A this morning to describe to you the results we have so far secured in our attempts to devise a practical test for determining and controlling the stability of asphalt paving mixtures. number of other investigators have been working on the same problem, and have devised various types of tests, all of which, I believe, if carried to the ultimate conclusion, would lead to the same general deductions. In other words, the exact method of test which may eventually be used for this purpose will depend more upon practicability, simplicity and ease with which the laboratory results can be correlated with service results than upon the method itself. RESEARCHES ON ASPHALT PAVING MIXTURES By Prevost Hubbard and F. C. Field. Owing to the generally acknowledged desirability of increasing, or at least intelligently controlling, the stability of our asphalt pavements so as to reduce their tendency to displacement under modern traffic, quite a number of investigators, in recent years, have been engaged on the problem of devising some practical test for determining the stability of compressed asphalt paving mixtures. Before describing our own investigations in this field it will be of interest to briefly consider the various tests which have so far been developed, without attempting to discuss their relative merits. The principles of these tests are roughly illustrated in Fig. 1. Besson's test consists in dropping a constant load from a given height upon a rod or pin of standard cross section resting upon the upper surface of a cylindrical test specimen of the compressed asphalt paving mixture. The character of failure is noted and the number of blows delivered is recorded. The method of rating relative stability of different test specimens is rather indefinite, as stability is not directly indicated by the number of blows. By comparing test specimens, however, and taking into account the number of blows delivered, some idea may be obtained of their relative stability. Thus a highly unstable specimen will show considerable distortion and will be penetrated to an appreciable depth by the rod, while a very stable specimen will show less penetration and distortion. McNaughton's test is made in a manner similar to the penetration test for asphalt. The cylindrical test specimen is held in the mold in which it was formed and stability is measured by the depth which a rod of standard cross section penetrates into the compressed asphalt paving mixture maintained at a constant temperature, when operated under a given load applied for a given length of time. Here the stability value, of course, varies inversely with the depth of penetration. Howe makes use of the same general method as McNaughton, but determines the depth of penetration of a steel ball. His test specimen is, however, free from the lateral support of any mold or container. Ulman and Milburn, working more or less independently of each other, apply a constant load to a cylindrical test specimen maintained at a constant temperature, and measure the decrease in height after a period of load application. The stability value varies inversely with decrease in height. In Skidmore's test stability is measured by the load required to shear off the free section of a cylindrical test specimen part of which is held in a frame or mold. The test is made at a constant temperature and the load applied in successive increments. Emmons and Anderton compress their test specimen in the form of a rectangular block which is placed in a close-fitting mold open at the top, but otherwise closed, except for three rectangular openings, one at the bottom and one at each of two opposite sides. Stability is measured in terms of maximum load uniformly applied to the upper surface of the specimen, which is required to force the mixture, maintained at a constant temperature, through the openings in the mold. The load is increased at a uniform rate until the maximum is reached. This is known as a squeeze test and was designed for testing both coarse and fine aggregate asphalt paving mixtures. The test of Hubbard and Field is a shear test made upon a cylindrical test specimen maintained at a constant temperature and held in a cylindrical mold equipped with a snugly fitting plunger. Stability is measured by the maximum load required to force the mixture through a circular orifice in the bottom of the mold, the load being increased at a uniform rate of application. It will be seen that although the methods described vary in certain details they all aim in some way to measure the resistance of compressed bituminous paving mixtures to permanent deformation under applied loads. Whether or not any of these tests are ultimately adopted for general use must depend upon a number of factors, among which the following are believed to be important. 1. Accuracy or ability to produce reasonably close check results. 2. Adaptability for use on test specimens taken directly from the pavement as well as on laboratory prepared specimens. 3. Correlation of test values with service results. 4. Development of a sufficiently wide range of test values 5. Compactness and simplicity of equipment and method. No attempt will here be made to evaluate the tests which have been described, as none of them are in a sufficiently advanced stage of development to warrant such an analysis. It is the purpose of the authors, however, to review the progress of their own investigations to date and to present a few test results and tentative conclusions which may be of interest to paving engineers, contractors and asphalt technologists. Their first efforts were confined to devising a method for determining the relative stability of mineral aggregates, such as used in the construction of sheet asphalt pavements, without reference to the asphalt cementing medium. This was done for the purpose of eliminating as many variables as possible, but the test itself was developed with the primary object of later adapting it for use in testing samples of complete asphalt paving mixtures. Such tests were therefore made with the same general type of apparatus and in the manner just described, except that no temperature control was found necessary. The mineral aggregate was first mixed with castor oil, an aggultinating but non-cementitious agent, in order to prevent segregation of the mineral particles. The mixture was then compressed in a 2-inch-diameter cylindrical mold with a circular orifice in the bottom which was closed with a metal plug. After the sample had been thoroughly compressed the plug was removed from the orifice and a gradually increasing load applied to the upper surface of the test specimen by means of a close-fitting plunger. The maximum load developed in forcing the mixture through the orifice was taken as a measure of the stability of the mineral aggregate. It was found that this test would produce very close check results, with a variation well within 5 per cent of the average. Appreciable differences in stability values were also obtained with sands of different mesh composition, and of the same mesh composition, but differing in shape and texture of grain. With every sand tested a marked increase in stability was produced by the addition of mineral filler in increasing amounts up to 30 per cent. The early results which had been obtained were presented in a paper at the Third Asphalt Paving Conference at Louisville, Ky., and therefore need not be given in detail in the present paper. Three. rather important indications were, however, obtained which while not yet conclusively confined by later work, still appear to be true. These are as follows: 1. A sand of supposedly inferior grading and showing a relatively low stability value alone or combined with a small percentage of mineral filler may, when combined with a larger percentage of filler, develop a higher stability value than a sand of supposedly better grading combined with the same percentage of filler. Thus alone, or with a small percentage of filler, the commonly accepted ideal light traffic sand developed less stability than the ideal heavy traffic sand, but when over 12 per cent of a commercial limestone filler was added to both, the light traffic sand gave higher stability values than the heavy traffic sand. 2. Stability values obtained from a number of sands of widely varying mesh composition mixed with 15 per cent or more of commercial limestone filler strongly indicate that our present specification limits for grading of sheet asphalt sands, while safe, are unnecessarily restrictive, and that many sands which are now eliminated from use may be made to produce good stable paving mixtures. 3. The general trend of test results obtained upon mineral aggregates mixed with castor oil are confirmed with mixtures of the same mineral aggregates with asphalt cement. This is important because in any study of the effect of variations in grading of mineral aggregate the castor oil mixtures may be much more quickly prepared and tested than the asphalt mixtures, and are free from a number of important variable factors which are introduced by the presence of asphalt cement. * Municipal and County Engineering, December, 1924. |