at the moment. The moist types of capacitors may undergo very 

 peculiar changes. In this case we find that the leakage begins to go 

 up as pressure increases. As the pressure increases further, the 

 capacitors begin to get very erratic and often short out. But if we 

 just keep with it, increasing the pressure further, we nnay find that 

 around 8, 000 p. s. i. the capacitor may begin to improve. In fact, 

 some of them have been known to improve up to pressures of 

 20,000 p.s.i. Interestingly enough, sometimes a capacitor which 

 has succeeded in operating satisfactorily at 20, 000 p. s.i. may for 

 all intents and purposes be a better capacitor when it is returned 

 to atmospheric pressure than it was before. It is recommended 

 that electrolytic capacitors which are to be operated under high 

 pressure be substantially derated to operate at about half the 

 rated voltage. 



Lights: Often it is necessary to use lights for various purposes. 

 Very cheap ones may be adequate. The ordinary flashlight bulb, 

 which the industry classifies as a G-3 1/2 envelope, will withstand 

 pressures of over 20, 000 p. s. i. However, about 20 percent of the 

 commercial supply is defective due to seal failures. 



Electron tubes: The electron tube causes much less trouble than 

 you would suspect. The subminiature series, the T-3 1/2 envelope 

 of some manufacturers -- not all, for reasons, incidentally, 

 which are not too clear -- will withstand pressures as high as 

 20, 000 p. s. i. 



Relays and switches: Relays operate perfectly well in fluids at 

 high pressure. Microswitches or snap action types are also easily 

 used in the fluid environment. Stepping switches operate well, but 

 one must be sure to use a fluid with an appreciable lubricity be- 

 cause otherwise the stepping switch will fail. 



Clocks: The automotive d.c. types, even using escapement mechan- 

 isms, will operate quite well if the viscosity of the fluid enveloping 

 thenn is low enough. 



Transistors: If transistors are fluid-filled, they become pressure- 

 equalized devices and we have been unable to measure any pressure 

 effects whatsoever. I must be very careful when I say "not able to 

 measure, " because someone may say we didn't go out far enough 

 into the decimal places. By this I simply mean that when the 

 transistor was used in an oscillating circuit (sine wave) which 

 tests many of its essential parameters and was subjected to 



34 



