100 ACCURACY. 



to certain errors, some of which are capable of remedy, 

 while some it is, in the nature of things, not possible to 

 avoid. 



Let us see to what the errors of testing machines are 

 due. First, there may be errors due to the distances between 

 the knife edges not being accurately measured in the first 

 instance, or to their distances being altered by wear. For 

 this reason levers with large fulcrum distances and small 

 mechanical advantage are preferable to those with very 

 small fulcrum distances : they are easier to measure in the 

 first place, and slight distortion or wear gives rise to a 

 smaller percentage of error than does the same distortion 

 in a lever of smaller fulcrum distance. In the Greenwood 

 machines two levers, each of about 10 to 1 leverage, are 

 used ; in the Adamson machine are four levers, each of 

 about the same mechanical advantage ; in the Wicksteed 

 vertical machine a single lever is used, with a 50 to 1 

 leverage. It might be thought that this latter leverage is 

 too great, but it must be remembered that the lever is a 

 very large one, and the fulcrum distance is comparatively 

 great, whereas in the multiple-lever machine the individual 

 levers are smaller, and their fulcrum distances correspond- 

 ingly so. It must also be pointed out that if there are 

 errors in the first lever fulcrum distance of a compound 

 machine, this error is multiplied b}- the number of consecu- 

 tive levers ; in the single-lever machine the error cannot 

 be multiplied by any subsequent levers. 



The second way in which an error can creep into a 

 testing machine lies in the fact that the weight or weights 

 themselves may be incorrect. A one-ton weight must be 

 exactly one ton, or else it is of no use. All weights used in 

 testing machines, whether great or small, can be and are 

 tested by the Standards Authorities. There is no difficulty 

 with the smaller weights. When it is required to check 

 the larger weights they may be lifted from the machine by 

 chain blocks, and weighed with a suspended weighing 

 machine, which has been previously calibrated with known 

 standards. This should be done from time to time in all 

 machines. 



The effect of the friction of the knife edges has already 

 been mentioned. Errors due to this cause are very small, 

 and are practically negligible. 



Stresses, which are not indicated, are sometimes caused 

 by inertia forces due to the movement of the machine 

 parts. Theoretically, these are greatest in those machines 



