a) a aay a) cal La Spare a | on Na) ww 7 a) 2 ru) “sh a) EE _ Ne ~ C—O SE er —_ : ? — a rr rr ae Y GEA Thtr0xal., Cun gr 7 Cet This thesis was contributed by oe a I, ' Mr. H. F. Rook under the date indicated by the department stamp,to replace the original which was destroyed in the fire of March 5, 1917. Thesis for The Degree of Bachelor of Science in ingineering. by #700 and Geo. Blackford. eee Michigan Agricultural College Rast Lansing, ‘ich. I9I 4. Strength of Concrete by Different Proportions of Aggregate. 108437 In selecting this subject for our thesis we had in mind the possible results which could be obtained by perfor- ming various experiments along some definite outlines as su- ggested and deseribed in practically all textbooks dealing with concrete. Every textbook on the market, dealing with this sub- ject, emphasises the importance of a thorough study of the aggregates and their proper proportions. However, when we lookéd for comparative results, or what had been done in this line, we could not find any written matter along this particular subject. The theory haa been well disoust but practical results as jet have not been published. For references we used various textbooks, the most noteworthy ones being: "freatise on Masonry Construction” by Baker, "Concrete Plain and Reinforced" by Tailor and Thompson, Report of Committee on Specifications and Methods of Tests for Conorete Materials” ,"American Civil Engineer's Pooketbook" ,and Trautwine’'s Civil ingineer's Pocketbook". The object of our thesis is to determine the strength of concrete of a given mixture by different proportions of aggre- gate. Conerete under digeusion is plain and not reinforced, for reinforced concrete an over cemented mixture is some times neces- sary, particularly in column construction. The experiments in our thesis are based upon the theory that the atrength of concrete depends entirely on the adhesion of the cement to the sand and stone. Since concrete is always desi- gned for compression only, the tensile strength was not taken into consideration and no tests were made for the tensile strength. The two theories which govern the proper proportioning of oonorete are: (I} For the same cement and the same sand the strength increases with the amount of cement in a umit of volume of the mortar. (2) For the same proportion of cement, in a given volume of mortar the strongest mortar is that which has the greatest density, i. e. contains the largest proportion of solid matter. In other words the ideal or theoretically perfect conorete is that in which the best aggregates are proportioned and graded in sise a0 as to re- duce the per centage of voids to a minimum, and giving the greatest density. The truth of the first law is obvious,because a clear cenent mixture is stronger than a cement mixture with some matrix. The second law,however, is not so easily analyzed. The first con- Gition ontering in deals mainly with the percentages of voids. If only one or two different grades of aggregates are used the process of determining the percentages of voids of each material is very simple, but it is more difficult where a large number of different Sizes of aggregates are used. Yet,it ia not so difficult in prac- tical work but that it can be used, when one is familiar with making acourate measurements and figuring percentages. Perhaps the best and easiest method of determining the peroentages of voids is by direct measurements. There are a largo number of methods used in different countries and all of them have their faults. Some of them are tedious and not well adapted for ordinary prac- tical work. Another condition that enters in the second theory is the density of the stone. If the aggregate varies in ph¥sical conditions, such as epecific gravity, the method of determining the peroentages of voids, or maximum density, becomes more compli- cated, particularly where the density ia obtained by specific gravity. Again a third condition to be considered is the size of the specimen. In small specimens a alight diviation might lead to erroneous conolusions. The most valuable use of the method of proportioning by mechanical analysis is in cases where the character of the work warrants employing several grades, that is several sizes of stone an@ sand. Such mixtures are being increasingly employed as engineers and contractors more fully aporeciate the necessity of so economically proportioning the materials as to produce a mixture of the greatest possible density, that is with the fewest possible voids, thereby reducing the quantity of cement and at the same time improving the quality of the conorete, in other words, making both a better and cheaper concrete. A second object in view was the variation in the strength of the oconorete by the use of different sands. Good sand oannot be easily defined, or an inflexible specification written, as sands Of various proportions and propertisa belonging to it may make equally good conorete. The usual specifications for sand ere:"The sand shall be sharp, clean and coarse". For the use in our testa the sand was screened thru two different soreens. The largest screen was one of 1/20 inch mesh and all the sand passing this screen and remaining on a 1/30 inoh soreen was used thruout one test. The sand passing the 1/30 inoh scrsen was used for a comparative teat. In order that we might observe the difference in the strength of the cement and if possible compare resulta of simi- lar tests we decided to use two different kinds of cement. For this purpose we selected a iichigan Rock Cement, manufactur ed by the Volverine Cament Co. for one series of tests and the Uni- versal Portland Cement, manufactured by the Universal Portland Cement Co. This latter cement was chosen heoause it is of known reputation and is a slag cement, made by mixing and grinding biast furnace slag and hydrated lime. Method of procedure: In order that we should obtain results that correspond to some standard specifications we adopted the rules laid down by the American Concrete Institute. These rules will be briefly stat ed below: (1) The best shape of a test piece is a cylinder. (2) Cubes, cylinders, or prisms not shorter in length than their least diameter, can be used for comparative tests. (3] The smallest dimension of the test piece should be at least four times the size of the largest particle of stone. (4) Teste of similar concrete made with different percentages of water confirm that an inorease in the quantity of water in mixing reduces the strength of the concrete to a marked degree, particularly in the early stages. The aggregate used was obtained from local gravel~banks. Gravel as found 14 this vicinity is from a terminal moraine,and consists of sandstone, trap, limestone ,granite,and some quarts. The latter being ohiefly present in the sand. This aggregate was then screened thru different sized screens, the largest screen having a one inch mesh. The other sizes of sorees were as follows: 3/4 inch, 1/2 inoh, 1/4 inch, 1/6 inch, 1/8 inoh, 1/20 inoh, and 1/20 inch mesh. The sand and gravel was perfectly dry and no moisture present. This was determined by weighing out 200 granmmsa of the sand and heating this amount over a gas burmer for an hour. No 4ifference in the weight was noticed. The next step was to determine the percent of impurities in the sand. This was done by measuring out 200. grams of the sand in a graduated tube and then adding water to it. By successive washings the silt, olay, and other inpu- rities were removed. These were then poured in another graduated tube and allowed to stand for an hour. YWhen they had settled to the bottom the nuxber of 6.0 were read. The number of c.c. thus obtained divided by 200 gave us the per cent of impurities. A considerable amount of impurities were found in the number 30 sand, This, however waa as expected, because it contained a lot of fine material. d3and Ho. 30 showed 8.5 per cent of impurities while sand No. 20 showed but 1 percent of impurities. When everything was ready for the actual mixing of the concrete different methods for obtaining the percentages 02% voids were considered. After some studying of all the methods outlined - fin various books and treatises we decided to use the method usually known as the determination of voids by measurements. This method is not as accurate as some others, but where the materials are carefully mixed and graded, and consist of a large number of dif- ferent sizes of aggregate, the error is quite small. The chief inaccuracy of this method of basing the proportions of the fine materials of a concrete mixture upon water contents of the voids in the larger, is due to the difference in the compactness of the Material under varied methods of handling. Another factor enter- ing in is the fact that the actual volume of voids in a coarse material may not, and usually ddbes not correspond to the quantity of sand required to ffll the voids. The reason for this is : the grains of sand thrust apart the particles of stone, and because with most aggregates a portion of the particles of sand or fine screenings are to coarse to enter the voids of the coarsest mater- ials. Keeping in mind that an exess of gand always increases the voida in the conorete, we decided to use a proportion of 1:2: 7. This was decided upon after several mixtures of different combi- nations were tried without the cement and a suitable percentage of votds obtained. Later, cement in the above stated proportion was added and the percentage of voids correctly determined. The method as followed in our experiments will be briefly described. We took an average sample of the mixture of 200 oc.ca. and put it in a graduated tube. In another graduated tabe 200 c.c. Of water were poured and into this the 200c.ca. of the mix- ture added. Care was taken to allow all the air to edcape. After the mixture stood long enough, so aa to settle, the number of 0.0. of settled matter were read and also the final volume. ‘The difference between the first two quantities put together i.e. 400 «a.6. and the final, divided by 200 gave us the percentage of voids. This method is short and simple and we used it because we were limited in time. The next step waa the mixing. The proportions of the various sizes of the aggregate were taken at random, measured out, and emptied in a large tin pan. The same wag done with the sand and cement but put in a separate pan. Here the two materials were mixed by hand until the mixture showed a uniform color through- out. The coarse material was then mixed and added to the fine and both mixed dry by shoveling the whole over not less than three times. Water was then added till a normal consistency was obtained as suggested by the American Conorete Institute. The concrete was then placed in the forms in layers of . n tamper. about two inches deep and carefully tamped with a woode i Another layer of the samo thickness was then placed on top of this and again tamped. This was then continued till the form waa full. It waa then leveled off with atrowel and covered with damp cloths and allowed to remain there for 24 hours. The forms were then removed and the blocks atored in a large tank filled with water. The water in the tank had a temperature from 67 to 70 degrees Farenheit. We made aight aamplesa, sach of eight different mixtures of aggregates. In the first set of Samples sand Wo.30 waa used. In the sacond set we used sand No. £0. In the third set and also in the fourth sand No. 30 and No. 20 waa used respectively, but a different cement was substituted. Due to the short period of time in which we had to per- form the testing we decided to break the cubes at regular inter- vals. Thase intervals were as follows: 4-7-10-14-17-21-24 and 28 days. The testing was done in a Amsler Hydraulic Testing Machine. This machine ia well suited for this kind of work. The pressure can be applied very slowly and uniformly. The machine has a capacity of 150,000} and is operated by hand. The specimens were allowed to remain in the water till they were ready for the testing. When the time was up and the specimen ready for testing 4t was taken out of the water and placed in the machine. Here it was imbeded in several layers of cushioned paper on the top and bottom. When it was carefully centered the pressure was applied and the sample tested to complete failure. The maximum pressure obtained waa then recorded. It was also noted that when the total pressure went up to about 18,000 # all the sand- stone failed. Diffioulties encountered. One difficulty that confronted ua was ths matter of normal consistency. Because of thse difference in the effeot of different sanda upon the consistency, it was impossible to speci- fy a definite percentages of water. For this reason we followed the suggestions of the Amerioan Concreta Inatitute. They reco- mend that in filling a conical form with concrete, immediatly inverting this, and by ropeated trials a conaistoncy can be found such that the amount of water willcoause the cone juat to bagin to slump when the form is romoved. A dry mixture is of sourse unsatisfactory while it is almost iupossible to desoribe a wet mix that will give uniform results. The advantage of this method lies in the fact that the original amount of moisturs in the sand and aggregate does not effect the final normal consaiatonoy. Another difficulty that we had to sncoounter was the matter of forms. ‘‘e expected to use oylindrical forms but it was found to be inconvenient to make them for our short testa. In order not to conflict with the other requireuenta of tests by the American Conorete Institute we had to use cubes at least five inches in diameter. There were no forms of that sige in the laboratory 830 we constructed wooden formas of that sige, each holding eight oubes. These forms were built out of one inch dresaged lumber and held together by clamps and wedges. Those forms wers thoroughly oiled with a heavy o11 before being usad. But perhaps tho biggest odds we had to fight waa tho short period of time. Dus to this faot our results are probahly low and do not represent an average oondition. Undoubtedly tests extending over a longer period of time would ha more reliable. Conclusions. The results az obtained show that the proper proporti- oning of the aggrogate, of any concrete, influence greatly the strength of the concrete. Inoidentally this is the same thing ag saying that by proper proportioning of aggrogates the voids are reduced to a minimum, and hence the first law, aa stated in the begining, should be sgatisficd. Again the second law was verified in az far as will bs seen from the curves. An analysis of the curves will show that the ideal mixture is the only one that will satis fy all the conditions which have been considered. . Further observations will show that the gand has a considerable influence on the strength of the concrete. In nearly all of the cases, where the concrete was made of the fine sand, a greater strength was obtained than where the ooarse sand was used. The latter complied theoretically to all requirements, but 414 not show aa good final results aa the other concrete, although the fine sand contained 8.5 per cent impurities. This particular subject has received considerable attention of late by many engineers and it is hoped that results and data will be available for comparative tests. It ia altogether a reasonable proposition to state that these proportions, made of selsoted ingredients, produce a conorete whioh is actually stronger than if thea are selected at random. A method of proper proportioning of aggregates, aa opposed to the usual practice of specifying arbitrary proportions regardless of the charactor of the available ingredients, or of the work to be done, has the advantage of offering an incentive to good workmanship. While the ingredients may in some cases prove more expensive, the resulting concrete actually costs less per cubic yard. With exp ensive aggregates the engineer will take less chances of waste, and therefore exert more care in mixing and placing. Arbitrary specifications, which simply state good gand and stone shall be used together with Portland Cement, meeting certain teats, mixed in proportions 1-2-4, or 1-2-6, as the case may be, may mean a very rich mixture, or again result in a very lean concrete. The large number of conditions, such as normal consis- tenoy, workmanship, care in placing, care after the concrete is placed, tesperature at the time of mixing and at the time of setting, eto. all effecting the final strength of the concrete, cannot be neglected. But if they are kept constant it is a reason- able proposition that the final strength of the concrete can be figured within a small margin, provided the character of the ingre- diente are known and the proportion of suoh aggregatea definitly stated which will give a maximum density. The thought of maximum density should be kept constantly in mind and the idea of arbitrary proportions eliminated, for the character of the aggregates will entirely govern the proportions which will give the strongest concrete. 11 First Mixture. parts ” Vv ve 2 1 1 1 a 1 2 1 “4” 17" 21" 24” Percentage of voids = 19.0 of 1 inch gravel n gon " a " "opm " " 1/6 * " "1/8" " " Ho. 20 sand " Wolverine Cement. strength at different ages. 226007 or 905#/sq. inch. 29500 * llso " ” 31600 " 12 " * 28800 " 1156 " * 28700 " ll4aeo" ” ea500 " lize" ° 30000 )6”®)=— -—1z00 =" " 56400 7” 1454 " " 12 Second Mixture. Percentage of voids = 18.5 2 parts of 1 inch gravel 2 " " 2 1 OO 1 " " + rm 1" 1 1 rv + i] "1 1 " ve 1/6 " 8 lo " 7 /8 7 " g ”" " No. 30 sand 1" " Volverine Cament™ Crushing Strength at different ages. 4 days 20200f or 808f/sq. inch. 7 " 28300 " 1131 " " lo " 32500 " 1300 .-" " 4%" 38600 " 1642 " " 17" 40700 " 1624 " ” 21" 38100 * 1623 " " 24 " #.}37000 ” %U76 " " 28 " =+$%39200 " 1570 #" " 13 Third Mixture. Percentage of voids = 12.5 part of 1 inch gravel " n 2 om " a " 1+ ” >» iT 10 e 1 l z 1 1." * i/o” " 1 waple - 2" " 50.30 sand 1 7. " Volverine Cement. Crushing Strength at different ages. 4 Rays 11800#,’or.4737/aq. inch. 7 " 18200 " 798 7 » lo ” 19600 ” 785 " " 4" 22800 * 912 ” " 17 * 18700 * 754 "* " 21" 25600 * 1024 " " 24" 22400 ™ 896 ” " 25" 25400 "1015 " # 14 Fourth Mixture. Percentage of voids = 26.0 2 parts of # inch gravel 2 * a n 2" "1/6 * " 1” * 1/8" iu 2 * * sand No. 3. 1 " " Wolverine Cemont. Crushing Strength at different ages. 4 days 6600 # or 264 #/sq. inch. 7" 9000 ”* 360 ” " lo" 10400 ” 416 " " 14” 12500 ” 500 " " 17” 11800 ” 472 * " 21°" 14000 " 560 " " 24” 14400 " 676 " " 28" 14900 " 596 " " 15 Fifth Mixture. Peroentage of voids = 17.5 4 parts of 1 inoh gravel. 1 7 r" 3 1 1" 1 rv v0 * ? rt 1 r 6 1/8 rt 1” 2 * " gand Zo. 23 2 ”* " Wolverine Cement. Crushing 3trength at different ages. 4 days 15900# or 636 #/s,. inch. 7 " 93800 " 962 " " 10 " 30400 " lsl7 * " 14 * 83500 " 940 ” " 17 " 98000 " 1120 " " 21 " 36000 " 1438 * " 24" 42600 " 1702 " " 28" 39700 " Ilsce " ” 16 Sixth hixture. Percentage of voids = 21.0 4 parts of $ inch gravel 1" " 3h 0 " 2” "fe" * i” "l/a» gz" " gand Ho.30 1" " Wolverine Cement. Crushing 3trength at different ages. 4 days 14900/ or 596j}}/s9. inch. 7 ™ 20600 * 832 ” " 1" 87600 " 905 " " 144" 27600 "llog " " 17" 92900 " 916 " * 21" 98100 * liege * ” 24" 26200 "1046 " " 28" 29200 "1166 ™~ " 17 Seventh Luixtiure. Parts A “ ” ba | 4 1 ee i 2 1 Percentage of voids = 23.5 ~ ‘ * of ¢ inch gravol. 7 + " "9 | 1 /6 1 " " 7 /8 " 18 " sand No. 30 " Wolverine Cement. Strength at different ages. 15100# or 604 #/sq. inch 20300 " gle " " 24400: " 976 " " 24100 " 965 " " £6600 ® 1062 ” " 84600 " 984 =" " 27000 " 1080 " " 32000 " 1280 " ” aghth Mixture, 4 L * 3 1 2 1 18 Percentage. of volds = 16.5 parts of l inch gravel. ry " $ iT " "2. n n # 1/6" " * 1/8" " " 30. 3 sand " Yolverinse Cement. Crushing Strength at different ages. 4 7 10 14 17 21 24 28 days 270004 ¢ 7 28000 31200 38600 358 200 38700 55800 42700 ” or 1080 #/sq. inoh " ll90* * " 1280 " " " 1545 " 8 "1525 " " " 154g” . " lage " 17 10 v tt Binth Mixture. 19 vVai:xturs Jo. 9 13 the sumg as mixture No.l except in placo of sand to. W/W w2 used dard NO. 20, Crushing 3trencthn at aifferont uges. 4 days 21000} or 840;/s.. inch. 7 lo" 14" 17" 21" 24" 20” £1700 22100 £2900 308C0 52200 36 400 54000 i Tenth Mixture. Mixture No. 867 88S. 916 1230 1287 1415 1360 re 19 9 1? wv "9 qs 10 i3 the same as mixture No.2 oxcept in place of sand No. 30 we ujed sand Ho. 20. Crushing 3trength at different ages. 4483/39. inoh 4 daya 11400; 7 ON 19 " 4" 17" 21" 24" 28." 17400 18800 17600 £0600 21600 18000 25600 or 4 " 9 " 8 6 663 752 712 824 865 720 1022 LA] 9 9 tf ba] ef t} "7 rt 1% Sleventh kizxtura. in place Crushing 4 7 10 14 17 21 24 28 daya i] Mixtures So. of sancé Strensth at differant ages. 11100# 14200 13100 22209 20 500 £2000 £5400 51600 Twelfth Mixture. place of sand No. 30 we used sam No. 0. Crushing strength at different ages. 4 7 10 14 17 21 24 28 Mixture No.12 is the same as mixture No. 4 except in y” ye oY * " % De ll is the Samo as mixture No. 20 we used sand Wo. 440#/aq. inch. 1 ¥9 tf ve vf > ? A vt rv rY % r? days 69007 or 276$/aq. inch. " rv vw 89000 12000 14500 16500 18600 20300 20 700 * "t 320 480 580 620 744 812 828 ? " I w 0 LA] A ve A 3 axcept 21 Thirteenth Mixture. Mixture No. 198 is the same as mixture No. 5 except in place of sand No. 30 we used sand Ho. 20. Crushing Strength at different ages. 4 days 12600 # or 504#/aq. inch. 7 * 15400 " 616 " * 10" 17500 " 70 " ®* 14" 20800 " el" * 17" 80200 " 808 " " 21” #19400 * 775 " ° 24" j295400 "ole ” * 28" #7300 "1090 " * Fourteenth xixture. Mixture flo. 14 is ths gamc as mixture NO.6 except in place of sand No. 30 wa used sand to. 20. Crushing Strength at aifferont agus. 4 days 8700f or 348f/aq. inch. 7" 13300 " 532 " "™ lo" 16000 " 600 " " 14" 44100 " 564 " " 17 " 173000 " 692 " " 21" 19600 "” 784 "* " 24" 24700 "*" 988 " ” 28 " 25400 " 1016 " " Fifteenth slxture. kixture Es. 15 is the sane as mixture To. 7 excs3pt in place of gand Xo. 3 we used sand ko. 20. Crushing Strongth at different ages. 4 days 105007 or 420/s7. inch. 7" 11600" 464 " " lo" 17200 " 688 " * Ma" 15500 " 620 7 " 17" 21000 " e40 " " 21" 100 " 644 " 7 24" e200 " 912 " * 26 22600 " 908 " " Sixteenth wixture. wixture 2O. 16 is tho same a3 mixturs Ho.S excapt ix plecs of sand xO. 30 w3 uscd sand Lo. 20, Crushing strength at difforent apes. 4 days 10600 or 424 3/s3-. inch. 7 " 17400 7 696 " " lo" 17400 " 696 " " 144° scoo " 742 * " 17" 91300 " 852 " " 21" 17300 " 691 " . e4" 21800 * e7e " " 28 " 21000 " 840 " " 25 Javonteenth ‘ixrture. aixtur? 30. 17 i3 th2 sems as mixture Jo. 1 except in placa of Wolverin: Coment we used Universal Cement. Crushing strenghh at different ages. 4 days 7400 # or 296 #/s4. inch. 7" 9500 “ 380 " " 19" 9002" 360 ” ” 144" e000 " 472 " " 17" 14000 " 560 " " gl" 1z70o ” 548 " " ea" 14900 " 596 " " pe" 16000 ” 640 " " Eighteenth iixturs. Mixture No. 18 is the same as mixture No.2 except in Place of Volverine Cament we used Universal Cement. | Crushing Strength at different ages. 4 days 8000 or %4203/sq. inch. 7" 102090 " 408 " " 10" 12800 " 512 " " 14" 14000 " §60 " " 17" 15000 " 600 " " 21" 16800 " 672 "” " 24" 17500 " 700 " " 26" 19000 " 760 "” " Rinetesnth .inture. Jiztur3 ro. 139 13 the samo a3 mixture fo.3 except in place of Wiolvarine Tanant wea used Universal Cement. Crushing strength at alfferent ages. 4 daya 75007 or 300 $/33. inch. 7" 9600 " 334 ” " 10" 4isoo " 476 " " 14." 13500 " 540 " " 17." 1zz00 " 528 ” " 21 " 15400 " 616 ” " 24" 16800 " 672 " " 23 "* 189000 " 72 © Twentieth [dixturs. ixtare Eo. 80 i3 the same as mixture No.4 exospt in place of Volverine Canent we used Universal Cement. Crushing Strength at different ages. 4 dsya s000f or £00 4/sq. inch. 7 * 6500 * 270 4" " 10 " 7700 " #& " " 14" 10809 " 432 #2" " 17" 11400 7” 456 " " 21" 40000 " 400 " " 24" 12000 " 488 '" " 28 " 12000 " #520 " " 26 Twenty-first ‘ixture. Aixturas 70, 2113 the osis? sa mixture Jo.5 exaent in plasa of “olverine Csemant wo usod Universel Cement. Crushing strongtn at different a:es, 4 d@aysa...110003 or 440 !/sq. inch. 7" 14600 ™“ 584 " " 19." 13200 2" 527" " 14." 15800 " 672 =" " 17" 19200 " 727 " " 21" 16200 " 727 " " £4" 19000 " 69 " " a3" £0300 " 312 " " Iwenty-sveconi