[ 343 ] 



XLI.— ON A NOTT-SPARKING DYNAMO. By GEOEGE 

 EEANCIS FITZGERALD, A.M., F.T.C.D. 



[Eead, June 9, 1884.] 



What are called unipolar dynamos liave no sudden change in tlie 

 nature of the sliding contact, which is essential in all dynamos, and 

 are consequently non-sparking. The arrangement I am ahout to 

 describe is somewhat different, and is an application to a dynamo 

 of the principle that Maxwell applied to Sir William Thomson's 

 electric doubler (Maxwell's Electricity and Magnetism, vol. i., 

 p. 298, s. 213) in order to prevent sparks in it, and which he 

 likened to the regenerator in a heat engine. 



Consider a series of coils arranged to rotate in a magnetic field 

 as a gramme ring, but with their wires unconnected with one 

 another, each coil being connected with a separate commutator 

 and pair of brushes, and half the brushes being connected with one 

 electrode, and the other half with the other. In this case the cur- 

 rent would in general go through a number of the coils in parallel 

 circuit. Now it is evident that by a suitable arrangement of the 

 magnetic field and of the positions of the brushes on the commuta- 

 tor, it could be arranged that the break of contact of the brushes 

 and commutator should take place when the current in the coil was 

 zero. This is evident, because if the brushes made contact through- 

 out a whole revolution, as in an alternating current dynamo, the 

 current would be reversed every half revolution ; and if the instant 

 of reversal were the instant of break of contact, this would happen 

 when no current was passing in the coils. It is not possible to 

 arrange this unless there are at least two parallel coils carrying 

 the current, as otherwise the whole current would be fluctuating, 

 and might as well be alternating. With a number of coils parallel, 

 it would evidently be quite possible to arrange that the time of 

 break of contact of each pair of brushes and its commutator should 

 occur when no current was running in the corresponding coil. One 

 way of doing this would be, as in Maxwell's electric doubler, to 

 arrange a subsidiary magnet system to alter the magnetic field 



