692 
Transactions of the Royal Society of South Africa. 
only a small portion is heated during the time required to start a motor, as 
there is no time for circulation to set in. Experiments carried out with 
freely suspended coils of No. 22 S.W.Gr. copper wire showed that the time 
elapsing before circulation takes place depends upon the current density. 
The results of the test are shown in fig. 6. When the current was 8 amperes 
H took 70 seconds, for 10 amperes 40, and for 12 amperes 30 seconds. The 
rates of temperature rise during these times are also indicated in the inset on 
the right. The quantity of oil affected was approximately the same in all 
cases — 1'25 kilograms out of 24 kilograms contained in the vessel. A large 
vessel for short duration starters thus means waste of material. 
Analysing the results, it was found that for the same temperature rise in 
air-cooled and oil-insulated rheostats for short duration loads, when cooling 
is negligible, the current density in the latter case may be about four times 
higher than in the former. It must, however, be remembered that for air- 
cooled resistances much higher temperature rises are allowable, depending 
upon the mechanical design, whereas the maximum temperature of oil should 
not exceed 100° C. if sludging and deterioration is to be avoided. The 
saving of material in oil-insulated starters is thus illusory, even if we neglect 
the cost of the vessel and the expense of the oil. As a matter of fact, the oil 
is used mainly for safety in places where sparking is dangerous, as in coal- 
mines, where all switching has to take place under oil. 
For rheostats embedded in sand or enamel the current density may be 
about twice as high as for air-cooled resistances, assuming the same tempera- 
ture rise, which may be considerably higher than oil permits, so that now 
resistance material is saved. 
Rheostats for Loads of Moderate Duration. — The general law of 
heating reads 
t . 
S^S^l -^'TJ 
As all heating curves have the same shape we may plot one for ^;„=1 and 
to—l, as has been done in fig. 7. We next determine t„ and the ratio j, 
whence for a given value of the rise is also determined. 
I . . . 16. 
for circular conductors 
_4IV 
~A'rf-5 
=57P generally 
_cad 
~4A' 
where d denotes the diameter. 
for circular conductors 
