I'll IV ^111 M \\.-l \|, MOT>|> 71 



'1 hotiiHoli meter the starting curi'<- < :iin|><-r--. the 



maximum currvnt h nni)Ti s. The curves are obtained 

 from data given by Mr < . \\ I >. Ki. k>. in a paper read 

 before the Mriti-h Association, Auiu- . -. the 



' 



If the t.-njue curve is not straight, a further error 

 will be introduced, tending to make the meter run slower 

 or faster according as the curve is concave or convex t.. 

 the current axis. In all these meters the cun'es an 

 cave to the current axis, and for small speeds t In- 

 due to the starting torque is the greater of the two. I < r 

 high speeds, when tin - very small, the error due 



to the bend of the curve becomes apparent, and the 

 accuracy curve turns down a^in. 



Hitherto we have 8up|>osed M to be constant. If M 



varies, the speed will depend on the product of If and -, 



.1 .ind M m< r.-a-.d. the speed will 



increase with M, >-r if lx>th and \l increase, the speed will 



increase with their \< 



If the magnetising coil is wound as a shunt across the 

 main line termin.i uit in which the current is 



flowing, the current in the shunt will !* proportional to 

 the tension, and if the magnetic circuit contains no iron 

 the induction curve will In- a straight Ir value of 



.17 will In* pn|>ortional to the tension, hence the speed will 

 be proportional to the watte in the circuit. Tin- is the 



!! of the Thomson watt-meter. For a i 

 nrrent the speed increases directly in pro|H.rtion to he 

 



\\ e hall now consider the case when two dynamos 

 are mechanically coupled HO that they rotate at the 

 same speed, and have their main terminals connected in 



