Marshall Stability Outlier in a 12.5 mm Surface Mix
A laboratory is verifying a Marshall mix design for a 12.5 mm surface mix at 5.7% AC. Three 4 in specimens from one batch are compacted with 75 blows per side at 275°F. Results: A height 2.49 in, air voids 4.1%, stability 1,860 lbf, flow 10 (0.01 in); B height 2.61 in, air voids 4.4%, stability 1,430 lbf, flow 16; C height 2.52 in, air voids 4.2%, stability 1,690 lbf, flow 12. The specification requires an average stability of at least 1,500 lbf, flow 8-16, and no individual stability below 1,400 lbf after thickness correction. The bath log shows A and C were held at 140°F for 35 min. B was removed after 22 min because the bath was needed for another test and sat on the bench for 3 min before breaking. The technician proposes to report the average stability of 1,660 lbf and flow of 12.7 and accept the design.
1. Should the laboratory accept the design using the reported average? Explain.
Read the worked answer
No. The average cannot be used because specimen B was not conditioned to the required test temperature for the required time. AASHTO T 245 requires Marshall specimens to be immersed in a 140°F water bath for 30 to 40 minutes before breaking. B received only 22 minutes and then cooled on the bench for 3 minutes. Testing a specimen that is too cool lowers stability and raises flow, which is exactly the pattern seen in B: the lowest stability and the highest flow. The problem is not just an outlier to be averaged away. B's height of 2.61 in also requires the T 245 thickness correction; applying the correction lowers its stability to about 1,390 lbf, below the 1,400 lbf individual floor. Even before correction, the test is invalid. The technician should reject B, not report the three-specimen average as if all were valid.
2. What is the correct corrective action, and what conditions must be controlled during the retest?
Read the worked answer
Discard the invalid result and retest with a new set of three specimens from the same batch. Verify the water bath is at 140°F ± 2°F and record the immersion time for each specimen; hold each specimen for 30 to 40 minutes and break it promptly after removal without allowing it to cool. Confirm the specimen heights and apply the T 245 thickness correction to stability before comparing to the 1,400 lbf individual floor. Check the Marshall hammer for the correct 75-blow count, drop height, and foot alignment, and verify the compaction temperature and specimen cooling before extrusion. Also check the flow meter and breaking head so the flow values are measured consistently. Report individual corrected stabilities and flows, not just the average.
3. If the retest again produces one specimen near the flow limit, what would you conclude about the mix?
Read the worked answer
A repeat high-flow specimen would require a mix-level review, not another automatic retest. First confirm the test was valid: correct bath time and temperature, correct specimen height, and no damage during extrusion. If the test is valid, high flow with acceptable air voids can indicate an over-asphalted or tender mix, excessive fines, rounded aggregate, or a compaction temperature problem. Review the asphalt content, dust-to-binder ratio, and aggregate angularity, and compare the retest average and individual values to the specification. If flow remains at or above 16 while stability drops, the design should be adjusted, for example by reducing asphalt content or modifying the gradation, rather than being accepted on a marginal average.