THS NND'ENC THE MAXIMUM FRACTTCABLE DENSITY OF COARSE AGGREGATE THROUGH SCIENTEFIC GRADING T310513. for ”‘48 Degree of B; S. ‘. 1'3. 'T'HC‘MPSON 1929 THESIS ‘.l'.l-l|\l I it ill: ‘Irl\ .Illlllll . ‘1‘ fl I I ‘ . Ur STATE W5? ‘ " ‘ . - :‘ri‘fifl A‘_*.,¢!_\) M.) (xx-wig!!!‘ ’ . .0 ‘1 u T ; '\ v ‘u 1., g .. L, . v ".. \ ‘ *m .'.‘ , a; ). ‘ .‘ _"_ *‘n'k . f ~ y t . .. ' , . ‘ ' t u- I . . a a J . . Q» \ \ , , . Finding the maximum Practicable Density of Coarse Aggregate Throu;h Scientific Grading A Thesis Submitted to The Faculty of IXicnigan State College of Agriculture and Applied Science . .. r 9 3-9;- 3:. [‘1’ J. P;'Taempsan w Candidate for the Degree of June 24, 1923 -1- In the November, l)23 issue of the Crusned Stone Journal; the United States Bureau of Public Roads made an annouceuent in Which they declared, that the results of their re earcn and tests, show that it is possible by appliin5 tne theories deveIOped by tne ey xpelilents; to produce concrete of mucn 5reater uniformity and increased stren5tn. Chief Thos. d. Lac Donald, laid the following principles by whicn tnese results may be obtained. 1. The abandonment of volumetric prOportionin5 of a55re— gates and the adoption of proportionin5 by 1ei nts as standard practice. Inundation will be re c05nized as a permissiole alter— nate method for fine e55re51to, but wei5hin5 is preferred. 2. is intenance of the lowest vaater-ceuent ratio wnicn, with the particular t pe, rading and proportions of a55'e5ate used, and the me hods of fi’ nishin5 employed, will produce a worxable, dense and unifor1 com1 rete . 3. The scientific 5radi15' of cor Mr e a55re5ate by conoi— nation of seg1arated sizes 11 each batch 11 the porportions whi ch will produce the maxi:num practicable de11sity. 4. The abandonhez1t of the hand finishin5 methods in favor of machine finishin5. The principle wnich is of very vital interest to all pro- ducers of either gravel or crusnedmto1e; is No.3. One of tne many reasons for the laid din5 down of tnis prin— ciple was that ce51,1ation of sizes of 55re 5:ates occurs in _stora5e piles wt1icL Lave oeen built up cs a cone by droppin5 material in the ce;nter. The lar5er particles roll to the bot— tom of the cone and the smaller led5e satewnere bet"e;n the top and the bottom with the s1allest a the top. The worst cases of 3e55re5ation are found in piles built up Q5 convejiu5 machinery w1ich disc1m15es in one spot only. If, in addition, material is remove d by a 1 oader that works around tne ed5e of the pile, the core an<71the too 0 the pile new becone alnost entirely fine material a17 ' e all 1.: r5: roez. A sieve analysis from sun h 3 p1 1‘3 20 f' high 5ave t.1e followin5 re— 31 Lva \’ I o o I ' " ' 1 1 1. 3“ us: \ 11.15 31.1“? 8118.1.) Ll; W113 zade D'f tile Portland Cengeut Assoc atio1) op Pv Q L) d b" .J Sieve size : for cent Pass1n5 ; : Iniches : Sa1ple tahe1 fPJJ : ; Center of : 2 tns. of : Jut;r ; file : blotudoo : ed5c of : z to out:r 0‘ e; rile : : of File : 23 ‘ - lOO : 130 : loo 2 3 94 : 04 g 30 l : 55 : 30 : O k : l7 : 1 : 0 1 PO 1 It is nearly impossiole to mane 5ood concrete with se5 re5ated materials. N333 the coarse rock is 5cin5 into tne nixer the concrete is harsh; when the center of a se5re51ted pile is be1n5 used there is too nuch fine material. The re— sult is porous concrete in the first case and wean, scaly concrete in the second. 3e5ro5ated haterials can be done away witr oy the application of Principle no.3. For the application of this principle the produccr “use ship his 5ra el, not accordin5 to a 5iven 5radation extenu— in5 fron the snallest to the lar5est size as is present—day pract ce,but, on the otz1er hzn1d, the 5ravel must be sepa— rated and s1ip11ed in several clifci 1t sizes, such, as il— lustration, as k to Z in., 5 to l} 11., and l} to 2} in. These separate sizes will then be conoi1ed at a centeral proportionin5 plant an2 the p1over qua11tities 0 each baoc of concrete made; Thi 1 V' “LA.“G. p .sethod of propo1't1o1115‘ 111. e se of the i s to produce con cre e of 5ueater uniformity1nd so uniformly low voids thus produce d in tie creased suunrrt of 5ravel.z1zui a decrea sed.zumm1nt of nortar is re dered poss'b e, with no sacrifice in stren5th. The e is no lost of stren5th but rather an increase ccoreinx to the principles of ‘ooo r p1o1ortio1i15, j Geor5e A. nool, Consultin5 en5ineer, Professor of true- ural 3n52, ineerin5 of the University of Nisco1sin. He states, For 5iven materials (includi15 water) and assinin5 worn— ble 11ixtures only, the followin5 factors govern the theory )rOper orooortio1i15, namely 1.5t1e115t1 of concrete is deterhined by tne ratio of the vol;1me of :1ixin5 water to the volume of cement, called water— e1e1t ratio. Stren5 th i1crease es as the water—cenent HIE ta! 2. 1or the cane 11011.51 is increased by usi15 c'a 5re5ate requires less water ratio. 3. For water-ti5ht work, the size as well as the peru centa5e of voids is i1portant; t1er 1‘ cm -2f..l attentiO1 should be 5iven to t‘ne 5radin5 of tne a55re ates. 4. The quality of cenent and the size and "radin5 of the a55re5ate affect t1e stren5th of the concrete only in so ity, the stren5tn of concrete r a55 re5ate, since coarser a5— and mi1ces the water-cs ant *SL)‘ ,3 {.1 far as they affect the quanity of nixin5 water needed. 5. For a iven hi“, the stren5th of concrete is increas— ed by inc1eas'.s c n nt content, or decreasin5 the water content. The fact that the Bureau of Public R\ads s enco1ra in5 these new methods for concrete orooortioxi1. will undoubt— edly result in their very wide adeption and as tine 5oes on, gravel producers will be called upon to snip 5ravel in two or nore, separated sizes to be coabined in their proper proportions at th: site of the work. The development of a HC: '3 ElJ ( I .3_ desire for more uniform concrete has been quite apparent and the present action on the part of the Bureau of_Public Roads was fo eseen by the Bureau of 3n5ineerin5 of the Crushed Ston Ass c1 iation as a stron5 possibility more than a 5ear a5o.oom1m1tly, work was immediately started in their la oorat riese aid the inf or: 1at101 5ain was given to the crushed stone producers. Therefore the 5ravel producer is naturally interested iA snow‘n5 what affect this new netnod of concrete proportion— in5 will have on his produ uction problems. Principle Ho. 3 states that the coars (-5.51 5ste is to be coaoi1ed in ea ch batch in those proportioAs which vill produce the aaxi— nun practicable density. This, in 5eAeral, Jeans the com— binations f the various sizes of 5ravel in sucn a way that the suallest percenta5e of voids will result. The purpose of this thesis is, therefore, to deternine what combina— tions of sizes will 5ive the lowest percenta5e of voids and also what variation may be made in the proportions of the different sizes wthout incre asin5 the voids appreciably over the minimum. method of Procedure For the purpose of this inve sti5ation 5ravel of a rather 5ood quality was us ed. It w as screened into ei5nt differeit s17 es as follows; 1/3 ~ 1/4 in., 1/4 — 3/3 1a., 3/3 - 1/2 in., 1/0 — 3/4 in., 3/4 - 1 in., 1 - 13 in., 11" -- 2’: ill. The 1/8 - 1/4 in., 1/4 - 3/8 in., 3/8- 1/2 in., and 1/2 — 3/4 in. sizes were then comoined in accordance with a strai5ht line 5radation. A strai5ht line on a mecnahical analyses dia5ram indicates a uniform 5radin5 size. The me— chanical analysis of this ecsnoination is sqown as curve No. 66 in F15. 1 . This required 20% of 1/8 to l/4in., 205'; of 1/4 to 3/3 111., 90,. of 3/3 to 1/2 111. and 4c}; of 1/2'to 3/4 in. sizes. Similarly the 3/4 to 1 in. and the 1 to 11 in. sizes were combined in the preportions of 33 1/3 per cent of the 3/4 to l in. and 66 2/3 per cent of the l to 1; s17 zes, this Mtrai ht line 5radation is shown in curve No. 11 in Fi5. l. The I}; to 2 in. and 2 to 23- in. sizes were co.nbined in equal proportions to produce a st rai5ht line 5radation shown in curve No. l. in Fi5. l. The ce three si7 zes, each havin5 a strai5ht line necnanical analysis curve were then combined in 66 different proportions havin5 mechanical aa nayscs all snown in Pkg. 1. It will be noted that these mechanical analyses curves cover the entire field of poss- ible combinations for these three different sizes. These combinations are also shown in the triaxial dia5ram of Ei5. 2. As this dia5ran is very convenient for showih5 the ef— fects resultin5 from the combinations of three sizes of -4- gravel in different prOportions it will be well to explain it briefly. Each point on the diagram represents a given combina- tion of the three different sizes of materials and the per- centage of each size is represented by the perpendicular distance between the point on the diagram and the side of the triangle which serves as the base line for the size under consideration. Thus, point Ho. 42 in Fig. 2 represents a material made up of 40 per cent of 1/3 to 3/4 in., 30 per cent of 3/4 to 15 in., 30 per cent of 1% to 2} in. size. Similarly, point No. ll represents 100 per cent of 3/4 to 1% in. and 0 per cent of the other sizes. From this diagram it can be seen again that all the possible conbinations of the three batcnes are conered. The 1/8 in. size was chosen as the smallest size used in this investigation because standard coarse aggrepatfi spe— cifications permit not more than 5 per cent of the aggre— gate to pass a sieve having Opening 0.185 in. square, and by extending the minimum size down to 1/3 in., the permis- sible amount passing the 1/4 in. screen will approximately be obtained. The procedure in testing was to weigh out 90 pounds of the conbined sizes wnich were then thorothly nixed. This mixture was snoveled into a standard American Society for Testing Haterials one—half cubic foot measure. The tests were made in two ways, first, by loose measurement and, second, by the standard A.S.T.J. rodded method. When tLe loose neasurenent tests were made, the material was snoveled into the measure and struck off. The rodding test calls for filling the measure one—third full, rodcing 25 tines with a 5/8 in. by 24 in. bullet-pointed netal rod. This process was repeated three tines and the container was leveled off and weighed. The we gnt per one-half cubic foot determination was repeated twice for eacn method of fillinf the measure and the average was obtained for eacn method. When one determination did not agree with the other within one—half pound, a tnird determination was made, and the a‘era;e tanen. In only two cases was this necessary. The percenta;e of the voids was calculated as fOl 1 CW 8 0 wt. per 3 cu. ft. Percentage of voids = 100( l - .«__——-_-—-—---u sp.gr. x 31.17773‘ A sample of the gravel was obtained by the quarter— ing method from the l to l} in. size. Then the specific gravity of each stone was found by weighing each one out of water and in water, the difference of the two weights gave the loss of weight in water. Then dividinp the weight of the stone out of water by the loss of weight in water gave the specifice gravity of that piece of gravel. The average of all was t.xen as the specific gravity of the gravel used. In Table 1 the weight pe half cubic foot ‘ and the percentage of voids are snovn for both loose and rodded condition. The results given in tnis table for voids are plotted in Figs. 3 and 4 on triaxial diagrams. -5- The lines are contour lines of equa l voids. The specific gravity of the gravel used in this in estigation was 2.79. Dicussion of Results Referring first to 313. 3 which represents the voids for the material in a loose condition in the half cubic foot measure, it is evident that tLe lO'fieSt percenta'e of voids is obtained with gradin' No. 3). This material is co nposed of 40 per cent 1/3 to 3/4 in. size, 0 per cent of 3/4 to 131 in. size and 60 per cent of 11; to 2} in. size. Ther are no in- termediate sizes in the nix. The percenta3e of vaids for this 3.adi1 was 32.5. It can be seen that as more and more in- ternedia te size is added to tie mixture there is a 3eneral tendency toward an increase in the percenta3e of voids. Obviously if every producer were equired to furnish gravel graded to have the very nininun of voids on a loose measurement No. 3) or na:ielr' 1/3 to 3/4 in., 40 per cent; 3/4 to l} in., on cent ; 11- to 2} in., 60 per cent, there would be the necessity of disposing of all of the 3/4 to if in. size for soz1e other use than for road concrete. This would be a wasteful procedure and would mane the production of gravel more expensive. On the other hand if a low per— centa3e of voids, a few per cent hore than the ni ni1un is permissible and it obviously would have to be permissible to mane the ’ Huer 1 scheme practical and wornable, the field is Opened up c)to a reasonable range of Si? zes. This point is well illustrated in Fig. 5 in which are plotted the neonani— cal analyses of mixtures having within 2 per cent ni3ner voids than the nininnn. These mixtures are as follow 5: 1/3 - 3/4 11-1. 3/4 — 13,,- 111. 11- .- 23— in. 30 o 7,0 4o 0 - 30 '00 - 30 7-; 50 o - lo 50 - 1+0 :2 60 O - 30 4o - 10;; 70 o - 20 30 — 10 7’ If the lowest pra ctic: ble oer 1e 1ta'e of vaids on a “rodded‘ basis is required a1d as word “p1acticaole” is interpreted to mean within 2.5 per cent of the mininua possiole percez1ta e of voids, a ran3e of gradations w1icn will satisfy the re1uirenents is snown in graphical fo1n as 313: . 6. This possible range of mixtures is seen to be as follows. f? 1/3 - 3/4 in. 3/4 _ 1‘; 111/ 13- - 23;,- 1:1. 110 on so to — 4o ,3 so 0 - so 40 — 20 ,3 The closer to the absolute mininun percenta3e of voids required by the specifications, the sore restricted is the range of cohbinations of sizes, but the dia ra1s in Ei3s. 3 and 5 or 4 and 6 will show the available cohoinations -0- .1 (1.8 k, necessary to produce an1 s ed percentage of voids. In the precedin3 discu 1s ion it sno 1d net be infer— red tnat the lowest percen1ta;e of voids referred to is the absolute 1i1iiu1 tziat can be obtained but is merely the mini1un wh ich was obta i ed with the three s1; es of gravel used, each of which had a strai3nt line adation. No doubt a still lower persenta3e of voids could rbe pro— duced by special 3racation of the three sirm used but t1is would be a rat1er difficult procedure fro1 the 01~rat11U sta ndpoint a1d the sli3ht decrease 1:1 voids wnicn n1 3nt thus be obtained would probably not warra1t tie ext 1a cost involved. ir s Table 1 ~ Weights Per Half Cubic Foot and Percentages of Voids of G) Gravel ( "S p: p. H. .3 *7 H WHw~JO¢n¢flxn>HCD i C' specific Gravity — 2.7) ) Graded as shown in Figures 1 and 2 Loose Rodded Wt. per Per cent Ht. per Per cent Cu. Ft. Sblids Voids Cu. Ft. Solids Voids 43.8 56.2 43.3 52.00 59.8 40.2 47.75 55.0 45.0 53.5 61.5 33.5 4).3 56.56 43.44 54.25 62.4 37.6 50.25 57.7 42.3 53.74 61.3 33.2 43.3 56.2 43.3 54.26 62.4 37.6 51.06 53.7 41.3 55.36 67.4 32.0 50.91 5 .5 41.5 54.37 62. 37.5 50.1 57.5 42.5 54.17 62.3 37.7 49.65 57.1 42.9 54.7 02.) 37.1 43.2) 55.5 44.5 5 .76 61.8 33.2 43.41 55.6 44.4 53.34 61.2 33.8 50.3 57.3 42.2 55.19 63.5 36.5 52.51 60.4 33.6 56.4) 63.0 33.0 52.52 00.4 3).6 55.7 64.1 35.9 52.76 60.6 39.4 56.9 65.5 34.5 53.42 61.5 33.5 56.34 65.4 34.6 52.35 60.2 39.3 56.2) 64.6 35.4 52.07 5).3 40.2 56.76 65.2 34.6 51.09 53.7 41.3 55.19 63.5 36.5 51051 5903 4007 55005 63.2 30.3 50.3 53.6 41.4 55.0 63.2 36.3 54,3 6204 J700 56.74 0502 3408 33.51 61.6 33.4 57.73 66.4 33.6 54.12 02.2 37.8 53.23 67.0 33.0 54.9 63.1 36.) 57.37 69.5 33.5 53.11 61.1 33.9 53.04 66.6 33.4 52.63 60.5 32.5 57.76 66.4 33.6 52.51 60.4 32.6 57.65 66.2 33.6 52.13 60.0 40.0 57.62 66.2 33.6 51.53 59.3 40.7 57.66 66.2 33.3 57018 d507 3403 U002} 0302 3303 54.7 62.9 37.1 00.53 03.0 30.4 54.35 62.5 37.5 53.91 63.9 31.1 54.77 63.0 37.0 53.56 68.6 31.4 54.84 63.1 36.) 60.55 6).6 30.4 Table 1 coatinucd Loose Rodded Grading Wt. per Per cent Wt. per Per cent No. Cu. Ft. Solids Voids Cu. Ft. Solids Voids 36 54.13 62.2 37.3 53.33 67.3 32.7 37 54.23 62.3 37.7 53.61 67.4 32.6 33 54.13 62.3 37.7 53.32 -67.3 32. 3 53.66 67.5 32.5 61.65 70.) 23.1 40 57.25 65.8 34.2 61.63 70.) 2).1 41 56.64 65.6 34.4 61.25 70.4 2).6 42 56.23 63.5 34.5 59.41 63.4 31.6 43 55.32 64. 35.2 59.94 6).0 31.0 44 55.72 64.0 36.0 5).63 63.6 31.4 45 54.53 62.3 37.2 53.97 07.) 32.1 46 53.55 67.3 32. 60.3 70.0 30.0 47 5306 $703 jgoé 60.66 6).? 3303 ‘400 60.73 6908 3002 4) 25.71 6-.0 36.0 53.95 03.0 31.0 50 5:.96 64.3 35.2 60.3) 62.5 30.5 51 55.51 63.9 36.1 5).2) 63.2 31.3 5- 57.0 65.6 34.4 63.2 72.7 27.3 54 57.74 66.3 33.5 61.32 70.5 21.5 55 57.47 66.0 34.0 60.36 69.6 30.4 56 56.2) 64.6 35.4 53.31 63.3 31.2 57 53056 @703 5207 00.3 70.0. 30.0 53 53.27 67.0 33.0 60.24 09.2 30.6 5- 57.37 65.9 34.1 53.42 63.4 31.6 60 53.4 63.7 36.3 53.77 63.7 31.3 41 55.33 63.7 36.3 60.3) 70.1 2).) 62 56.2 34.5 65.5 33.7 63.6 31.4 23 54.74 63.0 37.0 53.1) 66.9 32 1 64 54.1: 62.2 37.8 53.37 67.0 3:.0 63 £4.63 62/3 37.2 56.91 65.5 34.5 60 52.71 00.6 33.4 56.34 65. 34.6 ‘a'~- M/ \\ $§\ \ . 0 0L 1+ I5- .2 up. NI~ h \ [0 0 I00 I0 O Percentage of VJidS in gravel. Loose measure— ment in A.S.T.Z. standard oxe—nalf cubic foot measure. Eample of use of diagram. Point A represents the following gradation. (square Opening ) gs". g:— 20% M A o . Afi w 123- 2;!" “Z :0 49.3% 2. “"1 mm .. .. AVAYQX‘X .. 9o ’3 “'"V‘ ‘¢%%VA loo ?0 80 70 60 5’0 I entage of voids in gravel rodded in A.S.T.;. taadard one—half cubic foot measure. Example of use of diagram. Point A represents "_ 20 % [00 O ‘QWIIFK ‘w‘ ‘1 .. Av: A’M‘} .31, p o .. A7A§WAWA . A1493“ MY“ ,. "° mu ‘WA‘WA‘V w ‘WA ‘Va.A_vav_-=4’nvAvA .o ‘l.lm.__. 7 _____ “3‘.wa 1 1w ‘ “Wm-m ,. M ‘vAV‘V.gAV£:5 ' A _' F“ \W4WAWRW‘24A“ £44 W” 2,4.~\\ ' 1 \V’v\_/" V._ . m 3 via ’00 20 IO 0 o J V '- V!” «33 V37. V 37.1 If: 90 ea 70 60 60 40 do Jechanical afialvc , ‘ J“es 0f ”ravel ‘ .5 " sfigoids, loose measureznqt “JVlHQ 10h Percedtade o S are u 4, o _ _ luv. . with the 42;;01n 2 per cent of the lowest Possible 1% to 2L {A c Sifes 1/3 to 3/4 in., 3/4 to 1{ in e .1 . ., used 111 v._r~1ous cornbinatioqo ‘23 .. ‘ .30 loo 90 _ \ // / 60 3“ / //// Per— ' 1 .7 //, cent age ' .. , 5 _ . .- 337' b / ‘0 ’ .144 / . A” ,b Jo 4o 30 .‘G a! 20 [o O °‘L J. 1..L 1 .‘ O 4 a 2 34L ' '2'.- 2 Iechanical anaylses of gravel, centage of voids, saving low per— staadard A.S.T.H. rodded measurement. Voids are within 2 possible, 3/4 to 1} in., 1} t er cent of tne lowest ee sizes l/S to 3/4 in., L used in various 3 p with the tnr o 2 combinations. I00 90 60 Per- centage PaS- 70 sing 60 50 40 30 20 IO 4/ 27.3 / ,1 / / / / o VOIJS / - ___._-___—____ O \1 \ 0h *2 0»!- NF '11 H C C\ . TN -.-..AH. SIP»; - V :5.- .. nod: