Precracked Charpy slow bend tests and plane-strain fracture toughness (KIc) tests were made on the following alloys: (1) 4340 and 18Ni maraging steels heat treated to a wide range of strength levels; (2) several 2000 and 7000 series aluminum alloys; and (3) a metastable beta titanium alloy, Ti-8Mo-8V-2Fe-3Al. Load-deflection records for the precracked Charpy specimens were obtained from measurements of tensile machine screw rotation and directly from the specimen deflection. It was shown that records obtained from screw motion measurements contained large extraneous components of deflection arising primarily from elastic strains in the tensile machine. The magnitude of these components will be a function of the compliance of the testing machine and, depending on the method of record analysis, can have a substantial influence on the derived W̄/A values. Load-deflection record analysis procedures are presented which minimize the effects of these extraneous deflections. If the deflections are obtained from measurements made directly on the specimen, the influence of the tensile machine compliance for practical purposes may be neglected, and a simplified method of record analysis can be used which does not involve graphical integration of the load-deflection trace.
Relations between W¯E/Aσ ys2 and the crack size factor KIc2/σ ys2 are presented for the alloy conditions investigated. In all but a few cases there was a general upward trend in the Charpy energy values with increasing crack size factor. In some cases, excellent correlation was obtained. Where relatively large scatter was observed, the Charpy derived values in a few cases did not follow the trend of the crack size factor. Apparently no generalizations can be made concerning the degree of correlation that can be expected between precracked Charpy energies and plane-strain fracture toughness. Correlations can be useful but should be established carefully over the complete range of toughness of interest.
The authors recommend that an ASTM test method be developed for the slow bend precracked Charpy test and believe there is sufficient information available to do this now. Some suggestions along these lines are presented in this paper.