Summary
Although the compression test on cubes and cylinders is used very extensively as a measure of the general quality of concrete, there has been increasing evidence to show that such tests are not as reproducible as has normally been assumed. There are large differences in the strengths obtained from different testing machines, and often a considerable scatter in strengths from one machine. In a series of tests, the influence of such characteristics as spherical seating behaviour, misalinement, homogeneity of specimen, longitudinal stiffness and platen restraint on the strength of cubes and cylinders has been examined. Depending upon the behaviour of the spherical seating, differences of about 6% may be obtained with accurately centred specimens. If such specimens are also misalined, this difference becomes even larger, as enormous reductions in strength are observed with effectively pinned seatings. Cube and cylinder strengths are affected quite differently by the method of end loading, the cylinder/cube strength ratio being lowest when both ends are fixed. Longitudinal stiffness of the machine does not affect strength. Platen restraint, however, has a very significant influence, particularly if the end-blocks are small.
