High-Frequency Alternator, 305 



the saving in tangential pull on the pulleys, owing to their 

 larger radius. It is evident that there is in any given case a 

 best diameter of pulley which will give the least strain on the 

 belt; this diameter was found by experiment in the present case 

 to be the size of pulleys shown on the spindle in PL V. fig. 1. 

 The object of making the new pulleys of red fibre was to 

 prevent the heat developed by any slip of the belt being con- 

 ducted to the spindle, and so expanding it and causing it to 

 seize in the bearings. 



At these high speeds the centrifugal force on the belt 

 introduced a serious difficulty, as it tended to lift the belt 

 right out of contact with the small pulleys, so that the slip 

 became excessive. If, on the other hand, the belt was 

 tightened up sufficiently to prevent this, then the strain on it 

 became too great, and it broke after a short run. The joint 

 in the belt was also a source of trouble owing to the inertia 

 forces acting on it. Many different kinds of belts and ways 

 of joining them were tried, such as: gut with connectors or 

 splices; leather both round and flat with different kinds of 

 wire and sewn joints; hempen cord with various splices; 

 common cotton-tape twisted up and sewn together at the ends; 

 and ^g-inch diameter cotton-cord joined with a long splice. 

 Of all these, the last gave the best results. In order to get 

 sufficient friction between the belt and the pulleys, a small 

 quantity of thick solution of resin in turpentine was applied 

 to it ; and if it was then found to be too sticky, a trace of oil 

 was put on the belt with the finger whilst running. The life 

 of one of these belts was not very long, being only about 

 four hours continuous use ; but owing to their cheapness, the 

 ease with which a new one could be made, and to the fact 

 that they did no damage when they broke, this was not con- 

 sidered a very serious disadvantage. 



In spite of every care, all attempts to run the inductor at 

 1000 revolutions per second failed, owing to the heating of 

 the bearings ; but on removing the inductor from the spindle, 

 this latter could be run at 1000 revolutions per second with 

 comparative ease. This was evidently due to the axis of inertia 

 of the inductor not coinciding with its mechanical axis; all 

 attempts to adjust this, such as making new inductors, &c, 

 failed. 



To give some idea of the importance of the two axes coin- 

 ciding, suppose that they are parallel and L mm. apart; then 

 at 1000 revolutions per second the pressure on the two 

 bearings works out to about 0'8 metric ton; so that an error 

 of only j-Ju mm. would mean a prohibitive pressure on such 

 small bearings. 



Phil. Mag. S. 6. Vol. 9. No. 51. March 1905. X 



