ACCURACY 101 



with the greatest leverage, but in practice this is not so, 

 because the angular movement of the whole system is so 

 very much more restricted in high-power machines. 



It is not only desirable that the errors of a testing 

 machine shall be relatively small, but it should be possible 

 to ascertain whether these errors exist, and what is their 

 magnitude, or in other words, to verify the readings as 

 given on the scale. 



In this matter of verification, the vertical machine has 

 a great advantage. It is only necessary to hang dead- 

 weights from the shackles, balance these by altering the 

 position of the jockey-weight, and observe how the scale 

 reading agrees with the actual load. 



The accuracy of most horizontal compound machines 

 is made to depend on the accuracy of the knife-edge 

 distances, which are checked by careful measurement from 

 time to time, and on the correctness of the weights them- 

 selves, which can be adjusted to a very high degree of 

 accuracy. In some cases special subsidiary knee-levers are 

 provided, so that the leverage can be tested by actual loads 

 hung at carefully determined knife-edge distances. 



Another cause of error is due to the fact that the levers 

 of a testing machine deflect slightly under the loads. It 

 is therefore important that the levers shall be designed for 

 stiffness as well as for mere strength. For this reason 

 also high-power levers are clearly better than low-power 

 ones. 



It has already been mentioned that the lever should be 

 neutral, that is, that the centres of gravity of the lever and 

 iockey weight or weights should be in the plane of the 

 knife edges. If this is not so it will be found that small 

 errors are introduced, and this is especially noticeable when 

 taking elastic readings, when it will be observed that quite 

 small angular displacements of the beam cause variations to 

 the load, as shown by the measuring gear. 



56. Summary. From what has been said, the reader 

 should be able to draw his own conclusions. So far from 

 any one type of machine being pre-eminently the best, each 

 type has its good points. 



The vertical single-lever machine has the advantage as 

 regards simplicity, and ease of verification. 



Horizontal machines are best as regards accessibility and 

 adaptability. 



As regards adaptability, probably the horizontal Buck- 

 ton-Wicksteed machine is first. 



