PB 99 - 100182

REPORT FHWA/NY/SR-95/119

Compression Testing of Concrete:

Cylinders vs. Cubes
DAVID J. ELWELL
GONGKANG FU

SPECIAL REPORT 119

TRANSPORTATION RESEARCH AND DEVELOPMENT BUREAU
NEW YORK STATE DEPARTMENT OF TRANSPORTATION
George E Patald, Governor/John B. Daly, Commissioner
REPRODUCED BY: NTtS
U.S. Department of Commerce”
'"

National Technical Information Service

Springfield, Virginia 22161
STATE OF NEW YORK

George E. Pataki, Governor

DEPARTMENT OF TRANSPORTATION

John B. Daly, Commissioner

Michael J. Cuddy, Assistant Commissioner for Engineering and Chief Engineer
Paul J. Mack, Deputy Chief Engineer, Technical Services Division
Robert J. Perry, Director of Transportation Research and Development

The Transportation Research and Development Bureau conducts and manages the engineering research program
of the New York State Department of Transportation. The Federal Highway Administration provides financial
technical assistance for these research activities, including review and approval of publications.

Contents of research publications are reviewed by the Bureau's Director, and the appropriate section head.
However, these publications primarily reflect the views of their authors, who are responsible for correct use of
brand names and for the accuracy, analysis, and inferences drawn from the data.

It is the intent of the New York State Department of Transportation and the Federal Highway Administration that
research publications not be used for promotional purposes. This publication does not endorse or approve any
commercial product even though trade names may be cited, does not necessarily reflect official view or policies
of either agency and does not constitute a standard, specification, or regulation.

TRANSPORTATION RESEARCH PUBLICATJONS

A. Donald Emerich, Editor

Donna L. Noonan, Graphics and Production
'Nancy A. Troxel/, Copy Preparation
COMPRESSION TESTING OF CONCRETE: CYLINDERS VS. CUBES

David J. Elwell, Junior Engineer

Gongkang Fu, Engineering Research Specialist II

Conducted in Cooperation With

The U.S. Department of Transportation
Federal Highway Administration

Special Report 119

March 1995

PROTECTED UNDER INTERNATIONAL COPYRIGHT

ALL RIGHTS RESERVED.
NATIONAL TECHNICAL INFORMATION SERVICE
U.S. DEPARTMENT OF COMMERCE

Reproduced from
best available copy.

TRANSPORTATION RESEARCH AND DEVELOPMENT BUREAU

New York State Department of Transportation
State Campus, Albany, New York, 12232-0869
 ABSTRACT

Concrete cylinder and cube specimens for compression testing were compared through a survey of
past research, including testing procedures, factors affecting the cylinder/cube strength ratio, and
conversion factors and equations. The main difference between cylinder and cube testing procedures
is capping. Cylinder ends are usually not plane or parallel enough to mate properly with platens of
compression testing machines, and thus must be capped with sulphur, neoprene, or other suitable
material for proper distribution of the applied load. Cubes, however, are not capped but cast in rigid
molds with sides that are plane and parallel. When tested, they are flipped on their sides so that
machine platens mate properly with cube surfaces. Factors affecting the cylinder/cube strength ratio
are 1) casting, curing, and testing procedure; 2) specimen geometry; 3) level of strength; 4) direction
ofloading and machine characteristics; and 5) aggregate grading. Past efforts to determine empirical
conversion relationships and conversion factors have shown that it is difficult (if not impossible) to
predict relationships between cylinder and cube strengths. Past research has also shown the
cylinder/cube strength ratio to be between about 0.65 and 0.90, although ratios outside that range
have also been observed. Based on this survey of past research, replacing cylinder testing with cube
testing is not recommended.

Preceding page blank iii

 CONTENTS

I. INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

A. Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
B. Compression Test Specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

II. FACTORS INFLUENCING THE CYLINDER/CUBE STRENGTH RATIO . . . . . . . . 5

A. Effects of Casting, Curing, and Testing Procedures . . . . . . . . . . . . . . . . . . . . . . . . 5

B. Effects of Specimen Geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
C. Effect of Strength Level (fc') . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
D. Effects of Direction of Loading and Machine Characteristics . . . . . . . . . . . . . . . . . 9
E. Effects of Aggregate Grading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

III. STRENGTH CONVERSION FACTORS AND EQUATIONS . . . . . . . . . . . . . . . . . . . . 11

IV. ADVANTAGES OF CUBES AND CYLINDERS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

V. SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

REFERENCES ............................................................. 21

Preceding page blank v

 I. INTRODUCTION

A. Background

Concrete strength in structures is typically estimated by casting smaller specimens from the same
concrete and crushing them in the laboratory. Most countries have their own standards for concrete
compression testing, which differ in many ways but probably most significantly in type of specimen
used. Cylindrical specimens-150 mm (6-in.) in diameter by 300 mm (12-in.) in height- are used
in Australia, Canada, France, New Zealand and the United States. Cube specimens (150 or 100 mm)
are used throughout much of Europe, including Great Britain and Germany (1).

The two specimen types differ in several respects, both in testing procedure and results. One of the
most significant differences is the process of capping. When cast, ends of cylinder specimens are
seldom sufficiently plane and parallel to mate properly with testing machine platens. To counteract
this problem, they must be capped. Cubes, however, are cast in rigid molds with opposite walls that
are plane and parallel. When tested, cubes are flipped on their sides, and do not require capping.

This difference in capping requirements for cylinders and cubes initiated an investigation concerning
the two specimen types. This report surveys the literature and past research on the relationship
between cylinders and cube-shaped specimens, covering three areas of comparison:

1. Testing standards and procedures,

2. Factors affecting the cylinder strength/cube strength ratio,

3. Conversion factors and equations.

Following discussion on these areas, important advantages and disadvantages of using cubes and
cylinders are summarized, and a recommendation is made regarding testing of cubes versus
cylinders. First, standards from the United States and Britain used in this investigation will be
reviewed.

B. Compression Test Specifications

1. The U.S. Cylinder Test

Standards for concrete compression tests in the United States relevant to transportation structures
are provided by AASHTO (2) and ASTM (3). Standards for making, capping, curing, and testing
cylinders are provided in the following specifications:
2 Cylinders vs. Cubes

AASHTO
M 205 Molds for Forming Concrete Test Cylinders Vertically.
T 22 Compressive Strength of Cylindrical Concrete Specimens
T 23 Making and Curing Concrete Test Specimens in the Field
T 126 Making and Curing Concrete Test Specimens in the Laboratory
T 231 Capping Cylindrical Concrete Specimens

ASTM
c 31 Practice for Making and Curing Concrete Test Specimens in the Field
C39 Standard Test Method for Compressive Strength of Cylindrical Concrete
Specimens
C 192 Practice for Making and Curing Concrete Test Specimens in the Laboratory
C 470 Standard Specification for Molds for Forming Concrete Test Cylinders Vertically
C 617 Practice for Capping Cylindrical Concrete Specimens
C 1231 Practice for Use of Unbonded Caps in Determination of Compressive Strength of
Hardened Concrete Cylinders

Cylinder specimens are cast in individual molds, which can be either multi-use or single-use. Typical
multi-use molds are made of steel or cast iron, with separate bases clamped to the mold walls.
Typical single-use molds are made of sheet metal, cardboard, paper, or plastic. Specimens are cast
and compacted in multiple layers [three layers for a 300 mm (12-in.) cylinder]. Compaction is
achieved by rodding or by vibration of each layer.

Curing occurs in a controlled environment. Initially, cylinders must be held in a moist environment
between 16 and 27° C (60 and 80° F). After shipment to the laboratory within 48 hours of casting,
molds are removed. Specimens are then placed in a moist environment at a temperature of
23 ± 1.3 o C (73 .4 ± 3o F). Additional conditions and options for curing are given in AASHTO T 23.
Specimens are typically tested after 28 days, though tests are often performed at 24 hours, 3 days,
7 days, and 90 days.

AASHTO standards require the surfaces of compression test specimens to be plane within 0.050 mm
(0.002 in.), and perpendicular to the axis within 0.5°. Because most specimens do not meet this
requirement, the ends are usually capped before any load is applied. Examples of capping materials
are sulphur mortar and cement paste. More recently, neoprene pads with steel extrusion controllers
have been used in lieu of bonded capping materials.

Specimens are loaded between two bearing blocks, the lower of which is fixed. The upper block is
spherically seated, and rotates to mate with the surface of the specimen. Specimens are loaded to
failure at a specified rate. Compressive strength is then calculated as the maximum applied load
divided by the measured cross-sectional area of the specimen. Failure mode is also recorded.
Introduction 3

2. The British Cube Test

Standards for concrete compression tests in Britain are published by the British Standards Institution.
Standards for molding, curing, and testing cube specimens are found in various parts of BS 1881:
Methods of Testing Concrete (4) as follows:

Part 108: Method for Making Test Cubes from Fresh Concrete
Part 111: Method ofNormal Curing of Test Specimens (20 Degrees Celsius Method)
Part 112: Methods of Accelerated Curing of Test Cubes
Part 116: Method for Determination of Compressive Strength of Concrete Cubes

The standard casting and curing procedure for cube specimens is similar to that for cylinder
specimens. Concrete is sampled on site; placed in a clean, dry container; and, if necessary,
transported to the location where cubes will be cast. The concrete is re-mixed by hand, and cubes
are cast in rigid multi-use molds as soon as possible after sampling. Cube molds are usually steel
or cast iron, with the base rigidly clamped or bolted to the mold walls. Single-use molds are not
permitted. Specimens are cast and compacted in layers 50 mm (2-in.) thick. Rodding and vibration
are acceptable means of compaction. Cubes are cured in a moist environment under well controlled
conditions. Curing temperatures are 20 ± 2°C (68 ± 4° F) for specimens tested at less than seven
days, and 20 ±5° C (68 ± 9° F) for specimens tested at seven days or more.

Unlike cylinders, cubes need not be capped. When tested, cubes are flipped on their sides. Opposite
faces of rigid cube molds are assumed parallel and plane. The platens of the testing machine thus
will mate evenly with the specimen surfaces without need for a capping system to distribute the load.
Molds must be checked often for intended geometry to ensure that specified tolerances are met, or
strength test results could be misleading.

Testing machines are set up similarly for cubes and cylinders. The lower platen is fixed, and the
upper platen is free to rotate to ensure that the platen aligns with the cube surface. The upper platen
is then fixed after mating with the cube surface. Specimens are loaded to failure at a specified rate,
and compressive strength is calculated as the maximum applied load divided by the measured
cross-sectional area of the specimen. Mode of failure is also noted.
 II. FACTORS INFLUENCING THE CYLINDER/CUBE STRENGTH RATIO

The relationship between compressive strength of concrete cylinder and cube specimens is complex.
The effects of numerous factors, combined with inherent variations associated with concrete,
complicate comparisons between the two types of tests. Past research has identified effects of
several factors on the cylinder/cube strength ratio:

1. Casting and testing procedures

2. Specimen geometry

3. Level of strength (fc')

4. Direction of loading and machine characteristics

5. Aggregate grading

A. Effects of Casting. Curing. and Testing Procedures

Two main factors are associated with the compression testing process: 1) type of mold, and
2) capping and surface planeness. The cylinder/cube strength ratio can be affected by the type of
mold used to cast cylinder specimens. Cubes are always cast in rigid molds (i.e., made of steel or
cast iron), but cylinders can be cast in either rigid or non-rigid molds. Cylinders cast in cardboard
molds have shown a slight reduction in strength (up to 3.5 percent) from those cast in steel or cast
iron (5).

Method of capping and out-of-plane surfaces can affect the cylinder/cube strength ratio. Method of
capping is known to have some effect on cylinder strength (6, 7,8), thus affecting the cylinder/cube
strength ratio. Planeness of cube specimens is as important as capping of cylinder specimens.
Out-of-plane mold surfaces have been shown to reduce the strength of concrete cubes by as much
as 15 percent (9). Care must be taken to ensure that mold and specimen tolerances are met, or the
cube strength and cylinder/cube strength ratio could be misleading.

B. Effects of Specimen Geometry

Various aspects of specimen geometry are recognized to affect compressive strength test results.
The most significant factor is the specimen's height-to-maximum lateral dimension (hid) ratio. Other

5
Preceding page blank
 Figure 1. Approximate effects of multiaxial stresses in cylinder and cube specimens

\ Specimen affected
\
\ by lateral stresses
0.866d \ I 30o throughout its
\ I height - - - - - - _ .
\ I
+----t---v
2d 0.268d
Centeral region
unaffected by
lateral stress

Figure 2. General relationship between height/diameter ratio and strength ratio.

2.0

1.8
~
~

~ 1.6

\
5
iI
1:"1)
.....0 1.4

0
~
~
1:"1)

1.2
\'
~ ~ ........._
1.0

0.8
0 1.0 2.0
--
---- -
3.0
-
4.0
hid

6
Strength Ratio 7

geometrical factors are d alone, and overall specimen volume (V). These three factors, along with
cross-sectional shape, will now be discussed.

Although desirable, true uniaxial stress throughout a specimen is not possible in practice. Frictional
effects between the specimen and machine platens produce lateral stresses in specimens. This results
in a multiaxial state of stress that tends to increase the strength exhibited by test specimens. Lateral
stresses affect specimen stress state in a cone-shaped or pyramid-shaped region to a depth of about
(.r3/ 2 )d = 0.866d at each end (1). Such specimens as 150 x 300 mm (6 x 12 in.) cylinders with
h <~: 1.7d will have a region not usually experiencing these multiaxial stresses. Cubes are affected by
multiaxial stresses throughout the specimen. Both are shown in Figure 1. It can be expected that
cubes will exhibit greater strengths than cylinders for otherwise identical concretes. Past research
has shown this to be the case (1 0, 11).

The general relationship observed between hid and strength is shown in Figure 2 (1, 10). It is evident
from Figure 2 that the strength ratio is more sensitive to hldfor hid< 1.5 than for hid> 1.5. Cubes
thus are more susceptible than cylinders to variations resulting from slight differences in hid.

The effect of d is based on the probability of having a weak unit of concrete in any cross-section.
The larger a specimen's d, the more likely it contains a cross-section with an element of low strength
that governs failure of the specimen. Neville (12) suggests that the relationship between strength
P and lateral dimension dis P oc 1/d z. Because 150 x 300 mm (6 x 12 in.) cylinders and 150 mm
cubes have identical d, lateral dimension is not an issue when comparing these specimens. However,
variations in dimensions, such as the 100 mm cube allowed by BS 1881, make dan influencing
factor in the cylinder/cube strength ratio. Neville (J) has also observed that the effect of d becomes
insignificant as dis increased beyond 600 mm (24-in.), indicating that such reductions in strength
need not be extrapolated to real pavements and structures with much larger dimensions.

Specimen volume V affects strength in two ways. First, distribution of stress throughout a specimen
is relatively more uniform for specimens of larger volume . This is important, since high stress
concentrations can increase the chances of premature failure, and increase the variation of test results
(11). A 150 x 300 mm (6 x 12 in.) cylinder has a volume about 1.65 times that of a 150 mm cube,
making cubes more susceptible to stress concentrations. The second effect of V is that greater
volumes are more likely to contain a weak unit that reduces strength. This second volume effect is
closely related to the effect of d.

Cross-sectional shape is not believed to have a significant effect on strength test results. Tests on
cubes and cylinders with hid= 1 yield similar results (11, 13). A comparative study between 150 x
150 x 300 mm (6 x 6 x 12 in.) prisms and 150 x 300 mm (6 x 12 in.) cylinders showed little
difference ~etween strengths of the two specimen types (14). Thus, corrections based on shape are
not needed.
 Table 1. Evans data relating strength to cylinder/cube
strength ratio (1, 15).
Strength, MPa (ksi) Cylinder/Cube
Cubes Cylinders Ratio
9.0 (1.3) 6.9 (1.0) 0.77
15.2 (2.2) 11.7 (1.7) 0.77
20.0 (2.9) 15.2 (2.2) 0.76
24.8 (3.6) 20.0 (2.9) 0.81
27.6 (4.0) 24.1 (3.5) 0.87
29.0 (4.2) 26.2 (3.8) 0.91
29.6 (4.3) 26.9 (3.9) 0.91
35.8 (5.0) 31.7(4.6) 0.98
36.5 (5.3) 34.5 (5.0) 0.94
42.1 (6.1) 36.5 (5.3) 0.87
44.1 (6.4) 40.7 (5.9) 0.92
48.3 (7.0) 44.1 (6.4) 0.91
52.4 (7.6) 50.3 (7.3) 0.96

Figure 3. Direction of loading relative to casting direction.

Direction of
Loading

Direction of
Loading
....

Casting
Layers

8
Strength Ratio 9

C. Effect of Strength Level (fj_

Nominal strength of concrete has been shown to affect the cylinder/cube strength ratio. Research
by Evans (15) indicates that this ratio decreases with decreasing concrete strength. Table 1 indicates
that the cylinder/cube strength ratio ranges from 0. 77 to 0.96, depending on concrete strength level.
The relationship between strength and cylinder/cube strength ratio is noticeable, and should be
considered whenever the two geometries are compared.

D. Effects of Direction of Loading and Machine Characteristics

Compression loads are applied in the direction of casting for cylinders, and perpendicular to the
direction of casting for cubes. Because both cylinders and cubes are cast and consolidated in
multiple layers, direction of loading is an important factor in the relationship between cylinder
strength and cube strength. Figure 3 shows the difference. When cylinders are loaded, each casting
layer occupies an entire cross-section, and receives the total load from the testing machine. When
cubes are loaded, each layer extends from top to bottom, and receives a portion of the total load.

The effect of flipping a cube on its side depends on two related factors: aggregate segregation and
platen fixity. Segregation is important because it can cause variations in strength and elasticity
between the casting layers (16). When cubes are flipped, these variations exist in every cross-section
perpendicular to the direction of loading. Each casting layer receives its load from a different part
of the platen. If non-segregating concrete is used, then strength across a plane should not vary
significantly.

The second effect- platen fixity - comes into play only when segregation occurs. Tarrant (16)
showed that when the upper platen is not allowed to rotate, the strongest layers carry most of the
load. This results from weak layers yielding more than strong layers. Failure occurs only after the
strong layers are loaded to capacity, and the weak layers are forced to carry the remaining load. The
observed cube strength is higher than that of an equivalent cube loaded in the direction of casting.
If the upper platen is free to rotate, it adjusts to deformations that vary between casting layers. Each
layer carries the same amount of load, and failure occurs when capacity of the weakest layer is
reached. Strength of the stronger layers is never fully used. The observed strength is less than that
of an equivalent cube loaded in the direction of casting.

Experimental results have shown the effects of segregated concrete on cube and cylinder strength.
Sigvaldason (17) found that segregating concretes generally gave smaller cylinder strength/cube
strength ratios than more uniform concrete. His results showed the cylinder/cube strength ratio to
be 0.71 to 0.77 for segregated concrete, and 0.76 to 0.84 for non-segregated concrete. Also, cube
strength was more sensitive to the method of end loading. Sigvaldason observed strength variations
of about 7 percent for cubes and 1 percent for cylinders for varying end conditions (pinned-pinned,
pinned-fixed, and fixed-fixed). Neville (18) found no statistical difference between strengths of
cubes of non-segregating concrete loaded perpendicular and parallel to the direction of casting.
10 Cylinders vs. Cubes

Figure 4. Wall effect.

Aggregate
Mortar

Tarrant (15) studied effects of lubrication in the pin of the platen. He found that the type of
lubrication used determined the extent to which such a platen could be considered pinned. Type of
lubrication used in the pin, combined with cube flipping effects, also affects the cylinder/cube
strength ratio.

E. Effects of Aggregate Grading

Aggregate grading is known to affect concrete strength in any structure or specimen. The effect is
magnified for compression test specimens due to the relative size of aggregate particles to specimen
dimensions. Most standards set a lower limit for the ratio of d to maximum nominal aggregate size.
Typically, this allowable minimum is around 3 or 4.

Aggregate grading affects specimen strength through the "wall effect" (J), as shown in Figure 4. In
concrete specimens a greater volume of space exists between the aggregate and mold wall than
between aggregate particles within the specimen. The extra mortar at the walls increases the
specimen's compressive strength. Because this is a surface effect, compressive strength increase is
greater for specimens with larger surface/volume ratios.

Comparing the surface/volume ratio for standard cylinders and cubes indicates how aggregate
grading may affect the cylinder/cube strength ratio. The surface/volume ratio of 150 x 300 mm (6 x
12 in.) cylinders is 0.033/mm, but it is 0.040/mm for 150 mm (6-in.) cubes, suggesting that cubes
are more s~nsitive to changes in aggregate grading than cylinders. In fact, Gyengo's research (14)
showed 1) that cylinder strength is less affected by changes in aggregate grading, and 2) that the
cylinder/cube strength ratio decreases with increasing coarseness of aggregate grading.
 III. STRENGTH CONVERSION FACTORS AND EQUATIONS

Past studies have resulted in equations or factors for converting strength test results among specimen
types. In addition, some standards specify conversion factors for specimens of different sizes.
Conversion factors and equations from several studies and standards are presented here.

AASHTO/ASTM (2,3)

Conversion factors are provided in the AASHTO and ASTM codes for evaluating strength test
results when hid < 1.8. These factors convert strength test results to equivalent results for a
specimen having hid= 2. Generally, these factors are used for tests on cores taken from structures,
from which it is not always possible to obtain hid= 2 specimens. Factors from ASTM C 39 and
AASHTO T 22 are as follows:

hid: I. 75 1.50 1.25 1.00

Factor: 0.98 0.96 0.93 0.87

United States Bureau of Reclamation (19)

Conversion factors for variations in hid are provided in the U.S. Bureau of Reclamation's standards
for concrete. These factors also convert strength test results to 150 x 300 mm (6 x 12 in.) cylinder
strengths. The factors may be taken from a gt:aph similar to Figure 2 or from a table, both provided
in Bureau of Reclamation standards. Values from the table are given in Table 2. Also provided in
the standards are conversion charts for cylinders with hid= 2 and varying d.

UNESCO (20)

In UNESCO's reinforced concrete manual, conversion factors are provided for specimens of varying
size and shape. Table 3 provides the factors recommended by UNESCO.

11
 Table 2. U.S. Bureau of Reclamation conversion factors.
hid Factor,% hid Factor,%
2.25 101.4 1.6 96.8
2.2 101.1 1.5 96.0
2.1 100.6 1.4 95.2
2.0 100.0 1.3 94.5
1.9 99.2 1.2 93.1
1.8 98.4 1.1 90.0
1.7 97.6 1.0 85.0

Table 3. UNESCO conversion factors.

Specimen Shape Specimen Size Conversion Factor
Cylinder 15x30cm 1.00
10 x 20 em 0.97
25 x 50 em 1.05
Prism 15 x 15 x 45 em 1.05
20 x 20 x 60 em 1.05
Cube 10cm 0.80
15 em 0.80
20cm 0.83
30cm 0.90

12
Strength Conversion 13

Neville's Equations (11)

In 1966, Neville used data from several previous studies to develop a general relationship between
compression test specimens of varying shapes and sizes. He postulated that strength could be
determined as a function of V, d, and hid:

r: =1 cv. d. ~) (1)

The result was two equations relating specimen strength to that of a 150 mm (6-in.) cube. His first
equation includes all three variables:

p 0.697 (2)
- = 0.56 +
p6 v
-+-
h
6hd d

where P = compressive strength,

P6 =strength of a 150 mm (6-in.) cube,

V =volume,

h = height, and

d = maximum lateral dimension.

The strength ratio of any specimen type to a 150 mm (6-in.) cube is obtained by inserting values of
V, h, and d for the specimen of interest. Since these equations were developed using English units,
all dimensions must be specified in inches. According to this equation, the ratio of 150 x 300 mm
(6 x 12 in.) cylinder strength to 150 mm cube strength is 0.81.

When Neville dropped the variable hid, he could still obtain a good fit with the data. The resulting
equation is

(3)

where d6 =maximum lateral dimension of a 150 mm (6-in.) cube= 150 mm (6-in.),

A = cross-sectional area, and

A 6 =cross-sectional area of a 150 mm (6-in.) cube= 22,500 mm2 (36 in. 2).
 Table 4. Cylinder/cube strength ratios using L'Hermites
equation.
~be Strength, psi Cylinder/Cube Ratio
3000 0.765
4000 0.790
5000 0.809
6000 0.825
7000 0.838
8000 0.850

Table 5. Results of cylinder vs. cube comparative studies.

Reference Average Remarks
cyL/cube
ratio
Cormack (11) 0.87 Study focused on high strength concrete. Few data
were generated for fc'<41 MPa (6 ksi)

Evans (1, 15) 0.77-0.96 Lower-strength concretes had generally lower

cylinder/cube strength ratios
Sigvaldason (17) 0.71-0.77 Segregating concrete
0.76-0.84 Non-segregating concrete
Lysle and Johansen (21) 0.86

Gyengo (14) 0.65-0.84 Variation due to varying coarseness of aggregate

grading

Gonnerman (1 0) 0.85-0.88 Tests performed using standard cylinders and 6" and 8"
cubes
Plowman, Smith, and 0.74 Water-cured specimens} In all cases, portions of steel
Sheriff (13) 0.64 Air-cured specimens bars were embedded in
cyIinder specimens
Rajuand 0.61 Using 150 mm cubes
Basavarajaiah (22) 0.51 Using I 00 mm cubes

Lasisi, Osunade, and 0.67-0.76 Landcrete specimens (small aggregate from lateric soil)
Olorunniwo (23) 0.55-0.86 Concrete specimens

14
Strength Conversion 15

According to Equation 3, the ratio of 150 x 300 mm (6 x 12 in.) cylinder strength to 150 mm (6 in.)
cube strength is 0.79. Test results used to create these equations w~re compared with predictions
using these equations. Neville's equations predicted actual strength test results to within 10- to IS-
percent difference.

L'Hermite's Equation (J)

In 1955, R. L'Hermite proposed a simple equation for the cylinder/cube strength ratio as a function
of the cube strength:
(4)
Cylinder Strength = 0.76 + 0.2 xlog feu
10
Cube Strength 2840

where feu is cube strength in psi. Some strength ratios based on L'Hermite's equation are given in
Table 4.

Results of Other Comparative Studies

Numerous studies have sought to determine an experimental relationship between cylinder and cube
strengths, many of which appeared in the earlier discussion. Results for nine studies are summarized
Table 5. It is evident from these results that the relationship between cylinder strength and cube
strength cannot be easily represented by simple equations or conversion factors.
 IV. ADVANTAGES OF CUBES AND CYLINDERS

Several considerations may be evaluated in choosing the type of specimen to be tested. Cubes and
cylinders should be evaluated with respect to both testing procedures and quality of results.
Important advantages and disadvantages of each specimen type are listed here.

Advantages:

• Rigid molds and loading cubes on their sides eliminate need for a capping system.
• Cubes are lighter, making them easier to handle. Cube weight is 64 percent of cylinder
weight for cylinders with hld=2 and identical d. For 2400 kg/m3 (150 pcf) concrete, a 150
mm (6-in.) cube is 8.1 kg (18.8 lb), and a 150 x 300 mm (6 x 12 in.) cylinder is 12.7 kg
(29.5 lb).
• Less concrete is required to cast cubes.
• Cubes require less storage space in curing rooms.

Disadvantages:

• The higher compressive strength and greater cross-sectional area of cubes require a
higher-capacity testing machine.
• More cubes are needed due to the greater variability associated with smaller specimens.
• Cubes are more sensitive to changes~ aggregate grading, platen fixity, and variations in the
hid ratio.
• If concrete segregates, cube flipping may result in misleading test data.
• Only rigid multi-use molds are permitted for cubes.

Cylinders

Advantages:

• Cylinders with high hid ratios have a region free of friction effects at the platens, which
contributes to less variation in test results.
• Lower variations in test results reduce the required number of cylinders.
• Both single-use and multi-use molds are permitted for cylinders.
• Cylinders are tested in the direction of casting.

Preceding page blank 11

18 Cylinders vs. Cubes
Disadvantages:

• Cylinders have a weight more than 1.5 times that of cubes, when dis identical and hid=2.
• Capping is required, because cylinder surfaces seldom meet the tolerances set by AASHTO
and ASTM. Capping adds extra material and labor costs to the compression testing process.
For example, the cost of neoprene caps, based on NYSDOT experience, is $10 to $36 per cap
pair, with 100 specimens tested per pair. Additional costs appear for extrusion controllers
and labor.
 V. SUMMARY

Two main types of concrete specimens are used for compression testing. Cylinder specimens are
used in the United States, Australia, Canada, France, and New Zealand. Cube specimens are used
for compression tests throughout much of Europe, including Great Britain and Germany. The
significant differences between the two procedures are twofold. First, cylinders may be cast in
single-use or rigid metal molds, but cubes must be cast in rigid metal molds. Second, cylinder
specimens must be capped to distribute the applied load evenly over the ends, but cubes need not be
capped and are flipped on their sides for loading.

Numerous factors influence the cylinder/cube strength ratio, these five being the most important:

1. Testing procedure- type of cylinder mold and capping system, and planeness of cube
surfaces.

2. Specimen geometry- ratio of height-to-maximum lateral dimension (hid) is recognized as

the most important geometrical factor, and is inversely related to specimen strength. Other
factors are d and volume V.

3. Strength level- this is positively correlated with the cylinder/cube strength ratio.

4. Pattern ofload distribution- fixity of the loading platens and loading direction for cubes
perpendicular to the direction of casting are significant, for segregating concretes.

5. Aggregate Grading- changing aggregate grading affects cube strength more than cylinder
strength, and increasing aggregate coarseness decreases the cylinder/cube strength ratio.

Strength test results have not been accurately converted between specimen types. Substantial
variation exists among the numerous conversion factors and equations developed for this purpose.
Although results vary among investigations, the cylinder/cube strength ratio is between 0.65 and
0.90 for 150 x 300 mm (6 x 12 in.) cylinders and 150 mm (6-in.) cubes.

In conclusion, this survey of past research provides no strong support for recommending that cube
testing replace cylinder testing. It was suggested at the beginning of this investigation that should
previous research results lead toward a positive recommendation, then limited physical testing might
be performed to verify those results. In particular, verification was considered for empirical
conversion relationships, conversion factors, and affecting factors. As summarized here, numerous
factors influence the cylinder/cube strength ratio, and empirical relationships and conversion factors
are associated with relatively large scatter. A limited testing program would be insufficient to verify

19
20 Cylinders vs. Cubes

previous research results, and cube testing is not being recommended. Physical testing to support
such a recommendation thus was not undertaken.
 REFERENCES

1. Neville, A.M. Properties of Concrete. London: Pitman, 1981 (3rd ed.), pp. 529-604.

2. Standard Specifications for Transportation Materials and Methods ofSampling and Testing.
Washington, D.C.: American Association of State Highway and Transportation Officials, 1993.

3. Annual Book ofASTM Standards. Philadelphia: American Society for Testing Materials, 1993.

4. BS 1881: Methods ofTesting Concrete. London: British Standards Institution, 1983.

5. Price, W .H. "Factors Influencing Concrete Strength." Journal of the American Concrete
Institute, Vol. 47, Feb. 1951, pp. 417-432.

6. Richardson, D.N. "Effects of Testing Variables on the Comparison of Neoprene Pad and
Sulphur Mortar-Capped Concrete Test Cylinders." ACI Materials Journal, Vol. 87, No. 5
(September-October 1990), pp. 489-495.

7. Grygiel, J.S., and Amsler, D.E. Capping Concrete Cylinders with Neoprene Pads. Research
Report 46, Engineering Research and Development Bureau, New York State Department of
Transportation, April 1977.

8. Saucier, K.L. Effect of Method of Preparation of Ends of Concrete Cylinders for Testing.
Miscellaneous Paper C-72-12, U.S. Army Waterways Experiment Station, Vicksburg,
April1972.

9. Westley, J.W. "Some Experiments on Concrete Cubes with Non-Plane Surfaces." Magazine
of Concrete Research, Vol. 18, No. 54 (March 1966), pp. 35-37.

10. Gonnerman, H.F. "Effect of Size and Shape ofTest Specimen on Compressive Strength of
Concrete." Proc. ASTM, Vol. 25, Part 2, 1925, pp. 237-250.

11. Cormack, H.W. "Notes on Cubes versus Cylinders." New Zealand Engineering (Wellington),
Vol. 11, No.3 (March 1956), pp. 98-99.

12. Neville, A.M. "A General Relation for Strengths of Concrete Specimens of Different Shapes
and Sizes." Journal of the American Concrete Institute, Vol. 63, No. 10 (October 1966),
pp. 1109-1195.

21
22 Cylinders vs. Cubes

13. Plowman, J.M., Smith, W.F., and Sheriff, T. "Cores, Cubes, and the Specified Strength of
Concrete." The Structural Engineer, Vol. 52, No. 11 (Nov. 1974), pp. 421-426.

14. Gyengo, T. "Effect of Type of Test Specimen and Gradation of Aggregate on Compressive
Strength of Concrete." Journal of the American Concrete Institute, Vol. 34, Jan.-Feb. 1938,
pp. 269-282.

15. Evans, R.H. 1944. "The Plastic Theories for the Ultimate Strength of Reinforced Concrete
Beams." Journal of the Institution of Civil Engineers (London), Vol. 21, 1943-1944,
pp. 98-121.

16. Tarrant, A.G. "Frictional Difficulty in Concrete Testing." The Engineer, Vol. 198, No. 5159
(December 1954), pp. 801-802.

17. Sigvaldason, O.T. "The Influence of Testing Machine Characteristics Upon the Cube and
Cylinder Strength of Concrete." Magazine of Concrete Research, Vol. 18, No. 57
(December 1966), pp. 197-206.

18. Neville, A.M. The Influence of the Direction of Loading on the Strength of Concrete Test
Cubes. ASTM Bulletin No. 239, July 1959, pp. 63-65.

19. Concrete Manual, Part 2. Denver: United States Bureau of Reclamation, 1992 (9th ed.), p. 93.

20. UNESCO. Reinforced Concrete: An International Manual. London: Butterworths, 1971,

p.20.

21. Pancholi, V.R. "Design Considerations." Civil Engineering (London), September 1979,
pp. 51, 53, 81.

22. Raju, N.K., and Basavarajaiah, B.S. '"Experimental Investigations on Prismatic Control
Specimens for Compressive, Flexural, and Tensile Strength of Concrete." Journal of the
Institution ofEngineers (India). Civil Engineering Division, Vol. 56, May 1976, pp. 254-257.

23. Lasisi, F., Osunade, J.A., and Olorunniwo, A. "Strength Characteristics of Cube and Cylinder
Specimens of Laterized Concrete." West Indian Journal of Engineering, Vol. 12, No. I,
(January 1987), pp. 50-59.