Poor CBR tests in crushed rock gravel

[geoeng44]

I had two CBR tests run in a crushed rock gravel, gap-graded with about 20 percent fines. It was produced by a rock crusher on-site, and acts a lot like screenings. The CBR results are a concave shape for the entire curve up to 0.5 inch deflection. ASTM says to run the tests again, however, I have two tests of similar shape, so I think a rerun will be similar.

The CBR results range from 4 to 5 for one sample (wet at 18% moisture) for deflections of 0.1 to 0.5 inches,; the other test ranges from 10 to 20 (near optimum at 15% moisture).

For those who have run CBRs on crushed rock, do you find concave up curves common. How would you recommend evaluating the results?

 

[Ron]

Bearing tests on gap graded materials can give flaky results. Make sure you have scalped all material greater than 3/4" and have made appropriate replacement or compensation. The problem with these tests in such material is that you tend to point load a couple of large particles, causing excessive displacement (the reason for the curve shape). You end up shearing the soil in small areas rather than at the perimeter of the plunger.

 

[BigH]

Good point Ron.

 

[geoeng44]

Thank you, Ron. My lab manager pointed out the same thing, that the seating may be uneven. We did do the replacement of greater than 3/4" material, using material retained on the #4.

Since it looks like you agree not to rerun the tests, my question is, should I focus on the later data, 0.3 to 0.5 inch deflection, to calculate the CBR? Also, should I perform the correction for seating (in these tests, the correction displaces the data about 0.15 inch). The CBR is increased by the correction about 50 percent. Also, in the future, should I consider extending the test beyond 0.5 inches of deflection. If so, what would the factor be to make the result equivalent to 0.1 inch deflection?

 

[Ron]

If the center portion of your curve is relatively straight, apply the correction and move on. Initial seating, grain size and test procedure can influence initial readings.