EQUIPMENT FAULTS 



the output stage anode (or anodes), then transfer to the output stage grid, 

 then to the penuhimate stage anode, and so on. At first the appHcation of 

 the dips will suppress the fluctuations of the base line, but at length a point 

 will be reached at which hooking them on makes no difference. Thus if this 

 point is the second stage grids, suspect the second stage anode loads or 

 cathode resistors. To find out which component is the offender, disconnect 

 the capacitors and connect one of them across each suspect resistor in turn. 

 Resistors which, when so shunted, have no effect on the fluctuations may be 

 regarded as in order; those which, when shunted, suppress the fluctuations, 

 or at least round them off, should be replaced by good components in turn 

 until the faulty resistor is found. The culprit so detected may seem to behave 

 quite normally if tested on an ohmmeter, but if replacement proves it in 

 fact to have been responsible, it should be at once destroyed, preferably with 

 a hammer*. It is a bad practice to allow dubious components to lie about 

 where they may be 'rescued' and used by someone else. 



The same 'freezing' process may be used to detect the point of entry into 

 the amplifier of 50 or 100 cycle interference from within, e.g. because an 

 electrolytic decoupling or smoothing capacitor has become 'low' or open- 

 circuited. 



{4) Illusory faults 



The purpose of this section is really to utter a warning: not to assault 

 the apparatus with screwdriver and soldering iron until it is clearly established 

 that a fault condition really exists. So often, hours are wasted whilst a 

 'fault' is eventually found to be an accumulator in need of charging, or a HT 

 battery plug in the wrong tapping, or even not plugged in at all. As further 

 examples of the kind of thing which can baffle one for a time, a stimulator 

 whose delay period is accidentally set to longer than its repetition time will 

 seem to behave in an irrational manner. D.c. amplifiers may refuse to 

 balance simply because of an excessive standing potential developed by the 

 electrode system. Double-sided amplifiers are often used with a single 

 electrode, the other input being earthed. In this case remember that the 

 standing potential developed by the preparation-earth connection must be 

 taken into consideration too. Enormous and inexplicable base-line shifts 

 may be produced in direct-coupled apparatus in sympathy with the move- 

 ments of the experimenter, because his shadow is faUing across his pre- 

 amplifier valves. Some valves, excellent in other respects, are extremely 

 photoelectric, and should be shut up in light-tight boxes. Moving-iron 

 meters can give utterly meaningless readings because they are being used in 

 the neighbourhood of a magnetic field. The list could be extended indefi- 

 nitely, but to end this somewhat depressing chapter, I retail this story of 

 Professor Matthews'. 



A piece of apparatus was set up, away from home ground, for a demon- 

 stration. It included a pre- and a main amplifier; the former was stood 

 upon the top of the bench, and the latter on the floor. The cable joining 

 them was passed through a hole, which happened to have been drilled in the 

 bench-top on some previous occasion. Upon trying out the apparatus, the 



* Psychologically satisfying. 



669 



