Alternate Current- Condensers. 105 



seems to be homogeneous, and is not hygroscopic, yet it heats 

 considerably. 



For commercial condensers the choice of insulating material 

 is limited. Most work has been done on condensers with paper 

 and hydrocarbon insulation. It is difficult to know whether 

 a fault lies in the paper or in the hydrocarbon. The paper is 

 baked at a high temperature, and all temperatures and times 

 of baking have been tried. Paper goes on giving off water 

 till there is nothing but a charred brittle mass left. Many 

 hydrocarbons contain enough water to prevent, for instance, 

 their dissolving rubber. Phosphoric anhydride was used to 

 dry several samples, but seemed to do little good, in fact the 

 crude hydrocarbon was often better. Papers vary astonish- 

 ingly in their behaviour. Mr. Bourne, who has been working 

 at these things for some months, has tried almost every con- 

 ceivable way of making condensers. We are just beginning 

 to succeed, but it is more by chance than anything else. We 

 see no reason why one particular kind of paper and one 

 particular kind of hydrocarbon should be better than the 

 others. 



Want of homogeneity in the dielectric of a condenser, or 

 of any cable, may lead to disruptive discharges, or break- 

 downs. For instance, if a dielectric is made up of portions 

 with inductive capacities of 3 and 1 respectively, the parts 

 with a high inductive capacity may arrange themselves so 

 that the " electric displacement " in the other is so great that 

 a disruptive discharge ensues. For instance, if two conduct- 

 ing plates are 3 millim. apart in air, with enough pressure to 

 spark over 2 millim., and if a 2-millim. slab of a dielectric 

 with a specific inductive capacity of 3 is put in, the fall of 

 potential over the air is nearly doubled, and it breaks down 

 and starts a short circuit. 



Loss of power in dielectrics is not confined to condensers ; 

 it may be very serious in cables. A condenser that heated 

 excessively was made like the Deptford mains. The dielectric 

 was very much thinner in proportion to the pressure, and 

 allowing the loss per cubic centimetre to vary as the square 

 of the pressure on its sides, a Deptford main would have a 

 loss of, very roughly, 7000 watts in the dielectric. It does 

 not follow that this is the exact loss in a Deptford main. 

 Very small differences in the constitution of the dielectric 

 cause large variations in the power wasted. The loss may 

 therefore be very much greater or very much less. Taking 

 the specific inductive capacity of the dielectric as 2, a seven- 

 mile main has a capacity of 2 microfarads. Wi h 10,000 

 volts at a frequency of 80, this takes 10 amperes and 100,000 



