CENTRIFUGAL PUMPS 671 



Madison University l give the following values for the ratio of calculated 

 to observed manometric efficiency. 



Series 1. 24 vanes, area of waterways increasing . . '910 



Series 2. 24 ,, ., constant . . *910 



Series 3. 12 ,, ,, increasing . . *875 



Series 4. 12 ,, ,, constant . . '845 



Series 5. 6 ., ,, increasing . . '75 



Series 6. 6 ,, constant . . "68 



From these and similar results it appears that iv 3 is always less than 

 is given by the formula ' 3 = w 3 f< A cot y. This is probably due partly 

 to the fact that a dead water space which is not utilised for discharge is 

 formed on the rear side of each vane, thus increasing the radial com- 

 ponent/a, and partly to the fact that only those layers of water near to 

 the driving edge of the vane are discharged parallel to its tip. Thus the 

 true mean values of / 3 and of cot y are both greater than the apparent 

 values. It also appears that the ratio of the true to the calculated value 

 of - 3 increases with the number of vanes and hence as the guidance of 

 the water becomes more perfect, but diminishes as the width of the 

 passages is reduced, probably because the dead water space then forms a 

 proportionately larger part of the whole discharging area. 



It also diminishes slightly as the discharge increases, and as the angle 

 y increases. 



The following may be taken as approximate values of the ratio, in 

 pumps of normal design : 



w experimental. 

 w a calculated. 



Pump upwards of 20 in. diameter ; 12 or more vanes 



7 from 15 to 30 . '97 



30 to 60 . '95 



60 to 90 . -92 



Pump 10 in. to 20 in. diameter ; 12 vanes 



y from 15 to 30 . . '93 



30 to 60 . . -90 



60 to 90 . . -87 



8 to 12 vanes 



y from 15 to 30 . . '90 



30 to 60 . . -87 



60 to 90 . . -84 



1 "Bulletin of the University of Wisconsin," No. 173. 1907. 



