eee 
oA 
ON IMPACT TESTS. 27 
satisfactory results. If the anvil is supported on springs the results 
would compare more or less with the double pendulum, but the com- 
pression of the springs could not be determined as accurately as the move- 
ment of the pendulum anvil. 
To standardise the impact test, it would be desirable to agree as to 
the magnitude of the moving masses and the type of machine, as the 
blow would be much sharper in the direct fall machine in consequence of 
its more massive anvil, and in specifying the resilience it is necessary 
to assume that all the conditions affecting the results should be maintained 
practically constant. 
In connection with the notched bar tests made by the authors, the 
angle of rupture was measured after fracture (fig. 8). It is only possible 
to do this accurately when there is contact over the whole surface. In 
the tougher materials there is considerable sliding of the surfaces, 
one may be convex and the other concave. Again, in some of the larger 
pieces, both parts were convex. If the angles are plotted against energy 
of rupture, it will be found that a relationship exists :— 
Eee bere nee By orn 
Angle of rupture D 
In the various tables of results it will be seen that this ratio is more 
or less constant, and any difference is largely due to the difficulty in 
correctly measuring the angle of rupture. 
Mesnager found that for the smaller test-pieces K=0°375 D, and for 
the larger test-pieces K=1+40-58 D. 
He also found that if R denoted the tensile strength of the material 
in kilogrammes per sq. cm., then :— 
For the small pieces :—R+2°66 D=95, 
3? ” larger ” R+1°72 D=87. 
A similar relationship could be established between the angle of 
rupture and hardness by plotting a sufficient number of tests, but all 
such equations would be more or less approximate, and in the case of 
brittle steels the angle is too small to measure and the method fails entirely. 
The steels tested by the authors have the following composition :— 
Tassie II. 
Chemical Analysis. 
Mark C Si | Ss | 1B Mn Cr | V | Ni 
Bias. ay. e | | % 
Rll 29 083 027 ‘024 48 56 — 2°88 | 
R19 25 108 038 “031 71 — — See 
R31 | 30 121 | -027 | -028 65 se as oe 
5960 | 22 140 “032 028 50 81 17 —- 
3198 44 12 014 Trace 87 — _- _— 
1324 43 13 023 008 83 - — — 
2300 50 13 024 ‘015 68 -— — -— 
3184 45 074 ‘019 -012 | 72 — _— —_— 
