This international standard was developed in accordance with internationally recognized principles on standardization established in the Decision on Principles

for the
Development of International Standards, Guides and Recommendations issued by the World Trade Organization Technical Barriers to Trade (TBT) Committee.

Designation: E606/E606M − 21

Standard Test Method for

Strain-Controlled Fatigue Testing1
This standard is issued under the fixed designation E606/E606M; the number immediately following the designation indicates the year
of original adoption or, in the case of revision, the year of last revision. A number in parentheses indicates the year of last reapproval.
A superscript epsilon (´) indicates an editorial change since the last revision or reapproval.

1. Scope times at repeated intervals. The test method may be adapted to

1.1 This test method covers the determination of fatigue guide testing for more general cases where strain or tempera-
properties of nominally homogeneous materials by the use of ture may vary according to application specific histories. Data
test specimens subjected to uniaxial forces. It is intended as a analysis may not follow this test method in such cases.
guide for fatigue testing performed in support of such activities 1.5 The values stated in either SI units or inch-pound units
as materials research and development, mechanical design, are to be regarded separately as standard. The values stated in
process and quality control, product performance, and failure each system may not be exact equivalents; therefore, each
analysis. While this test method is intended primarily for system shall be used independently of the other. Combining
strain-controlled fatigue testing, some sections may provide values from the two systems may result in non-conformance
useful information for force-controlled or stress-controlled with the standard.
testing. 1.6 This international standard was developed in accor-
1.2 The use of this test method is limited to specimens and dance with internationally recognized principles on standard-
does not cover testing of full-scale components, structures, or ization established in the Decision on Principles for the
consumer products. Development of International Standards, Guides and Recom-
mendations issued by the World Trade Organization Technical
1.3 This test method is applicable to temperatures and strain
Barriers to Trade (TBT) Committee.
rates for which the magnitudes of time-dependent inelastic
strains are on the same order or less than the magnitudes of
2. Referenced Documents
time-independent inelastic strains. No restrictions are placed
on environmental factors such as temperature, pressure, 2.1 ASTM Standards:2
humidity, medium, and others, provided they are controlled A370 Test Methods and Definitions for Mechanical Testing
throughout the test, do not cause loss of or change in dimension of Steel Products
with time, and are detailed in the data report. E3 Guide for Preparation of Metallographic Specimens
E4 Practices for Force Verification of Testing Machines
NOTE 1—The term inelastic is used herein to refer to all nonelastic
strains. The term plastic is used herein to refer only to the time-
E8/E8M Test Methods for Tension Testing of Metallic Ma-
independent (that is, noncreep) component of inelastic strain. To truly terials
determine a time-independent strain the force would have to be applied E9 Test Methods of Compression Testing of Metallic Mate-
instantaneously, which is not possible. A useful engineering estimate of rials at Room Temperature
time-independent strain can be obtained when the strain rate exceeds some E83 Practice for Verification and Classification of Exten-
value. For example, a strain rate of 1 × 10−3 sec−1 is often used for this
purpose. This value should increase with increasing test temperature. someter Systems
E111 Test Method for Young’s Modulus, Tangent Modulus,
1.4 This test method is restricted to the testing of uniform and Chord Modulus
gage section test specimens subjected to axial forces as shown E112 Test Methods for Determining Average Grain Size
in Fig. 1(a). Testing is limited to strain-controlled cycling. The E132 Test Method for Poisson’s Ratio at Room Temperature
test method may be applied to hourglass specimens, see Fig. E177 Practice for Use of the Terms Precision and Bias in
1(b), but the user is cautioned about uncertainties in data ASTM Test Methods
analysis and interpretation. Testing is done primarily under E209 Practice for Compression Tests of Metallic Materials at
constant amplitude cycling and may contain interspersed hold Elevated Temperatures with Conventional or Rapid Heat-
ing Rates and Strain Rates
1
This test method is under the jurisdiction of ASTM Committee E08 on Fatigue
and Fracture and is the direct responsibility of Subcommittee E08.05 on Cyclic
2
Deformation and Fatigue Crack Formation. For referenced ASTM standards, visit the ASTM website, www.astm.org, or
Current edition approved June 1, 2021. Published July 2021. Originally approved contact ASTM Customer Service at service@astm.org. For Annual Book of ASTM
in 1977. Last previous edition approved in 2019 as E606/E606M – 19ɛ1. DOI: Standards volume information, refer to the standard’s Document Summary page on
10.1520/E0606_E0606M-21. the ASTM website.

Copyright © ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959. United States

&RS\ULJKWE\$670,QW
ODOOULJKWVUHVHUYHG:HG-XQ*07
1
'RZQORDGHGSULQWHGE\
<LOGL]7HNQLN8QLYHUVLWHVLSXUVXDQWWR/LFHQVH$JUHHPHQW1RIXUWKHUUHSURGXFWLRQVDXWKRUL]HG
 E606/E606M − 21

NOTE 1—* Dimension d is recommended to be 6.35 mm [0.25 in.]. See 7.1. Centers permissible. ** This diameter may be made greater or less than
2d depending on material hardness. In typically ductile materials diameters less than 2d are often employed and in typically brittle materials diameters
greater than 2d may be found desirable.
NOTE 2—Threaded connections are more prone to inferior axial alignment and have greater potential for backlash, particularly if the connection with
the grip is not properly designed.
FIG. 1 Recommended Low-Cycle Fatigue Specimens

E337 Test Method for Measuring Humidity with a Psy- E1245 Practice for Determining the Inclusion or Second-
chrometer (the Measurement of Wet- and Dry-Bulb Tem- Phase Constituent Content of Metals by Automatic Image
peratures) Analysis
E384 Test Method for Microindentation Hardness of Mate- E1823 Terminology Relating to Fatigue and Fracture Testing
rials
E399 Test Method for Linear-Elastic Plane-Strain Fracture 3. Terminology
Toughness of Metallic Materials
3.1 The definitions in this test method are in accordance
E466 Practice for Conducting Force Controlled Constant
with Terminology E1823.
Amplitude Axial Fatigue Tests of Metallic Materials
E467 Practice for Verification of Constant Amplitude Dy- 3.2 Definitions: Additional definitions associated with time-
namic Forces in an Axial Fatigue Testing System dependent deformation behavior observed in tests at elevated
E468 Practice for Presentation of Constant Amplitude Fa- homologous temperatures are as follows:
tigue Test Results for Metallic Materials 3.2.1 hold period, τh—the time interval within a cycle
E691 Practice for Conducting an Interlaboratory Study to during which the stress or strain is held constant.
Determine the Precision of a Test Method
3.2.2 inelastic strain, εin—the strain that is not elastic.
E739 Practice for Statistical Analysis of Linear or Linearized
Stress-Life (S-N) and Strain-Life (ε-N) Fatigue Data 3.2.2.1 Discussion—For isothermal conditions, εin is calcu-
E1012 Practice for Verification of Testing Frame and Speci- lated by subtracting the elastic strain from the total strain.
men Alignment Under Tensile and Compressive Axial 3.2.3 total cycle period, τt—the time for the completion of
Force Application one cycle. The parameter τt can be separated into hold and
E1049 Practices for Cycle Counting in Fatigue Analysis non-hold (that is, steady and dynamic) components:

&RS\ULJKWE\$670,QW
ODOOULJKWVUHVHUYHG:HG-XQ*07
2
'RZQORDGHGSULQWHGE\
<LOGL]7HNQLN8QLYHUVLWHVLSXUVXDQWWR/LFHQVH$JUHHPHQW1RIXUWKHUUHSURGXFWLRQVDXWKRUL]HG
 E606/E606M − 21
sively in isothermal, constant-rate testing, in the analysis of hysteresis
τt 5 ( τ 1( τ
h nh (1)
loops. In such cases, a value for E* can best be determined by cycling the
where: specimen prior to the test at stress or strain levels below the elastic limit.
E* is NOT the monotonic Young’s modulus.
∑τh = sum of all the hold portions of the cycle and
∑τnh = sum of all the nonhold portions of the cycle.
4. Significance and Use
τt also is equal to the reciprocal of the overall frequency
4.1 Strain-controlled fatigue is a phenomenon that is influ-
when the frequency is held constant.
enced by the same variables that influence force-controlled
The following equations are often used to define the instan-
fatigue. The nature of strain-controlled fatigue imposes distinc-
taneous stress and strain relationships for many metals and
tive requirements on fatigue testing methods. In particular,
alloys:
cyclic total strain should be measured and cyclic plastic strain
ε 5 ε in1ε e (2) should be determined. Furthermore, either of these strains
typically is used to establish cyclic limits; total strain usually is
σ
εe 5 ~ see Note 2 ! controlled throughout the cycle. The uniqueness of this test
E*
method and the results it yields are the determination of cyclic
and the change in strain from any point (1) to any other stresses and strains at any time during the tests. Differences in
point (3), as illustrated in Fig. 2, can be calculated as fol- strain histories other than constant-amplitude alter fatigue life
lows: as compared with the constant amplitude results (for example,
periodic overstrains and block or spectrum histories).
S
ε 3 2 ε 1 5 ε 3in1
σ3
E* D S
2 ε 1in1
σ1
E* D (3) Likewise, the presence of nonzero mean strains and varying
environmental conditions may alter fatigue life as compared
All strain points to the right of and all stress points above with the constant-amplitude, fully reversed fatigue tests. Care
the origin are positive. The equation would then show an must be exercised in analyzing and interpreting data for such
increase in inelastic strain from 1 to 3 or: cases. In the case of variable amplitude or spectrum strain
σ1 σ3 histories, cycle counting can be performed with Practice
ε 3in 2 ε 1in 5 ε 3 2 ε 11 2 (4) E1049.
E* E*

Similarly, during the strain hold period, the change in the 4.2 Strain-controlled fatigue can be an important consider-
inelastic strain will be equal to the change in the stress di- ation in the design of industrial products. It is important for
vided by E*, or: situations in which components or portions of components
undergo either mechanically or thermally induced cyclic plas-
σ2 2 σ3 tic strains that cause failure within relatively few (that is,
ε 3in 2 ε 2in 5 (5)
E* approximately <105) cycles. Information obtained from strain-
NOTE 2—E* represents a material parameter that may be a function of
environment and test conditions. It also may vary during a test as a result controlled fatigue testing may be an important element in the
of metallurgical or physical changes in the specimen. In many instances, establishment of design criteria to protect against component
however, E* is practically a constant quantity and is used rather exten- failure by fatigue.
4.3 Strain-controlled fatigue test results are useful in the
areas of mechanical design as well as materials research and
development, process and quality control, product
performance, and failure analysis. Results of a strain-controlled
fatigue test program may be used in the formulation of
empirical relationships between the cyclic variables of stress,
total strain, plastic strain, and fatigue life. They are commonly
used in data correlations such as curves of cyclic stress or strain
versus life and cyclic stress versus cyclic plastic strain obtained
from hysteresis loops at some fraction (often half) of material
life. Examination of the cyclic stress–strain curve and its
comparison with monotonic stress–strain curves gives useful
information regarding the cyclic stability of a material, for
example, whether the values of hardness, yield strength,
ultimate strength, strain-hardening exponent, and strength
coefficient will increase, decrease, or remain unchanged (that
is, whether a material will harden, soften, or be stable) because
of cyclic plastic straining (1).3 The presence of time-dependent
inelastic strains during elevated temperature testing provides
the opportunity to study the effects of these strains on fatigue

FIG. 2 Analyses of a Total Strain versus Stress Hysteresis Loop 3

The boldface numbers in parentheses refer to the list of references at the end of
Containing a Hold Period this standard.

&RS\ULJKWE\$670,QW
ODOOULJKWVUHVHUYHG:HG-XQ*07
3
'RZQORDGHGSULQWHGE\
<LOGL]7HNQLN8QLYHUVLWHVLSXUVXDQWWR/LFHQVH$JUHHPHQW1RIXUWKHUUHSURGXFWLRQVDXWKRUL]HG
 E606/E606M − 21
life and on the cyclic stress-strain response of the material. 6.3 Fixtures:
Information about strain rate effects, relaxation behavior, and 6.3.1 To minimize bending strains, specimen fixtures should
creep also may be available from these tests. Results of the be aligned such that the major axis of the specimen closely
uniaxial tests on specimens of simple geometry can be applied coincides with the force axis throughout each cycle. It is
to the design of components with notches or other complex important that the accuracy of alignment be kept consistent
shapes, provided that the strains can be determined and from specimen to specimen. Accuracy and repeatability of
multiaxial states of stress or strain and their gradients are alignment are functions of the load frame alignment and
correctly correlated with the uniaxial strain data. specimen-to-specimen geometric variability. Alignment shall
be checked by means of a trial test specimen with longitudinal
5. Functional Relationships strain gages placed at four equidistant locations around the
5.1 Empirical relationships that have been commonly used minimum diameter. The trial test specimen shall be turned
for description of strain-controlled fatigue data are given in about its axis, installed, and checked for each of four orienta-
Appendix X1. These relationships may not be valid when large tions within the fixtures. The maximum bending strains so
time-dependent inelastic strains occur. For this reason, original determined shall not exceed 5 % of the minimum axial strain
data should be reported to the greatest extent possible. Data range imposed during any test program. For specimens having
reduction methods should be detailed along with assumptions. a uniform gage length, it is advisable to place a similar set of
Sufficient information should be developed and reported to gages at two or three axial positions within the gage section.
permit analysis, interpretation, and comparison with results for One set of strain gages should be placed at the center of the
other materials analyzed using currently popular methods. gage length to detect misalignment that causes relative rotation
5.2 If use is made of hourglass geometries, original data of the specimen ends about axes perpendicular to the specimen
should be reported along with results analyzed using the axis. An additional set of gages should be placed away from the
relationships in Appendix X2. gage-length center to detect relative lateral displacement of the
specimen ends. The lower the bending strain, the more repeat-
6. Methodology able the test results will be from specimen to specimen. This is
6.1 Testing Machine—Testing should be conducted with a especially important for materials with low ductility where
tension-compression fatigue testing machine that has been much better alignment may be needed.
verified in accordance with Practices E4 and E467, unless more NOTE 6—This section refers to Practice E1012 Type A tests.
stringent requirements are called for in this specification. The NOTE 7—Four strain measurements, 90° opposed to each other, are
testing machine, together with any fixtures used in the test required to ensure that bending strains are not large. Utilization of a single
program, must meet the bending strain criteria in 6.3.1. The extensometer with dual axial outputs will allow for only two specimen
machine should be one in which specific measures have been loadings to gather the required four strain readings, without the necessity
of strain gaging specimens.
taken to minimize backlash in the loading train.
6.3.2 Several commonly used fixturing techniques are
NOTE 3—Force measuring capability of 45 kN [approximately 10 kips]
or greater would be sufficient for the recommended specimens (Section 7)
shown schematically in Fig. 3. The selection of any one
and most test materials. The machine force capacity used for these fixturing technique depends primarily upon the user’s speci-
specimens would not be required to exceed 110 kN [approximately 25 men design. Fixtures should be constructed of hardened steel
kips]; however, large-capacity fatigue machines may be beneficial because for high strength and abrasion resistance. The collet type grip
of increased axial stiffness and decreased lateral deflection of these shown, or another fixturing technique that provides high
systems. Achieving a change in axial concentricity of less than or equal to
0.05 mm [0.002 in.] TIR (total indicator reading), as measured between precision lateral stiffness to hold precise alignment is accept-
the top and bottom specimen fixture under cyclic force, is a measure of able. Fixtures not capable of high alignment may be coupled
success with respect to minimizing lateral deflection of the loading train. with the Field’s metal pot (2, 3) of Fig. 4 or a similar device.
6.2 Strain Control—Testing machine controls should permit Such a device may help to compensate for misalignment in the
cycling between constant strain limits. If material behavior loading train that would induce bending strains in the specimen
permits (for example, aging effects do not hinder), control during fixturing. Placement of the fixtures within die-set or flex
stability should be such that the strain maximum and minimum bars reduces relative lateral motion of specimen ends and
limits are repeatable over the test duration to within 1 % of the increases lateral stiffness that is important in machines that do
range between maximum and minimum control limits. not provide adequate safeguards against compressive buckling
of the test specimen.
NOTE 4—See 6.4.1 and 6.5 on use of force and strain transducers in
relation to repeatability requirements.
6.3.3 For elevated-temperature testing it is usually neces-
NOTE 5—For strain control under long-life conditions it is sometimes sary to provide some means for cooling the fixtures to prevent
advantageous to run a pseudostrain control test under force control. The damage to other loading-train components such as force
test could be started in strain control and switched to force control after tranducers. One method commonly used employs water-
cyclic stabilization of the stress response occurs. In these cases strain cooling coils attached to the fixtures or to other appropriate
should be monitored (directly or indirectly) and adjustments made in force
control to maintain strain limits within an envelope of 2 % of the desired locations in the loading train. Care must be taken to avoid
strain amplitude of the specified waveform. Practice E466 provides affecting the force tranducer calibration or the loading-train
additional details on force controlled axial fatigue testing. alignment by the addition of cooling coils.

&RS\ULJKWE\$670,QW
ODOOULJKWVUHVHUYHG:HG-XQ*07
4
'RZQORDGHGSULQWHGE\
<LOGL]7HNQLN8QLYHUVLWHVLSXUVXDQWWR/LFHQVH$JUHHPHQW1RIXUWKHUUHSURGXFWLRQVDXWKRUL]HG
 E606/E606M − 21

NOTE 1—The clamping force should be greater than the cyclic force to avoid backlash within the specimen fixture.
FIG. 3 Schematic Examples of Fixturing Techniques For Various Specimen Designs

6.4 Extensometers—Extensometers should be employed for should be suitable for control purposes, readout, and recording.
the purpose of measuring deformation in the gage section. The extensometers should qualify as Class B-2 or better in
They should be suitable for dynamic measurements over long accordance with Practice E83.
periods of time.
6.4.1 The non-self contained extensometer may be of two NOTE 8—For best results, the extensometer system (mechanical and
electrical) should have a maximum nonlinearity of 0.3 % of full-scale
major types: Contacting (for example, the more frequently range. Thus, the extensometer design should minimize sources of me-
used strain gage or LVDT type as shown in Fig. 5) or chanical hysteresis. The more effective designs have a low activation force
noncontacting (for example, optical types). The output of the that eliminates slippage of the contacts and a low mass to provide high
extensometer or auxiliary device of the extensometer system natural frequency for improved dynamic response characteristics.

&RS\ULJKWE\$670,QW
ODOOULJKWVUHVHUYHG:HG-XQ*07
5
'RZQORDGHGSULQWHGE\
<LOGL]7HNQLN8QLYHUVLWHVLSXUVXDQWWR/LFHQVH$JUHHPHQW1RIXUWKHUUHSURGXFWLRQVDXWKRUL]HG
 E606/E606M − 21
minimum diameter. Calibration of extensometers should be
conducted before and after each test program.
NOTE 10—Care should be taken in the measurement of diametral strains
for materials such as cast materials that possess large grains or a large
degree of preferred orientation. These, as well as hexagonal close-packed
materials, tend to be anisotropic and therefore may require special
methods of strain measurement and interpretation because Poisson’s ratio
changes substantially with the orientation of the extensometer with respect
to the crystallographic orientation of the specimen. Cyclic hardening or
softening also might alter the apparent value of Poisson’s ratio, thereby
complicating data analyses and interpretation.
6.5 Force Transducers—A force transducer should be
placed in series with the test specimen for the purpose of
measuring magnitude and sense of the axial force transmitted
through the specimen. Force transducer capacity should be
selected to adequately cover the range of forces to be measured
in the test being conducted, but not so large as to render larger
errors (that is, greater than 1 % of the difference between
maximum and minimum control limits). Force transducer
calibration should be verified in accordance with Practices E4
and E467.
NOTE 11—The force transducer should be designed specifically for
NOTE 1—Field’s metal pot is used to provide initially zero stress in the fatigue testing and possess the following characteristics: high resistance to
specimen during fixturing. This pot may be within a die-set to combine bending; high axial stiffness; high linearity; accuracy and sensitivity; low
zero fixturing stress with rigid alignment. Field’s metal is a non-toxic hysteresis; high overturning moment stiffness; and high lateral stiffness.
alternative to Wood’s metal, which was referred to in earlier versions of For best results, it is recommended that the maximum force transducer
this standard. nonlinearity and hysteresis should not exceed 0.5 % and 0.3 % of
FIG. 4 Schematic of Field’s Metal Pot Showing Principle of Op- full-scale range, respectively.
eration
6.6 Data Recording Systems—Analog strip chart and X-Y
recorders or their digital equivalent should be considered a
6.4.2 Extensometers should measure longitudinal deforma- minimum requirement for data collection.
tion when a uniform-gage specimen, such as shown in Fig. NOTE 12—Accuracy of recording systems should be kept within 1 % of
1(a), is tested. Generally, these extensometers are attached as full scale. Analog/digital devices are available that include maximum and
shown in Fig. 5(a). minimum limit detection, maximum-minimum memory, and underpeak
detection.
NOTE 9—Care should be exercised when installing the longitudinal NOTE 13—Data acquisition system characteristics such as sampling
extensometer so as to prevent damage to the specimen surface and frequency and data skew between force and deformation (stress and strain)
consequential premature fatigue failure at the contact points. It is very channels can affect hysteresis loop presentation on an X-Y recorder used
important to ensure a secure attachment of the extensometer to the test in digital recording systems. It is recommended that these characteristics
specimen. Damaged or worn contacts or flexure in the attachment be taken into consideration along with the strain rate or frequency of
apparatus can induce significant hysteresis errors in the measurement. cycling to determine that the hysteresis plots are within the required error
Often, small strips of transparent tape, copper bondable strain gage limits.
terminals, or other such protective tabs are adhered to the specimen’s
uniform section at the locations where extensometer tips would contact the 6.6.1 X-Y Recording—Some means of X-Y recording should
material. Use of the tape or tabs tends to “cushion” the attachment. be used for the purpose of recording hysteresis loops of force
Another alternative is the use of quick-drying epoxy. Light force springs versus deformation or stress versus strain. A potentiometric
or small rubber bands are often employed to hold the extensometer to the
specimen. Dulling the tips for softer material is also commonly done.
X-Y recorder or an oscilloscope equipped with a camera or
Extensometer slippage can be observed after the first several cycles from data storage capability is an acceptable alternative. The poten-
X-Y traces or strip chart recordings by observing the stress-strain tiometric X-Y recorder should be used only when the rate of
response. Unusual shifts in mean values of stress in response to imposed cycling results in a pen velocity that is less than one-half of the
strain ranges are an indication of such slippage. recorder’s slewing speed. At higher frequencies, the oscillo-
6.4.3 Extensometers should measure diametral deforma- scope may be used. Alternative devices include: digital X-Y
tions when specimens having hourglass profiles are tested. A plotters for real time recording or to plot stored data and data
typical method of diametral displacement measurement is logging devices that store data in a host computer system or
shown schematically in Fig. 5(b). Curved extensometer tips, transmit data to a printer.
convex in the longitudinal plane, can provide point contact 6.6.2 If digital-type recording devices are used, it is recom-
during testing. Care should be exercised during installation of mended that a sufficient number of simultaneous data pairs
the diametral extensometer to prevent damage to the test (such as stress and strain) be taken for both the ascending
specimen surface. Extensometer tips should be adjusted prop- segment and descending segment of the hysteresis loop to
erly to minimize the force they impose on the specimen. When adequately determine the shape of the loop.
installing the extensometer, gently move its tip longitudinally 6.6.3 Strip Chart Recording—Strip chart recorders may be
along the specimen while watching the gage readout to find the used to monitor force (or strain). If used, the frequency of the

&RS\ULJKWE\$670,QW
ODOOULJKWVUHVHUYHG:HG-XQ*07
6
'RZQORDGHGSULQWHGE\
<LOGL]7HNQLN8QLYHUVLWHVLSXUVXDQWWR/LFHQVH$JUHHPHQW1RIXUWKHUUHSURGXFWLRQVDXWKRUL]HG
 E606/E606M − 21

FIG. 5 Extensometer Schematic

&RS\ULJKWE\$670,QW
ODOOULJKWVUHVHUYHG:HG-XQ*07
7
'RZQORDGHGSULQWHGE\
<LOGL]7HNQLN8QLYHUVLWHVLSXUVXDQWWR/LFHQVH$JUHHPHQW1RIXUWKHUUHSURGXFWLRQVDXWKRUL]HG
 E606/E606M − 21
test should be such that the recording pen velocity never
exceeds one-half of the recorder’s slewing speed. It is recom-
mended that these recorders be calibrated at the testing
frequencies used. Storage oscilloscopes also may be used to
record the force versus strain loops. Force or strain peaks also
may be monitored by devices that detect, display, and retain
maximums and minimums in memory or that reproduce these
data at predetermined periods.
6.7 Cycle Counter—A cycle counter shall be used to indi-
cate total accumulated cycles of loading or straining. An
elapsed time indicator is a desirable adjunct to the cycle
counter to provide an excellent check of both frequency and the FIG. 6 Block Diagram of Strain Computer (See Appendix X2 for
current cycle count. Two types of counters are generally Discussion of Mathematical Relationship)
available, mechanical or electronic. A minimum requirement is
that a counter have typically five or six digits and × 10, × 100,
and × 1000 range multipliers. Digital counters with 1 count (see Note 10 and Note 14). Both of these recommended
resolution with 1 count resolution (no multipliers) are avail- specimens possess a solid circular cross section and minimum
able. Counters are often equipped with a “preset count” feature diameters of 6.35 mm [0.25 in.] in the test section. Specific
that may be used to stop a test for examination of the specimen, cross-sectional dimensions are listed here only because they
to command a recorder to take data, or to end a test after a have been dominant in the generation of the low-cycle fatigue
specific number of cycles. database that exists in the open literature. Specimens possess-
ing other diameters or tubular cross sections may be tested
6.8 Calibration—The calibration interval of all electronic successfully within the scope of this test method; however,
recording and transducer systems should be performed in crack growth rate, specimen grain size, and other consider-
accordance with the manufacturer’s recommendations; in the ations might preclude direct comparison with test results from
absence of these, the interval shall be no greater than one year the recommended specimens (see Note 15). While design of
and even more frequently if necessary to maintain required specimen end connections is primarily dependent upon user
accuracy. Calibration should be checked whenever accuracy is preference (see Note 16), a number of commonly used con-
in doubt. All calibrations should be traceable to the Interna- figurations are shown in Fig. 1(c), 1(d), 1(e) and 1(f). Care
tional System of Units (SI) through a National Metrology must be exercised in the machining of uniform-gage specimens
Institute (NMI) or an International Metrology Institute. When to blend the shoulder radius at the specimen ends with
calibrating a transducer system, it is important that it be minimum diameter so as to avoid undercutting. So that stress
performed using the same setup and arrangement of compo- concentrations are minimized, the shoulder radius should be as
nents as used in the test. As an example, when calibrating a large as possible, consistent with limitations on specimen
force transducer used on an automated system, it is necessary length.
to calibrate the output from the computer, not from any
intermediary electronics. NOTE 14—Lives determined using tubular specimens are less than those
for solid specimens, the extent of which depends on the failure criteria and
6.9 Strain Computer—An analog (or digital) computer is specimen configuration. Differences in excess of a factor of two are not
recommended for use in low-cycle fatigue tests of hourglass unusual for failure criteria based on separation, whereas for failure defined
by crack size contained within the tube wall there will be much less
specimens whenever appreciable cyclic hardening and soften- difference.
ing occurs during the test. Such a computer is useful when used NOTE 15—Selection of either the uniform-gage section or hourglass
in the real-time mode with servocontrolled testing machines profile is commonly based upon the magnitude of strain range to be
and can be used for limit control of screw-driven machines. imposed. The recommended uniform gage specimen is frequently suitable
for strain ranges up to about 2 %. Above 2 % hourglass specimens may be
The computer should be designed to convert diametral strain
necessary. Soft materials or elevated temperatures may dictate lower strain
and axial force signals into an axial strain signal. See Appendix ranges. The maximum strain range may be increased by appropriate lateral
X2 for conversion relations. In the case of servocontrolled restraints and through the use of short loading trains. Options to increase
machines, this axial strain signal may be used as a feedback stiffness to avoid the use of hourglass specimens should be exhausted
signal for control purposes, thus simulating axial strain control. before adopting the configuration shown in Fig. 1(b). If these options fail,
the recommended hourglass specimen possesses a profile ratio of 12:1 for
A block diagram for the analog (or digital) computer program radius-of-curvature to minimum radius-of-specimen. If the user wishes,
is shown in Fig. 6. different ratios between the limits of 8:1 and 16:1 may be employed.
Lower limits will increase stress concentration and may affect fatigue life;
7. Specimens higher ratios limit the specimen’s buckling resistance. For some materials
tested in the low-life range, hourglass specimens might give different
7.1 Specimen Design—Fig. 1 shows two basic specimen results from similarly stressed uniform-gage specimens. It is very difficult
configurations. Fig. 1(a) shows a recommended uniform-gage to determine axial strains from measurements of diametral strain in
specimen. When the choice of an hourglass configuration is hourglass specimens for many anisotropic as well as cast materials.
NOTE 16—Design of specimen end connections is dependent upon user
deemed necessary, the profile recommended is as shown in Fig. preference, fixturing, or availability of material, or a combination of all
1(b). Use of Fig. 1(b) should follow careful consideration of three; it is constrained principally by proper considerations of axial
problems of data interpretation, and anisotropy and buckling alignment and backlash. Button-head end connections, such as those

&RS\ULJKWE\$670,QW
ODOOULJKWVUHVHUYHG:HG-XQ*07
8
'RZQORDGHGSULQWHGE\
<LOGL]7HNQLN8QLYHUVLWHVLSXUVXDQWWR/LFHQVH$JUHHPHQW1RIXUWKHUUHSURGXFWLRQVDXWKRUL]HG
 E606/E606M − 21
shown in Fig. 1(d) and 1(e), permit precise alignment with a specimen end section hourglass specimen in Fig. 7(b), see Ref (4) for other
clamping preload (to avoid backlash in the grip). The threaded connection, designs. The geometries that are adequate for resisting buck-
shown in Fig. 1(c), is useful where the available material is not thick
enough to provide for the larger diameter button-head ends. As a
ling and/or incremental bending collapse at short lives often
cautionary note, threaded connections are more prone to inferior axial will lead to grip failures at long lives. The investigator may find
alignment and have greater potential for backlash particularly if the it convenient to employ two geometrically similar specimen
connection with the grip is not properly designed. The efficiency button- designs for development of a strain-life curve.
head connection, shown in Fig. 1(e), provides the button-head preloading
feature without requiring larger diameter ends. The button-head design is 7.2 Specimen Preparation—Specimens should be prepared
useful at elevated temperatures, as it does not suffer the “oxidation- by a specific set of procedures that is known to provide
sticking” experienced with threaded ends, but it may produce some consistent test results. Agreement between the testing organi-
specimen failures in the fixture when used at room temperatures. The
design shown in Fig. 1(f) is convenient for use in collet-type hydraulic zation and the user of the test results concerning preparation
grips. This configuration eliminates long life thread failures often associ- procedures should be obtained. The following provides recom-
ated with Fig. 1(c) type specimens. mended guidelines.
7.1.1 Alternative Specimen Design for Sheet Specimens— 7.2.1 Specimen Coupons and Materials—Coupons from
Often, it is desirable to obtain test specimens from sheet which specimens are machined should either be nominally
material that is less than 6.0 mm [0.24 in.] thick. In general, the homogeneous or sampled from the source material, or both, so
considerations discussed in other sections apply equally to as to be representative of the properties sought in the applica-
sheet testing. However, special specimen geometries and tion of the material to its end use. Thus, when material
gripping arrangements, as well as more sensitive force and requirements allow, specimens should be removed from the
strain transducers, are necessary. It is strongly recommended same material and product form that will be used in the
that torques introduced by actuator rod rotations be eliminated fabricated component of interest. Any material orientations,
by use of rotational restraints or similar devices. Typical such as rolling direction or casting direction, should be
specimen designs that have been used successfully are shown identified with respect to the orientation of the specimen axes.
in Fig. 7. The specimens in Fig. 7(a) have a rectangular cross Orientation notation used in accordance with Test Method
section and are suitable up to strain amplitudes of at least 1 % E399 is acceptable such as L, T, S, LT, TL, ST, and the like.
applied to sheets as thin as 2.54 mm [0.10 in.]. For higher strain 7.2.2 Specimen Surface Preparation—Specimens prepared
amplitudes, antibuckling restraints can be adapted to the from coupons will possess a “surface preparation history” as a
specific geometry and extensometer used. In using such consequence of machining operations, heat treatments, and the
restraints, care must be taken to avoid increased resistance to effects of environment during the storage period prior to
axial force influenced by the restraints. When restraints cannot testing. Unless the purpose of testing is to determine the
be adopted, it may be necessary to use the cylindrical cross influence of specific surface conditions on fatigue life, it is

FIG. 7 Sheet Fatigue Specimens—Alternative to Fig. 1 Specimens

&RS\ULJKWE\$670,QW
ODOOULJKWVUHVHUYHG:HG-XQ*07
9
'RZQORDGHGSULQWHGE\
<LOGL]7HNQLN8QLYHUVLWHVLSXUVXDQWWR/LFHQVH$JUHHPHQW1RIXUWKHUUHSURGXFWLRQVDXWKRUL]HG
 E606/E606M − 21
recommended that specimen surface preparation be performed 8.1.1.1 For materials that are fatigue tested at temperatures
in a manner that will have a minimum influence upon the other than ambient, all temperatures throughout the gage
variability in fatigue lives exhibited by the specimen group section (for uniform gage specimens this is the region with
tested. Ordinarily, this would be accomplished by: constant cross-sectional area) shall be:
7.2.2.1 consistently machining specimens to be as smooth T n 6∆T (6)
and uniform in surface finish (in the gage region) as feasible for
the subject material and the machining techniques available, where:
and by employing as a final operation a machining or other Tn = nominal test temperature in °C and
“finishing” procedure that would introduce minimal surface ∆T = 2 °C [3.6 °F] or 1 % × Tn, °C, whichever is greater.
metal distortion (see Note 17), and by NOTE 18—The temperature variability in the gage section can become
7.2.2.2 ensuring, through the use of protective atmospheres, a critical issue, particularly if material properties (for example, major
alterations of strength, modulus of elasticity, ductility, etc.) or metallur-
that surface attack, such as oxidation and corrosion, does not gical stability (for example, microstructure, crystal structure, etc.) are
occur, either during heat treatments or during specimen affected significantly. For these reasons as well as others, the temperature
storage, for all specimens within a program. variability within the gage section should be maintained as small as
possible. Because temperature effects can be significant, the actual
NOTE 17—Appendix X3 presents an example of a machining procedure temperature variability should be reported with the test results, as should
that has been employed on some metals to minimize variability of the heating method (induction heating, resistance heating, infrared lamp,
machining and heat treatment influences upon fatigue life. etc.).
The exact procedure of specimen preparation and handling 8.1.1.2 For the duration of the test, the controlled tempera-
should be clearly and carefully documented. It also would be ture of the specimen should be Tn 6 2 °C [3.6 °F].
prudent to determine and record the surface residual stresses
and the residual stress profile of at least one exemplary NOTE 19—If the temperature cannot be maintained within limits
specimen. mentioned above, then temperature deviations should be reported. If
possible, the effect of temperature should be demonstrated throughout the
7.3 Specimen Storage—Test specimens that may be suscep- range of test temperatures.
tible to corrosion in moist room-temperature air should be 8.1.2 Elevated temperatures may be imposed by any of
protected immediately after preparation and stored until they several methods: (1) high-frequency induction (Note 20), (2)
are tested. Specimens may be stored in a suitable protective resistance or radiant furnace, or (3) immersion in an inert
environment, such as dry inert gas (as might be conveniently heated gas or liquid. In (1) and (2) above, an enclosure is
employed in a laboratory desiccator) or a vacuum. The method recommended to prevent air currents in the vicinity of the
of storage should be clearly and carefully documented. specimen from causing undesirable temperature gradients.
7.4 Materials Description—A complete material description Specimens tested at room temperature also should be in
is desirable. It is recommended that the following microstruc- draft-free surroundings. Temperatures below room temperature
tural and mechanical properties be obtained. may be imposed by placing the specimen and gripping appa-
7.4.1 Microstructural Characteristics—Composition, grain ratus in a refrigerated chamber that may be either of the liquid
size (see Test Methods E112), crystallographic structure, pre- or gaseous type, depending on temperature requirements and
ferred orientation if present, general shape of grains (that is, other possible environmental considerations. Liquefied gases,
equiaxed or elongated), second-phase particles (see Practice such as liquid nitrogen, or solidified gases, such as dry ice
E177), heat treatment (whether at the mill, during fabrication, placed in a liquid medium, provide possible means for low-
in the laboratory, or a combination of all three), position in temperature testing.
ingot or sheet roll, and specification designation (ASTM,
NOTE 20—When inductively heating magnetic materials (those mate-
ASME, AISI, Military, SAE, etc.). rials having relative permeabilities significantly greater than unity), it
7.4.2 Mechanical Properties—For purposes of performing should be recognized that a varying stress in the specimen can affect the
the test and calculating results it is desirable to have available distribution of eddy currents in the specimen and may change the
the following representative mechanical properties, obtained at temperature profile. This effect is influenced by the specimen material,
design and heat transfer characteristics, the temperature magnitude, the
the appropriate temperature and measured in accordance with stress magnitude and distribution, the cyclic waveform, and the testing
the applicable standards such as Test Methods E8/E8M, E9, frequency (strain rate). The most pronounced effect is generally produced
E111, E132 and Practice E209; tensile or compressive yield when conducting tests at low frequencies or with tests containing hold
strength or yield point, or both; ultimate tensile strength; periods. In any case, the temperature profile of magnetic specimens should
percent elongation; percent reduction of area; Poisson’s ratio; be evaluated throughout the straining cycle. When the effect is severe, it
may be necessary to use a susceptor with the induction coil or to use an
and Young’s modulus. The following true stress-strain proper- alternate heating method.
ties also may be desirable: true fracture strength, true fracture NOTE 21—Use of glass insulation may avoid difficulty with wires
ductility, strain hardening exponent, and strength coefficient. submerged in a cooling solvent.
Hardness also may be determined in accordance with Test
8.1.3 If testing is performed in air, relative humidity may be
Methods A370 or E384, or both.
measured in accordance with Test Method E337, unless it has
already been determined that moisture has little or no effect on
8. Procedure
fatigue life for the material under test. If an effect is present,
8.1 Test Environment: relative humidity should be controlled; when uncontrolled it
8.1.1 Temperature: should be carefully monitored and reported.

&RS\ULJKWE\$670,QW
ODOOULJKWVUHVHUYHG:HG-XQ*07
10
'RZQORDGHGSULQWHGE\
<LOGL]7HNQLN8QLYHUVLWHVLSXUVXDQWWR/LFHQVH$JUHHPHQW1RIXUWKHUUHSURGXFWLRQVDXWKRUL]HG
 E606/E606M − 21
8.2 Measurement of Test Specimen Dimensions—For the a hold on diametral strain will permit the total axial strain to
purpose of making an accurate determination of specimen change during each cycle and will not produce correct relax-
cross-sectional area, measure the reduced section as follows: ation information.
8.2.1 Measure the diameter at the center of the gage section 8.3.3 Other Control Methods—Fatigue testing machines
by means of an optical comparator or other optical means to an that do not provide continuous closed loop control of either
accuracy of 0.0125 mm [0.0005 in.] or better. A precision specimen force or specimen displacement generally have the
micrometer may be used in place of the optical comparator if capability to impose limits on the chosen test variable.
its use does not damage the gage section surface in a way as to However, they do not control that variable throughout the
affect specimen performance. For uniform-gage specimens, fatigue cycle. Limit control is a special case of closed loop
check diameters for at least two other positions within the control. Thus, force and displacement signals may be handled
specimen gage length. The minimum cross-sectional areas in a manner similar to that of 8.3.2 to determine strain limits.
should be used for computing the stresses in the specimen It is not necessary to use a computer for limit control of
during the test. The area at temperature should be used in hourglass specimens if periodic adjustments are made to the
calculating stress. This area can be obtained by correcting the diametral strain amplitude in such a manner as to maintain
room-temperature result using the coefficient of thermal ex- constant axial strain limits. These adjustments are necessary
pansion. for materials that undergo significant cyclic hardening and
softening because of attendant changes to the relationship
8.3 Test Machine Control—It is necessary to control one (or between axial strain and induced diametral strain. The addi-
more) variable(s) (for example, stress, strain, force, tional complications of complex waveforms and time-
displacement, or other appropriate parameters) in a manner that dependent inelastic strains severely curtail the acceptability of
is in keeping with the test objectives. limit control techniques. If the technique of limit control is
8.3.1 Control Mode—Total axial strain amplitude is the used, the intra-cycle and inter-cycle variation of the parameter
most commonly utilized control variable in a low-cycle fatigue of interest should be monitored, and if necessary, periodic
test. Total axial strain is often controlled continuously through- adjustments should be made to the testing machine to produce
out each fatigue cycle in a manner prescribed in 8.4. It also is the desired response. Such changes should be reported.
acceptable to control only the limits of either total axial strain 8.4 Waveform—The strain versus time waveform should be
or plastic axial strain. In such cases, vary another variable, such consistent throughout a test program unless test objectives are
as diametral strain, displacement, or force, between these limits to determine waveform effects. In the absence of specific
in some cyclically consistent manner under either closed loop waveform requirements or equipment limitations, a triangular
or other control means. For long-life fatigue tests that exhibit waveform for continuous cyclic tests and trapezoidal wave-
low levels of plastic strain, it is acceptable to control force forms for hold period tests are preferred as this provides for a
while monitoring strain and making periodic adjustments of constant strain rate. A sinusoidal waveform is also acceptable
mean force and force range in order to maintain the desired where consistent with the test objectives.
strain limits. Similarly, tests may be initiated in strain control 8.4.1 The oscillating strain should be maintained at all times
and switched to force control using the stabilized force peaks within an envelope of 2 % of the desired strain amplitude of the
as limits. When time dependent effects are present, it may no specified waveform. If systematic or anomalous deviations are
longer be acceptable to control only the limits of the required identified they should be reported.
strain. Continuous control of the parameter of interest may be 8.5 Strain Rate and Frequency of Cycling—The strain rate
necessary to obtain the desired intra-cycle response. For or frequency of cycling should not be changed for the duration
example, if the force is controlled between total axial strain of each test as well as for the duration of a test program, unless
limits in this regime, a quite different material response will be the test objective is specifically to determine either strain rate
produced than if the total axial strain is continuously con- or frequency effects, or with the agreement between the parties
trolled. involved. In all cases, the details of waveform and of strain rate
8.3.2 Closed Loop Method—Fatigue testing machines of the or frequency shall be reported with the test data.
closed loop servocontrolled type often are capable of continu- NOTE 22—While constant strain rate testing is often preferred in
ously controlling specific test variables such as force or research due to the possible effect of strain rate on material fatigue
displacement through appropriate selection of feedback sig- properties, constant frequency testing is often preferred in testing pro-
nals. Application of scale factors to these signals thereby grams for commercial applications. Constant strain rate testing may be
permits continuous control of stress or strain. Axial stress may experimentally more tractable than constant frequency testing since
long-life, small-strain tests in the former mode may be completed in
be scaled directly from the force transducer signal. Axial strain shorter periods of time than tests conducted in the latter mode. Fatigue
may be scaled directly from an axial extensometer signal when properties derived from constant frequency testing should be so noted,
uniform-gage specimens are tested. When hourglass specimens particularly if there is a selective transfer from strain-control to force-
are tested, an axial strain signal must be determined from a control in the longer life regimes as a means to decrease testing time.
diametral extensometer signal and the force signal by means of 8.5.1 Other means of rate control are available. One ac-
a computer (see 6.9) if closed loop control of diametral strain cepted procedure is to maintain constant average strain rate
changes during cycling hardening and softening. Additional (twice the product of strain range and frequency) throughout
precautions should be observed when hold times are employed each test and for the duration of the test program. Another
and time-dependent inelastic strains are present. For example, acceptable procedure, is to maintain constant average plastic

&RS\ULJKWE\$670,QW
ODOOULJKWVUHVHUYHG:HG-XQ*07
11
'RZQORDGHGSULQWHGE\
<LOGL]7HNQLN8QLYHUVLWHVLSXUVXDQWWR/LFHQVH$JUHHPHQW1RIXUWKHUUHSURGXFWLRQVDXWKRUL]HG
 E606/E606M − 21
strain rate, that is, a constant value of twice the product of lengthy test durations or the limited availability of recorders, intermittent
plastic strain range and frequency. records or alternative sampling of the recorded variables is acceptable.
8.9 Determination of Failure—The definition of failure may
NOTE 23—Constant average plastic strain rate may be achieved by: (1)
applying plastic strain limit control; or (2) applying a cyclic stress-strain vary with the ultimate use of the fatigue life information (see
curve from preliminary tests to estimate the expected plastic strains and Note 25). Acceptable alternatives are as follows:
setting test frequencies on that basis. 8.9.1 Separation—Total separation or fracture of the speci-
8.5.2 Strain rates or frequencies should be sufficiently low men into two parts at (1) some location within the uniform
as to preclude specimen heating in excess of 2 °C [3.6 °F]. In section of a uniform-gage specimen, or (2) the vicinity of the
using a servocontrolled testing machine, make a comparison of minimum diameter in an hourglass specimen. All failure
the program and feedback signals to ensure that the selected locations should be recorded.
rates or frequencies are and remain within system capabilities NOTE 25—A post-mortem failure analysis should be performed to
and accuracy requirements. Frequency response of extensom- uncover any unusual causes of failure. Reporting the actual failure
eters (depending upon their design) is often a limiting factor in location is important. Inclusions, voids, defects, etc., that are not repre-
the system. sentative of the bulk material or its application may render fatigue life
determination invalid (see 8.11.3). Also, consistent failures at one position
8.5.3 If the test objective includes investigating a wide may signal alignment problems or “knife-edge” failures caused by
range of fatigue lives extending to cases of predominantly extensometer attachment.
elastic strain, then test rates or frequencies may be increased as
8.9.2 Modulus Method—For any specified number of
convenient to obtain failure in a reasonable test time, within the
cycles, N, during the test, the modulus for unloading following
requirements just stated, and with the details reported for each
a peak tensile stress is defined as ENT and the modulus for
test.
loading following a peak compression stress is ENC (see Fig.
8.6 Test Commencement: 8). Failure is defined when the ratio:
8.6.1 Begin all tests in the same direction of initial straining, Q N 5 E NT/E NC (7)
tensile or compressive, unless the purpose of testing is to study
initial loading effects. For some materials, it is acceptable to reaches one-half the value of Q for the first cycle (see Fig. 8).
increase strain amplitude gradually and continuously over a Q N f 5 0.5 Q 1 (8)
period no greater than approximately 20 cycles or 2 % of
anticipated life (whichever is less). Care should be taken in The number of cycles where this occurs is designated as the
selecting the target strain for the initial cycle to avoid over- number of cycles to failure, Nf. However, if total separation
shooting while minimizing the number of cycles to obtain the occurs first, as in 8.9.1, the life is Nf.
desired strain limits. Depending on the material, temperature, 8.9.3 Microcracking—The existence of surface microcracks
strain range, loading range, and dynamic response of the test (for example, as observed optically or by replicas) that are
systems, the peak strain of the initial cycle may vary. If larger than some preselected size consistent with test objec-
overshooting does occur, the control limits should not be tives.
adjusted downward. 8.9.4 Force (Stress) Drop—It is acceptable to define failure
8.6.2 In reverse cyclic testing of thin sheet specimens it may in a manner related to the ability to sustain a tensile force
be advisable to always begin in tension, particularly with coil (stress). Failure is often defined as the point at which the
products of materials where natural out-of-flatness occurs. In maximum force (stress) or elastic modulus (as measured when
other materials such as flat sheet products, either tension or unloading from a peak tensile stress) decreases by approxi-
compression going starts are appropriate. mately 50 % because of a crack or cracks being present. The
exact method and the percentage drop should be documented.
8.7 Number of Specimens—It is suggested that a minimum
of ten specimens be used to generate a fatigue strain-life curve. 8.10 Test Duration—Conduct testing at least until failure
It also is suggested that the replication guidelines given in STP and preferably until fracture when needs dictate and economics
588 (5) be followed, especially if subsequent statistical analy- allow. Record total accumulated cycles to failure (and fracture)
sis is planned. Reference also is made to Practice E739. by means of a cycle counter and check against a measure of
elapsed time.
8.8 Recording—Unless computerized data retrieval systems
are employed continuously and it is convenient to record in 8.11 Analysis of Data—While it is not the purpose of this
logarithmic increments of fatigue life (that is, 1, 2, 5, 10, 20, test method to specify data analysis techniques, the following
50, ...), record the initial series of hysteresis loops of axial example represents a common procedure utilized when time
stress (or force) versus total or plastic axial strain (versus total dependent strains are insignificant.
or plastic diametral strain if an axial strain signal is not 8.11.1 Determination of the Cyclic Stress-Strain Curve—
available). Record hysteresis loops thereafter at successively Generate a cyclic stress-strain curve from paired values of
larger increments of a cycle count. For tests of 100 cycles or stress amplitude and strain amplitude typically at material
more, a minimum of ten additional hysteresis loops is desir- half-life. When practical, assume a simplifying mathematical
able. When practical, continuously record the dependent vari- expression for the cyclic stress-strain relationship (6).
ables (for example, axial stress and plastic axial strain in a total NOTE 26—See Appendix X1 for expressions.
axial strain control test) as a function of time.
8.11.2 Determination of the Strain-Life Relationship—
NOTE 24—When continuous recording is not practical either because of Generate a strain-life curve from paired values of total strain

&RS\ULJKWE\$670,QW
ODOOULJKWVUHVHUYHG:HG-XQ*07
12
'RZQORDGHGSULQWHGE\
<LOGL]7HNQLN8QLYHUVLWHVLSXUVXDQWWR/LFHQVH$JUHHPHQW1RIXUWKHUUHSURGXFWLRQVDXWKRUL]HG
 E606/E606M − 21

FIG. 8 Definitions of Tension and Compression Modulus for a

Determination of Failure

versus life or plastic strain versus life and elastic strain versus 9.1.2 Specimen Materials Description (including process-
life. When practical, assume a simplifying mathematical rela- ing) *.
tionship. In a long fatigue life test program, if all the fatigue 9.1.2.1 All available mechanical properties including: yield
tests were started in strain control and switched to force control strength or yield point, or both, ultimate tensile strength,
after cyclic stabilization of the stress response (Note 5), then percent elongation and gage length, percent reduction of area,
Practice E468 could be used to generate a fatigue life relation- Poisson’s ratio, elastic modulus, true fracture strength, true
ship. fracture ductility, monotonic strain hardening exponent, mono-
8.11.3 Post-Mortem Examinations—Metallographic exami- tonic strength coefficient, hardness number, and degree of cold
nation of the failed specimens is desirable for a variety of work.
purposes depending in part on user interests. Of foremost 9.1.2.2 All available metallurgical characteristics: certified
importance is a fractographic examination of the two surfaces composition, grain size, crystallographic structure, preferred
to determine any unusual causes of failure that might invalidate orientation with respect to specimen axis, general shape of
the test results. Dimensional instability of the specimen (unin- grains (that is, equiaxed or elongated), second phase particles,
tentional changes in specimen geometry) may occur during the and heat treatment. Include photomicrographs when possible to
test as a result of time-dependent inelastic deformations. The document the above properties.
post-mortem examination should include an evaluation of this 9.1.3 Specimen Description:
occurrence. 9.1.3.1 Drawing of the specimen design, or reference to a
8.11.3.1 Scanning electron microscopy and transmission geometry illustrated in this test method*.
electron microscopy of fracture replicas are two common 9.1.3.2 Specimen fabrication and surface preparation proce-
methods used in such an investigation. Ref (7) provides a dures. If specimens were heat treated after fabrication, details
useful basis for fractographic analysis. The techniques of light must be provided*.
metallography and transmission electron microscopy are fre- 9.1.3.3 Deviations from recommended specimens configu-
quently used when studying structural changes that occur ration and specimen preparation procedures, if any.
during fatigue or the effects of metallurgical structure on 9.1.4 Description of Equipment:
fatigue behavior. 9.1.4.1 Specimen fixtures and the method used to maintain
column rigidity during compression loading.
9. Report 9.1.4.2 Testing machine.
9.1 The list of items of information that follows is suggested 9.1.4.3 Transducer system (that is, force transducer, defor-
for inclusion in any report. When publishing results in the open mation transducer).
literature, include as much information as possible, indepen- 9.1.4.4 Recorders and recording equipment.
dent of the author’s purpose. Routine laboratory reports need 9.1.5 Description of Testing Environment:
include only information pertinent to the end use of the test 9.1.5.1 Gas, liquid, or vacuum; chemical composition of
data. Minimum recommended requirements are indicated by an medium*.
asterisk (*). 9.1.5.2 Humidity of gaseous environment*.
9.1.1 Specific Objective of Testing. 9.1.5.3 Test temperature and temperature control method*.

&RS\ULJKWE\$670,QW
ODOOULJKWVUHVHUYHG:HG-XQ*07
13
'RZQORDGHGSULQWHGE\
<LOGL]7HNQLN8QLYHUVLWHVLSXUVXDQWWR/LFHQVH$JUHHPHQW1RIXUWKHUUHSURGXFWLRQVDXWKRUL]HG
 E606/E606M − 21
9.1.5.4 Temperature sensing devices, location of tempera- failure was performed during 1988 and 1989. The experimental
ture measurements, temperature variations in the gage section, program was conducted on uniform gage length specimens
any temperature variations in the cycle caused by adiabatic machined from 304 stainless steel rod material. Eight labora-
heating or magnetomechanical effects*. tories participated in the room temperature testing and five
9.1.6 Testing Conditions and Procedures: laboratories participated in the elevated temperature (538 °C
9.1.6.1 Deviations from recommended procedures, if any. [1000.4 °F]) testing. Each laboratory completed two to four
9.1.6.2 Frequency of cycling (or cyclic strain rate) and tests on specimens at each condition, with three replicates
description of waveform*. being generated in the majority of cases. The design of
9.1.6.3 Mode of control, that is, force or stress, continuous experiments and the within- and between-laboratory analysis
strain control, strain limit control, axial strain feedback, diame- of variance of the fatigue life data were conducted in confor-
tral strain feedback, etc*. mance with Practice E691.
9.1.6.4 Ratio of axial strain limits (minimum and maxi- 10.2 Test Program Results—The precision information
mum) and total axial strain range*. given below, for strain-controlled fatigue cycles to failure, is
9.1.6.5 Procedure for maintaining constant axial strain lim- stated as a percentage of the logarithm (base 10) of cycles to
its. failure. Two strain ranges, 0.70 % and 1.50 %, were tested at
9.1.6.6 Sign of strain at first quarter cycle, tensile or both the room and elevated temperatures during the interlabo-
compressive. ratory program. Fully reversed (R= -1.0) strain cycling condi-
9.1.7 Test Results—Tabulate the results for all test speci- tions were used for all of the experiments. The repeatability
mens. When used for purposes of structural analysis, the and reproducibility of the method was similar at the two
following three items are most important: temperatures tested, so the results presented below were
9.1.7.1 Initial, stabilized, or half-life values, or a combina- computed based on a weighted average of these results.
tion of all three, of the dependent variables from the list of
stress range, strain range, and inelastic strain range. Complete 10.3 Precision:
curves of these quantities throughout the specimen life. When High Strain-Range Conditions (log Nf ~ 3.0 to 3.5)
95% repeatability limit (within laboratory) 3.83%
complete curves are impractical, curves through intermittent 95% reproducibility limit (between 12.03%
values are acceptable*. laboratories)
9.1.7.2 Relaxation or creep information also should be Intermediate Strain-Range Conditions (log Nf ~ 4.0 to 4.5)
included for hold-time tests. This should include the values of 95% repeatability limit (within 2.41%
laboratory)
the relaxed stress or of creep strain, the total amount of 95% reproducibility limit (between 6.32%
relaxation or creep, and the change in the amount of inelastic laboratories)
strain during the hold period. Periodic stress-time recordings Example 1 – Given high strain-range conditions where
obtained with an expanded time scale may be necessary to average log Nf ~ 3.00
95% repeatability limits in log Nf ~ 3.00 * (1 ± 0.0383)
obtain the desired resolution and to fully characterize the - within-lab variability in Nf ~ 770 to 1,300 cycles
relaxation or creep behavior*. 95% reproducibility limits in log Nf ~ 3.00 * (1 ± 0.1203)
9.1.7.3 Total number of cycles to failure, Nf, and some - between-lab variability in Nf = 435 to 2,300 cycles

additional life measure indicative of crack formation, including Example 2 – Given intermediate strain-range conditions
where average log Nf ~ 4.50
the definition used*. 95% repeatability limits in log Nf ~ 4.50 * (1 ± 0.0241)
9.1.8 Results of Analysis for Cyclic Stress-Strain - within-lab variability in Nf ~ 24,600 to 40,600 cycles
95% reproducibility limits in log Nf ~ 4.50 * (1 ± 0.0632)
Properties—If data analysis is performed by means of the - between-lab variability in Nf = 16,400 to 60,900 cycles
relationships in Appendix X1, a tabulation of results should
include the cyclic strain hardening exponent and the cyclic The above terms (repeatability and reproducibility limits)
strength coefficient. are used as specified in Practice E1245. It is important to
9.1.9 Results of Analysis for Strain-Life Properties—If data consider that the above results were obtained using only one
analysis is performed by means of the relationships in Appen- specimen geometry and one heat of material. Substantially
dix X1, a tabulation of results should include the fatigue different repeatability and reproducibility results may have
strength exponent, the fatigue ductility exponent, the fatigue been found with other materials and testing conditions allowed
strength coefficient, and the fatigue ductility coefficient. by Test Method E606/E606M.
9.1.10 Brief Description of the Fracture Characteristics— 10.4 Bias—This method has no bias because strain-
Results of post-test metallography and scanning electron controlled fatigue properties of materials are defined in accor-
microscopy, identification of fracture mechanisms, and the dance with this method.
relative degree of transgranular and intergranular cracking.
Identify the mechanism or mode of cracking at the region of 11. Keywords
crack initiation, as well as in the region of crack growth, and 11.1 accuracy; bending strain; cycle; cyclic strain hardening
note any differences. exponent; cyclic strength coefficient; cyclic stress-strain prop-
erties; extensometer; fatigue; fatigue ductility coefficient; fa-
10. Precision and Bias tigue ductility exponent; fatigue life; fatigue strength coeffi-
10.1 Interlaboratory Test Program—An interlaboratory cient; fatigue strength exponent; fatigue testing; fatigue testing
study of the variability in strain-controlled fatigue cycles to system; force; gage length; hysteresis diagram; inelastic strain;

&RS\ULJKWE\$670,QW
ODOOULJKWVUHVHUYHG:HG-XQ*07
14
'RZQORDGHGSULQWHGE\
<LOGL]7HNQLN8QLYHUVLWHVLSXUVXDQWWR/LFHQVH$JUHHPHQW1RIXUWKHUUHSURGXFWLRQVDXWKRUL]HG
 E606/E606M − 21
load range; plastic extension; preload; rate of straining; rate of
stressing; reduced section; reduction of area; reversal; sensi-
tivity; strain; strain amplitude; strain range; strain-control;
strain-life; stress

APPENDIXES

(Nonmandatory Information)

X1. FUNCTIONAL RELATIONSHIPS

X1.1 For many metals, the following empirical relation- ∆ε p /2 5 ε' f ~ 2N f ! c (X1.5)
ships (8) have been used for a convenient description of b
∆ε/2 5 @ σ' f /E # ~ 2N f ! 1ε' f ~ 2N f ! c
(X1.6)
low-cycle fatigue data. Please note that the subscript used is
that for“ plastic strain” rather than “inelastic strain.” They are where the variables are:
used interchangeably in these fatigue equations. ∆σ = true stress range,
∆ε = true strain range,
X1.1.1 Cyclic Stress-Strain Behavior: The total strain am- ∆εp = true plastic strain range,
plitude for a completely reversed, strain-controlled test may be Nf = cycles to failure, and
expressed as: 2Nf = reversals to failure;
∆ε ∆σ ∆ε p
5 1 (X1.1) and the constants are:
2 2E 2
n' = cyclic strain hardening exponent,
Recognizing that: b = fatigue strength exponent,
∆σ/2 5 K' ~ ∆ε p /2 ! n' (X1.2)
c = fatigue ductility exponent,
K' = cyclic strength coefficient,
we may express the cyclic stress-strain curve by the consti- σ'f = fatigue strength coefficient,
tutive equation: ε'f = fatigue ductility coefficient, and
E = Young’s modulus (modulus of elasticity).
∆ε ∆σ
2
5
2E
1
∆σ
2K' S D 1/n'
(X1.3) NOTE X1.1—The user is cautioned that the equations presented in these
appendices are not readily amenable to direct use in conventional
X1.1.2 Fatigue-Life Relationships: statistical analysis procedures. Care should be taken in transforming the
variables to logarithmic base and in defining the dependent and indepen-
∆σ/2 5 σ' f ~ 2N f ! b (X1.4) dent variables.

X2. CONVERSION FROM DIAMETRAL STRAIN TO AXIAL STRAIN FOR ISOTROPIC MATERIALS AND
UNIFORM STRAINS

X2.1 Conversion of diametral strain to axial strain requires ε 5 2ε de/v e 2 ~ ε d 2 ε de! /v p (X2.5)
first the separation of the elastic and plastic components from The diametral elastic strain, εde, is related to axial stress by
the total strain by: means of Poisson’s ratio and Young’s modulus,
ε 5 ε e 1ε p (X2.1) ε de 5 2 ~ ν e σ ! /E (X2.6)
Thus,
ε d 5 ε de1ε dp (X2.2)
ε 5 σ/E 2 ε d /v p 2 ~ ν e σ ! / ~ ν p ! E (X2.7)
where: Assuming that plastic deformation occurs under a constant
e = elastic component, volume condition:
p = plastic component, ν p 5 1/2 (X2.8)
d = diametral component, and such that:
ε = total axial strain.
ε 5 ~ σ/E !~ 1 2 2ν e ! 2 2ε d (X2.9)
Axial and diametral components of strain are related through In an experiment using a diametral strain gage and an axial
Poisson’s ratio, ν, as: force transducer, analogs of σ and εd are continuously avail-
able. Young’s modulus, E, can be determined in accordance
ε e 5 2ε de/v e and ε p 5 2ε dp/v p (X2.3)
with Test Method E111. The elastic portion of stress, σ, versus
The above expressions may be rearranged to yield: diametral strain, εd, provides σ/εde, which permits νe to be
ε dp 5 ~ ε d 2 ε de! (X2.4) calculated.

&RS\ULJKWE\$670,QW
ODOOULJKWVUHVHUYHG:HG-XQ*07
15
'RZQORDGHGSULQWHGE\
<LOGL]7HNQLN8QLYHUVLWHVLSXUVXDQWWR/LFHQVH$JUHHPHQW1RIXUWKHUUHSURGXFWLRQVDXWKRUL]HG
 E606/E606M − 21

X3. EXAMPLE OF MACHINING PROCEDURE

X3.1 The following procedure was developed for machin- wherein material can be smeared over tool marks, thereby creating a
ing high-strength materials and results in minimal surface potentially undesirable influence on crack initiation during testing.
damage and alteration. It can also be applied to lower strength X3.2.3 After polishing (see Note X3.2), all remaining grind-
materials. As a conservative general measure, this procedure is ing and polishing marks should be longitudinal. No circumfer-
recommended unless: (a) the experimental objective is to ential machining should be evident when viewed at approxi-
evaluate another given surface condition, or (b) it is known that mately 20× magnification under a light microscope.
the material under evaluation is relatively insensitive to surface X3.2.4 If specimen material is soft (for example, copper,
condition. aluminum, lead, etc.) at room temperature, final material
X3.2 Procedure: removal can be performed by means of turning (rather than
grinding) and subsequent polishing.
X3.2.1 In the final stages of machining to within 0.025 mm
X3.2.5 Degrease the finished specimen. Caution should be
[0.001 in.] of the final diameter, remove small amounts of
exercised to assure that the degreasing agent does not alter
material and reduce the gage diameter 0.125 mm [0.005 in.] by
material behavior (for example, methanol on titanium alloys).
cylindrical grinding at a rate of no more than 0.005 mm
[0.0002 in.]/pass. X3.2.6 If heat treatment is necessary, conduct it before final
machining or in such a manner as to avoid any surface damage;
NOTE X3.1—Some cast materials will not benefit from successive
removal of material in small amounts, although this procedure is probably
employ an inert protective atmosphere to eliminate surface
not to their detriment. oxidation.
X3.2.2 Remove the final 0.025 mm [0.001 in.] by polishing X3.2.7 If surface observations are to be made, the test
(see Note X3.2) longitudinally to impart a maximum of 0.2-µm specimen may be electropolished in accordance with Methods
[8-µin.] surface roughness. E3.
NOTE X3.2—Extreme caution should be exercised in polishing to X3.2.8 Imprint specimen numbers on both ends of the test
ensure that material is being properly removed rather than merely smeared section in regions of low stress, away from grip contact
to produce a smooth surface. This is a particular danger in soft materials surfaces.

REFERENCES

(1) Mitchell, M. R.,“Fundamentals of Modern Fatigue Analysis for Analysis for Fatigue Experiments, ASTM STP 588, ASTM, 1975. See
Design,” Fatigue and Microstructures, ASM Handbook, Vol. 19, also Little, R. E. and Jebe, E. H., Statistical Design of Fatigue
Fatigue and Fracture, ASM International, 1996, pp. 227–249. Experiments, Applied Science Publishers, Ltd., London, 1975.
(2) Raske, D. T. and Morrow, JoDean, “Mechanics of Materials in Low (6) Landgraf, R. W., Morrow, JoDean, and Endo, T., “Determination of
Cycle Fatigue Testing,” Manual on Low Cycle Fatigue Testing, ASTM the Cyclic Stress-Strain Curve,” Journal of Materials (JMLSA), Vol
STP 465, ASTM, 1969, pp. 1–25. 4, No. 1, March 1969, pp. 176–188.
(3) Feltner, C. E. and Mitchell, M. R., “Basic Research on the Cyclic (7) “Fractography and Atlas of Fractographs,” Metals Handbook, Am.
Deformation and Fracture Behavior of Materials,” iden, pp. 27–66. Soc. Metals, Vol 9, Eighth Edition, 1974.
(4) Marsh, G. M., Robb, A. D., and Topper, T. H., “Techniques and (8) Ellis, J. R., “Results of an Interlaboratory Fatigue Testing Program
Equipment for Axial Fatigue Testing of Sheet Steel,” SAE Paper No. Conducted on Alloy 800 at Room and Elevated Temperatures,”
730578, May 1973. Journal of Testing and Evaluation (JTEVA), Vol 15, No. 5, September
(5) Little, R. E. and Jebe, E. H., Manual on Statistical Planning and 1987, pp. 249–250.

ASTM International takes no position respecting the validity of any patent rights asserted in connection with any item mentioned
in this standard. Users of this standard are expressly advised that determination of the validity of any such patent rights, and the risk
of infringement of such rights, are entirely their own responsibility.

This standard is subject to revision at any time by the responsible technical committee and must be reviewed every five years and
if not revised, either reapproved or withdrawn. Your comments are invited either for revision of this standard or for additional standards
and should be addressed to ASTM International Headquarters. Your comments will receive careful consideration at a meeting of the
responsible technical committee, which you may attend. If you feel that your comments have not received a fair hearing you should
make your views known to the ASTM Committee on Standards, at the address shown below.

This standard is copyrighted by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959,
United States. Individual reprints (single or multiple copies) of this standard may be obtained by contacting ASTM at the above
address or at 610-832-9585 (phone), 610-832-9555 (fax), or service@astm.org (e-mail); or through the ASTM website
(www.astm.org). Permission rights to photocopy the standard may also be secured from the Copyright Clearance Center, 222
Rosewood Drive, Danvers, MA 01923, Tel: (978) 646-2600; http://www.copyright.com/

&RS\ULJKWE\$670,QW
ODOOULJKWVUHVHUYHG:HG-XQ*07
16
'RZQORDGHGSULQWHGE\
<LOGL]7HNQLN8QLYHUVLWHVLSXUVXDQWWR/LFHQVH$JUHHPHQW1RIXUWKHUUHSURGXFWLRQVDXWKRUL]HG