Question: I have a further question regarding the rounding of compressive strength results ("How Precisely Should Labs Round off Compressive Strength Results?" April, page 26).

At our lab, we test 4x8-inch and 6x8-inch concrete cylinders for compressive strength, and we round to the nearest 10 psi according to ASTM C 94, -Specification for Ready Mixed Concrete. We also test 2x2-inch cement cubes, which are the subject of my question.

According to ASTM C 109/C 109M, "Standard Test Method for Compressive Strength of Hydraulic Cement Mortars (Using 2-in. or [50-mm] Cube Specimens)," when calculating compressive strength of the cubes, a technician must average the compressive results of three test specimens. If the deviance of a specimen's result strays from the average by more than a specified maximum, it should be discarded. The average of the remaining two is taken, and the calculation procedure is repeated, but the maximum specified deviance from the average is lower between two specimens than among three specimens. In other words, after we round the compressive values from the 3 cubes, additional calculations often follow.

Why intermediately round after determining the ratio of applied load and surface area (compressive strength) when I can measure the cubes to 0.001 accuracy, and our loading machine measures to the nearest 1 psi? Doesn't intermediate rounding introduce greater and unnecessary error? I understand rounding to the nearest 10 psi on the final result when no other calculations will be needed. It seems like this goes against everything we learn in the classroom about significant figures and propagation of error.

Jack Lapham
Engineering, Oregon State University

Answer: To answer this question, we consulted Willy Morrison, supervisor of physical testing at Construction Technology Laboratories.

Morrison reminds readers that the method for testing and calculation of compressive strength of concrete cylinders is ASTM C 39, "Standard Test Method for Compressive Strength of Cylindrical Concrete Specimens." ASTM C 94 is a standard for ready-mixed concrete that references ASTM C 39. ASTM C 94 has a section on strength as a basis for acceptance of test results for design strength.

With regard to cement cube testing, Morrison notes, C 109 requires that the compressive strength of all 'acceptable test specimens' made from the same sample and tested during the same period be averaged and reported to the nearest 10 psi. Rounding to the nearest 10 psi is consistent with the requirements of ASTM C 150, "Standard Specification for Portland Cement." If there is agreement among test values for each of the individual cubes, you should report the rounded result. In effect, all test specimens are acceptable."

However, to account for potential discrepancies among results from the individual cubes, C 109 provides a means to check for "outliers" (unacceptable specimens) and determine whether a retest is needed. This is covered in the section on ôFaulty Specimens and Retests,ö which provides limits for the maximum permissible range of the test results.

When checking for faulty specimens in accordance with ASTM C 109, you should not round the data.