TESTS OF PLAIN AND REINFORCED CONCRETE MADE WITH HAYDITE AGGREGATES I. INTRODUCTION 1. Introductory Statement.-New structural problems have come into being through the ever increasing length of bridge spans and height of buildings and through frequent necessity of adding to existing structures. The solution of these problems has led to the development of numerous light-weight concretes, some of these being obtained by the use of light-weight aggregates. The investigation reported herein was made to secure information regarding one of these materials, Haydite, and the work was further limited to a study of the structural properties of poured Haydite concrete as distin- guished from the dry-tamped Haydite mixtures used in the manufac- ture of concrete building blocks, tile, pipe, and other pre-cast units. Haydite is the material produced by burning shale in a rotary kiln to the point of incipient fusion (about 2000 deg. F.), whereupon the oxidation of its carbon content forms gases and causes the material to expand into a light-weight cellular clinker. This clinker is crushed and screened to produce the gradations desired for concrete making. The manufacturing process was developed and patented by Mr. S. J. Hayde in 1918, and the material has gradually come into general use. Although material of this type was used during the war in building reinforced concrete ships, its use since has been confined largely to the shop manufacture of concrete products. During the last few years there has been a considerable develop- ment of the use of reinforced Haydite for bridge floors and for build- ings, and this has led to a demand for information as to desirable proportions of mixture to produce a given weight, strength, and workability, as well as for data on the elastic properties of the material and on its action in structural members. This investigation was planned to secure some of the information needed. Progress reports have already been published* by the authors. Throughout the investigation the tests of Haydite concrete have been paralleled by tests of sand and gravel concrete, to furnish com- parisons of the properties found and to give an idea of the uniformity of the materials, test methods, and workmanship employed. *"Tests of Plain and Reinforced Haydite Concrete," Proc. A.S.T.M., Vol. 30, Part II, p. 674, 1930. "Construction and Design Features of Haydite Concrete," Jour. Am. Cone. Inst., October, 1930. "Tests of Bonding Floor Finish to Slabs of Haydite and Gravel Concrete," Jour. Am. Cone. Inst., December, 1930. ILLINOIS ENGINEERING EXPERIMENT STATION TABLE 1 OUTLINE OF PRINCIPAL TESTS "All-Haydite" concrete was made with fine and coarse Haydite. "C-Haydite" concrete was made with coarse "C" Haydite and natural sand. Gravel and limestone concretes were made with natural sand as fine aggregate. Standard 6 by 12-in. compression test cylinders were used. Pull-out specimens, 2-ft. bar embedded in 8 by 8-in. cylinder. Series Properties Types of Number and Kind of No. Studied Concrete Used Test Specimens A-2 Properties of Plain Concrete All-Haydite 162 Cylinders A-3 Properties of Plain Concrete All-Haydite 162 Cylinders G-2 Properties of Plain Concrete Gravel 162 Cylinders S Properties of Plain Concrete C-Haydite 162 Cylinders 1 Properties of Reinforced Beams of Ordinary All-Haydite 32 Beams, 6 by 12-in., 8 ft. long Design C-Haydite 192 Cylinders Gravel Limestone B Bond Resistance All-Haydite 18 Beams, 8 by 12-in., 6 ft. long C-Haydite 27 Pull-out Specimens Gravel 54 Cylinders D.T. Diagonal Tension Resistance All-Haydite 24 Beams, 8 by 24-in., 8 ft. long C-Haydite 72 Cylinders Gravel C Column Strength All-Haydite 30 Columns, 8-in. diam., 5 ft. long C-Haydite 90 Cylinders Gravel X Bonding of Floor Finish to Base All-Haydite 12 Beams, 12 by 5 in., 4 ft. long Gravel 12 Cylinders T Effect of Age on Compressive Strength All-Haydite 72 Cylinders C-Haydite la Effect of Age on Compressive Strength C-Haydite 30 Cylinders Gravel 2. Acknowledgment.-The Investigation of Concrete made with Light Aggregates was conducted as a research project of the Engi- neering Experiment Station under a cooperative agreement with the WESTERN BRICK COMPANY, Danville, Illinois, manufacturers of Hay- dite aggregates. The Western Brick Company was represented in the outlining and planning of the investigation by an advisory com- mittee, consisting of Mr. F. W. BUTTERWORTH, President, Mr. I. N. DOUGHTY, Manager of Research, and Mr. FRANK PAYNE, Engineer. The investigation was conducted under the administrative direc- tion of DEAN M. S. KETCHUM, Director of the Engineering Experi- ment Station, and PROF. M. L. ENGER, Head of the Department of Theoretical and Applied Mechanics. 3. Outline and Scope of Tests.-The investigation may be divided into two groups of tests: (1) those concerned with the properties of the Haydite aggregates themselves, and (2) those concerned with the properties of concrete made with these aggregates. In the first group TESTS ON PLAIN AND REINFORCED CONCRETE 9 TABLE 2 UNIT WEIGHT AND MOISTURE CONTENT OF AGGREGATES Lot No. C-1.......... No. C-2.......... No. C-3.......... No. C-4.......... No. C-6.......... No. C-7.......... No. C-8.......... No. C-9.......... No. A-1.......... No. A-2t........ No. A-3......... No. A-4.......... No. A-5......... No. A-6.......... No. G-1 ......... No. G-2 . ...... No. G-3......... No. L-l.......... No. S-1.......... No. S-2 .......... No. S-3.......... No. S-4.......... Unit Weight, lb. per cu. ft. Kind of Aggregate Coarse Haydite Coarse Haydite Coarse Haydite Coarse Haydite Coarse Haydite Coarse Haydite Coarse Haydite Coarse Haydite Fine Haydite Fine Haydite Fine Haydite Fine Haydite Fine Haydite Fine Haydite Gravel Gravel Gravel Limestone Natural Sand Natural Sand Natural Sand Natural Sand Aggregate Used in Series B X 1, A-2, A-3, S, B, X, T D.T. C. Absorption A-2 A-3, B, T X, 1, D.T., C Absorption Absorption 1 S, G-2, B, T C, D.T., X 1 1 G-2, S, T, B C, D.T. X Loose Moist 46 45 5i *A.S.T.M. Standard Method of Test for Unit Weight of Aggregate for Concrete (C29-27). tThis lot from outdoor stockpile. are microscopic studies and determinations of unit weight, gradation, soundness, absorption, specific gravity, and moisture content. These are discussed in Chapter II. In the second group are the tests out- lined in Table 1 and discussed in Chapters III to VIII. While the tests listed in Table 1 do not include the determination of properties such as permeability, shrinkage, and plastic flow, it is believed that the most fundamental properties of the material have been covered. All of the Haydite tested was the product of a single plant, that of the Western Brick Company at Danville, Illinois, and the test data and discussion to follow apply particularly to this material. II. PROPERTIES OF CONCRETE MATERIALS 4. General Description of Materials.-The Haydite aggregates used were the fine or "A" size, passing a e16-in. round-hole screen, and the coarse or "C" size, passing a 34-in. screen and retained on a %2-in. screen. Shipments of aggregates were received at intervals during the investigation and are designated by lot numbers in Table 2 and in the tabulations of test results. A summary of data on unit weight, moisture content, and sieve analysis of these aggregates is Loose Dry 42 43 41 45 45 44 42 50 55 54 56 95 100 88 105 107 105 105 Rodded Dry* 46 47 45 49 48 48 46 44 60 64 57 106 102 108 98 112 113 111 Moisture Content when Re- ceived, per cent by weight 7.6 5.8 3.3 5.9 6.0 6.3 4.5 9.8 27. Ot 2.9 8.3 ILLINOIS ENGINEERING EXPERIMENT STATION TABLE 3 SIEVE ANALYSES OF AGGREGATES Kind of Aggregate Coarse Haydite... Coarse Haydite.... Coarse Haydite.... Coarse Haydite.... Coarse Haydite.... Coarse Haydite.... Coarse Haydite.... Coarse Haydite.... Fine Haydite...... Fine Haydite ...... Fine Haydite...... Fine Haydite...... Fine Haydite...... Fine Haydite...... Gravel. .......... Gravel ........... Gravel ........... Limestone ........ Natural Sand..... Natural Sand..... Natural Sand ..... Natural Sand .... Per cent by Weight Retained on Tyler Sieve No. 16 30 50 100 8 TABLE 4 TENSION TESTS OF REINFORCING STEEL Each figure represents the average of 3 or more tests. Used in Yield Point Ultimate Strength Elongation in Reduction of Size Series No. lb. per sq. in. lb. per sq. in. 8 in., per cent Area, per cent Y/ in. Round ............ 1 39 100 70 000 23 51 IV1 in. Square ........... B 63 100 110 200 15 11 Sin. Round ............ B 48 000 72 700 23 34 Y in. Square ............ B 34 000 54 900 29 40 1 in. Round ............. D.T. 37 600 58 900 31 35 j in. Round ............ D.T. 52 800 83 900 18 41 M in. Round ............ Col. 54 900 86 100 18 45 6 in. Square ............ Col. 48 300 78 000 21 51 No. 5 Wire ............. Col. 49 400 81 800 15 62 given in Tables 2 and 3. Special tests of Haydite aggregates are described in Sections 5, 6, and 7. Gravel, limestone, and sand were used in tests to parallel those made with Haydite aggregates. The gravel and most of the sand were from pits on the Wabash River at Covington, Ind. For the adhesion tests described in Chapter VIII, a rather fine sand from Lincoln, Illinois, was used. The broken limestone came from Kan- kakee, Illinois. Properties of these aggregates are given in Tables 2 and 3. Two lots of Universal Portland cement were used, though nearly all of the tests were made from the first lot. Average tensile strengths Fineness Modulus 5.89 6.30 6.00 6.30 3.01 2.80 2.81 2.59 2.33 2.12 6.24 6.13 6.12 5.92 3.07 3.07 3.11 2.70 TESTS ON PLAIN AND REINFORCED CONCRETE of 1:3 standard Ottawa sand mortar briquettes in pounds per square inch are as follows: at 7 days, Lot 1, 260; Lot 2, 290; at 28 days, Lot 1, 370; Lot 2, 390. Properties of the reinforcing steel used are given in Table 4. 5. Absorption and Specific Gravity.-(a) Preliminary Tests. In studies involving concrete voids or water-cement ratio it is necessary to determine the absorptive properties, and in some cases the specific gravity, of the aggregates. The absorption of water, a minor factor with ordinary aggregates, is important with Haydite aggregates, in which the absorptive capacity is appreciable in comparison to the amount of mixing water used. A preliminary series of specific gravity tests was made on Haydite aggregates for use in connection with strength tests. It was found that the apparent specific gravity varied widely, increasing with fine- ness of particle. Absorption tests made with the same samples showed that in a given period of time the fine Haydite absorbed con- siderably more water than the coarse material. The tests were made on dry material and also on samples of material having various initial moisture contents. The procedure used with the coarse aggregate was the usual immersion method,* extended to permit determinations of absorption as well as specific gravity. Briefly, the method requires that a given weight of dry material shall be immersed in water for 24 hours, surface-dried, and re-weighed, then weighed immersed. The weight of dry material, divided by the difference in weight of the surface-dried material in air and in water, gives the apparent specific gravity. While the period of immersion should have no effect on the apparent specific gravity, the 24-hour period was used. Intermediate readings were taken to determine the absorption at definite intervals. The absorption was taken as the difference between the weight of the dry sample and that of the soaked, surface-dried sample. The mate- rial was submerged in a closed wire basket to keep the lighter particles from floating, and was surface-dried by rolling it between paper towels until a change in color was noted. The foregoing method could not be used with the fine Haydite due to difficulty in immersing the sample without entrained air. The method used was to place a soaked sample in a centrifuge which was used to remove surface water. The criterion of dryness was the ability of the material to run freely through the fingers. The amount of *"Standard Method of Test for Apparent Specific Gravity of Coarse Aggregates," (D 30-18). A.S.T.M. Book of Standards, Part II, 1930. Following the terminology of this standard, the property measured has been called "apparent specific gravity" throughout the bulletin. It would be more accurate to call it "bulk specific gravity," since what is determined is the weight per unit of bulk or overall volume of aggregate, including both permeable and impermeable pore space. Due to the current use of this A.S.T.M. standard, its nomenclature has been followed to avoid confusion. ILLINOIS ENGINEERING EXPERIMENT STATION K T/I;e 1,7 110o4rS TI;77e //? Da'ys Fla. 1. ABSORPTION OF WATER BY INITIALLY DRY HAYDITE AGGREGATES absorbed water was found by oven-drying to constant weight, at about 200 deg. F. With the absorption known, the specific gravity was easily determined by use of the Le Chatelier flask. Figure 1 shows the absorption determined for dry fine and coarse Haydite for various periods of immersion. Figure 2 shows similar data for both dry .and initially moist materials for various immersion periods. It is clear that Haydite does not have a fixed amount of absorption in a given time, as usually assumed for ordinary aggre- gates, but that the amount depends upon the initial moisture content, at the beginning of the absorption test. Aggregates containing mois- ture at the beginning of the test show a larger total absorption in a given time than initially dry ones, especially in the case of the coarse material. The fine material reaches its absorptive capacity more quickly than the coarse and is less affected by initial moisture conditions. There is little information available regarding the period of time that aggregates should be immersed in water to produce an absorption equal to that taking place during the mixing and placing of concrete. Periods varying from 15 minutes to 3 hours have been assumed by various investigators, without great difference in result when ordinary aggregates were used. Observation of fresh concrete in forms indi- TESTS ON PLAIN AND REINFORCED CONCRETE /1r't/A7/ Mo/sture Co7f'e/l /, , Percentag-7 e of 0rL, We/g' FIG. 2. ABSORPTION OF WATER BY HAYDITE AGGREGATES HAVING VARYING INITIAL MOISTURE CONTENTS cates that in so far as water content fixes the voids in the concrete, the mass takes its final volume (neglecting subsequent shrinkage of the order of 0.1 to 0.2 per cent of volume) within 15 minutes after the concrete is placed. Changes in water-cement ratio due to further absorption by the aggregates may affect the process of hydration and thus affect the strength. An estimate of this effect has led to the use in this investigation of an absorption allowance corresponding to a one-hour immersion test. Values for one-hour absorption for material with various initial moisture contents are shown by the heavy curves of Fig. 2. It will be noted that Fig. 2 refers to initial absorbed mois- ture, not total moisture content. The proportions of absorbed and free moisture in a sample will depend entirely upon how the moisture content has been taken on. If a wet sample is drying out the amount of absorbed moisture will be high; a sample just wet down will contain mostly free surface water. (b) Later Tests. The foregoing preliminary tests were followed by a rather careful study of both absorption and specific gravity. The aggregates used were fine and coarse Haydites from composite sam- ples consisting of a mixture of lots taken daily from the plant output ILLINOIS ENGINEERING EXPERIMENT STATION over a period of two weeks. The material was oven-dried and sepa- rated into sizes by the use of Tyler sieves. The sieves used were the 3%, Y2, and 3%-in., Nos. 4, 8, 16, 30, 50, 100, and 200. For the largest three separation sizes, 3-2-in., 12-38-in., and 38-in.-No. 4, it was possible to use the same method for specific gravity and absorption determinations as was used in the preliminary tests. After one absorption determination was made on a sample, further absorption was found by successive weighings of the submerged material. In studying the fine Haydite, a number of methods were tried. The problem of determining the absorption of water by fine Haydite is not difficult when the specific gravity is known. To determine the specific gravity with the apparatus used for sand, such as the Le Chatelier or Chapman flask or the Jackson apparatus, a non- absorbent sample or one that contains absorbed water but no surface water is required. Fine Haydite in some sizes will absorb up to 14 per cent by weight in three minutes, and it is difficult to surface-dry such a sample without losing absorbed water. The fact that some particles float, even in kerosene, complicates the problem. The Rea method,* in which the aggregate was allowed to absorb kerosene (or, in a modification of the method, colored gasoline) and immersed in water to remove the surface coating, did not give consistent results. Neither did a method of determining the volume of the aggregate by displacement in a dry, powdered material. The Pearsont method of detection of surface moisture by the ability of the aggregate to stick to the sides of a glass jar was tried with the ordinary gradation of fine Haydite. Using the method of adding water by small increments until particles adhered to the sides of a Mason jar gave unsatisfactory results. The method produces a variable time interval during which absorption takes place. There was found to be no definite critical point at which material first began to adhere to the sides of the vessel, the finer particles adhering first. A modification of Pearson's method was applied to material be- tween the No. 8 and No. 16 sieves. A weighed quantity of dry aggregate was placed in a calibrated specific gravity bottle of the bulb type having a small neck and a ground glass stopper with a capillary bore. Alcohol was added until the aggregate was covered, and ab- sorption of the alcohol was permitted for about 24 hours. Excess alcohol was then drained off and the surface alcohol evaporated by introducing a stream of air into the neck of the bottle. Drying was continued until only a few grains of Haydite adhered to the sides of *A. S. Rea, "Apparent Specific Gravity of Non-Homogeneous Fine Aggregates," Proc. A.S.T.M., Vol. XVII, Part II, 1917. tJ. C. Pearson, "A Simple Titration Method for Determining the Absorption of Fine Aggre- gates," Rock Products, Vol. 32, p. 64, May 11, 1929. TESTS ON PLAIN AND REINFORCED CONCRETE the bottle when it was rolled. A weighing was made immediately. The bottle was then filled with water, trapped air was released by rolling, and another weighing was made. The apparent specific gravity was calculated from the dry weight of the aggregate and the volume displaced by the aggregate with its pores filled with alcohol. Alcohol was used in place of water in this procedure in order to bring the time required for surface drying down to a reasonable figure, about 12 to 15 minutes being required. All weighing was done on an analytical balance. The specific gravity found for the No. 8-16 material by this method was 1.24, three samples of four giving results of 1.24, and the fourth giving 1.25. An unsatisfactory feature of the method was the cooling of the inside of the bottle due to the rapid evaporation of the alcohol, affecting the adhesion of particles to the sides of the bottle. Further, the contact of water and alcohol produced heat soon after the water had been added. For material between the No. 16 and No. 30 sieves this method gave specific gravities varying from 1.26 to 1.39 and averaging 1.32. For finer material the method did not give satisfactory results. As a check on the value of apparent specific gravity found for the No. 16-30 size aggregate, a weighed sample was placed in a specific gravity bottle as before. After adding and stirring into the mass a few drops of water, the aggregate was inspected through the neck of the bottle under a microscope. Such inspection was made after each successive addition of water until surface water was visible. When the pore spaces had been filled, as indicated by the presence of surface water, the bottle with its aggregate was weighed, filled with water, and weighed again as in the previous method. Two samples of No. 16-30 material showed apparent specific gravities of 1.31 and 1.33 by this method, confirming the value of 1.32 found previously. This microscopic method, with the substitution of alcohol for water to fill the pores, was applied to the material between the No. 30 and No. 50 mesh sieves. Apparent specific gravities of 1.41 and 1.49 were obtained for two samples. Increments of absorption for fine aggregates beyond the first three minutes were obtained by means of the specific gravity bottle. This was done simply by releasing air bubbles, bringing the water level back to the top of the capillary tube, and weighing at each time interval. Weighings were made at 3, 15, and 30 min., 1, 3, and 24 hours, and at 4, 7, 14, 21, and 28 days, except where the increases became negligible. With the apparent specific gravity known, the absorption that took place during the first time interval of three minutes, as well as for succeeding intervals, was determined. This ILLINOIS ENGINEERING EXPERIMENT STATION known rate of absorption for increments of time after the first 3 minutes was useful in checking the reliability of the results found with the material finer than the No. 100 and 200 sieves. With these fine sizes the surface drying and removal of entrapped air in the specific gravity bottle were rather difficult, and the results found are probably less reliable than those for the larger sizes. The values given, however, are consistent among themselves. The material passing the No. 200 sieve, with 1 per cent absorption in 3 minutes, had a specific gravity of 2.50. Had there been no absorption the specific gravity would have reached an upper limiting value of 2.55 for this material. The apparent specific gravity and the absorption for the various sizes of Haydite aggregate are shown by curves in Fig. 3. The curves for material finer than the No. 100 size are shown by dotted lines to indicate that the data for these tests are not very definitely estab- lished. The shape of the absorption curves may be explained by the fact that while the porosities of the largest 5 or 6 sizes are nearly the same (as indicated by the curve for apparent specific gravity), the finer sizes have more surface area exposed for a given weight and so may be expected to permit a more rapid absorption of water. Figure 3 shows increasing absorption with decrease in particle size until the No. 16-30 or 30-50 sizes are reached, when the absorption begins to decrease rapidly. This decrease is evidently due to the decrease in porosity of the finer particles. In addition to the data of the foregoing tests, a curve is shown in Fig. 3 for 3-minute absorption, based on tests using a centrifuge to surface-dry the soaked material. The Haydite sample, after 3- minute immersion in water, was placed in a centrifuge and subjected to a force about 250 times its weight for 11 minutes or more. This was sufficient to surface-dry all material retained on the No. 50 sieve, though there is some uncertainty as to loss of absorbed water. How- ever, the two curves for 3-minute absorption agree fairly well and to some degree substantiate one another. The apparent specific gravity varies from 1.17 for the coarse material to 2.50 for the finest. If the absolute specific gravity is taken at 2.55 this indicates a porosity as great as 54 per cent for the Y8-y2-in. material. The data of Fig. 3 were collected with the expectation that the absorption of any Haydite sample could be computed from its sieve analysis as the summation of the weights of each size times the corre- sponding absorption percentage from the figure. To test this scheme various artificial mixtures of fine aggregates were made up and their TESTS ON PLAIN AND REINFORCED CONCRETE 17 N. K U) 'ZPC4 Size of A#gregate FIG. 3. VARIATION IN ABSORPTION AND SPECIFIC GRAVITY WITH SIZE OF HAYDITE AGGREGATE L ILLINOIS ENGINEERING EXPERIMENT STATION TABLE 5 ABSORPTION BY COMBINED FINE AGGREGATES Absorption is given in per cent, by weight. Particle size is indicated by sieves between which the material lies. Per cent Absorption Percentage of Sample of Given Size Per 3 orption Sample No. 4-8 8-16 16-30 30-50 50-100 100-200 200-0 Calculated By Test .................. .. 32 42 26 .. .. .. 14.4 14.4 2 ................ .. 62 .. 38 .. .. .. 14.0 12.9 3..... ......... . .. .. 61 39 .. 14.8 14.7 4.................. 27 40 18 15 .. 11.7 9.4 5 .................. 70 .. .. .. 30 .. .. 10.3 8.3 6 ..................2 .. .. 28 12.3 10.0 7................ 3 24 21 14 10 7 21 10.5 5.6 absorptions found experimentally by use of the specific gravity bottle. Since the fine material was difficult to handle and small weights were used, weighings were made after the material had been placed in the bottle; consequently, each experimental value represents a single test. Absorptions found for combined materials are given in Table 5. For combinations of the coarser particles there is good agreement between calculated and experimental values but a divergence appears whei very fine material is included. Keeping the value of absorption and specific gravity consistent for the finer sizes and adjusting them to lower absorptions will not account for the divergence. The only explanation seems to be a retarding of the penetration of water when the particles of a graded fine aggregate are in close contact as occurs in an inundated mass. In the case of an ordinary gradation of the fine aggregate, an appreciable time was required for the water to percolate through the voids and coat the surface of the particles, while for an aggregate of a single size the wetting was almost instan- taneous. While all of the foregoing tests presage considerable diffi- culty in determining accurately the absorption of water by Haydite in a concrete mixture, they do serve to emphasize the importance of wetting the aggregates considerably in advance of using. The greater the portion of the absorptive capacity of the aggregate that has been satisfied (before mixing) the less the later absorption becomes and the more nearly the true absorption can be predicted. Other advantages of "pre-wetting" the aggregates will be noted in sections to follow. 6. Soundness.-In a study of the soundness and durability of Haydite aggregates, several lots of Haydite, gravel, and limestone were subjected to the sodium sulphate test.* This test, intended to *Proposed Method of Test for Soundness of Fine and Coarse Aggregates by Use of Sodium Sulphate. National Sand and Gravel Bulletin, February 1931. (a) Material between No. 4 and No. 8 Sieves (x2) (b) Material between No. 30 and No. 50 Sieves (x25) FIG. 4. MICROGRAPHS OF HAYDITE AGGREGATES (a) Material passing No. 200 sieve (x250) (b) Natural size FIG. 5. MICROGRAPH OF HAYDITE AGGREGATES AND VIEW OF SECTION THROUGH HAYDITE CONCRETE TESTS ON PLAIN AND REINFORCED CONCRETE produce an effect similar to freezing and thawing, consists of soaking the material in a saturated solution of sodium sulphate at 70 deg. F. for about 19 hours, then drying at 212 to 230 deg. F. for 4 hours and cooling for one hour. The formation of crystals of sodium sulphate within the pores of the aggregate produces a bursting effect similar to that in a freezing test. Five of these 24-hour cycles were employed. The Haydite aggregates passed the test satisfactorily. Samples of No. 8-4, No. 4-3/8-in. and Y3-y-in. sizes; 250 grams each, when rescreened after the test showed a loss of material through the finer sieve of 2.5 to 6 per cent. Of three samples of the larger size, 12-3/-in., in which the number of pieces before and after the test was recorded, in one case one piece out of 102 split in two, in the second, three out of 120 split up, and in the third there was no change. Compared with the behavior of the other materials tested, the Haydite aggregates gave results as good as those for gravel and better than those for lime- stone, although the latter is a good durable aggregate which has been used in concrete in this locality for many years. These tests raise no question as to the soundness of these Haydite aggregates. 7. Microscopic Studies of Haydite Particles.-A microscopic exam- ination of Haydite particles of various sizes furnished interesting information as to the pore size and the shape and surface texture of particles. Figure 4a shows a view of a sample of No. 8-4 material, magnified to twice natural size. This view gives a good idea of the porous structure and indicates a considerable range in quality from the denser to the more porous particles. The particles have the edges rounded off to a considerable extent by the crushing and grinding process of manufacture, only the very porous ones retaining decidedly irregular shapes. This is also true of the No. 4-38-in. and %-4-in. material, which appears quite similar, except as to size, to that shown in Fig. 4a. Figure 4b is a micrograph of No. 30-50 material, magnified 25 times. It shows very clearly the irregular shape and rough, porous texture of the particles of this size. Here again a considerable range in quality of material is seen. The glassy nature of the partially fused material and the sharp irregular projections typical of the fine Hay- dite particles are apparent. It will be noted that while the average particle diameter, based on the limiting sieve sizes, is less than 0.02 in., the size of visible pores and surface cavities is relatively very much smaller. Figure 5a is a micrograph of material passing the No. 200 sieve (width of square opening, 0.0029 in.), magnified 250 times. This ILLINOIS ENGINEERING EXPERIMENT STATION TABLE 6 TESTS OF GRAVEL CONCRETE, SERIES G-2 Aggregates used: Sand, S-2; gravel, G-2. See Tables 2 and 3. Water allowed for absorption: sand and gravel, 1 per cent by weight. Proportions by loose dry volume; dry aggregates used. Quantities required per cubic yard of concrete based on fresh concrete. No allowance has been made for settlement and waste. All values given are the average of 3 tests. Mix 1:1J :2 ½ .. 1:2:2 ...... 1:22:1%.. 1:2:3 ...... 1:2M:22.. 1:3:2 ...... 1:2%:3%.. 1:3:3 ...... 1:32:2%.. { Slump, in. material exhibits properties not observable in the larger sizes. It consists almost entirely of translucent, glassy particles of irregular shape, with many of exceedingly small size, probably less than 0.0002 in. in diameter. Many of the finest particles appear drawn together in clusters. While the surfaces of these particles can not be clearly seen, the material does not appear to have any appreciable amount of closed pore space, or space not instantly accessible to water. This is indicated by the translucency of the material and is in agreement with the results of the absorption studies of Section 5 from which it Flow 7-Day Cylinders Corn- Modulus pressive of Elas- Strength, ticity, lb. per lb. per sq. in. sq. in. 3680 3 460 000 2600 3 260 000 2020 2 660 000 3630 3 850 000 2770 3 170 000 2140 3 150 000 3410 3 620 000 2860 3 170 000 2120 2 840 000 2880 3 430 000 2170 3 320 000 1660 2 310 000 2740 3 290 000 2220 2 500 000 1730 2 330 000 2200 2 860 000 1850 2 690 000 1420 2 190 000 1860 2 890 000 1290 2 360 000 1090 1 900 000 1840 2 890 000 1390 2 530 000 1100 1 850 000 1430 2 360 000 1210 2 450 000 990 1 920 000 3.4 7.7 8.0 Unit 'Weight of Fresh Concrete, lb. per cu. ft. 148.3 148.3 146.2 147.6 147.6 148.1 145.9 144.3 143.8 148.1 147.3 146.3 147.4 146.8 146.6 145.3 144.5 143.7 147.3 147.4 144.3 147.7 144.7 145.3 143.7 144.0 142.8 1.1 1.2 1.3 28-Day Cylinders Corn- Modulus pressive of Elas- Strength, ticity, lb. per lb. per sq. in. sq. in. 5400 4 310 000 4020 3 850 000 3580 3 460 000 4600 4 140 000 4420 3 890 000 3580 3 710 000 4480 4 000 000 4410 3 770 000 3590 3 430 000 4390 4 060 000 3790 3 720 000 3090 3 330 000 3990 3 950 000 3830 4 030 000 3190 3 440 000 3740 3 750 000 3390 3 500 000 2770 2 970 000 3080 3 800 000 2740 3 460 000 2290 3 140 000 3190 3 650 000 2710 3 530 000 2210 2 900 000 2570 3 160 000 2160 2 880 000 2020 2 970 000 Quantities Required per Cubic Yard of Concrete Ce- Sand, Gravel, ment, cu. yd. cu. yd. bbl. loose, loose, dry dry 1.85 0.41 0.69 1.83 0.41 0.68 1.79 0.40 0.66 1.83 0.54 0.54 1.81 0.53 0.53 1.79 0.53 0.53 1.79 0.66 0.40 1.75 0.65 0.39 1.72 0.64 0.38 1.54 0.46 0.68 1.52 0.45 0.68 1.50 0.44 0.66 1.52 0.56 0.56 1.50 0.56 0.56 1.49 0.55 0.55 1.47 0.65 0.44 1.45 0.65 0.43 1.43 0.64 0.42 1.31 0.48 0.68 1.30 0.48 0.67 1.26 0.47 0.65 1.30 0.58 0.58 1.26 0.56 0.56 1.26 0.56 0.56 1.25 0.65 0.46 1.24 0.64 0.46 1.22 0.63 0.45 TESTS ON PLAIN AND REINFORCED CONCRETE TABLE 7 TESTS OF C-HAYDITE CONCRETE, SERIES S Aggregates used: Sand, S-2; coarse Haydite, C-6. See Tables 2 and 3. Proportions by loose dry volume. Dry aggregates used. Water allowed for absorption: Sand, 1 per cent: coarse Haydite, 7 per cent. Quantities required per cubic yard of concrete based on fresh concrete. No allowance has been made for settlement and waste. All values given are the average of 3 tests. 28- Cyl Com- pressive Strength, lb. per sq. in. 4300 3810 2940 4370 3960 3190 4150 4030 3430 3640 3110 2530 3490 3030 2700 3190 2590 2360 2530 2050 1850 2620 2100 1920 2180 1880 1440 Day inders Modulus of Elas- ticity, lb. per sq. in. 2 610 000 2 440 000 2 370 000 3 010 000 2 780 000 2 600 000 3 040 000 3 020 000 2 660 000 2 600 000 2 370 000 2 150 000 2 680 000 2 530 000 2 310 000 2 670 000 2 470 000 2 320 000 2 250 000 2 210 000 2 000 000 2 450 000 2 220 000 2 250 000 2 470 000 2 320 000 2 210 000 7-Day Cylinders was concluded that the 0-No. 200 material had practically no absorption. These micrographs give some quantitative data regarding the size of particle and pore space; they also show the transition in shape and texture from the porous, partially rounded, dull-gray particles of the larger sizes to the extremely irregular, glassy, translucent material of the smallest sizes. Presumably the large particles are made up of the samematerial as the fine ones, but do not show the fused, glassy structure without magnification. Quantities Required per Cubic Yard of Concrete Coarse Sand, Haydite Ce- cu. yd. C-6, ment, loose, cu. yd. bbl. dry loose, dry 1.83 0.41 0.68 1.79 0.40 0.66 1.75 0.39 0.65 1.81 0.54 0.54 1.78 0.53 0.53 1.73 0.51 0.51 1.77 0.66 0.39 1.74 0.64 0.39 1.71 0.63 0.38 1.51 0.45 0.67 1.50 0.44 0.67 1.47 0.44 0.65 1.50 0.56 0.56 1.48 0.55 0.55 1.43 0.53 0.53 1.44 0.64 0.43 1.44 0.64 0.43 1.41 0.63 0.42 1.29 0.48 0.67 1.26 0.46 0.65 1.24 0.46 0.64 1.27 0.56 0.56 1.25 0.56 0.56 1.24 0.55 0.55 1.24 0.64 0.46 1.22 0.63 0.45 1.21 0.63 0.45 Slump, in. Flow Mix 1:12:2½ 1:2:2.... 1:2M:1% 1:2:3 .... 1:2A:2% 1:3:2.... 1:23:3M 1:3:3.... 1:3M:2½ { Water- Cement Ratio, by Volume 0.77 0.87 0.97 0.76 0.86 0.96 0.74 0.84 0.94 0.89 0.99 1.09 0.87 0.97 1.07 0.96 1.06 1.16 1.10 1.20 1.30 1.09 1.19 1.29 1.17 1.27 1.37 Unit Weight of Fresh Concrete, lb. per cu. ft. 114.6 113.6 112.7 121.2 120.5 119.0 126.2 125.3 124.8 113.3 113.9 113.0 119.0 118.5 116.3 122.3 123.0 121.7 113.9 112.0 112.3 117.5 117.4 117.2 120.8 120.8 121.2 Com- pressive Strength, lb. per sq. in. 2700 2150 1430 2820 2310 1660 2810 2470 1890 1950 1500 1170 2110 1610 1290 1710 1370 1150 1240 950 800 1300 980 920 1100 890 770 Modulus of Elas- ticity, lb. per sq. in. 2 120 000 1 760 000 1 780 000 2 520 000 2 090 000 1 930 000 2 460 000 2 330 000 2 050 000 2 030 000 1 730 000 1 600 000 2 250 000 1 780 000 1 780 000 2 070 000 1 850 000 1 710 000 1 480 000 1 320 000 1 070 000 1 690 000 1 470 000 1 410 000 1 540 000 1 270 000 1 410 000 22 ILLINOIS ENGINEERING EXPERIMENT STATION TABLE 8 TESTS OF ALL-HAYDITE CONCRETE, SERIES A-3, MADE WITH DRY AGGREGATES Aggregates used: Fine Haydite, A-3; coarse Haydite, C-6. See Tables 2 and 3. Proportions by loose dry volume. Dry aggregates used. Water allowed for absorption: Fine, 14 per cent of dry weight; coarse, 7 per cent of dry weight. Quantities required per cubic yard of concrete based on fresh concrete. No allowance has been made for settlement and waste. All values given are the average of 3 tests. 28-Day Cylinders Mix 1:1 :2 j 1 1:2:2..... 1:2½:1% { 1:2:3..... 1:2%:2% { 1:3:2..... 1:2M:3% 1:3:3..... 1:3%:2% { Water- Cement, Ratio, by Volume 0.86 0.96 1.06 0.88 0.98 1.08 0.90 1.00 1.10 1.01 1.11 1.21 1.03 1.13 1.23 1.14 1.24 1.34 1.25 1.35 1.45 1.27 1.37 1.47 1.39 1.49 1.59 7-Day Cylinders Another idea of the structure of aggregate particles may be ob- tained from Fig. 5b, which shows an almost plane surface of cleavage through a piece of All-Haydite concrete. Besides showing the full size and shape of particles, the figure shows the porous structure of the Haydite particles. It is to be remembered, however, that all of this material has been crushed and ground to size, so that the larger particles which have withstood the grinding are in general of unusual strength and toughness. 8lump, in. 5.4 6.6 9.6 5.5 7.5 9.9 3.1 6.6 9.7 1.9 5.3 8.4 1.7 5.4 7.7 3.0 7.2 8.2 1.0 .7.4 7.6 2.7 5.3 6.7 2.7 3.3 6.9 Flow 232 251 300 230 249 300 192 251 300 175 240 300 187 240 285 230 272 300+ 187 250 288 200 243 298 218 232 297 Unit Weight of Fresh Con- crete, lb. per cu. ft. 99.0 96.8 96.3 99.5 97.6 97.1 99.6 98.6 97.7 95.9 95.1 94.8 96.2 96.9 95.7 97.1 96.0 94.8 95.0 94.7 92.6 95.2 95.8 93.0 95.7 94.7 93.9 Quantities Required per Cubic Yard of Concrete Fine Coarse Haydite Haydite Ce- A-3, C-6, ment, cu.yd. cu.yd. bbl. loose, loose, dry dry 1.86 0.41 0.69 1.79 0.40 0.66 1.75 0.39 0.65 1.83 0.54 0.54 1.77 0.52 0.52 1.73 0.51 0.51 1.79 0.66 0.40 1.75 0.65 0.39 1.70 0.63 0.38 1.53 0.45 0.68 1.50 0.44 0.67 1.47 0.44 0.65 1.51 0.56 0.56 1.50 0.56 0.56 1.46 0.54 0.54 1.47 0.66 0.44 1.44 0.64 0.43 1.40 0.62 0.42 1.30 0.48 0.68 1.28 0.48 0.67 1.24 0.46 0.64 1.29 0.57 0.57 1.28 0.57 0.57 1.23 0.55 0.55 1.26 0.65 0.47 1.23 0.64 0.46 1.21 0.62 0.46 Com- pressive Strength, lb. per sq. in. 3700 3060 2440 3470 2890 2490 3320 2770 2150 2530 2300 1920 2620 2310 2060 2350 1960 1310 1580 1680 1310 1760 1620 1280 1470 1220 1110 Modulus of Elas- ticity, lb. per sq. in. 2010000 1 860 000 1800000 1 950 000 1 670 000 1 770 000 1 890 000 1 840 000 1 660 000 1 820 000 1 730 000 1 670 000 1 850000 1 810 000 1 630 000 1 770 000 1 660 000 1 450 000 1 330 000 1 620 000 1 540 000 1 690 000 1 310 000 1570 000 1 550 000 1 630000 1 320 000 Com- pressive Strength, lb. per sq. in. 1990 1500 1160 1880 1460 1210 1850 1470 1060 1370 1160 870 1250 1070 900 1120 870 670 750 760 540 820 710 540 640 560 470 Modulus of Elas- ticity, lb. per sq. in. 164000( 1 500000C 121000( 1 60000( 1480 00( 1320 00C 1530 00( 1 400 00C 1300 00C 1 410 OOC 1 300 OOC 1410 00C 1490 OOC 1 160 OO0 1 200 00C 1 230 OOC 1 140 OOC 1 030 00C 1 240 OOC 1 190 00( 1000 00C 1 320 OO0 1080 00C 970 OOC 970 OOC 1 020 0OC 840 OOC TESTS ON PLAIN AND REINFORCED CONCRETE III. PROPERTIES OF PLAIN HAYDITE CONCRETE 8. Outline of Tests of Plain Concrete.-Experience has shown a considerable saving in weight over that of ordinary concrete through the use of coarse Haydite and natural sand aggregates, and at the same time the harshness found with some mixes containing fine Hay- dite is avoided. In the tests, therefore, two types of Haydite mixture were employed, one using as aggregates natural sand and coarse "C" Haydite and the other using fine "A" Haydite and coarse "C" Hay- dite. As noted in Table 1, reference is made to the first of these as "C-Haydite" concrete and to the second as "All-Haydite" concrete. Several series of compression tests of 6 by 12-in. machine-mixed concrete cylinders were made. Three cylinders of a kind were tested at ages of 7 and 28 days. In addition to compressive strength, other properties of the concrete mixtures were found, such as workability, unit weight, yield, and modulus of elasticity. Five principal series of tests were made as follows: SERIES G-2-This series consisted of 162 cylinders, made with sand and gravel aggregates, for comparison with Series S and A-3. Nine proportions of mix and three consistencies were used. The pro- portions were determined by the use of loose, dry volumes. All cylin- ders were moist cured and tested moist. Test data are given in Table 6. SERIES S-This series was identical with Series G-2 except that the aggregates were natural sand and coarse Haydite. Test data are given in Table 7. SERIES A-3-This series was also identical with Series G-2 except that the aggregates were fine and coarse Haydite. Test data are given in Table 8. SERIES A-2-This series was made with fine and coarse Haydite aggregates. While the preceding three series were proportioned by loose, dry volume, this series was proportioned by loose, moist vol- ume. The fine aggregate carried about 27 per cent of total moisture; the coarse, 6 per cent. Since the aggregates were bulked considerably, the mixes of this series had a considerably higher cement content than the corresponding mixes of the previous three series. Series A-3 and A-2 furnish an extreme range in moisture conditions, the first using oven-dried aggregates and the second a supply of aggregate with an unusually high moisture content. The test data of Series A-2 are given in Table 9. TIME SERIES-This series consisted of five groups of cylinders in each of which the age at time of test varied from 7 days to 1 year. Details of these tests are given in Section 14. ILLINOIS ENGINEERING EXPERIMENT STATION TABLE 9 TESTS OF ALL-HAYDITE CONCRETE, SERIES A-2, MADE WITH MOIST AGGREGATES Aggregates used: Fine Haydite, A-2; coarse Haydite, C-6. See Tables 2 and 3. Moisture present in aggregates when used: C-6, 6 per cent; A-2, 27 per cent of dry weight. Water allowed for absorption: A-2 supplied 7 per cent of dry weight as free water; C-6 required 3 per cent of dry weight additional water. Proportions by loose moist volumes. Quantities required per cubic yard of concrete based on fresh concrete. No allowance has been made for settlement and waste. All values given are the average of 3 tests. 7-Day Unit Cylinders Wailrht Mix 1:1i:2% 1:2:2..... 1:20:10 { 1:2:3..... 1:2%:2M 1:3:2 ..... 1:23:3% 1:3:3 ..... 1:32:2% { Water- Cement Ratio, by Volume 0.77 0.87 0.97 0.77 0.87 0.97 0.78 0.88 0.98 0.89 0.99 1.09 0.90 1.00 1.10 1.00 1.10 1.20 1.10 1.20 1.30 1.11 1.21 1.31 1.22 1.32 1.42 28-Day Cylinders Slump, In. 3.0 6.5 9.5 2.3 8.3 10.0 1.2 8.1 10.5 1.4 2.5 7.5 1.3 3.5 7.2 2.8 6.5 8.6 1.4 5.2 7.7 1.5 5.0 7.0 4.0 7.7 8.8 Modulus of Elas- ticity, lb. per sq. in. 2 070 000 2 020 000 1 790 000 Quantities Required per Cubic Yard of Concrete Fine Coarse Haydite Haydite Ce- A-2, C-6, ment, cu. yd. cu. yd. bbl. loose, loose, moist moist 2.12 0.47 0.79 2.05 0.46 0.76 1.95 0.43 0.72 Flow 182 243 290 170 243 300 146 220 300 144 192 235 145 193 240 201 229 292 175 215 280 160 220 270 207 273 293 Information concerning the properties of plain concrete was also obtained from the tests of the control cylinders which were made in connection with tests of reinforced structural members, to be described later. 9. Compressive Strength.-The compressive strength and its rela- tion to water-cement ratio has been studied for the various types of concrete tested. In determining the water-cement ratios, absorption allowances have been varied for the Haydite aggregates depending Modulus of Elas- ticity, lb. per sq. in. 1 940 00( 1 690 00( 1 190 00( 2 020 00( 1 660 00( 1 400 00( 1 900 00C 1 680 000 1 370 OOC 1 680 OOC 1 540 0OC 1 390 00 1 670 000 1 640 000 1 590 000 1 550 000 1 690 000 1 430 000 1 330 000 1 320 000 1 140 000 1 500 000 1 260 000 1 200 000 1 370 000 1 030 000 1 090 000 Com- pressive Strength, lb. per sq. in. 4510 3690 2710 4550 4050 2870 4810 4550 2990 3690 3250 2820 3780 3600 3060 3330 3130 2430 2620 2040 1790 2880 2270 1970 2450 1890 1770 of Fresh Con- crete, Com- lb. per pressive cu. ft. Strength, lb. per sq. in. 104.9 2850 103.3 2230 99.8 1340 105.8 3180 105.1 2410 101.5 1490 106.3 3240 106.2 2660 103.6 1460 101.9 2150 101.5 1670 100.1 1350 102.8 2320 102.9 1900 101.1 1590 103.8 1810 103.0 1600 102.3 1220 98.9 1360 99.3 1000 99.1 800 100.7 1460 99.2 1090 99.8 1040 99.8 1150 100.2 900 100.9 860 2 110000 2.15 0.64 0.64 2 050 000 2.09 0.62 0.62 1900 000 1.98 0.59 0.59 2220000 2.16 0.80 0.48 2180000 2.11 0.78 0.47 1 770000 2.02 0.75 0.45 1970000 1.76 0.52 0.78 1920000 1.73 0.51 0.77 1 780 000 1.68 0.50 0.75 2 010 000 1.79 0.66 0.66 2 080 000 1.76 0.65 0.65 1 910 000 1.70 0.63 0.63 1 940 000 1.78 0.79 0.53 1 780 000 1.70 0.76 0.51 1620000 1.70 0.75 0.51 1 880 000 1.48 0.55 0.77 1 660000 1.47 0.55 0.76 1 550 000 1.45 0.54 0.75 1 910000 1.51 0.67 0.67 1 780 000 1.47 0.65 0.65 1 590000 1.46 0.65 0.65 1 810 000 1.48 0.77 0.55 1 690 000 1.46 0.76 0.54 1520000 1.46 0.76 0.54 TESTS ON PLAIN AND REINFORCED CONCRETE V500 K N40C  300 ea X ^0 206 '100 K '0 0 '0 0 0 I'll o -Series "G-2" Gravel Concrete - * -Ser/es "3" C-HaydiSte (Ory) - - Series "A 3" A//-Haaydife (Ory') - + - Series "At-2" A//-Haydife (Moist) "G-2"and, '1-2" rLL~ 0 -/1i~ and Con&ter'c~' L/ar/ah'~' - -- ---- 1-f I I I j I I ___ ___ 0.7 0.8 0.9 0 // /.e 1.3 /.4 1S 16 Wa/er- Cem,7ent Ra/'o by Vo//m7?e FIG. 6. RELATION BETWEEN WATER-CEMENT RATIO AND COMPRESSIVE STRENGTH FOR THREE TYPES OF CONCRETE upon the initial moisture content of the aggregate when used. The absorption allowances made for each series are given in the headings of Tables 6 to 9. Data from the 28-day tests have been plotted in Fig. 6. All of the points fall within a narrow zone on the diagram, though there are small differences in the average curves for the four groups of tests. The All-Haydite concrete made with moist aggregates appears to be consistently stronger than that made with dry Haydite, and as strong as that made with sand and gravel. The difference in strength of All-Haydite concretes may be due to a better bond between the moist particles and the cement paste than that obtained with the dry particles, to inaccurate absorption corrections in the two cases, or to differences in the water-cement ratio relation. It is to be noted that the strengths obtained are considerably greater than those indicated by the strength-water-cement ratio relations commonly quoted. This is due largely to the fact that the strength of cement produced in the last few years has been much greater than that on which most test data are based. However, in the design of Haydite concrete to be placed under ordinary conditions of supervision where careful absorp- tion determinations are not made it would be advisable to assume strengths of perhaps 80 per cent of those given in Fig. 6. + .5~ere~ + N Seres A3" * I ~ OK/C ~/,na~ers j ~ " I -~ .~eres + ILLINOIS ENGINEERING EXPERIMENT STATION K' Ns K: 4\ /60- 4~ MIV eo -4=- - - - KGrve/ Cm Crere -- -- --e d Co=__ _rs - / 'ne -- -- ;:--- C:S: Coq,rrse 1-1&ýIdfean S s Ccrf_ _ I*- €7ne ana Co arse, Hcri/d/ft Con 'ref'e 1:41 lO4O 80 1-l l - - - I- 1-1 35 40 45 so 55 6'0 65 i/a/'o of lwe 4ggreyaf/e to Toh/ Ag egaes in Per CAe e FiG. 7. UNIT WEIGHT OF GRAVEL AND HAYDITE CONCRETES The indication in Fig. 6 that greater strength is obtained through the use of moist Haydite aggregates than with dry ones, even when a logical correction is made for differences in absorption, is borne out by several other groups of tests. This favors the practice of pre-wetting of aggregates in the stock-pile previous to placing them in the mixer, which has been done, however, for the primary purpose of securing better and more uniform workability and to prevent undue drying out of the concrete after it was placed. 10. Unit Weight.-Data on unit weight of freshly placed concrete as found from test cylinders are given in Fig. 7. It is seen that the weights of gravel and of All-Haydite concretes do not vary greatly with changes in proportions and consistencies corresponding to those of Tables 6 to 9. The values shown in Fig. 7 are for concrete made with dry aggregates. The weight of All-Haydite concrete made with moist aggregates was 3 to 6 lb. per cu. ft. greater than that of Fig. 7, and this is further evidence that the total absorption of moist aggre- gate is greater than that for aggregates initially dry. The increase in weight of C-Haydite concrete with increasing amounts of natural sand is self-explanatory. To give information on the unit weight of specimens somewhat larger than the 6 by 12-in. cylinders, and to study the changes in weight of the concrete after hardening, a group of 95 blocks were TESTS ON PLAIN AND REINFORCED CONCRETE TABLE 10 UNIT WEIGHTS OF CONCRETE AFTER VARIOUS STORAGE CONDITIONS Special unit weight blocks used-6 by 12 by 12 in. Blocks made with cylinders of Series G-2, S, A-2 and A-3, Tables 6-9. Blocks weighed at age of 1 day, when forms were removed. Blocks weighed at age of 7 days, after 1 day in forms, 6 days in moist room. Blocks weighed at age of 28 days, after 1 day in forms, 6 days in moist room, 21 days in air of laboratory. Data of concrete mixtures given in Tables 6-9. Dry aggregates used, except in Series A-2. Each weight given is the value from a single block. Unit Weight of Concrete, lb. per cu. ft. Mix Gravel-Series G-2 C-Haydite-Series S All-Haydite-Series A-3 All-Haydite--Series A-2 1-day 7-day 28-day 1-day 7-day 28-day 1-day 7-day 28-day 1-day 7-day 28-day 148.0 148,4 146.8 113.0 112.3 110.8 96.2 96.8 92.7 103.0 104.3 99.0 1:1 :2 . .. . . ... .. . 114.3 114.1 112.2 93.9 96.9 91.9 .. . ... . ..... : 146.3 147.7 145.0 112.2 114.0 110.3 98.7 99.3 92.8 99.4 101.9 96.4 145.7 146.3 144.7 119.8 120.2 118.5 98.4 98.1 94.2 103.8 104.7 99.5 1:2:2 ....... 119.1 119.1 116.8 97.5 97.2 92.7 ..... 144.4 146.1 143.3 121.7 121.2 119.0 98.1 97.8 92.4 101.5 103.8 97.7 1143.7 144.4 142.3 123.9 123.8 122.2 98.0 99.0 93.4 103.7 104.5 99.0 1:2% :l . .. ... ..... ..... 124.3 124.4 122.4 96.6 97.5 91.5 ..... ..... ..... 143.7 145.2 142.5 125.0 124.3 122.2 99.0 100.3 94.5 105.4 107.8 101.0 1147.6 148.7 146.5 113.0 113.3 111.0 93.2 93.2 88.8 98.4 100.9 93.7 1:2:3....... . ... . ..... 115.7 117.6 114.3 94.7 96.1 91.5 ... . ..... S 146.2 147.3 144.4 114.2 114.6 110.6 94.3 94.7 89.1 99.2 101.6 96.0 146.5 147.2 145.0 118.0 117,6 115.6 95.0 94.6 89.7 100.3 102.5 95.5 1:2 :2 .... . . ... . 119.0 118.4 116.4 94.2 96.2 92.3 ..... ..... :145.0 145.7 143.6 120.8 120.5 117.5 94.5 95.3 90.8 100.6 102.8 95.8 143.3 144.2 141.3 122.8 122.8 120.6 94.5 96.8 93.0 102.1 104.4 96.3 1:3:2....... ..... ..... . .... 123.0 122.3 119.3 94.8 95.1 91.2 .. . ... . ..... [143.3 144.4 141.3 124.7 125.2 121.6 94.2 95.6 88.2 101.7 104.2 96.8 146.8 146.9 145.2 114.3 116.3 111.6 92.3 92.3 86.7 98.6 ..... 93.0 1:2%:3%... 145.7 146.6 144.0 114.1 116.1 111.0 94.8 94.3 88.8 96.2 98.3 90.5 1146.9 147.7 144.3 113,8 116.0 110.3 90.0 93.5 85.3 99.2 101.0 92.8 1144.5 144.6 142.5 118.0 119.5 115.2 93.7 95.8 88.4 99.7 ..... 93.7 1:3:3....... 143.5 144.0 141.6 112.0 112.1 113.2 94.5 95.7 89.2 97.3 99.3 91.3 (144.6 145.4 142.0 118.3 119.7 114.6 90.7 92.2 84.1 99.2 100.3 91.7 142.8 142.9 140.3 120.8 122.2 117.5 92.7 95.6 86.1 100.7 ..... 95.3 1:3%:2... . 120.3 121.6 116.7 92.3 95.3 85.7 98.9 101.4 92.3 1 44.2 145.0 140.7 122.5 122.8 117.6 91.7 94.8 85.1 100.7 101.7 91.0 Average Differences from Unit Weights of Freshly Molded Concrete given in Tables 6, 7, 8, and 9 -0.9 -0.1 -2.6 +0.3 +0.8 -2.3 -1.4 -0.2 -6.1 -1.1 +0.7 -6.4 made from the same concrete as used in the cylinders of Series G-2, S, A-3, and A-2. The blocks were 12 in. high, 6 in. wide, and 12 in. long. They were molded in wooden forms, and were tamped in the standard manner used in making cylinders. They were left with the top surface exposed to the air for 24 hr., when the forms were removed and the blocks were weighed. They were then stored 6 days in the standard moist room and weighed again. This was followed by 3 weeks storage in the air of the laboratory and a final weighing. ILLINOIS ENGINEERING EXPERIMENT STATION 4 -- --Wr ana' Cons/sfencF - - o - - - Ivara I/e o ooo 0 0 - Gave/ Co,-,e're S wu A -F'n'e and Coarse H'yd/// 0ConKre/e L - 7000 2000 3000 4000 5000 Co7,vressive Stre7ngqh, ( ? /f, b. ,ver sl. s /'. FIG. 8. RELATION BETWEEN INITIAL MODULUS OF ELASTICITY AND COMPRESSIVE STRENGTH FOR THREE TYPES OF CONCRETE The unit weights of the concrete of these blocks are given in Table 10, arranged to correspond with the unit weights of freshly placed concrete given in Tables 6 to 9. At the bottom of the table is given the average change in unit weight, based upon the weight of freshly molded cylinders of Tables 6 to 9. In general, it appears that the unit weight decreased a fraction of a pound while the blocks stood in the forms, but increased slightly more during the 6 days of moist storage. Relatively large changes were produced, however, by the 3 weeks of air storage, the gravel and C-Haydite concretes showing a decrease in unit weight of 2.6 and 2.3 lb. per cu. ft. and the two All- Haydite concretes, 6.1 and 6.4 lb. per cu. ft. It will be remembered that the large absorption allowances made for the fine Haydite aggre- gates produced a large total water content in the All-Haydite mixtures. Table 10 also shows a tendency toward an increase in loss of mois- ture in air storage as the mixes are made leaner and wetter, but these variations are not altogether consistent. 11. Modulus of Elasticity.-Experimental values of initial tangent moduli of elasticity of concretes made with gravel and sand, coarse Haydite and sand, and fine and coarse Haydite are shown in relation to compressive strength in Fig. 8. Each value is the average of 3 tests. Mixes varied from 1:4 to 1:6, percentages of fine aggregate to total aggregate from 37.5 to 62.5, and slumps from 1 inch to 10 inches. V.S t^ 1^ ^ ^ ^ II 1 l-^ TESTS ON PLAIN AND REINFORCED CONCRETE It will be seen that for each kind of concrete the variation in E with strength may be represented by a straight line for strengths between 2000 and 4000 lb. per sq. in. It will also be noted that for a given strength the moduli for C-Haydite and All-Haydite concretes average about 75 and 55 per cent, respectively, of the values for gravel concrete. For a more complete study of the possibilities of variation of secant modulus of elasticity and to give information concerning rela- tive deformations of the three concretes, the stress-strain curves of Fig. 9 are presented. These are typical curves taken from individual 6 by 12-in. cylinders of varying mix and consistency. The curves were obtained by means of an attached extensometer and a semi- autographic recording device. A constant ultimate unit deformation of about 0.0015 was found for gravel concrete of all strengths except those considerably below 2000 lb. per sq. in., this value being in agreement with numerous previous tests. Ultimate unit deformations of the Haydite concretes are shown to be greater than those of gravel concrete, the proportional increase being greater for the concretes of higher strength. Deformation of All-Haydite concrete shows an in- crease over that of C-Haydite concrete similar to the increase in deformation of C-Haydite over that of gravel concrete. A considera- tion of the nearly straight stress-strain curve for high-strength gravel concrete will show that concretes having initial moduli 75 and 55 per cent as great must necessarily receive higher deformations in order to develop the same stress as the gravel concrete. The curves of Fig. 9 have been replotted in Fig. 10 with the ab- scissas for the Haydite concrete reduced by the differences in hori- zontal intercepts of the tangents which represent the initial moduli of these concretes and gravel concrete. In other words, all curves have been brought approximately to a common tangent, that for an inter- mediate grade of gravel concrete. The curves as represented thus are scarcely more varied than is to be expected from a group of individual cylinders made with a single kind of aggregate. The similarity of these curves leads to the conclusion that for deformations up to 50 per cent of the ultimate deformation the relations of secant moduli of Haydite and gravel concretes are essentially the same as those of initial moduli. A further conclusion is that the primary difference in deformations is the difference in "elastic" deformation (deformation up to the line for initial modulus) as indicated by the ratios 75 and 55 per cent, and that the difference in "plastic" deformation (remaining deformation) is slight and of secondary importance. Since the "plastic" deformation gives an indication of the rate of breaking down ILLINOIS ENGINEERING EXPERIMENT STATION i ^ ^ OS d fi L >Ti ~ ~ ~ ; n] ll. ^ lc« ^ ' Z// 0ýs-01so 'y1 41 A00e7 -11a1 TESTS ON PLAIN AND REINFORCED CONCRETE Un1/t Oeformaf'ion /n? Tbhousan7dts FIGa. 10. MODIFIED STRESS-STRAIN CURVES FOR THREE TYPES OF CONCRETE of the structure it would appear that the Haydite particles themselves do not affect the rate of internal local failure and final complete rupture of plain concrete in compression. 12. Proportions of Materials.-Some users of Haydite aggregates have had difficulty in an attempt to use as large a proportion of coarse Haydite as would be used of 12-in. gravel. It has been established ILLINOIS ENGINEERING EXPERIMENT STATION by experience and tests that the amount of any coarse aggregate usable decreases as the size of particle decreases, at least below the 1-in. size. While an aggregate of l2-in. size may contain the same percentage of voids as one of 1 2-in. size, it is obvious that the size of the voids is smaller in the 12-in. material and will not accommodate sand grains without separation of the coarse aggregate particles. Proportions of C-Haydite (which is composed of material passing a 3/-in. round-hole screen) varying generally from 40 to 60 per cent of the total were used in the tests represented by Fig. 6. Considering the better workability obtained with the use of an excess of fine aggregate and also the accompanying increase in weight of the sand and C-Haydite mixture, it is believed that a mixture of 45 per cent fine and 55 per cent coarse is the most desirable combination, at least with the sand and Haydite mixtures. For All-Haydite concretes it may be just as economical to use equal parts of fine and coarse material. The data given in Tables 7 and 8 on C-Haydite and All-Haydite mixtures, made from materials initially dry, may be found useful as a guide to proportioning mixtures of this type. Similar data for mixtures made with moist Haydite aggregates are given in Table 9. As noted previously, due to the bulking of the aggregates, the mixtures used in the tests recorded in Table 9 are actually richer in cement than those reported in Table 8. In general, Table 9 should be used as a guide to proportions when moist aggregates are to be used, although the moisture content of the fine aggregates of Table 9 is unusually high. Another method for determining proportions when moist aggre- gates are to be used is to make bulking and moisture content deter- minations on the aggregates, and then to use volumes of the moist materials containing solids equal to those of the dry aggregates of Table 8. The bulking effect is small for the coarse Haydite and usually will not exceed 5 or 6 per cent, but for the fine Haydite the bulking may be important. Thus the fine Haydite weighing 55 lb. per cu. ft. by dry loose measure, weighs 51 lb. per cu. ft. with 27 per cent of moisture present. This 51 lb. consists of 40.1 lb. of dry Hay- dite and 10.9 lb. of water. To obtain 55 lb. of dry Haydite in the 55 mixture requires . or 1.37 cu. ft. of the moist material. With this 37 per cent bulking in the fine aggregate and a 5 per cent bulking assumed in the coarse, for a mix of 1:2:3 concrete as listed in Table 8, the volume proportions with the moist materials should be 1:2.74:3.15. Similarly, the quantities required per cubic yard of con- TESTS ON PLAIN AND REINFORCED CONCRETE crete as given in Table 8 would be multiplied by the ratios of 1.37 and 1.05 for the fine and coarse materials, respectively. The use of moist aggregates just described will naturally affect the water requirements. In a one-bag batch the 2.74 cu. ft. of fine aggre- gate contain 10.9 lb. of water per cu. ft. or 29.9 lb. This must be considered in computing the amount of water for mixing and absorption requirements. The quantities of materials in Tables 6 to 9 contain no allowance for waste or for settlement of material in forms. 13. Workability.-A comparison of the data of Tables 6, 7, and 8 (Series G-2, S, and A-3), in which dry aggregates were used, shows that for a given mix and water-cement ratio the slump is greatest for the gravel concrete, intermediate for the C-Haydite concrete, and least for the All-Haydite concrete. An illustration shows that for a 1:2:3 mix with a water-cement ratio of 1.0 the slumps found were 9.2, 8.3, and 1.9 inches, respectively, for gravel, C-Haydite, and All- Haydite concretes. To obtain a slump for Haydite concrete equal to that of gravel concrete, requires either some sacrifice of strength through an increase in the amount of water used, or an increase in the richness of mix. The change required in the mix is greater for All- Haydite than for C-Haydite concrete. Such a comparison of like mixtures is valid only if the gradation of the different aggregates is the same. The sand and gravel used in these tests were similar in gradation to the fine and coarse Haydites, but no attempt was made to produce identical gradations in the two materials. It is evident that to the gradation and texture of the Haydite aggregates is due their requirement of a greater water-cement ratio for a given slump than that required in similar mixes with sand and gravel aggregates. Where water contents were varied for a constant mix to bring all three types of concrete to approximately equal slumps no appreciable difference was observed in the work required to place the concretes. Concrete was made with Haydite aggregates with no particular de- parture from ordinary methods. The same mixing procedure was used with Haydite concretes as with gravel and limestone concretes. Uniformly good appearance of the molded surfaces of beams was ob- tained with the consistencies used in the various beam series. For beams with fairly heavy reinforcement, very satisfactory placing of the concrete was accomplished with slumps of 6 and 7 inches. There was practically no honey-combing or segregation of the materials, except that the Haydite beams presented a rather harsh top surface as the material settled into place in the forms. ILLINOIS ENGINEERING EXPERIMENT STATION TABLE 11 EFFECT OF AGE ON COMPRESSIVE STRENGTH AND MODULUS OF ELASTICITY All cylinders moist cured, tested moist. Aggregates used: Series 1: Sand, S-1; gravel, G-1; coarse Haydite, C-6. Series T: Sand, S-2; fine Haydite, A-3; coarse Haydite, C-6. See Tables 2 and 3. Wet consistencies, 6 to 9-in. slumps. All values given represent the average of 3 tests. Series la Series T Age at Sand and Coarse Fine and Coarse Moist Coarse Dry Coarse Time of Sand and Gravel, Haydite, Haydites, Haydite and Sand, Haydite and Sand, Test 1:2:3 Mix, 1:2:3 Mix, 1:1:1% Mix, 1:l%:2% Mix, 1:lV:24 Mix, Water-Cement Water-Cement Water-Cement Water-Cement Water-Cement Ratio, 1.00 Ratio, 1.05 Ratio, 0.66 Ratio, 0.99 Ratio, 1.02 Compressive Strength of 6 by 12-in. Cylinders, lb. per sq. in. 7 Days ........ 2430 1650 3350 2160 1920 14 Days ......... .... 4350 2990 2920 28 Days ......... 3980 3450 5120 3870 3770 3 Months....... 4670 4500 5680 4850 4640 6 Months....... 4730 4590 6800 4870 5180 12 Months....... 5700 5070 7340 5790 5570 Compressive Strength in Terms of 28-Day Strengths, per cent 7 Days ......... 61 48 65 56 51 14 Days ......... ... ... 85 77 78 28 Days......... 100 100 100 100 100 3 Months....... 118 130 111 125 123 6 Months....... 119 133 133 126 137 12 Months....... 145 147 143 149 148 Initial Modulus of Elasticity, lb. per sq. in. 7 Days......... 2 970 000 1 830 000 2 010 000 2 120 000 1 900 000 14 Days ......... ... ... . ....... 2 100 000 2 190 000 2 440 000 28 Days......... 3 910 000 2 480 000 2 180 000 2 680 000 2 660 000 3 Months....... 4 500 000 2 710 000 2 270 000 2 890 000 2 850 000 6 Months....... 4 660 000 2 590 000 2 630 000 2 920 000 3 020 000 12 Months....... 4 900 000 2 790 000 2 630 000 2 930 000 3 200 000 Unit Weight of Fresh Concrete, lb. per cu. ft. 147.4 115.4 102.1 116.3 116.3 14. Effect of Age on Compressive Strength.-Two groups of cylin- ders were made in connection with the beam tests of Series 1, Section 15. These cylinders were of a 1:2:3 mixture. One group was of gravel concrete and the other of C-Haydite concrete. The aggregates were proportioned by loose volume, the sand being dry and the Hay- dite containing about 6 per cent of moisture. Three cylinders from each of these groups were tested at ages of 7 and 28 days and 3, 6, and 12 months. Three later groups of cylinders, Series T, were made with richer mixes, using oven-dried Haydite aggregates. One group was made TESTS ON PLAIN AND REINFORCED CONCRETE Ha/71/f El/evaf/o,7 SectOOoE FIG. 11. DETAILS OF BEAMS OF SERIES No. 1 with fine and coarse Haydite to which a definite amount of moisture was added and the material stored in tight containers one day pre- vious to the time of mixing. Of the other two groups, one used coarse Haydite to which a definite moisture content had been added and the other used the dried Haydite directly. The absorption allowance and water-cement ratio were the same for the two groups. These three groups of cylinders provided for tests at ages of 7, 14, and 28 days, 3 and 6 months, 1, 2, and 5 years. All specimens were cured in the moist room until test, and were tested moist. Table 11 gives values of the strengths of the five lots of cylinders at ages up to 1 year, as well as values of the initial modulus of elasticity and of the unit weight of each kind of concrete as freshly placed. The table also shows the strengths at different ages expressed as ratios of the standard 28-day strength. It appears from Table 11 that the increase in strength of the four Haydite concretes compares very well with that of the sand and gravel concrete. IV. BEAM TESTS, SERIES I 15. Purpose of Series.-One of the first series of tests was made on beams and cylinders of machine-mixed concrete to secure information on mixtures suitable for ordinary reinforced concrete construction, with particular attention given to securing workability, surface finish, and freedom from segregation. The selection of proportions of mix, consistency, and method of mixing and placing was made with a view to producing light-weight concrete in a considerable range of strengths and to placing it easily in the forms and around reinforcing steel. The beams of this series consisted of three groups made with torpedo sand, using as coarse aggregates gravel, limestone, and C- Haydite of comparable size and gradation, and a fourth group made with Haydite fine and coarse aggregates. Data of the aggregates are given in Tables 2 and 3. ILLINOIS ENGINEERING EXPERIMENT STATION TABLE 12 RESULTS OF BEAM TESTS, SERIES 1 Aggregates used: Sand, S-1; gravel, G-1;limestone, L-1; fine Haydite, A-4; coarse Haydite, C-6. See Tables 2 and 3. Proportions by loose, dry volumes, except for coarse Haydite which contained 6 per cent moisture. Additional moisture allowances were; for sand, S-1, and gravel, G-1, 1 per cent; for limestone, L-1, 3 per cent; for coarse Haydite, C-6, 3 per cent; for fine Haydite, A-4, 18 per cent. Values for cylinders are average of 3 tests. 28-Day Tests, 28-Day Tests, Beam 6 by 12-in. Cylinders Unit Mix and Weight Water-Cement Slump, Flow of Fresh Measured Corn- Modulus Ratio . Concrete, Ultimate Load* at Deflection f at pressive of Elas- cul. ft. lb.', Point, 13,000 lb., lb per lb. per lb. p r lb. lb. in lb. per lb. per lb. per I-sq. in. sq. mn. sq. in. Concrete Made with Natural Sand and Gravel Aggregates 1:1%:2Y.. f 8.8 243 145.8 16 800 15 500 0.151 38 100 4790 3920000 W 8.7 230 146.0 15 700 15 100 0.158 38 100 4710 4010000 =0.90.. 9.8 246 145.5 ...... ...... ..... ...... 4690 3910000 Average..... 8.8 237 145.8 16 300 15 300 0.155 38 100 4730 3950000 1:2:3 ......I 8.9 234 146.4 15 700 14 600 0.157 38 400 3300 3800000 W 8.7 254 146.4 17 900 14 700 0.169 43 800 4100 3690000 -=1.0. 8.0 240 147.4 15 900 14 800 0.155 37 500 4200 4100000 Average..... 8.5 243 146.7 16 500 14 700 0.160 39 900 3870 3860000 1:2 4:3%.. 1 8.4 238 146.8 16500 15 800 0.163 41 400 2990 3490000 W 8.5 247 146.5 16 700 15 600 0.169 42 000 3130 3 700000 -- 1.20.. 8.0 243 146.0 15 800 15 300 0.171 39 300 3300 3910000 Average..... 8.3 243 146.4 16300 15 600 0.168 40 900 3140 3700000 Concrete Made with Natural Sand and Coarse Haydite 1:1%:2%..1 8.5 253 117.1 16000 14 500 0.176 37 200 4150 2630000 W , 8.7 256 117.4 15400 14 100 0.179 31 800 3920 2680000 S-=0.94.. 8.8 252 118.3 15 700 14 500 0.180 37 200 4400 2550000 Average..... 8.7 254 117.6 15 700 14 400 0.178 35 400 4160 2620000 1:2:3...... 6.8 222 115.0 14 900 14 700 0.193 39 600 3500 2620000 W 7.7 250 115.6 15 000 14 100 0.161 41 400 3530 2330000 j1.02.. J 8.0 248 116.0 16 200 15 500 0 185 39 000 3190 2570 000 Average..... 7.5 240 115.5 15 400 14 800 0.180 40 000 3410 2510000 1:2%:3%..1 7.2 237 115.1 15 400 14 700 0.193 39 900 2590 2200000 W 1 6.6 222 114.4 15400 14500 0.199 38 100 2730 2130000 -1.25.. 7.4 225 114.4 15 000 14 400 0.199 39 600 2540 2360000 Average..... 7.1 228 114.6 15 300 14 500 0.197 39 200 2620 2230000 *Includes weight of beam. 16. Type of Specimen and Method of Testing.-Thirty-two beams, 6 by 12-in. in section, 8 ft. 6 in. long, were tested on an 8 ft. span with one-third point loading. Details of the test beams are shown in Fig. 11. The principal data of the concrete mixtures used are given in Table 12. Three proportions of concrete were used: 1:14 :21y, 1:2:3, TESTS ON PLAIN AND REINFORCED CONCRETE TABLE 12 (Concluded) Aggregates used: Sand, S-1; gravel, G-1; limestone, L-1; fine Haydite, A-4; coarse Haydite, C-6. See Tables 2 and 3 Proportions by loose, dry volumes, except for coarse Haydite which contained 6 per cent moisture. Additional moisture allowances were: for sand, S-1, and gravel, G-1, 1 per cent; for limestone, L-1, 3 per cent; for coarse Haydite, C-6, 3 per cent; for fine Haydite, A-4, 18 per cent. Values for cylinders are average of 3 tests. 28-Day Tests, 28-Day Tests, Beam 6 by 12-in. Cylinders Unit Mix and Weight Water-Cement Slump, Flow of Fresh Measured Coin- Modulus atio in. Concrete, Ultimate Load* at Deflection f. at presiv of Elas- lb. per Load, Yield at Load of YieldPoint, Strength, ticity, cu. ft. ,lb. Point, 13,000 lb., lb. per lb. per lb. per lb. i sq. in. sq. in. sq. inm. Concrete Made with Natural Sand and Limestone 1:1%:2y4.. 7.3 218 145.0 16 300 15 300 0.157 37 500 4720 3 420 000 W8.5 229 144.7 15 500 14600 0.164 36 600 4720 3720 000 0.90.. 1 7.3 220 145.1 16 600 15 700 0.151 36 600 5040 3 630 000 Average..... 7.7 222 144.9 16 100 15 200 0.157 36 900 4830 3590 000 1:2:3 ...... 6.4 210 143.9 16200 15 200 0.163 40200 3710 3080000 W 0 5.1 197 144.2 15700 14800 0.152 42600 4100 3710000 = 1.00.. 4.0 200 145.5 15 600 14 900 0.152 35 100 3890 3 740 000 Average..... 5.2 202 144.5 15 800 15 000 0.156 39 300 3900 3510000 1:2A:3 1.. 1 5.6 208 146.6 16 100 14 600 0.184 37 800 2900 3240 000 W 7.0 222 145.8 15 500 15 000 0.175 36 600 2890 3 120000 = 1.20.. 6.6 205 145.6 16 800 15 600 0.163 40 200 3190 3 700 000 Average..... 6.4 212 146.0 16 100 15 100 0.174 38 200 2990 3350 000 Concrete Made with Fine and Coarse Haydite Aggregatesa 1:1%:2h..\ . 9.0 300 100.0 16 300 13 800 0.208 34 800 3310 1730 000 W/C 0.97J . 8.6 300 100.5 18 700 14 600 0.192 35 400 3200 1580000 Average..... 8.8 300 100.2 17 500 14 200 0.200 35 100 3250 1650000 1:2:3 ......1 8.4 300 97.4 15 200 13 400 0.217 39 600 2400 1290 000 W/C= 1. 10 7.6 294 98.2 17 700 14 700 0,202 39 600 2380 1460000 Average..... 8.0 297 97.8 16 400 14 000 0.210 39 600 2390 1370 000 1:2%:33..J f" 7.1 270 96.4 15 400 14 700 0.223 41 400 1690 1390000 W/C=1.31J 7.8 269 96.9 15 100 14700 0.218 41 400 1780 1440000 Average.... 7.5 270 96.6 15 200 14 700 0.220 41 400 1740 1410000 *Includes weight of beam. tMade with lot 2 of cement, all others of lot 1. and 1:2y2:31, by loose volume, the fine aggregate being 40 to 44 per cent of the total aggregates. It was first intended to make the four groups of beams of Table 12 of concretes having the same net water- cement ratios (exclusive of absorption allowances for the different aggregates) but the resulting workabilities differed too greatly be- cause of variations in gradation and surface texture of the aggregates. The amounts of mixing water were therefore chosen to bring all mix- tures to approximately equal workabilities, as shown by slump tests. ILLINOIS ENGINEERING EXPERIMENT STATION FIG. 12. TEST BEAM OF SERIES NO. 1 IN TESTING MACHINE Slump and flow tests were made in accordance with the standard method* of the American Society for Testing Materials, using fifteen Y2-in. drops of the table in the flow test. The concrete was mixed in a one-bag batch mixer and was placed in the forms with shovels and spaded uniformly. With each beam, 6 by 12-in. control cylinders were made from the same batch of con- crete. These cylinders were weighed to determine the unit weight of the freshly molded concrete. The beams and control cylinders were cured in the moist room in a saturated atmosphere at about 70 deg. F., and were tested at the age of 28 days. Removal from moist storage a day in advance of testing was necessary for preparation of gage holes on the beams. The cylinders were removed with their companion beams. All beams were tested in an Olsen testing machine of 200 000-lb. capacity. Load was applied to the test beam at a speed of about 0.03 in. per minute. The beam supports at the reaction points per- mitted longitudinal adjustment during the application of the load, and the blocks at the load points permitted both longitudinal and transverse adjustment. The arrangement of the apparatus is shown in Fig. 12. *"Standard Method of Making Compression Test of Concrete," (C39-27) 1930 Book of Standards, A.S.T.M., Part II, p. 143. TESTS ON PLAIN AND REINFORCED CONCRETE FIG. 13. TYPICAL LOAD-STRAIN AND LOAD-DEFLECTION CURVES OF BEAMS OF SERIES No. 1 1 t C,) K ILLINOIS ENGINEERING EXPERIMENT STATION FIG. 14. TYPICAL BEAMS OF SERIES NO. 1 AFTER TEST Initial strain and deflection measurements were taken at a load of 1000 lb. and subsequent readings at 1000-lb. increments of load. Intervals of less than 10 minutes were required for application of an increment of load and measurement of strains. The deflection of the beam at mid-span was read by use of Ames dials as shown in Fig. 12. Strain measurements taken on the gage lines shown in Fig. 12 were made with an 8-in. Berry strain gage having a multiplication ratio of 5. Values of modulus of elasticity and compressive strength were obtained from tests of the 6 by 12-in. cylinders. 17. Discussion of Results.-All of the beams failed by exceeding the tensile yield point of the longitudinal steel. The principal results of the tests of Series 1 are given in Table 12. Load-deflection and load-strain curves are given in Fig. 13. The measured steel stresses and deflections shown in Table 12 do not include the effect of the weight of the beam and the initial load of 1000 lb. A view of three beams of this series, shown in Fig. 14, indicates that there was little difference in the appearance of beams made with limestone, Haydite, and gravel concrete. As noted in Section 14, there was little difference observed in the work required to place concrete of corresponding mixes. Since all of the beams failed by tension in the horizontal steel, there should be little difference in the maximum loads for beams alike in size, span, and reinforcement, except as the arm of the resisting couple, jd, might vary with the elastic properties of the different con- cretes. The relatively low value of the modulus of elasticity for the Haydite concretes tends to decrease the value of jd slightly, and the TESTS ON PLAIN AND REINFORCED CONCRETE 04 4 '-0" Zetgf /alf E/ep-a//- Sec- t\o \ FIG. 15. DETAILS OF BEAMS OF BOND SERIES B ultimate loads on the beams made with sand and coarse Haydite are seen to be 3 to 4 per cent less than those for the beams of gravel and limestone concrete. The load-strain curves of Fig. 13 show very similar action of Hay- dite and gravel or limestone beams. The similarity of the positions of the points of sharp curvature of the load-strain curves for steel show that cracking of the concrete on the tension side of the beams took place at approximately the same load for each type of concrete tested. As noted previously, the values of modulus of elasticity given in Table 12 are from compression tests. The result of the low modulus of elasticity of Haydite concrete beams is seen in the relatively large deflections at the 13 000-lb. load, as given in the table. A further consequence of the low modulus of elasticity is to cause the neutral axis to lie lower in the Haydite concrete beams than in those of gravel or limestone. This result was verified by the strains which were measured at mid-span on all beams. V. TESTS OF BOND RESISTANCE, SERIES B 18. Outline of Tests.-With the object of furnishing needed data for the design of reinforced Haydite concrete members, a series of bond tests was made. Tests were made on the usual tensile pull-out specimens, and on simple beams. It is felt by many engineers that the test of beam specimens is much more informative as regards the critical amount of slip in a bond test than is the pull-out test. Three grades of concrete were used: gravel, C-Haydite, and All-Haydite; each of these was made of three mixtures: 1:134:2Y4, 1:21Y:2%, and 1:2Y4:314, by loose volume. Two beams, three pull-out specimens, ILLINOIS ENGINEERING EXPERIMENT STATION FIG. 16. BEAM OF SERIEs B IN TESTING MACHINE and six companion 6 by 12-in. cylinders were made of each mixture and tested when 28 days old. 19. Beam Tests.-The beams, shown in detail in Fig. 15, were 8 by 12 in. in cross-section by 6 ft. 6 in. long, and were tested on a 6-ft. span with one-third point loading. The reinforcement consisted of a single straight longitudinal bar 1Y in. square, projecting Y1 in. beyond the ends of the beam, and four 2-in. round U-stirrups in each outer third of the span. The longitudinal bars used were deformed bars of rail steel, and had very gradually tapering diamond-shaped projections. Identical beams were made on different days, each beam being made with two batches of machine-mixed concrete. The concrete was shovelled into wooden forms and spaded around the steel and away from the sides of the form. The forms were stripped after 24 hours. All beams were cured in the moist room until one day before testing when it was necessary to prepare gage holes and to whitewash the beam surfaces to aid in the detection of cracks. A general view of a beam in the testing machine, loaded in an inverted position to facilitate strain measurements, is shown in Fig. 16. The location of the gage lines, on which strains were meas- ured by means of a 4-in. Berry strain gage, was indicated in Fig. 15. Deflections were obtained as in Series 1. End slip of the longitudinal bar was observed by means of two 0.0001-in. Ames dials, attached as TESTS ON PLAIN AND REINFORCED CONCRETE FIG. 17. INSTRUMENT USED TO DETERMINE END-SLIP OF LONGITUDINAL REINFORCEMENT shown in Fig. 17. The beams were tested in a 300 000-lb. Olsen uni- versal testing machine, using the test speed of 0.05 in. per min. The load increments were applied as rapidly as strain observation would permit. Figure 18 shows side and top views of beams made of All-Haydite, C-Haydite, and gravel concrete respectively. 20. Pull-Out Tests.-The pull-out specimen consisted of a 1 -in. square deformed bar, of the type used in the beams, embedded axially in an 8-in. concrete cylinder, 8 in. high. As the specimen was made, the bar projected about 2 ft. below the concrete cylinder and 1/Y in. above. The concrete was placed and rodded in the mold in the stand- ard manner used in making compression cylinders. The pull-out specimens were made from the same concrete as the beams, and were subjected to identical curing conditions. The pull-out specimens were tested in a 200 000-lb. Riehle testing machine. The set-up of a specimen in the machine is shown in Fig. 19. The specimen was permitted to center itself by means of a spherical bearing block containing a hole through which the long end of the reinforcing bar extended. Load was applied from the pulling head of the machine to the bar. The concrete cylinder was supported by machined bearing plates separated by a cushion of linoleum, in turn supported by the spherical block resting on the upper head of the ILLINOIS ENGINEERING EXPERIMENT STATION (a) Side Elevation (b) Bottom View FIG. 18. TYPICAL BEAMS OF BOND SERIES B AFTER TEST testing machine. End slip of the bar was measured by means of a 0.0001-in. Ames dial attached to a clamp and held in position above the projecting end of the bar in a manner similar to that shown in Fig. 17. 21. Discussion of Results.-The principal results of the bond tests are given in Table 13. For comparative purposes the minor varia- TESTS ON PLAIN AND REINFORCED CONCRETE FIG. 19. PULL-OUT SPECIMEN IN TESTING MACHINE tions in the lever arm of the resisting couple, jd, due to different qualities of concrete, have been neglected and a value of 8.75 inches has been used in the computations of the average nominal bond stress, u, as found from the beam tests. This value of u was calculated from the well-known equation V U mojd wherein V is the total vertical shear and mo is the circumference of the bar, or 412 in. It is probable that the values of jd for the Haydite concrete beams were actually 3 or 4 per cent lower than for those of gravel concrete, and that the values of u for the Haydite beams should be correspondingly higher than the values shown in Table 13. Figure 20 shows average bond stresses developed in the pull-out specimens at end slips up to 0.01 in. In general, the curves are nearly ILLINOIS ENGINEERING EXPERIMENT STATION TABLE 13 DATA OF BOND TESTS, SERIES B Aggregates used: Sand, S-2; gravel, G-2; fine Haydite, A-3; coarse Haydite, C-6 used with natural sand, C-3 used with fine Haydite. See Tables 2 and 3. Age of specimens, 28 days ± 1 day. Nominal = % assumed for all beams. All specimens moist cured. Proportions by loose dry volume. Dry aggregates used. Water allowed for absorption: Sand and gravel, 1 per cent; fine Haydite, 14 per cent; coarse Haydite, 7 per cent. 6 by 12-in. Cylinders Beam Tests Pull-Out Tests (Average of 3) Unit Average Average Mix and Weight Avrg Avere Water-Cement Slump, Fl of Fresh Load* Bond Bond Stress Initial Ratio, by in. Flow Concrete at End Ultimate Stress pCorni Modulus Volume lb. per Slip of Load,* at At End At prUsive of Elas- . ft1 . 0.001 lb. Ultimrate Slip of Ultimate Strength, ticity, in., lb. Load, 0.001 in., Load, lb. per lb. per lb. per lb. per lb. per sq. in. sq. in. sq. in. sq. in. s sq. in. Concrete Made with Natural Sand and Gravel Aggregates 1:1%:2%y1 8.6 250 147.7 19 700 28400 358 445 620 4040 3780000 W 628 700 = 0.80 '" 8.5 250 146.9 27 500 28 100 355 397 577 3960 3 730 000 Average...... 8.5 250 147.3 23 600 28 300 357 457 633 4000 3 750 000 1:2y:2%y 1 7.7 242 145.7 20 500 25 400 320 494 633 3180 3 660 000 81 480 53 = 0.90. '"1 7.8 245 146.1 24200 26600 336 517 578 3410 3640000 Average...... 7.7 244 145.9 22 300 26 000 328 497 598 3300 3 650000 1:2Y%:3 | 7.0 251 145.7 24 400 26 200 330 458 530 2990 3 250 000 W .. 483' 525 = 1.10 7.8 244 146.6 22 000 25 200f 317 525 616 2720 3 380 000 Average...... 7.4 247 146.1 23 200 25 700 324 489 557 2860 3 310 000 Concrete Made with Natural Sand and Coarse Haydite 1:1%:2%Y 7.3 243 119.4 21 800 24 500 309 436 625 3970 2 670 000 W 431 171 ~ 0.87" 7.8 243 117.1 21 900 24 800 313 447 567 3880 2 500000 Average...... 75 243 118.3 21 800 24 700 311 438 589 3930 2580000 1:23:2%0 7.4 250 119.6 22 500 26 300 333 405 505 3390 2 500 000 E 9 ... 319 461 = 0.98 " 7.5 261 118.4 22 300 24 500 309 442 605 3400 2 450 000 Average...... 7.5 255 119.0 22 400 25 400 321 389 524 3400 2 470 000 1:28/:3%1 7.3 274 118.5 24 500 28 900f 365 389 466 2560 2 250 000 = 1.19" 7.0 260 117.9 19 500 21 800 275 367 458 2520 2 160000 Average...... 71 267 118.2 22 000 25 400 320 393 481 2540 2200 000 Concrete Made with Fine and Coarse Haydite Aggregates 1:13:2MY 6.2 253 97.9 18 100 20 400 258 241 403 2820 1880000 IW 5 308i 500 = 0.97 "' 5.4 247 97.9 18 300 22 400 283 292 427 2910 1990 000 Average...... 5.8 250 97.9 18 200 21 400 270 280 443 2860 1 930000 1:23:2% 3.6 233 95.4 16 500 19 200 243 356 497 2580 1 760 000 2 8 367 472 = 1.12 ' 8.0 287 92.6 20 200 22 700 287 381 530 2250 1630000 Average...... 5ý8 260 94.0 18 300 21 000 265 368 483 2420 1 690 000 1:2%:3%1 f 6.6 265 94.6 15 900 18 400f 233 280 405 1760 1490000 5. 286 422 = 1.36 " ' 5.2 252 93.9 14 400 20 600 260 319 472 1720 1640 000 Average...... 258 94.3 15 100 19500 247 25" 433 1740 1 560000 *Including weight of beam. tStirrups in these beams were W-in. square instead of round. No effect on failure of beams. TESTS ON PLAIN AND REINFORCED CONCRETE CI) CI ^1 S.- (I FIG. 20. RELATION BETWEEN BOND STRESS AND END SLIP FOR CONCRETES OF VARYING STRENGTH IN PULL-OUT TESTS (Specimen numbers 4, 5, and 6 indicate 1:4, 1:5, and 1:6 mixes) horizontal at an end slip of 0.01 in., indicating that the bond resist- ance of the deformed bars used was similar in nature to that of plain bars. At the maximum load either the concrete cylinder split or the bar merely continued to slip with a decrease in load. The average values of bond stress developed by the pull-out specimens, at an end slip of 0.001 in. and at maximum load, are shown in Fig. 21 and in Table 13. The maximum bond strength developed varied from 15 to 25 per cent of the compressive strength of the companion cylinders. Figure 21 also gives values of the nominal bond stress developed in the beam tests at an end slip of 0.001 in. and at maximum load. The maximum bond stress varied from 8 to 14 per cent of the com- pressive strength, with the higher percentages corresponding to the leaner mixtures. In a similar way, the bond stress at a slip of 0.001 in. varied from 7 to 11 per cent of the compressive strength. Apparently the relation between bond strength and compressive strength is essentially the same for gravel concrete and Haydite con- crete, so that design rules in which the working stress in bond is expressed as a percentage of the cylinder strength would apply equally ILLINOIS ENGINEERING EXPERIMENT STATION ComfOressive Stre7ng/h, f, //? er 5. /s FIG. 21. RELATION BETWEEN BOND STRESS AND COMPRESSIVE STRENGTH well for the two materials. However, the most important result of the tests is seen by comparing the stresses developed with allowable working stresses in common use. Current usage permits a working bond stress in deformed bars equal to 5 per cent of the compressive strength at the age of 28 days; this rule, if applied to the test beams of the 1:1V4:24 mixture, would result in a factor of safety of 1.8 for gravel and sand concrete, 1.6 for C-Haydite concrete, and 1.9 for All-Haydite concrete. Considering the fact that, in other investiga- tions in this laboratory, beams under sustained loading have failed at a load which was just sufficient to produce end slip, it appears that the true factor of safety is even less than the values based upon maximum loads. It is evident that bond strength does not vary directly with compressive strength, so that the use of a working stress equal to 5 per cent of the compressive strength gives a variable factor of safety against bond failure, and one that is decidedly too low for the richer mixtures. Typical diagrams showing the variation in measured steel stresses along the reinforcing bar and the measured slip at the ends of the bar at various increments of load are given in Fig. 22. The curves apply to beams of sand and gravel, C-Haydite, and All-Haydite concrete of the 1:2Y4:234 mixture. The steel stresses given were calculated from strains measured at points along the length of the beam as indicated TESTS ON PLAIN AND REINFORCED CONCRETE I:: >1 NJ Distance a/ong Beam in /ncihes Enda S/ of Sor in /nches Fia. 22. TYPICAL LOAD-STRESS AND LOAD-END SLIP CURVES FOR BEAMS OF BOND SERIES B at the bottom of the figure. The variation in stress is about what would be expected with one-third point loading. End slips are plotted in Fig. 22 against the total applied loads. It is seen that failure occurred very soon after an end slip of 0.005 in. was reached, and, in general, slipping developed quite rapidly after anp'end slip of 0.001 in. had been reached. VI. DIAGONAL TENSION TESTS, SERIES D.T. 22. Description of Beams and Method of Testing.-Twenty-four large beams, designed to fail by diagonal tension, were tested in Series D.T. Eighteen of these beams were of the mixes and propor- tions used in the bond tests and contained no web reinforcement, ILLINOIS ENGINEERING EXPERIMENT STATION H//fl E/evea'/o7 Sect/on FIG. 23. DETAILS OF BEAMS OF DIAGONAL TENSION SERIES D.T. FIG. 24. TYPICAL REINFORCING UNIT OF SERIES D.T. WITH 1.0 PER CENT OF WEB STEEL while the remaining six were of one proportion of C-Haydite concrete, with three different amounts of web reinforcement in the form of stirrups. Three companion 6 by 12-in. cylinders were made with each beam. The details of the beams, which were 8 by 24 in. in section and of 8-ft. span, are shown in Fig. 23. To minimize the chances for tensile failure, 2.8 per cent of longitudinal steel was used, six 1-in. plain round bars, hooked at their ends, providing the necessary area. For the beams with web reinforcement, 0.5, 1.0, and 1.5 per cent of verti- cal web steel was provided in the form of 12-in. round U-stirrups at spacings of 10, 5, and 31Y inches, respectively. This web steel had a yield point of 52 800 lb. per sq. in., which was higher than was con- templated in the original design. The yield point of the longitudinal TESTS ON PLAIN AND REINFORCED CONCRETE FIG. 25. BEAM OF SERIES D.T. IN TESTING MACHINE steel was 37 600 lb. per sq. in. A typical view of a reinforcement unit, with 1.0 per cent of web steel, is shown in Fig. 24. Large corks, shown tied to the stirrups, left holes in the concrete to provide access to the steel for the legs of the strain gage. Each layer of longitudinal steel was kept at its proper depth by a pair of short bars, extending across the beam and projecting into the side forms. The concrete was machine-mixed, four batches being required for each beam and its companion cylinders. Two beams of different types of concrete were made in one day. The beams and cylinders were moist cured together on the laboratory floor by keeping them covered with several layers of burlap and by wetting them twice daily. The beams were tested in an inverted position in a 300 000-lb. Olsen testing machine as shown in Fig. 25. Strain measurements were taken on 4-in. gage lengths on the concrete at midspan, on the longitudinal steel at midspan and near the ends, and on stirrups about ILLINOIS ENGINEERING EXPERIMENT STATION TABLE 14 DATA OF DIAGONAL TENSION TESTS, SERIES D.T. Aggregates used: Sand, S-3; gravel, G-3; fine Haydite, A-4; coarse Haydite, C-7. See Tables 2 and 3. Test beams, 8 by 24-in., 9 ft. long, moist cured 28 days. Nominal value of j = V assumed for all beams. Proportions by loose dry volume. Aggregates used moist. Total allowance for absorption: Sand and gravel, 1 per cent; fine Haydite, 18 per cent; coarse Haydite, 8 to 11 per cent. 6 by 12-in. Cylinders Unit Shearing Computed (Average of 3) Mix and Weight Web Ultimate Stress at f, at Water-Cement Slump, Flow of Fresh Steel Load,* Ultimate Ultimate Corn- Modulus Ratio, by in. Concrete, percent lb. Load, Load, pressive of Elas- Volume llb. per Strength, ticity, cu. ft. sq. in. sq. in. lb. per lb. per sq. in. sq. in. Concrete Made with Natural Sand and Gravel Aggregates 1:1%:2%1 .... 6.0 213 147.7 0 142 900 484 26 100 4760 3 800 000 W/C = 0.80..1 7.0 242 148.7 0 159 700 540 29 200 4620 4 100 000 Average ...... 6.5 227 148.2 151 300 512 27 600 4690 3 950 000 1:2y:2.4...1 f 3.8 202 148.6 0 151 800 514 27 700 4290 3 830 000 W/C= 0.0..J0. 6.1 224 148.1 0 134 100 423 24 500 3860 3 570 000 Average....... 5.0 213 148.3 142 900 463 26 100 4050 3 700 000 1:2%:3%....1 8.0 276 146.4 0 105 800 357 19 300 2230 3 090 000 W/C = 1.10..J 6.3 257 146.9 0 116 500 394 21 300 2630 3 180 000 Average....... 7.1 266 146.6 111 100 375 20 300 2430 3 130 000 Concrete Made with Natural Sand and Coarse Haydite 1:14:2Y....1 J 7.6 240 1109.9 0 132 900 450 24 300 4550 2 850 000 W/C = 0.84..j1 7.7 255 119.1 0 133 700 453 24 500 3930 2 630 000 Average....... 7.6 247 119.5 133 300 451 24 400 4240 2 740 000 1:2:2/. .... J 4.0 221 120.5 0 124 900 423 22 900 3730 2 510 000 W/C = 0.94..J f 7.0 253 121.0 0 124 900 423 22 900 3460 2 550 000 Average....... W.5 237 120.7 124 900 423 22 900 3590 2 530 000 1:2Y4:2% ...1 J 6.3 235 118.7 0.5 155 300 526 28 500 3480 2 500 000 W/C = 0.92.. I. 5.3 228 118.8 0.5 169 000 572 30 900 3870 2 520 000 Average....... 5.8 231 118.8 0.5 162 100 549 29 700 3670 2 510 000 1:2Y:2Y.... 1 6.6- 242 119.0 1.0 191 000 644 34 800 3640 2 390 000 W/C = 0.96..J 6.8 260 119.2 1.0 201 200 682 36 800 3480 2 500 000 Average....... 6.7 251 119.1 .0 196 100 663 35 800 3560 2 440 000 1:24:2% ...1 f 6.7 250 119.6 1.5 204 600 693 37 500 3550 2 500 000 W/C= 0.96..J 6.9 256 120.3 1.5 205 700 698 37 700 3400 2 330 000 Average....... 6.8 253 119.9 1.5 205 100 695 37 600 3470 2 410 000 1:2%:3%4t...1 f 7.5 280 118.5 0 132 400 448 24 300 2420 1 930 000 W/C = 1.17t J 1 4.0 231 120.6 0 142 500 482 26 100 3280 2 430 000 Averaget....... 5.7 255 119.5 137 500 465 25 200 2850 2 180 000 Concrete Made with Fine and Coarse Haydite Aggregates 1:1%:2%....1 J 7.1 257 102.0 0 95 500 324 17 500 3400 1 950 000 W/C = 0.88..J I 8.5 266 102.3 0 98 400 333 18 000 3090 1 850 000 Average....... 7.8 261 102.1 96 900 328 17 700 3240 1 900 000 1:2Y4:2% i..... 6.5 250 100.1 0 89 100 308 16 600 2450 1 740 000 W/C = 0.98..J 2.0 200 100.4 0 107 000 363 19 600 2990 1 650 000 Average....... 4.2 225 100.2 98000 335 18 100 2720 1 690 000 1:2Y:3%.... 5.6 235 99.9 0 86 100 291 16 700 1790 1 510 000 W/C = 1.19..J ( 4.6 213 99.5 0 67 200 227 12 300 1800 1 460 000 Average....... 5.1 224 99.7 76600 259 14 500 1790 1 480 000 *Including weight of the beam. tMade with lot 2 of cement, all others of lot 1. TESTS ON PLAIN AND REINFORCED CONCRETE I) K 600,/ / ,/ / / / '0 7 500 / ---/ / /<< Grovel Con~crele / .^^ yC-Hw/leife Con'cre/e 400--i- 1 ZY / ! "'-A//-t,'/Y/ Con4-crete 300- /00 IO ,/,,_ _ ___/__ 0 00/ 0000 3000 4000 SOOO 61000 7- 6oompre~ssve Stegholf Caro/ Cy/1ndkes /,7 l e sý7 n FIG. 26. RELATION BETWEEN MAXIMUM SHEARING STRESS AND COMPRESSIVE STRENGTH FOR BEAMS WITHOUT WEB REINFORCEMENT midway between support and load point. Initial readings were taken at a load of 1000 lb., and thereafter at loads calculated to produce increments of shear of 30 lb. per sq. in. Load was applied at the slowest testing speed of the machine. Successive increments of load were applied at intervals of about six minutes. 23. Discussion of Results.-Table 14 gives the essential results of the tests. For the beams without web reinforcement, those of gravel concrete developed a nominal shearing unit stress, v, equal to from 11 to 15 per cent of the compressive strength, f', of the 6 by 12-in. control cylinders. For the sand and Haydite concrete the ratio varied from 11 to 16 per cent, while for the All-Haydite concrete it varied from 10 to 14 per cent. These beams failed initially by diagonal tension. Values of the maximum shearing unit stress developed in beams without web reinforcement are plotted in Fig. 26 against the com- pressive strength of the companion cylinders. It is seen that the diagonal tension resistance, like bond strength, is proportionately ILLINOIS ENGINEERING EXPERIMENT STATION lower for the strong concretes than for the weaker ones. The average value of the ratio v/f' is about 0.12 and the lowest value is 0.10. Compared with the current allowable shearing stress of 0.02 f,, this indicates an ample factor of safety in all of the beams of this group. Of the beams containing web reinforcement, those with 0.5 and 1.0 per cent failed by slip of longitudinal steel and crushing of con- crete at the support; those with 1.5 per cent failed by tension in the longitudinal steel followed by crushing of the concrete. The strength of the stirrup steel used was unexpectedly high and it is evident from the maximum loads carried and from the measured stirrup stresses that the full strength of the web reinforcement was not developed in any beam of this group. The shearing stresses listed in Table 14 therefore do not represent the full web resistance of these beams, and no very definite estimate of the potential web strength of these beams can be made. VII. TESTS OF REINFORCED CONCRETE COLUMNS 24. Outline of Tests.-Since Haydite concretes have a lower mod- ulus of elasticity than gravel concretes, the value of n (the ratio of the modulus of elasticity of steel to that of concrete) is correspond- ingly higher for the Haydite concrete. The ratio, n, has been consid- ered to be of particular importance with regard to reinforced com- pression members, since if the concrete and steel are elastic and deform alike, the unit stress in the steel must be n times the unit stress in the concrete. A relatively high value of n in Haydite con- crete columns would presuppose a relatively high steel stress. If the column design were based upon a maximum allowable stress in the concrete, a Haydite column might be expected to carry a greater load at this concrete stress than would a gravel concrete column. How- ever, conditions of elasticity do not obtain at ultimate loads, and it is principally upon ultimate loads that allowable design stresses must be based. To furnish experimental data bearing upon the carrying capacity of Haydite and gravel concrete columns, a series of tests was made. Thirty columns were tested in this series, together with their con- trol cylinders. Gravel, C-Haydite, and All-Haydite concretes were used, all having the same mix, 1:1%:2k1, by loose volume. Plain columns, tied columns having 1.47 and 3.74 per cent longitudinal steel, and spiral columns having 1.22 per cent spiral and 1.60 and 4.08 per cent longitudinal steel were included in the series. Two identical TESTS ON PLAIN AND REINFORCED CONCRETE FIG. 27. COLUMN IN 3 000 000-LB. TESTING MACHINE columns of each kind were made. Three standard control cylinders were made with each column. 25. Type of Specimen and Method of Testing.-The test columns were nominally 8 in. in diameter and 60 in. long. The ties and spirals were placed close to the surface of the column, giving a sectional area of the plain and tied columns of 53.5 sq. in. and a core area of spiral columns, out to out of spirals, of 49.0 sq. in. The two percentages of longitudinal steel, roughly 1.5 and 4.0, were provided by four 2-in. deformed round bars and eight 12-in. deformed square bars, respec- tively. The spiral reinforcement, of 1.2 per cent was made of No. 5 gage hot-rolled rod of intermediate grade, coiled with a pitch of 1.35 in. Ties used were of No. 9 wire spaced 8 in. apart. The reinforcing ILLINOIS ENGINEERING EXPERIMENT STATION bars were milled to exact length and care was taken to bring the milled ends flush with the plane ends of the concrete column. This was done by holding the bars in contact with a planed cast-iron base plate, by means of an arrangement of collars and springs, during the placing and setting of the concrete. A neat cement cap at the top of the column was also brought flush with the ends of the column bars. The concrete was machine-mixed and the columns were molded in a steel form. Gage holes were drilled in the steel and gage plugs set in the con- crete within a period of 24 to 48 hours after making the column. The columns were kept covered with wet burlap for a day or two until gage holes had been prepared; they were then stored with the cylin- ders in the standard moist curing room until time of test. The columns were tested in a Riehle testing machine of 300 000-lb. capacity or in a Southwark-Emery machine of 3 000 000-lb. capacity. A spherical block was used at the base of the column in order to set it axially in the machine. After the column had been plumbed, the bottom spherical block was wedged so that it became in effect a fixed block. An upper spherical block was used between the column and the movable head of the machine. A view of a column in the Southwark-Emery machine is shown in Fig. 27. Load was applied in from ten to fifteen increments, with a time interval between increments sufficient only to take strain measure- ments. Twelve longitudinal and six lateral measurements were made on the concrete and twelve longitudinal measurements on the steel. Longitudinal measurements were made with a 10-in. Whittemore strain gage having a dial reading directly to 0.0001 in. Lateral defor- mations were measured with a special diameter gage of 8-in. gage length equipped with an Ames dial reading to 0.001 in. and having a multiplication ratio of 5 to 1. 26. Discussion of Results.-The principal results of the column tests are given in Table 15. As had been found in previous tests, these tests showed that the unit strength of an 8 by 60-in. plain con- crete column is somewhat less than that of a 6 by 12-in. cylinder of the same material. The ratio of plain column strength to cylinder strength was higher for Haydite concrete than for gravel concrete, the value being 82 per cent for gravel, 87 per cent for C-Haydite, and 88 per cent for All-Haydite concrete. It is difficult to compare directly the loads carried by the rein- forced columns in which similar reinforcing units were combined with gravel, C-Haydite, and All-Haydite concretes. This is because the TESTS ON PLAIN AND REINFORCED CONCRETE TABLE 15 PRINCIPAL RESULTS OF COLUMN TESTS, SERIES C Aggregates used: Fine Haydite, A-4; coarse Haydite, C-8; sand, S-3; gravel, G-3. See Tables 2 and 3. Columns 8 in. in diameter, 5 ft. high. Concrete proportions: 1:1%:2h, by loose volume. Column ties, No. 9 wire, 8-in. spacing; spirals, No. 5 hot-rolled rod, 1.35 in. pitch. Yield point of spiral, 49 400 lb. per sq. in. Column section area: gross area of plain and tied columns, core area of spiral columns. Columns, 6 x 12-in. Cylinder Yield Unit 28-Day Tests 28-Day Tests Long. Spiral Point Weight Column Area Steel Steel Long. Slump Flow Fresh Unit Corn- No. sq. in. per per Steel in. Concrete Total Ultimate pressive cent cent lb. per lb. per Ultimate Load, Strength E/1000 sq. in. cu. ft. Load, lb. per lb. per lb. sq. in. sq. in. Concrete Made with Natural Sand and Gravel Aggregates 00G01......... 53.5 0 0 ...... 8.9 281 149.1 180 100 3370 4290 3870 00G04....... 53.5 0 0 ...... 4.5 222 148.8 229 500 4290 5070 4070 1.5G01......... 53.5 1.47 0 54 500 5.9 228 148.3 201 100 3760 5260 4150 1.5G04......... 53.5 1.47 0 53 500 6.7 220 147.4 229 700 4290 5020 4040 S4G01......... 53.5 3.74 0 50 000 6.7 228 147.6 281 500 5250 5270 4080 4G04......... 53.5 3.74 0 46 500 7.8 238 148.9 273 000 5100 5290 3880 1.5G11......... 49.0 1.60 1.22 53 500 7.3 228 147.4 307 000 6260 5040 4110 1.5G14......... 49.0 1.60 1.22 50 500 6.5 237 147.4 282 000 5750 4890 4100 4G11......... 49.0 4.08 1.22 46 400 2.5 192 148.2 371 500 7580 5400 4410 4G14......... 49.0 4.08 1.22 48 200 6.7 228 146.7 348 000 7100 5280 4190 Concrete Made with Natural Sand and Coarse Haydite OH01......... 53.5 0 0 ...... 9.1 276 120.4 224 000 4180 4960 2940 OH04......... 53.5 0 0 .... 7.0 246 118.1 237 700 4440 4890 2410 1.5H01......... 53.5 1.47 0 52 000 8.2 285 119.8 203 500 3800 4890 2440 1.5H04......... 53.5 1.47 0 54 100 8.0 254 118.7 199 600 3730 5080 2810 4H01......... 53.5 3.74 0 46 700 7.7 258 119.0 259 000 4840 5200 2420 4H04......... 53.5 3.74 0 50 000 8.7 270 119.5 259 000 4840 5210 2660 1.5H11......... 49.0 1.60 1.22 57 600 8.1 244 121.6 293 000 5980 5150 2890 1.5H1114......... 49.0 1.60 1.22 57 600 8.0 242 121.3 279 500 5700 5050 3000 4H11......... 49.0 4.08 1.22 49 000 6.5 234 121.5 351 500 7170 5280 2940 4H14......... 49.0 4.08 1.22 46 900 6.9 221 121.9 359 500 7340 5280 3030 Concrete Made with Fine and Coarse Haydite Aggregates OA01......... 53.5 0 0 .... 7.0 251 99.5 180 200 3360 4050 1760 0A04......... 53.5 0 0 .... 7.0 268 99.9 192 600 3600 3880 1720 1.5A01......... 53.5 1.47 0 54 100 5.9 257 98.7 179 500 3350 3940 1830 1.5A04......... 53.5 1.47 0 55 600 7.6 271 99.3 180 600 3370 4120 1790 4A01......... 53.5 3.74 0 50 800 7.0 270 101.3 228 000 4260 4070 1860 4A04......... 53.5 3.74 0 51 500 7.9 253 99.5 239 000 4470 3760 1780 1.5All......... 49.0 1.60 1.22 57 600 7.5 269 100.4 216 000 4410 3980 2030 1.5A14......... 49.0 1.60 1.22 57 600 7.5 270 101.2 192 200 3920 3960 1970 4A1......... 49.0 4.08 1.22 45 800 7.3 238 101.5 271 000 5530 4610 2040 4A14......... 49.0 4.08 1.22 48 000 6.9 251 100.8 287500 5870 4370 2000 strengths of the three concretes differed considerably. An indirect comparison may be made through the use of the hypothesis, con- firmed by other tests,* that the portion of the strength of a reinforced concrete column attributable to the concrete is equal to the strength *F. E. Richart and G. C. Staehle, Progress Report on Column Tests at the University of Illinois, Jour. Am. Cone. Inst., Feb., 1931., p. 731-60. ILLINOIS ENGINEERING EXPERIMENT STATION TABLE 16 RELATIVE STRENGTHS OF COLUMNS Values from Table 15, corrected on the basis of a cylinder strength of 5000 lb. per sq. in. for all types of concrete. Column Strength-lb. per sq. in. Relative Column Strengths Reinforcement Gravel C-Haydite All-Haydite Gravel C-Haydite All-Haydite Concrete Concrete Concrete Concrete Concrete Concrete Ties, 1.5 per cent Long........ 3910 3780 4200 1.00 0.97 1.08 Ties, 3.7 per cent Long........ 4955 4670 5285 1.00 0.94 1.07 1.2 per cent Spiral, 1.6 per cent Long .................. 6030 5760 4990 1.00 0.95 0.82 1.2 per cent Spiral, 4.1 per cent Long.................... 7100 7040 6095 1.00 0.99 0.86 Average...........:......... .... .... .... 1.00 0.96 0.96 of a plain column of the same net section of concrete. If a gravel concrete having a cylinder strength of 5000 lb. per sq. in. is used in a spirally reinforced column with 4 per cent of vertical reinforcement, the concrete should contribute to the load-carrying capacity of the column an amount equal to 5000 x 0.82 x 0.96 lb. per sq. in. of core area. The average strengths of Table 15 have been corrected on the basis of a cylinder strength of 5000 lb. per sq. in. throughout and the corrected values are compared in Table 16, using gravel concrete columns as the basis of comparison. From Table 16 the general conclusion may be drawn that for columns made with identical reinforcement and with concretes of the same compressive strengths, the average effectiveness of the Haydite columns is only slightly less than that of corresponding columns of gravel concrete. The greatest difference between the two is found in the case of spiral columns of All-Haydite concrete, for which the strength is only 84 per cent of that of columns of gravel concrete. While the comparison between different types of concrete is fairly definite, the effectiveness of the reinforcing steel does not compare well with other test results. Tests of similar columns showed* that the portion of column strength attributable to the vertical steel was equal to the steel area times the yield point strength of the material, and that the spiral steel was about twice as effective as an equal weight of vertical steel. In comparison with the plain columns, the reinforced columns of both Haydite and gravel concrete of Table 15 show a relatively low effectiveness of the reinforcing steel. For the *F. E. Richart and G. C. Staehle, Progress Report on Column Tests at the University of Illinois, Jour. Am. Cone. Inst., Feb., 1931., p. 731-60. TESTS ON PLAIN AND REINFORCED CONCRETE 59 (c) (d) FIG. 28. GRAVEL, C-HAYDITE, AND ALL-HAYDITE CONCRETE COLUMNS AFTER TEST ILLINOIS ENGINEERING EXPERIMENT STATION Longi/a'd/?a/ U1n/t Deformatt/on FIG. 29. STRESS-STRAIN CURVES FOR PLAIN AND TIED COLUMNS 'NJ TESTS ON PLAIN AND REINFORCED CONCRETE 7000 6000 5000 4000 3000 ecooo /000 o 6" 000 5000 4000 3000 ooo0 /000 i 4, 7 F 2 / / -9 Longitud/ina/ Unit Deforati/on FIG. 30. STRESS-STRAIN CURVES FOR SPIRALLY REINFORCED COLUMNS 7 77 2 0.0004 S. 1_S% Spir/l &e/forcert'e - .% Lro t'ko '/ we farc S_ oira//y Renfored S-3 0.0004 Conlvwe 4./% LonZ'/l.A /^ ? "t 'e/n7/ 0 L. ?• i^y^/?7^^7: ý-Cev:;Ve,'7, 71r/1 i "rreT/ror^/^ce/? %- Z116 ILLINOIS ENGINEERING EXPERIMENT STATION tied columns, one explanation is found in the settlement cracks that occurred beneath the ties near the tops of the columns, as shown in figures to follow. These cracks certainly reduced the effective area of the concrete to the net area within the tie. However, even when allowance is made for a reduction in area of from 53.5 to 49.0 sq. in., the effectiveness of the steel in the tied columns is comparatively low. By a similar comparison, the effectiveness of combined longitudinal and spiral reinforcement is less than would be expected. This result is not peculiar to Haydite concrete, since it occurs also with the gravel concrete columns. As shown by Table 16, however, the Haydite columns show slightly less, rather than greater, strength than the gravel concrete columns, regardless of the values of modulus of elas- ticity. The results thus justify the conclusion that values of the modular ratio, n, have little or no bearing upon column strength. Figure 28 shows typical columns after testing. The manner of failure of all plain columns was the sudden shattering failure typical of plain high-strength concrete in compression. The tied columns failed by buckling of vertical bars between ties. The spirally rein- forced columns reached and passed the maximum load slowly without breakage of the spirals. Stress-strain curves for longitudinal gage lines of all columns are given in Figs. 29 and 30. These curves reflect the low values of modulus of elasticity for the Haydite concrete columns in the relative- ly high strains shown for these columns. It is also noticeable that the total deformation for Haydite concrete columns is considerably greater than for those of gravel concrete. VIII. TESTS OF BONDING OF FLOOR FINISH 27. Outline of Tests.-The following tests of the adhesion of a finish course of sand and gravel concrete to reinforced base slabs of Haydite and gravel concrete were made to see what effect was pro- duced by the difference in elasticity of the finish and base courses or by the tendency of Haydite particles to rise to the surface of the base course. While these tests do not include the effect of differential shrinkage of the two courses, of heavy impact or abrasion, or of a long time interval between the placing of the two courses, they should show any obvious defects in the bonding of the finish course to the two types of base. Twelve reinforced concrete slabs, of dimensions shown in Fig. 31, were tested in flexure on a 4-ft. span, with one-third point loading. This loading produced a relatively high horizontal shear tending to TESTS ON PLAIN AND REINFORCED CONCRETE Half E/evation Sec/ion FIG. 31. DETAILS OF TEST SLABS OF ADHESION SERIES X separate the base and finish courses in the outer thirds of the span. The 3-in. finish course was in all cases a 1:1:2 mixture of Universal cement, Lincoln sand and Covington pea gravel. It had a water- cement ratio of 0.67. The 5-in. base slabs were of two types: one made with Covington sand and gravel and the other with fine and coarse Haydite. Two mixtures were used: a structural concrete, 1:21:2%3, by volume, and a lean concrete, 1:42:51, approximating floor fill. Three slabs and three control cylinders were made of each mix. Each test slab was reinforced with three 12-in. deformed round bars of intermediate grade, running full length of the slab and ending in hooks bent to a 12-in. radius. These bars were tied together in a unit by a double system of Y4-in. round U-stirrups of structural grade spaced 22 in. apart in the outer thirds of the span. None of the reinforcement projected into the finish course of the slab. 28. Making and Testing of Slabs.-The slabs were poured in wooden forms. The base concrete was mixed in a one-bag batch mixer, hauled to the form in wheelbarrows, shovelled into place and spaded uniformly around the reinforcing steel. The concrete was struck off to a level %4-in. below the edges of the form. Some settle- ment occurred after this screeding. After the concrete had stiffened, the surface was struck with a wood float and left exposed to the air until the following day. The finish course was made and placed in general accordance with the recommended practice given in the Report* of Committee 802 on Concrete Floor Finish, of the American Concrete Institute. The chief departure from the recommended practice was in the use of a rather fine sand instead of the coarser one specified. While this might *John G. Ahlers, "Good Practice in Concrete Floor Finish," Jour. Am. Cone. Inst., Mar., 1930, p. 520. ILLINOIS ENGINEERING EXPERIMENT STATION FIG. 32. ADHESION SLAB IN TESTING MACHINE affect the wearing quality of the surface, it probably had little effect upon the adhesion to the base course. Twenty-four hours after the base slab was poured, it was prepared for the finish course. This was done by wetting the surface thoroughly and then brushing on a thin coat of neat cement mortar with a wire brush. The finish was placed immediately after the application of the mortar coat. It was tamped uniformly over the surface of the slab and into the corners of the form to insure a full surface of adhesion in every slab. The surface was then struck off level with the edges of the forms and compacted and smoothed with a wood float. The surface was trowelled with a steel trowel one to two hours after being struck off. The average measured thickness of finish used with the four con- cretes was as follows: 1:5 gravel, 0.74 in.; 1:10 gravel, 0.75 in.; 1:5 Haydite, 0.75 in.; 1:10 Haydite, 0.90 in. The forms were stripped one day after the finish course was placed, and the slabs were stored under wet burlap for about five days; they were then removed to the moist room for standard moist curing at 70 deg. F. All cylinder molds were stripped after 48 hours and the cylinders cured in the moist room until the time of test, when they were tested moist. The slabs were removed from the moist room when 28 days old and tested TESTS ON PLAIN AND REINFORCED CONCRETE TABLE 17 RESULTS OF ADHESION TESTS Aggregates used: Fine Haydite, A-4; coarse Haydite, C-6; sand, S-4; gravel, G-3. See Tables 2 and 3. Slabs 5% by 12 in., 4-ft. span, one-third point loading. Thickness of base, 5 in.; finish, % in. Longitudinal reinforcement, 1 per cent. Slab 10 Ax3 failed initially by diagonal tension, all others by tension in longitudinal steel. Load on Slab, lb. Data of Concrete Used From 6 by 12-in. Cylinders Slab No. At Yield At Mix, by W Slump, Unit Corn- Initial Point Maximum Volume by V . Flow Weight pressive Modulustof Volume Fresh Strength Elasticity Concrete lb. per thousands of lb. per cu. ft. sq. in. lb. per sq. in. Finish Course, Sand and Pea Gravel, All Slabs 1:1:2 0.67 5.0 ... 149.2 5540 4030 Base Course, Sand and Gravel 5 Gxl.... 15 700 19 700 1:2Y:2% 0.90 5.8 220 146.8 4390 3720 2.... 15 800 18 900 3.... 16 200 18 700 15 900 19 100 10 Gxl.... 15 700 16 700 1:4%:5Y 1.70 1.5 213 142.9 1260 3220 2.... 15 800 17 900 3.... 16 200 18 100 15 900 17 600 Base Course, Fine and Coarse Haydite 5 Axl.... 16 000 19 200 1:24:2%y 1.02 5.3 235 97.1 3260 1860 2.... 16 400 20 300 3.... 15 900 20 100 16 100 19 900 10 Axl.... 15 700 18 300 1:4I :5% 2.02 1.9 ... 90.5 620 920 2.... 15 500 17 400 3.... ...... 16 300 15 600 17 300 within 6 hours after removal. Before the test a coat of whitewash was applied to the sides and ends to aid in the detection of cracks. In the 1:10 Haydite concrete slabs the moisture present in the base of the slab kept coming to the sides so rapidly that the whitewash failed to dry during the day of the test. A view of the testing apparatus is seen in Fig. 32, which shows a typical slab after test. The work was done in an Olsen testing machine of 300 000-lb. capacity. The load was applied through a ILLINOIS ENGINEERING EXPERIMENT STATION FIG. 33. SLABS OF 1:5 MIX AFTER TEST spherical bearing block to four 4-in. channels and thence through rollers and bearing plates to the third-points of the span. Load was applied in increments of 500 to 1000 lb. at the slowest testing speed of the machine. A close watch was kept for cracking of the slab and any indication of loosening of the topping from the base, but no strain measurements were taken. After the steel had passed its yield point, as indicated by a dropping of the weighing beam and opening of tension cracks, a faster test speed was maintained until final rupture occurred. The modulus of elasticity of the concrete was determined from stress-strain measurements on the control cylinders. 29. Discussion of Test Results.-With one exception the slabs failed initially by tension in the longitudinal steel. The single excep- tion was a Haydite slab of 1:10 mix which failed initially by diagonal tension at a load of 13 900 lb. In this slab the yield point of the longitudinal steel was reached at a load of 15 700 lb. and the ultimate load carried was 16 300 lb. At failure, a piece split off the side of the slab (No. 10 AX3, Fig. 34) near the end and there was evidence of slip of the hooked bar and compression of the concrete within it. TESTS ON PLAIN AND REINFORCED CONCRETE FIG. 34. SLABS OF 1:10 MIX AFTER TEST There was also a marked distortion of the weak concrete due to high bearing pressures at the support. Complete test data are given for all slabs in Table 17. In no test did the finish crack loose from the base before the ultimate load was reached. The finish in both Haydite and gravel concrete slabs re- mained integral with the base with deflections considerably in excess of one inch on a 4-ft. span. At final failure, with deflections of about 11 in., crushing of the 4-in. finish course occurred in the middle third near one of the load points and a section of the finish split off. The sections which split off from the Haydite slabs had portions of the base adhering to the topping, the fracture being mostly below the topping. Where the base was of gravel concrete the fracture occurred at the bonding surface between the base and topping. A general view of all the 1:5 slabs is shown in Fig. 33 and of all the 1:10 slabs in Fig. 34. In each figure the upper three slabs are of Haydite concrete and the lower three of gravel concrete. The tests show no marked difference in the adhesion of the floor finish to the Haydite and Gravel concrete bases, though the plane of ILLINOIS ENGINEERING EXPERIMENT STATION cleavage at the point of failure was more definite with the gravel than with the Haydite concrete. This separation at the crushing failure of the slab might be expected in view of the great deflection and dis- tortion of the slab; it is noteworthy that there was no such separation in the outer thirds of any slab, even at failure. An unexpected feature of the tests was found in the fact that the slabs of lean concrete carried nearly as much load as those of the rich mixture. Evidently the strong 3%-in. finish course furnished most of the compressive flexural strength of the slabs made with lean con- crete. However, since so much depended upon the quality of this finish course, it is doubtful if it should be depended upon in design, beyond the inclusion of the finish thickness in computing the effective depth, which seems justified by these tests. Considering the low strength of these 1:10 concretes, it is surprising that the longitudinal steel received sufficient anchorage to develop its tensile strength. As noted previously, these tests do not give information as to the effect of shrinkage or of heavy service conditions on a large floor slab, but within their limitations they show no obvious defects in the bonding of a finish course to bases of either Haydite or gravel con- crete. Consequently, they give support to the practice of including the finish thickness in computing the effective depth of the slab for design purposes. IX. GENERAL DIscussION 30. Applications of Test Data to Design.-The advantages to be obtained in design with the use of light-weight concrete are at present accompanied by a relatively high cost of the material so that the user will generally find it desirable to make a careful study of the overall economy of a particular structural design. The suitability of Haydite concrete for use in tall office buildings and in floors and sidewalks of large bridges is fairly obvious, and some general conclusions have been drawn regarding such structures.* The saving in dead load naturally affects all parts of a building structure, so that considerable economy may result indirectly, as when a cheaper type of foundation may be used than that required with a heavier dead load. Other features, not directly involved in the structural design, such as insulating and fireproofing qualities, may influence the economy of the Haydite concrete design. In considering solely the strength requirements of structural mem- bers, two principal points of difference between Haydite and ordinary *Frank A. Randall, "Economics of Light Weight Concrete in Buildings," Jour. Am. Cone. Inst. Vol. 2, No. 7, p. 925, March, 1931. TESTS ON PLAIN AND REINFORCED CONCRETE concrete are found; Haydite concrete combines light weight with a low modulus of elasticity The way in which these two items affect design will be shown by an analysis of several types of structural members. It may be noted that the comparisons to follow involve values of strength and modulus of elasticity from tests of cylinders rather than of beams. The limited amount of beam test data available is in accord with the analyses given; Haydite concrete beams show larger values of k and consequent smaller values of j, and increased strains and deflections at a given load, as compared to gravel concrete beams. Haydite concrete beams tested in studies of diagonal tension showed no initial failures at the extreme fiber in compression, even though percentages of longitudinal reinforcement (structural grade) as great as 2.8 were used, and tensile failure of the steel was developed. Obviously, beam and slab tests corresponding to the range of designs considered here would be impracticable, and such tests have not been made. Instead, the authors have felt justified in determining prop- erties of the material and in applying the results by means of well- established theory of the behavior of such members. 31. Effect of Low Modulus of Elasticity on Design of Beams and Slabs.-The following analysis covers only the effect of low modulus of elasticity and is independent of the question of weight of concrete. The following standard symbols and notation are used: A, = cross-sectional area of tensile reinforcement. b = width of rectangular beam or slab. d = effective depth of beam or slab. Ec = modulus of elasticity of concrete. f' = ultimate compressive strength of concrete. fc = compressive unit stress in extreme fiber of concrete in flexure. f, = tensile unit stress in longitudinal reinforcement. jd = moment arm of resisting couple. kd = distance from compression face to neutral axis of beam. M K = bd - resisting moment factor. M = resisting moment. n = ratio of modulus of elasticity of steel to that of concrete. A. p - = - ratio of effective area of steel to that of concrete in bd beam. The subscript, g, is used in connection with terms referring to gravel concrete. ILLINOIS ENGINEERING EXPERIMENT STATION TABLE 18 COMPARATIVE DESIGN OF RECTANGULAR BEAMS AND SLABS BASED UPON EXPERIMENTAL VALUES OF E Balanced Reinforcement, f. = 20 000 lb. per sq. in., f, = 0.40/',. Prop- Sec. erty Gravel Concrete C-Haydite All-Haydite f' 2000 2500 3000 3750 2000 2500 3000 3750 2000 2500 3000 3750 fe 800 1000 1200 1500 800 1000 1200 1500 800 1000 1200 1500 A n 10.1 9.4 8.8 8.0 13.4 12.7 11.9 11.0 18.4 17.1 16.0 14.5 k 0.288 0.319 0.346 0.375 0.348 0.388 0.417 0.452 0.423 0.461 0.490 0.520 J 0.904 0.894 0.885 0.875 0.884 0.871 0.861 0.849 0.859 0.846 0.837 0.827 0.0058 0.0080 0.0104 0.0140 0.0070 0.0097 0.0125 0.0169 0.0085 0.0115 0.0147 0.0195 104 143 184 246 123 169 216 288 145 195 245 322 Slab Comparison Based upon Constant Resisting Moment. b constant, d varies d B *, 100 100 100 100 92 92 92 91 84.5 86 87 87.5 per cent A. T-. 100 100 100 100 111 111 110.5 110 123.5 124.0 122.5 121.5 per cent Slab and Rect. Beam Comparison Based upon Constant b and d for All Concretes K C -"' 100 100 100 100 118 118 117 117 139 136 133 131 per cent A, A.-, 100 100 100 100 121 121 120 121 146 144 141 139 per cent Rectangular Beam Comparison Based upon Constant Resisting Moment and upon Proportional b and d for Haydites and Gravel bd D (-Io 100 100 100 100 89.0 89.5 90.0 90.0 80.0 80.5 82.5 83.5 per cent A, ,' 100 100 100 100 107 108 108 109 117 116 117 116 per cent Note: Subscript g denotes gravel concrete. Comparative properties and design factors for rectangular beams and slabs have been computed and tabulated in Table 18 for gravel, C-Haydite, and All-Haydite concretes having compressive strengths of 2000, 2500, 3000, and 3750 lb. per sq. in., using experimental values of E from Fig. 8. In Section A of Table 18, the factors n, k, j, p, and K are given for the three kinds of concrete, based upon the use of f, = 20 000 lb. per TESTS ON PLAIN AND REINFORCED CONCRETE sq. in. and fc = 0.40 fJ. The low value of Ec for Haydite concrete, while causing a slight decrease in j and a considerable increase in k and p, had the effect of increasing considerably the resisting moment factor K. The slab and rectangular beam comparisons which follow in the table are based upon various design limitations. In Section B, Table 18, under the heading "Slab Comparison Based upon Constant Resisting Moment, b constant, d varies," are given figures which are intended to show the net effect of the changed design factors for Haydite concretes. These have been computed as follows: For a slab of variable depth, d, with balanced reinforcement, the resisting moment M = Kbd2 = 12 Kd2 d= M 12K For a constant resisting moment the ratio of depth, d, required for Haydite concrete slabs to depth, d,, required for gravel concrete is d, K d Ratios of - for C-Haydite concrete are shown in Table 18 to vary from 0.90 to 0.91 and for All-Haydite concrete to vary from 0.845 to 0.875. Percentages of total volume of concrete required would be from 1 to 3 per cent higher than these percentages of depth due to a constant thickness of covering. Ratios of steel areas, A., required for Haydite slabs to steel areas, A.,, required for gravel concrete are net ratios based on the increased p required for balanced reinforcement in Haydite and the reduced d d as shown by the ratio -" Thus: A, d p A., dg pg Percentages of A. (Table 18, part B) show for this comparison a requirement of from 111 to 110 per cent of steel area for C-Haydite ILLINOIS ENGINEERING EXPERIMENT STATION concrete and from 123.5 to 121.5 per cent for All-Haydite concrete. The total effect is then to develop the same resisting moment with C-Haydite concrete with d reduced 8 to 9 per cent and A, increased 11 to 10 per cent, and with All-Haydite concrete with d reduced 15.5 to 12.5 per cent and A, increased 23.5 to 21.5 per cent. Section C of Table 18, which is headed "Slab and Rectangular Beam Comparison Based upon Constant b and d for All Concretes," applies to the case in which a possible reduction in depth of slab is not utilized. It shows that with C-Haydite concrete an increase of 18 to 17 per cent in resisting moment is obtained by the use of 21 to 20 per cent more steel, and that with All-Haydite concrete an increase of 39 to 31 per cent in resisting moment is obtained by an increase of 46 to 39 per cent in steel. The exact value of the increased carrying capacity or possibility of increased span length which is purchased in this case by the use of a higher percentage of steel is not at once easily determined as a simple percentage. If, however, no increase in resisting moment is desired, a constant resisting moment with con- stant depth of section may be provided with a Haydite concrete slab by the use of slightly more steel than is required by gravel concrete. This is due to the slight difference in j for the two materials, an effect due only to variations in modulus of elasticity. The compensating effect of a reduction in weight in reducing the steel required with Haydite concrete is considered in Section 32. In Section D of Table 18 comparisons of rectangular beams are based upon the condition that it is possible to reduce both b and d, and that b and d for Haydite concrete beams are proportional to b and d for those of gravel concrete. Calculations were made as follows: For rectangular beams with balanced reinforcement M = Kbd2 b = constant X d = Cd bd = Cd2 M = C Kd3 d2 UK For a constant resisting moment d2 = constant X bd d9 3 2 bd, d 2 ^\K) TESTS ON PLAIN AND REINFORCED CONCRETE bd These area ratios (- for C-Haydite concrete are shown to vary (bd),g from 89 to 90 per cent, and for All-Haydite concrete to vary from 80 to 83.5 per cent, of gravel concrete areas. Net ratios of steel areas, A., required for Haydites to areas, Ao, required for gravel concrete are 107 to 109 per cent for C-Haydite and 117 to 116 per cent for All-Haydite concretes. For these compu- tations A, bd p A., (bd) pg giving ratios of steel areas based on the reduced beam section as was done in the case of slabs of variable depth. Studies similar to those of Table 18 have been made by the use of two other assumptions as to the values of Ec: (a) by assuming the value for the three concretes to be 100, 75, and 55 per cent of the value specified by the A.C.I. Joint Building Code, and (b) by assum- ing constant values of E of 3 400 000, 2 500 000, and 1 900 000 lb. per sq. in. for gravel, C-Haydite, and All-Haydite concretes, respec- tively. These different assumptions produced relatively small diver- gences from the corresponding derived quantities listed in Table 18. 32. Combined Effect of Lightness in Weight and Low Modulus of Elasticity on Slab and Beam Design.-By reference to Fig. 7 it may be estimated that C-Haydite concrete is roughly 80 per cent, and All-Haydite 67 per cent, as heavy as gravel concrete. The effect of a reduction in dead load of a structure depends upon the percentage of dead to total load and upon what advantage can be taken of the reduced load. If beam sizes, spans, or slab thicknesses are limited by considerations other than stress, it may not be possible to obtain full advantage. Section A, Table 19, shows, for various ratios of dead to total load, the total load per square foot for floors of a given thickness made with the three grades of concrete. This is expressed in terms of resisting moment, taking the value for gravel concrete as 100 per cent. This reduction in applied load may now be combined with the data of Table 18 to show the combined effect of decreased load and low modulus of elasticity in the case of Haydite concretes. Choosing, for example, a concrete for whichf, is 3000 lb. per sq. in. the compari- son of Section B, Table 19 is made, similar to the corresponding section of Table 18. The percentages given for slabs with balanced ILLINOIS ENGINEERING EXPERIMENT STATION TABLE 19 COMBINED EFFECT OF LIGHTNESS IN WEIGHT AND Low VALUE OF E Comparative design of slabs and rectangular beams based upon experimental values of E. Balanced reinforcement, f, = 20 000 lb. per sq. in., f, = 0.40f/'* Compressive strength of concrete, f' = 3000 lb. per sq. in. For values of n, k, j, p, and K see Table 18. Sec. A B C reinforcement show a required slab depth for C-Haydite concrete of 87 to 90 per cent, and for All-Haydite concrete, 79 to 84 per cent, of that required by gravel concrete. The corresponding steel areas required for balanced reinforcement are 105 to 108 per cent for C- Haydite concrete, and 112 to 118 per cent for All-Haydite concrete, as compared to 100 per cent for gravel concrete. The comparisons of rectangular beams of Section C, Table 19, in which it is assumed that a reduction may be made in both b and d, show smaller requirements in both concrete and steel section than those of Section B. Thus for the case in which the dead load is one- half of the total load the cross-section required for an All-Haydite beam is only 73 per cent of that for a gravel beam, while the former requires 4 per cent more reinforcing steel than the latter. Prop- erty Gravel Concrete C-Haydite Concrete All-Haydite Concrete Dead Load Total Load' 50 40 30 20 50 40 30 20 50 40 30 20 per cent M M- 100 100 100 100 90 92 94 96 83 87 90 93 per cent Slab Comparison Based upon Reduced Resisting Moment Required. b constant, d varies d 7,' 100 100 100 100 87 88 89 90 79 81 82 84 per cent A, 100 100 100 100 105 106 107 108 112 114 116 118 per cent Rectangular Beam Comparison Based upon Reduced Resisting Moment; Proportionally Reduced b and d bd , 100 100 100 100 84 85 86 88 73 75 77 79 per cent -. 100 100 100 100 101 102 104 106 103 106 109 111 per cent TESTS ON PLAIN AND REINFORCED CONCRETE The analysis upon which these figures are based is as follows: (a) Slab comparison: d _ K M depth ratios, K M do K M, A,8 d p steel ratios, A, - dp As, d, p, (b) Rectangular beam comparison: . bd _ S/K0 M2 concrete areas ratio, (bd)- = -(- M (bd), K M, A, bd p steel ratios, A. = bd p A80 (bd) ,p The foregoing studies have not taken into account the reduction in dead load due to the reduction in concrete section for Haydite concretes, which would lead to a further slight decrease in both steel and concrete requirements. 33. T-Beams and Members with Compressive Reinforcement.-The usefulness of light-weight concrete in T-beams in which the depth is frequently limited by shear or because of architectural considerations is somewhat restricted. Using Haydite concrete with a given depth of beam, the decrease in bending moment due to reduced dead load will in general be somewhat greater than the decrease in value of j, resulting in a slight saving of steel. In Haydite beams of the same size as gravel concrete ones the efficiency of compressive steel, according to present theories, would be increased due to the large values of n and k produced. However, recent studies of the effect of plastic flow in concrete indicate that the values of these factors based on elastic action of the member have little influence upon the ultimate capacity of the member. Without further experiments, this apparent advantage of the Haydite over gravel concrete remains unproved. As noted in Section 26, a similar conclusion may be drawn in the case of reinforced concrete columns, for which the apparent advantage produced by a low modulus of elasticity for Haydite concrete was not verified by the tests. The principal advantage to be obtained with Haydite concrete columns is with regard to reductions in load. Many building codes permit the use of a reduction factor on column live loads, particularly in tall buildings. Since the column is then ILLINOIS ENGINEERING EXPERIMENT STATION designed for the reduced live load plus the dead load, a decrease in the latter due to the use of Haydite concrete is of even greater impor- tance than it was in the design of beams and slabs. The same argument applies to footings, which are commonly designed for reduced column live loads. 34. Deflections.-The reduction in concrete section made possible through the use of Haydite concrete may introduce increased bond and shearing stresses and greater deflections. These stresses may govern the size of the concrete section; it is also possible that in some cases it will be necessary to guard against undue deflection. A low modulus of elasticity, in combination with the usual stresses, at once indicates a large deflection. For homogeneous beams of constant section and span, the relative deflections at a given load should be inversely proportional to their moduli of elasticity. This direct relation does not hold for reinforced concrete, as is seen by referring to Maney's equation* which gives the deflection D as a function of the separate deformation of the extreme fiber of steel, e,, and of concrete, ec. K is a constant depending upon the loading, 1 the span length and d the effective depth of section. 12 D = K d (e, + ec) It is seen that while the value of ec would be directly affected by a low value of Ec for the concrete, the value of e, would be practically unchanged. Since 6c is generally somewhat less than e,, the increase in deflection of a reinforced concrete beam due to a decrease in Ec will be less than half that to be expected for a homogeneous beam. Referring to the tests of Series 1, beams made with gravel, C-Haydite, and All-Haydite concretes having relative values of E of about 1.00, 0.75, and 0.55, developed relative deflections of 1.00, 1.15, and 1.30, respectively, under a given load. This verifies the foregoing statement. The point to be emphasized in connection with reinforced Haydite concrete is that both the low modulus of elasticity and the oppor- tunity for reducing concrete section produce a tendency toward increased deflections. While deflections have usually received little attention in building design, the present tendency is toward the use of large flexible floor panels in which large deflections are a natural consequence. With such construction, the matter of allowable deflections may require attention. *G. A. Maney, "Relation between Deformation and Deflection in Reinforced Concrete Beams," Proc. A.S.T.M., Vol. XIV, Part II, 1914. TESTS ON PLAIN AND REINFORCED CONCRETE X. CONCLUSIONS 35. Conclusions.-In summarizing the principal features of the tests described in the bulletin, the following conclusions have been formulated: (1) The average unit weights of the fine and coarse Haydite aggregates used were, respectively, 54 and 43 lb. per cu. ft. by dry, loose measure, and 62 and 47 lb. per cu. ft. by standard dry, rodded measure. (2) The weight per cubic foot of the Haydite concretes used (when freshly molded) varied from 93 to 106 lb. for All-Haydite con- crete and from 112 to 126 lb. for concrete made with sand and coarse Haydite. The range in weights is due to variations in mixtures and in proportions of fine and coarse aggregates used. (3) The proportion of coarse aggregate that may be used depends upon its size, whether the material be Haydite, gravel, or limestone. The gravel and limestone used in these tests were of nearly the same size as the C-Haydite. The usable proportion of any of these aggre- gates was limited by the 3/-in. maximum size. The ease of working and placing of concrete is also very largely dependent upon the aggregate size, but it is evident that the gradation and texture of the Haydite aggregates required a somewhat greater water-cement ratio than gravel aggregates in similar mixes, as indicated by slump and flow tests. It is recommended that to secure workability the proportion by volume of coarse Haydite, of the sizes described herein, shall not exceed 55 per cent of the total aggregate. With this proportion, the unit weight of the usual range of structural concretes will be 115 to 120 lb. per cu. ft. for C-Haydite concrete and 95 to 100 lb. per cu. ft. for All-Haydite concrete. (4) Concrete was made with Haydite aggregates with no particu- lar departure from ordinary methods. The same mixing procedure was used with Haydite concrete as with gravel and limestone con- cretes. With the consistencies used in Series 1, no difference was noted in the work required to place the concrete or in the appearance of the molded surfaces of beams made with gravel, limestone, and coarse Haydite. The Haydite beams presented a harsher top surface, making finishing more difficult where no topping mortar was used. (5) Because of the high water absorption by the Haydite, the calculation of the water-cement ratio involves an accurate determina- tion of the amount of absorbed moisture. The tests indicate that the absorption allowance for a Haydite aggregate is not a fixed quantity, but depends upon the initial moisture content of the aggregate as ILLINOIS ENGINEERING EXPERIMENT STATION used, when mixed with cement and water. The total possible absorp- tion is greater if the material is initially moist than if it is dry. The allowances used for concrete mixtures were found by test on aggre- gates immersed in water for one hour; for dry Haydites, this amounted to 7 per cent for the coarse, and 14 per cent for the fine; for initially moist aggregates, varying percentages, as indicated in Fig. 2. These percentages are based upon the dry weight of the light aggregates, which are about one-half the weight of sand and gravel aggregates. (6) While concrete made with Haydite aggregate generally re- quired greater water-cement ratios for similar mixes and equal slumps, the relation between compressive strength and water-cement ratio of Haydite concrete does not differ greatly from that for gravel concrete, all test results falling within a narrow zone as shown in Fig. 6. Concretes made with initially moist Haydite were consist- ently a little stronger than those made with dry Haydite aggregates. The strength differences found are probably due to the effect of errors in absorption allowances upon the water-cement ratio. There is no indication that the strength of any of the mixtures was limited by the strength of the aggregate particles. (7) The increase in compressive strength with age of moist stored Haydite concrete compares well with that for gravel concrete for periods up to a year. (8) The ratio of bond resistance to compressive strength at the age of 28 days was essentially the same for like mixes of gravel and Haydite concrete. (9) For beams without web reinforcement, which failed by diag- onal tension, the ratio of the shearing unit stress to the compressive strength of control cylinders was practically the same for corre- sponding mixtures of gravel and Haydite concrete. (10) The ultimate strengths of reinforced columns made with Haydite and gravel concrete were compared, after adjusting to com- pensate for differences in concrete strength. The relative strengths of Haydite columns as compared with columns of gravel concrete with similar reinforcement are: Tied columns, C-Haydite... ........... 96 per cent Tied columns, All-Haydite............. 108 per cent Spiral columns, C-Haydite ............. 97 per cent Spiral. columns, All-Haydite. ........... 84 per cent The only appreciable weakness is seen in the spiral columns of All- Haydite concrete, where the spiral reinforcement appears to be less effective than with other types of concrete. There is no justification TESTS ON PLAIN AND REINFORCED CONCRETE here for increasing the allowable unit stress in the Haydite columns because of their low modulus of elasticity. (11) No apparent defects in the adhesion between a base slab of Haydite concrete and the finish course were found, in spite of differ- ences in modulus of elasticity of the two courses, nor was any effect produced by the floating of Haydite particles to the surface of the base slab. The tests showed that a satisfactory bonding of finish to base course could be obtained. (12) A low modulus of elasticity is one of the outstanding char- acteristics of Haydite concrete. For the considerable range of mix- tures and consistencies tested, the values of the initial modulus of elasticity for C-Haydite concrete were about 75 per cent, and for All-Haydite concrete about 55 per cent, of the values for gravel concretes of corresponding strength. (13) The design of reinforced Haydite concrete differs from the usual design principally because of two features, lightness in weight and low modulus of elasticity. The first permits large reductions in dead load over ordinary concrete, thus making possible a reduction in size of members of a floor system, or increases in spans, as well as decreasing the loads to be carried by columns and footings. The effect of low modulus of elasticity in a reinforced concrete flexural member is to lower the neutral axis and to require a higher percentage of steel as balanced reinforcement. At the same time the depth of member may be decreased so that the saving in concrete quantities will more than offset the increase in steel requirements. Any study of the economy of using Haydite concrete in construc- tion must be based upon actual costs, as well as of relative amounts of materials used, and no attempt is made here to present such in- formation. Other features to be considered are the possibility of increasing the number of stories of a building or the size of floor panels; or conversely, the ability to decrease the cost of supporting trusses, columns, footings, or other sub-structures.