Kansas Geological Survey, Bulletin 109, Part 2, originally published in 1954

Application of the Tri-Axial Type Diagram to the Study of Concrete Masonry Batch Mixes in Plant Practice

by Ronald G. Hardy

Originally published in 1954 as Kansas Geological Survey Bulletin 109, Part 2. This is, in general, the original text as published. The information has not been updated.

Abstract

A study of the effects and results of using various combinations of a lightweight aggregate, Portland cement, and water on the properties of 8 by 8 by 16 inch masonry units was undertaken. For this work two series of tests were performed; on one series a Stearns No. 9 Joltcrete block machine was utilized and on the other a Besser Vibrapac. The aggregate gradation used was checked at each batch to note any variations. The tri-axial method of plotting and organizing the batch compositions and data was employed for the study and proved to be an excellent means for correlating variations. Data indicate that there is an optimum water content which must be maintained for best overall results. At this water content the cement-aggregate proportions can be varied widely yet produce very satisfactory blocks. It was also noted that variations in water content affected block texture markedly. While the data presented could apply only to the particular aggregate studied, this method is suggested as an easy means for any block manufacturer to study mixes to work toward optimum conditions.

Introduction

The production of concrete masonry units has developed into an industry of major importance in the United States. Within the past few years the use of lightweight aggregates in such units has increased markedly. The increased production of lightweight units has resulted from a growing appreciation of the value of decreasing the overall weight in masonry structures and the desirability of better heat and sound insulation.

For a number of years large quantities of pumice and scoria, and some expanded shale aggregate have been shipped into Kansas for use in lightweight concrete masonry units. In the spring of 1952 the production of an expanded shale aggregate by Buildex, Inc., Ottawa, Kansas, and a sintered clay aggregate by Mineral Products, Inc., Kansas City, Kansas, greatly increased the supply of high-grade lightweight aggregate in this State (Plummer and Hladik, 1951).

Concrete masonry units made with lightweight aggregates vary greatly in yield, unit weight, workability, strength, and product color and texture. These variations are caused to a considerable extent by variations in the quantities of the batch components--lightweight aggregate, cement, and water. In general the method used to determine the optimum quantities of these materials have been haphazard, and in many cases the results have been unsatisfactory.

During the course of a study of the application of a sintered lightweight aggregate to concrete masonry units an investigation was undertaken wherein the effects of a systematic variation in the quantities of the materials making up the normal concrete masonry batch were observed. The tri-axial method of plotting and organizing the batch compositions and data was used.

In using the tri-axial diagrams three ingredients can be plotted so that the resulting batch compositions represent all possible combinations of these ingredients (Fig. 1). Each apex of the triangular grid represents 100 percent of one of the ingredients, and the base of the grid represents none (0 percent) of that same ingredient. The exact center of the triangular grid represents equal proportions of the three ingredients, or 33 1/3 percent of each. The batches plotted may be spaced very closely on the grid to represent variations of less than 1 per cent, or spaced widely to represent larger variations. The selection of spacing depends on the significant differences in results produced by the variations represented by the spacing.

It is common practice to have wide variations in batch components plotted on the tri-axial diagram for the first series of test batches (Fig. 1). The first series of test batches will indicate a relatively small area in the tri-axial plot which contains all the mixes suitable for further testing. Within this relatively small area smaller step-by-step variations in composition are plotted in order to obtain more precise data on optimum mixes (Fig. 2.).

This method proved to be an excellent means for working out and correlating systematic variations, and should be a useful tool for the block manufacturer who wishes to study variations in mixes with the objective of determining the optimum quantities of materials to be used.

Tests

Materials Used

The aggregate used was a sintered Kansas loess having a rather rounded particle shape and well-defined cell structure. This material, when sized for concrete masonry use, had a unit weight of 1,400 pounds per cubic yard in a loose dry condition. Normally this material when supplied to the customer contains 8 to 10 percent moisture which is added at the manufacturing plant to minimize dust and help prevent segregation. This added moisture is believed to improve mixing of the ingredients in batches. Using an aggregate compaction test described by the Housing and Home Finance Agency (1949, p. 6), this aggregate showed the compressive loads for indicated compactions as follows: 1 inch compaction, 3,000 psi, 1 1/2 inch compaction, 10,000 psi. A representative screen analysis of the aggregate as furnished to the masonry manufacturers is given below.

The cements were regular Portland cement as used at the plants wherein tests were conducted.

Methods and Equipment Used in Testing

Series of tests were performed at two different plants; one plant was equipped with a Stearns No. 9 Joltcrete, whereas the other plant had a Besser Vibrapac machine. It was noted that the intensity of all vibration applied to block forming by the Vibrapac machine was considerably greater than at the other plant. The effects of the greater vibration were quite pronounced.

The test procedure employed for each mix combination follows. (1) All the aggregate was weighed, the necessary corrections for moisture content were made, and it was fed into the mixer. (2) After the mixer had run for approximately 1 minute, a sample of the aggregate was taken for screen analysis. (3) All the water (either weighed or measured by volume) was added and the mixture run for 3 minutes. (4) All the cement (after weighing) was then added and the entire mixture stirred f or 3 minutes. (5) Each batch was made into 8 by 8 by 16 inch standard three-cell units. (6) Three blocks from each batch were weighed directly from the block machine to determine yields. (7) All blocks were steam cured according to the standard practice at the respective plants.

Test Data

As a systematic scheme for proportioning the cement-aggregate-water combinations, the tri-axial type of plot was used. Each point thus selected represents some combination of the three component materials used in the study. Figure 1 shows the location of this particular field of study in relation to the points representing 100 percent of each of the ingredients. Proportions indicated are in percentages by weight and are based on dry aggregate. From this diagram it can be seen that the field studied lies in a triangle whose points represent the following batch compositions:

It is clearly shown also that the field of workable mixes occurs within fairly narrow limits.

Figure 1--Tri-axial diagram showing location of field of study (triangular area I-II-III) in relation to a complete variation of composition from 100 percent of each of the ingredients.

Figure 2 is an enlargement of the area of mixes investigated showing the composition of the various batches. These figures are based on dry aggregate; consequently corrections in water and aggregate batch weights are necessary when moisture is present in the aggregate.

Figure 2--Tri-axial diagram showing the composition of the various batches in the area of mixes investigated. This diagram is an enlargement of triangle I-II-III shown in Figure 1. Underscored numbers indicate mixes tested.

Table 1 summarizes the data assembled from the test runs on the batches formed on the Stearns No. 9 Joltcrete. The data are shown graphically in Figure 3. It was possible to run all the batches, as indicated from the data summary, on this machine. However, above 15 percent water the blocks slumped quite badly when removed from the die box and were badly smeared. The green weights of the blocks as received from this machine were relatively low.

Table 1--Summary of data obtained on 8 x 8 x 16 inch lightweight blocks made on Stearns 9 Joltcrete.

Figure 3--Graph showing changes in properties with variations in composition of standard 3-cell blocks manufactured on a Stearns Joltcrete machine.

All the mixes containing 14 percent water gave indications of having the best all-around workability regardless of the relative amounts of cement and aggregate. Increasing or decreasing the water content by only 1 percent from 14 percent affected the workability of the mix and the appearance of the block markedly. This variation in water content had more effect on the color and texture of the finished block than any variation that might be ascribed to aggregate gradation variations.

Table 2 summarizes data collected on the test runs performed using a Besser Vibrapac block machine. The data are shown graphically in Figure 4. The effects of the increased vibration intensity are very marked, resulting in lower yields, heavier blocks, and increase in compressive strengths. It was impossible to run mixes having more than 15 percent water on this machine; the material for the most part would not vibrate into the feed box and any block that was formed was very badly smeared.

Table 2--Summary of data obtained on 8 x 8 x 16 inch lightweight blocks made on Besser Vibrapac.

Figure 4--Graph showing changes in properties with variations in composition of standard 3-cell blocks manufactured on a Besser Vibrapac machine.

Figures 3 and 4 are detailed plots of these data to analyze the effects brought about by relative variations in the three batch ingredients--i.e., cement, aggregate, and water. Study of these curves points up the following trends.

1. Constant cement; increasing aggregate with decreasing water content. (a) Yield of blocks per sack and blocks per cubic yard of aggregate decrease; (b) compressive load-bearing ability decreases markedly.
2. Constant aggregate; water increasing with decreasing cement. (a) Yield of blocks per sack increases whereas the yield per cubic yard of aggregate decreases. (b) Strength in compression of the block tends to increase with increasing water content up to 14 to 15 percent water and then to fall off.
3. Constant water; increasing aggregate with decreasing cement. (a) Yield of blocks per sack increases with yield of blocks per cubic yard of aggregate decreasing. (b) Block compressive strengths show an increase to a maximum, then drop; this trend is true for all water contents. Greatest strengths occur at 14 percent or more water content.

It is also interesting to note that for this aggregate, at a content of 73 percent aggregate there seems to be generally less variation in yield and strength plus a more uniform rise in strength with decreasing cement and increasing water.

At or near a water content of 14 percent the greatest number of satisfactory blocks were produced with regard to texture. Below this percentage the blocks tended to be smooth as though the aggregate was of a very fine size; their appearance was definitely "dry." When more than 14 percent water was used, especially where the cement content was the highest, smearing became evident and where the vibration was particularly intense the blocks were entirely smoothed over so that no texture was visible. The so-called "dry" textured blocks were also quite dark in color, having a very unpleasing appearance regardless of cement and aggregate proportions.

Assuming, as the data illustrate, that the best all-around results for this particular aggregate are achieved at 14 percent water, Table 3 summarizes the results with regard to yield, strength, and weight of 8 by 8 by 16 inch masonry units of this study.

Table 3--Data on mixes containing 14 percent water.

The screen analyses of the aggregate samples taken at each batch show fairly uniform gradation. There is little, if any, correlation between the slight gradation variations of the various mixes and the final block properties; other factors were strongly overriding. An outstanding example is the smooth-textured block produced by lack of water in the run which is entirely independent of the gradation as it was for these studies. Within the limits of practicality it is entirely reasonable to assume that the results that have been illustrated herein have not been influenced by the slight aggregate gradation variations. Table 4 gives screen analyses of the aggregates used in the batch mixes.

Table 4--Screen analyses of aggregate used for tests.

Conclusions

Although only one specific type of aggregate was used for the tests summarized in this report, it is reasonable to assume that any tests involving systematic variations in the proportions of aggregate, cement, and water will give comparable results. The results of the above tests indicate that the water content is by far the most important of the three variables. For example, 14 percent water was shown to produce optimum results, whereas an increase to 15 percent water markedly lowered the quality of the block. Batches in which the cement content was constant, the aggregate used in increasing amounts, and the water in decreasing amounts showed decreasing yields in blocks per sack of cement and per cubic yard of aggregate. Batches in which the aggregate remained constant, with increasing water and decreasing cement, showed increasing yields of blocks per sack of cement and decreasing yields in blocks per cubic yard of aggregate. The strength of the blocks tended to increase with increasing water content (and decreasing cement) up to 14 to 15 percent water. In the case of batches in which the water content was kept constant and the aggregate was increased with a decreasing cement content, the yield of blocks per sack of cement increased and the yield per yard of aggregate decreased. The compressive strengths increased to a maximum and then decreased. This is a result obviously to be expected because a composition containing no cement and all aggregate and water was being approached. The important result of these tests was not so much the determination of optimum proportions of aggregate, cement, and water for block manufacture, but the proof that the systematic investigation of varying proportions by means of the tri-axial type diagram yields easily interpreted results.

References

Housing and Home Finance Agency (1949) Lightweight aggregate concretes, pp. 1-28.

Plummer, Norman, and Hladik, W. B. (1951) The manufacture of lightweight concrete aggregates from Kansas clays and shales: Kansas Geol. Survey, Bull. 91, pp. 1-100.

Kansas Geological Survey, Tri-Axial Type Diagram, Concrete Masonry Batch Mixes
Placed on web Aug. 7, 2009; originally published in March, 1954.
Comments to webadmin@kgs.ku.edu
The URL for this page is http://www.kgs.ku.edu/Publications/Bulletins/109_2/index.html