the stage is working properly, find out how susceptible it is to 

 variations in circuit parameters, particularly supply voltages, so that an 

 estimate of the required power pack stability may be made. If it is a pulse 

 circuit, make sure that it does not 'work, but only just'. For example, if 

 in the electrophysiological unit which is our example the master oscillator 

 ceases to trigger the time base if the former's output pulse amplitude falls 

 by 10 per cent, then the design is clearly unsatisfactory. The effect of 

 production spreads on component values is such that in subsequent attempts 

 to reproduce the apparatus, a weak oscillator is liable to be combined with 

 a refractory time base, and the gear will not work. Even if it does, it may 

 soon cease to as the valves age. The trigger pulse should be made at least 

 twice the minimum required, either by stepping up the oscillator output or 

 by reducing the suppressor bias which holds the sanatron Miller valve 

 cut off". 



In building the first working model the mechanical construction should 

 be of 'lash up' form. At the moment we are concerned with the electrical 

 design. Lay-out is unimportant except in so far as haphazard placing of 

 components may lead to instability due to unwanted capacitive or inductive 

 couphngs. The object is simply to provide sufficient mechanical support 



674 



