The issue of concrete strength often arises in civil engineering practice, either due to quality control of new constructions or due to the assessment of existing structures. To this aim, one of the most widely spread techniques is the rebound hammer (Schmidt hammer) test, for which calibration is still related to the original Schmidt curve dating back to the early 50's. In spite of the large amount of research work performed in the last decades, the uncertainties of the rebound test are still not clearly quantified and open to further insight. This paper presents and discusses a wide research campaign on laboratory specimens and on third-party specimens delivered to the Laboratory for Building Materials of the University of Genoa, Italy, for standard quality controls. While it is well known that moisture content, surface finishing, and concrete maturity strongly affect the test result, the effect of the stress state has not yet been studied and is found in this research to be a further parameter affecting the test reliability. The final outcome of all the uncertainties is variability in estimated concrete strength as large as +/- 70%; additionally, some issues are discussed on the intrinsic uncertainty of this test. As already demonstrated by many authors, the results of this research also show that a universal calibration curve to be used for any concrete, in any condition, conceptually does not exist.
