RECENT INVESTIGATIONS 5 



limits the effective pull that can be developed. It causes 

 the horsepower that can be transmitted, within the range 

 of initial tensions that are recommended, to reach a maxi- 

 mum at a velocity of about 4000 ft. per minute. Any increase 

 in velocity beyond this point actually decreases the horse- 

 power that will be transmitted by the belt, and at a velocity 

 in the neighborhood of 6800 ft. per minute the belt will fail 

 to transmit any power whatever. The effect of centrifugal 

 tension is entirely neglected in the rules of thumb ; according 

 to them the horsepower transmitted will increase indefinitely 

 with an increase in velocity. Also, in the rules of thumb 

 the coefficient of friction is assumed to be a constant quantity 

 at all velocities. As a matter of fact, it is shown by the studies 

 of the experiments of Wilfred Lewis by Carl G. Barth that 

 the coefficient of friction is a variable quantity that increases 

 with an increase in velocity. 



It is thus evident that the rules of thumb are unreliable 

 even for determining the horsepower that a belt will transmit, 

 and that they give no information as to the tension under 

 which belts should be put upon the pulleys. It is also evident 

 that the design of a belt drive is not a simple matter of sub- 

 stitution in a formula, but requires the careful consideration 

 of many factors. 



Recent Investigations on Belts. Mr. A. F. Nagle l was 

 one of the first to present a means of determining the horse- 

 power that could be transmitted by a belt, which took into 

 consideration the different variables enumerated above. 

 Mr. Nagle, however, assumed a value for the coefficient of 

 friction that was constant at all velocities of belt, and also 

 based the value of the effective pull which could be allowed 

 in the belt upon the strength of the joint. This procedure 

 gave working stresses in the belt 275 Ib. per square inch for 

 laced belts and of 400 Ib. per square inch for riveted belts. 



1 Transactions, American Society of Mechanical Engineers, vol. ii, p. 91. 



