Module Six

Destructive Testing
Introduction
In Module 5, we discussed in detail many of the important properties of metals. Once it
is recognized that these properties may be important to the suitability of a metal or a
weld, we then become interested in determining the actual values for those properties.
That is, now the designer would like to put a number on each of these important
properties so he can effectively design some structure using materials having the desired
characteristics.
s mentioned in Module 5, there are numerous tests which are used to determine
the various mechanical and chemical properties of metals. While some of these
tests provide us with values for more than one property, most are designed to
determine the value for a specific characteristic of the metal. Therefore, it may be
necessary to perform several different tests to determine all of the desired information.
It is important that the welding inspector understand each of these tests at least to
the extent that he knows when the test is applicale, what results it will provide,
and how to determine if the test results are in compliance with the specification
re!uirements. It may also be helpful if the welding inspector understands some of the
methods used in the testing, even though he is not directly involved with the actual
testing.
"s a group, these tests are referred to as destructive tests. They bear this name, as
you would e!pect, because they render the material useless for service once the test has
been performed. In general, they destroy, or fail, the material to learn how it behaves in
the course of resisting that failure. Some of these tests are referred to as destructive
tests, even though they don not ruin the part to the extent that it cannot e used
#e.g. harness testing.$ "owever, that depends on the shape, size and re#uired surface
condition of the part.
Throughout this discussion there will be little mention made as to whether a specific
destructive test is being used to determine a base metal or weld metal property. $or the
most part, this does not represent a significant change in the manner in which the test is
performed. There will be occasions when a test is performed to specifically test the base
metal or weld metal, but the mechanics of the testing operation will vary little, if any.
%
s we discussed these various methods you should try to relate the actual test method
with the appropriate material property&ies'. In that way, it will be easier for you to
understand and relate the values determined from the test with the properties which they
describe. To aid in this understanding, the various tests will be discussed in essentially
the same order that the properties which determine were presented in Module 5.
Tensile Testing
The first material property discussed was strength, so our first destructive test method
will appropriately be the tensile test. This one test provides us with the greatest
amount of information aout a metal of all the tests which we%ll discuss. Some of
the properties that can e determined as the result of the tensile test include&
'ltimate Tensile Strength
(ield Strength
Ductility
)ercent *longation
)ercent +eduction of "rea
Modulus of *lasticity &(oung)s Modulus'
)roportional ,imit
*lastic ,imit
Toughness
*ome of these values can be determined through a direct reading of a gage, while others
can be #uantified only after analysis of the stress+strain diagram which is produced
during the test. The values for ductility can e found y making comparative
measurements of the tensile specimen efore and after testing. The percentage of
that difference then descries the amount of ductility present.
,hen we perform a tensile test, one of the most important aspects of that test involves
the preparation of the tensile specimen, or coupon. If this part of the test operation is
conducted carelessly, the validity of the test results will be severely reduced. Small
imperfections in the surface finish, for example, can result in significant reductions
in the apparent strength and ductility of the tensile specimen.
$igure % shows an e!ample of a machine used for tensile tests.
-
-igure .
*ometimes, the sole purpose of the tensile test is to simply show if the weld zone will
perform as well as the base metal. $or such an evaluation, all that is necessary is to
remove a specimen &sometimes referred to as a strap' transverse to the longitudinal a!is
of the weld having the weld roughly centered in the specimen. The two cut sides should
be as parallel as possible using a saw or cutting flame, but no further surface treatment
is essential, including the removal of any e!isting weld reinforcement.
$igure - shows e!amples of transverse and longitudinal tensile locations.
-igure /
.
This approach is used for procedure #ualification and welder #ualification testing in
accordance with /I %%01. " successful tensile test is descried as a specimen which
fails either in the ase metal, or in the weld metal as long as it fails at some
strength level aove the specified ase metal strength.
$or most cases in which the tensile test is re#uired, there is a need to determine the
actual amounts of strength and other properties e!hibited by that material, not 2ust if the
weld is as strong as the base metal. ,hen the determination of these values is
necessary, out specimen 2ust be prepared in some configuration which provides a
reduced section somewhere near the center of the length of the specimen. This reduced
section is intended to locali0e the failure. Otherwise the failure might tend to occur
preferentially near the grips, making the measurements for percent elongation e!tremely
difficult. lso, this reduced section results in the increased uniformity of the stresses
throughout the cross section of the specimen. The reduced section must exhiit the
following three features in order that valid results can e otained&
.. The entire length of the reduced section must e of a uniform cross
section.
/. The cross section should e of some configuration which can e easily
measured so that a cross sectional area can e calculated.
.. The surface of the reduced section should e free of surface
irregularities, especially of they lie perpendicular to the longitudinal
axis of the specimen.
$or these reasons, as well as considering the actual mechanics of preparing a specimen,
the two most common cross sectional configurations for tensile specimens are
circular and rectangular. 3oth are readily prepared and measured. $igure . shows
both configurations.
1
-igure 1
5
If re#uired to actually perform a tensile test, the welding inspector may have to be able
to calculate the actual cross sectional area of the reduced section of the tensile
specimen. The e!amples below show hot these calculations are done for both common
cross sections.
*xample .& "rea of a circular cross section
The area of a circle is best remembered as & pie 4r5 s#uared' r
-
.
6emember the radius of a circle is one+half the diameter.
The diameter of a circle is .500 in.
The radius would be the diameter divided by - 7 .500 8 - 7 .-50.
*o, the area is e#ual to r
-
. &use ..%1 for '
rea 7 ..%1 9 .0:-5 7 .%;:- in
-
. &s#uare inches'
*xample /& "rea of a rectangular cross section

The area of any rectangle is it)s width times the thickness.
The width of a test piece is .;00 in.
The thickness of the test piece is .<50 in.
rea 7 , ! T
rea 7 .:<5 in
-
&s#uare inches'
The determination of this cross sectional area before performing the tensile test is
essential to finally determining the strength of the test piece. Tensile strength is
calculated y dividing the applied load y the original cross sectional area. 2r, the
pounds per s!uare inch used to reak the piece divided y the original cross
sectional area of the piece is e!ual to the tensile strength.
:
*xample 1& 3alculation of tensile strength
=oad 7 %-.500 pounds per s#uare inch &psi'
rea 7 0.%;:- &taken from e!ample %'
%-,500 8 .%;:- 7 :.,<%0 psi.
If we go back to $igure . we see that for circular tensile specimens there are some
standard sizes. These sizes were fi!ed so as to give a cross sectional area of 0.- in
-
.
This was very convenient since dividing some numer y 4./ is the same as
multiplying that same numer y 5. s an e!ample>
%-,500 8 0.- 7 :-,500
%-,500 ! 5 7 :-,500
If a standard tensile specimen was used the tensile strength could be obtained by
multiplying the load used to break the specimen by 5. The use of the standard sized
specimens was most popular years ago before the invention of the calculator. It was
easier to accurately machine a tensile specimen to this e!act size and do the
multiplication then to do the more complicated math. Today it is easy to calculate the
e!act tensile strength no matter what the actual area happens to be.
nother operation which must be performed prior to testing is to accurately mark a gage
length on the reduced section. This gage length is normally marked using a pair of
center punches held at some prescribed distance apart. The most common gage
lengths are / and 6 inches. fter testing, the new distance between these marks is
compared to the original distance to enable a calculation for the amount of elongation or
stretch, undergone by the specimen when stressed.
3efore applying the tensile load, sometimes, a devise known as an e!tensometer is
connected to the specimen at the gage length marks. ?uring the application of the load,
the e!tensometer will measure the amount of elongation which results from the load
being applied. 3oth the load and elongation data are fed into a strip chart recorder to
result in a plot of the variation in the elongation as a function of the applied load. This
is referred to as a load versus deflection curve. 7ormally tensile tests results are
expressed in terms of stress and strain. &stress strain curve'
<
Stress is e!uivalent to strength, since it is the applied load at any time divided y
the cross sectional area. The strain is simply the amount of stretch apparent in a
given length.
=ooking at the stress+strain diagram &$igure 1' there are several important features to be
noted.
-igure 8
s the load is applied, the amount of strain increases linearly with stress. This is what
is called elastic behavior, where the stress and strain are proportional. On the chart the
line is straight and at about a 15 degree angle. This slope is referred to as the modulus
of elasticity or (oung%s modulus.
$or steel, the modulus of elasticity is appro!imately e#ual to .0,000 psi as compared to
%0,500 for aluminum. What this numer represents is the stiffness of the metal.
The higher the modulus of elasticity the stiffer the metal.
@ventually the strain will begin to increase faster than the stress, meaning that the metal
is stretching more for a given amount of applied stress. This change marks the end of
elastic ehavior and the onset of plastic, or permanent deformation. The point on
the curve showing the extent of the linear ehavior is referred to as the elastic, or
proportional limit. If the load has been removed at any time up to this point the
specimen would have returned to it)s original length. Aow, there is some permanent
deformation even if the load was released.

Many metals tend to e!hibit a drastic change from elastic behavior to plastic. The stress
may actually drop or remain steady while the strain increases. This characteristic is
known as yielding in ductile steel. In, metals showing this ehavior the yield point is
noted and that level of stress is referred to as the yield strength. ?uring a tensile
B
test, the yield point can be seen as a drop in the gage or recording device. The yield
point can be determined by simply observing and noting this load reduction. ,hen this
method is utilized, we refer to it as the 4drop+of+beam5 techni#ue.
?uring this yielding phenomenon, the plastic flow of the metal is increasing at such a
rate that stresses are being relieved faster than they are formed. When this plastic flow
occurs at room temperature, we refer to it as cold working. This action caused the
metal to ecome stronger and harder and it is said to e work hardened. The
yielding will therefore continue until the metal becomes work hardened to the e!tent
that it now re#uires additional stress to produce any further elongation and the curve
begins to climb in a non+linear fashion.
The stress and strain continue to increase at varying rates until some ma!imum stress is
reached. This point is the maximum or ultimate tensile strength. &Tensile strength'
The final information which can e gotten from a stress9strain diagram is the
amount of toughness exhiited y the metal. +ememer, the toughness is a
measure of a metal%s aility to asor energy. Toughness is defined y the area
under the stress9strain diagram, so, a metal having higher values for stress and
strain is considered to e tougher. $igure 5 shows a comparison between the stress+
strain diagrams of two different types of steel.
-igure 5
If you compare the areas under the two curves, it is evident that the area under the
structural steel is larger primarily due to the greater overall elongation even though the
spring steel e!hibits a higher tensile strength. The structural steel is therefore a tougher
metal.
;
Once a tensile test is complete it is now necessary to make a determination of the
metal%s ductility. This is done in one of two ways, both of which involve making
measurements both before and after the testing. 6emember, there are two ways in
which ductility is expressed, percent elongation and percent reduction of area.
To determine the percent elongation, it is necessary to have placed gage marks on the
specimen before pulling it. "fter the specimen has failed, the two pieces are placed
together and the new distance etween the gage marks is measured. Cnowing this
information , plus the original gage length, it is possible to calculate the percent
elongation.
*xample 5& Determination of percent elongation
Original length 7 -.0 inches
$inal length 7 -.: inches.
/ercent elongation 7 &final length + original length' 8 original length ! %00
D elongation 7 &-.: + -.0' 8 -.0 ! %00
D elongation 7 0.: 8 - ! %00
D elongation 7 .. ! %00
D elongation 7 .0D
(ou can also determine a metal)s ductility by how much it necks down during the
tensile test. This is referred to as percent reduction of area, where the original and final
areas of the tensile specimen are measured and compared.
*xample :& Determination of percent reduction of area
Original area 7 0.- in
-
$inal area 7 0.% in
-
/ercent reduction of area 7 &original area + final area' 8 original area ! %00
D reduction of area 7 & 0.- + 0.% ' 8 0.- ! %00
D reduction of area 7 0.% 8 0.- ! %00
D reduction of area 7 .5 ! %00
%0
D reduction of area 7 50 D
,hile both percent elongation and percent reduction of area represent e!pressions
for the amount of ductility e!hibited by a tensile specimen, their values will seldom, if
ever, e e!ual. )ercent reduction of area will e approximately twice the value for
percent elongation. )ercent reduction of area is thought to e a representation for
determining the ductility of a metal in the presence of some notch. Most often
percent elongation will e the most used method for determining ductility.
;ardness Testing
,e have talked already about the relationship between hardness and the strength of a
metal. "ardness tests is most often done to simply give us a measurement of the
metal)s hardness but, there is a direct connection between hardness and strength.
;ardness is the aility of a metal to resist penetration or indentation. "ardness
tests are performed , for the most part, using some type of penetrator which is forced
against the surface of the test ob2ect. ?epending upon the type of hardness test being
used, you then measure either the diameter or depth of the resulting indentation.

lthough there are many methods with which to measure a metal)s hardness this
discussion will concentrate on three basic groups of hardness tests> <rinell, +ockwell
and microhardness. The three groups differ from one another in the si0e of the
indentation which is produced, with <rinell eing largest and microhardness
smallest.
<rinell testing is commonly used for determining the hardness of metal stock. It is
well suited for this purpose ecause the indentation covers a relatively large area,
eliminating prolems associated with locali0ed hard or soft spots in the metal. The
higher loads used for 3rinell tests also assist in reducing errors produced by surface
irregularities. 3efore 3rinell tests can be performed it is necessary to prepare the
surface of the test piece so that test error will be minimized. This includes grinding or
sanding the surface to achieve a relatively flat test area. The surface has to be smooth
enough so as not to interfere with accurate measuring of the indentations. To perform
the 3rinell test a penetrator is forced into the surface of the test ob2ect at some
prescribed load. Once this load is removed, the diameter of the indentation is then
measured using some graduated magnifier. 3ased on the size and type of the indenter,
the applied load, and the resulting diameter of the impression, a 3rinell "ardness
Aumber &3"A' can be determined. *ince this is a mathematical relationship, we can
determine a 3"A for a variety of indenter types and loads. The <;7 is related to the
%%
actual tensile strength of the metal, in fact, the <;7 times 544 is approximately
e!ual to the metal%s tensile strength. *ome test conditions, such as specimen hardness
and thickness, will re#uire variations in both the type of bell and the amount of the
applied load. Other types of balls which can be used include the 5mm hardened steel
ball and the %0mm tungsten carbide ball. $or soft metals, loads as low as 500kg may be
used. Other loads between 500 and .000kg can also be used with e#ually good results.
Ese of different loads and balls is possible because the 3"A is simply an e!pression for
the load divided by the area of the impression. The 3"A can be calculated for any
combination of test parameters using the following formula>
3"A 7 /
&? + ?
-
F d
-
' ?8-
,here> p 7 load, kg
? 7 diameter of ball, mm
? 7 diameter of impression, mm
,e normally determine the 3"A by simply measuring the diameter of the impression
and reading the value from some table which has results already calculated, as in the
table below.
%-
The steps for 3rinell testing are as follows>
.. )repare test surface.
/. "pply test load.
1. ;old at load for prescried time.
8. Measure impression diameter.
5. Determine <;7 from tale or direct calculations.
2ne important feature to note in the aove procedure is the re!uirement to hold at
the test load for some specified time. -or iron and steel, this will e .4 to .5
seconds. Softer metals re!uire holding times of approximately 14 seconds.
It is evident from this simple procedure how easily the 3rinell test can be applied.
Often, there is a need for the testing of ob2ects too large to be placed in a test machine.
In such cases, we can use a portable test machine. There are a number of types and
configurations, but the same test principals are used.
The ne!t group of hardness tests to be discussed will be the 6ockwell tests. Gust as with
the 3rinell test, we can modify the basic test by using different indenters as well as
different test loads. +ockwell tests result in smaller indentations than for <rinell
testing. This allows for the localized testing of relatively small areas of a metal.
=ike the 3rinell test, +ockwell testing uses different indenters for different hardness
ranges. The indenters utili0ed are the diamond <rale, .=.:>, .=6>, !uarter inch,
and one9half inch diameter hardened steel alls. Esing one or the other of these
indenters, various loads can also be utilized. The loads are much lower than those used
for 3inell testing, ranging from :0 to %50kg.
There are also a group of 6ockwell tests designated as superficial. These are primarily
used to determine the hardness of thin metals, therefore the loads are much lower than
those for other types of 6ockwell tests.

Gust as with 3rinell testing, the test surface must be properly prepared prior to applying
a 6ockwell test. Then, the proper scales must be selected based on the appro!omate
range of hardness e!pected. The ?<> and ?3> scales are the most commonly used
scales for steel, with the ?<> scale chosen for softer alloys and the ?3> scale used
for the harder types. Then in dout as to which scale might e chosen for some
unknown alloy the ?"> scale could e used ecause it includes a range of hardness
covering oth the ?<> and ?3> scales.
%.
Once the proper scale has been selected, the test ob2ect is placed on the anvil in the
6ockwell testing machine. The anvil can be of various shapes depending upon the
configuration of the test piece. The ob2ect must be ade#uately supported, otherwise, test
errors will result. If the specimen moves during testing inaccurate readings will
result. " variation in the depth measurement of only 4.44446 inch will result in a
change of one +ockwell numer. The basic steps in 6ockwell testing are as follows>
.. )repare the test surface.
/. )lace the test specimen in the +ockwell tester.
1. "pply the minor load using the elevating screw.
8. "pply the ma@or load.
5. +elease the ma@or load.
:. +ead the dial.
A. +elease the minor load and remove the part.
The minor load is used simply to take the slack out of the system, improving the
accuracy of the test. $igure : shows each of the steps 2ust discussed.
-igure :
%1
The result gained from the +ockwell test can then e related to <rinell values and
therefore the tensile strength of the metal. $igure< shows how 3rinell, 6ockwell and
tensile strength values are related.
-igure A
=ike 3rinell testing, there are also portable devices which can be utilized to determine
the 6ockwell hardness of a metal. lthough their operation may vary slightly from that
of the stationary models, the results will be e#uivalent.
%5
The next group of hardness tests to e discussed are those which are referred to as
microhardness tests. They bear this name because their impressions are so small that
high magnification is re!uired to facilitate their measurement. ,e find
microhardness testing very beneficial in the investigation of metal microstructures,
because these harness tests can be performed on single grains of metal to determine the
hardness in that microscopic region. Therefore, the metallurgist is primarily interested
in this type of harness testing.
There are two ma@or types of microhardness tests& Bickers and Cnoop. 3oth utilize
diamond indenters, but their configurations are slightly different. ?iagrams of the two
types as well as the resulting indentations are shown in $igure B.
-igure 6
s can be seen, the s#uare+based Hickers indenter provides an indentation in which the
two diagonals are appro!imately e#ual. The Cnoop indenter, however, makes an
indentation having a long and a short dimension. $igure shows comparative
indentations made with the two types of indenters, using identical metals and loads.
s with the other test methods, you have a selection of test loads as well as indenter
types. The term microhardness implies that the applied loads will range from % to %000
g. "owever, the ma2ority of microhardness tests utilize loads in the range of %00 to 500
g.
To perform either Bickers or Cnoop microharness testing, the preparation of the
surface is of utmost importance. *ince we are looking at the harness of minute areas,
%:
even the smallest surface irregularity could cause inaccuracies. Aormally, we will
prepare the surface for microharness testing 2ust as we would for other metallographic
investigations. The importance of this surface finish increases as the applied test load is
reduced. The surface finish is brought to a polish.
Once prepared, the specimen is securely clamped in some type of test fi!ture or holder
so that the indentations can be accurately placed. Many microhardness machines
employ some type of moving stage which facilitates accurate movement of the
specimen without the need for its removal and ad2ustment. *uch a device is handy when
taking a number of readings across some region of the metal. n e!ample of this type of
application would be determination of the hardness variation across the weld heat
affected zone. The result would be referred to as a microharness traverse.
The steps used in the microharness testing of some ob2ect would be as follows>
.$ )repare test surface.
/$ )lace specimen in holding fixture.
1$ ,ocate area of interest, using microscope.
8$ Make indentation.
5$ Measure indentation using microscope.
:$ Determine harness using tale or calculation.
The use of harness testing will provide us with a great deal of useful information about
a metal. "owever, as can be seen from the preceding discussion, we must specify which
harness method will be used for a given application. If not, the desired results may not
be obtained.
Toughness Testing
nother metal property of interest is toughness. (ou have learned that this property
descries the material%s aility to asor energy. ,hen tensile testing was discussed,
you learned that the toughness of a metal could be described as the area under the
stress+strain curve. This is a value for the amount of energy that can be absorbed by a
metal when the load is applied gradually.
"owever, in the discussion of toughness in Module 5, you remember that when the load
is applied rapidly, the concern is with notch toughness, or impact strength. The
discussion which follows will concern itself with tests which can be used to determine
this particular property. Therefore, the various tests used to determine the notch
toughness of a metal will utilize a specimen containing some type of machined notch
%<
and the load will be applied in a very rapid manner. (ou further recall that the
temperature of the specimen has a drastic effect on the test results, so the testing must
be performed at some prescribed temperature.
*ince the advent of interest in the notch toughness of metals, numerous different tests
have been developed to measure this important property. When you talk aout the
energy asortion capailities of a metal, you must understand that the metal
asors energy in steps. -irst, there is a certain amount of energy re!uired to
initiate a crack. Then, some additional energy is re!uired to cause that crack to
grow, or propagate.
*ome of the notch toughness tests can measure the energy of propagation separate from
the energy of initiation while other methods simply provide us with a measure of the
combined energy of initiation and propagation. It will be up to the engineer to specify
that test method which will provide him wit the desired information.
"lthough numerous types of notch toughness tests exist, proaly the most
commonly used in this country is the 3harpy B9notch test. The standard specimen
used for this test is a ar 55 mm long and .4 mm s!uare. One of the long sides of the
specimen has a carefully machined H+shaped notch 0.0<; inch or -mm deep. t the base
of this notch, there is a radius of precisely 0.0%0 inch or 0.-5 mm. The machining of this
radius is e!tremely critical, since tiny inconsistencies will result in drastic variations in
test results. sketch of a standard Iharpy specimen is shown in $igure ;.
lso there are configurations for other subsize Iharpy specimens which may be
employed. Aormally, the results from these subsize specimens are corrected based on
their sizes compared to the standard size specimen.
%B
-igure D
3harpy impact specimens
Once the specimen hs been carefully machined, it is then cooled to the prescribed test
temperature, if it is some temperature below room temperature. This can be
accomplished using either li#uid or gaseous medium. fter the specimen is stabilized at
the re#uired temperature, it is then put into the testing machine. The machine consists
essentially of a pendulum with a striker head, an anvil, a release lever, a pointer and
scale. *ince we are essentially attempting to measure at the amount of energy which is
absorbed during the fracturing of a specimen, we supply a given amount of energy by
raising the pendulum to a specified height. Epon release, the pendulum will fall and
continue through its stroke until it reaches a ma!imum height of this swing is indicated
by the pointer on the scale. *ince this scale is calibrated, we can read the amount of
energy re#uired to break indicated by the pointer on the scale, directly from the scale.
This value, referred to as breaking energy, is the primary piece of information gained
from the Iharpy impact test. This energy is e!pressed in terms of foot+pounds of
energy. ,hile most Iharpy results are e!pressed in terms of foot+pounds of energy
absorption, there are other means of descriing the notch features of the failed
3harpy specimen. These values are& lateral expansion and percent shear. ,ateral
expansion is a measure of the amount of lateral deformation produced during the
fracturing of the specimen. It is measured in terms of mils, or thousandths of an inch.
)ercent shear is an expression for the amount of the fracture surface which failed
in a ductile, or shearing, fashion.
Ao matter which of these methods of measurement is employed, we are usually
concerned with the results from a whole series of tests. Once we have tested a number
of specimens at various temperatures, we can determine how the values change with
%;
temperature. If we plot these values versus temperature, we will get curves which have
upper and lower horizontal shelves. $or each category, there is some temperature at
which the values drop rather aruptly. These temperatures are referred to as
transition temperatures, which means that the ehavior of the metal changes from
relatively ductile to rittle at that temperature. The designer then knows that the
metal should behave satisfactorily above that temperature.
In addition to the Iharpy test are others which can be employed for various
applications. 2ther tests which may e weight nil9ductility, explosion ulge,
dynamic tear, and crack tip opening displacement #3T2D$. These tests employ
different types of specimens as well as various ways of applying the load to the
specimen.
Soundness Testing
This group of tests are designed to aid in the determination of the metal%s soundness,
or its freedom from imperfections. *oundness tests are routinely used for the
#ualification of welding procedures and welders. fter a test plate has been welded,
specimens are removed and then sub2ected to some soundness test to determine if the
weld contained any imperfections or defects.
There are three general types of soundness tests& end, nick9reak and fillet reak.
The first type, bend testing, can be performed in a number of different ways. This is
probably the most common test used to 2udge the ade#uacy of a welder)s #ualification
test coupon.
There are several different types of bend tests, depending on the orientation of the weld
with respect to the bending action. There are three types of transverse weld end
specimens& face, root and side ends. ,ith these three types, the weld lies across the
longitudinal a!is of the specimen and its name refers to the side of the weld which is
placed in tension during the test. That is, the face of the weld is stretched in a face bend,
the root of the weld is stretched in a root bend, and the side of a cross section of the
weld is stretched in a side bend. $igure %0 shows sketches of these three bend
specimens.
-0
-igure .4
<end tests are normally performed using some type of end @ig. There are three
asic types& guided9end, roller9e!uipped guided end, and wraparound guided9
end. The standard guided+bend test 2ig, show in $igure %%, consists of a plunger &also
referred to as mandrel or ram' and a matching die which form the previously straight
bend specimen into a E+shape.
To perform a bend test, the specimen is placed across the shoulders of the die with the
side to be placed in tension facing down toward the inside of the die. The plunger is
then situated over the area of interest and forced toward the die causing the specimen to
be bent %B0J and become E+shaped. The specimen is then removed and evaluated.
-%
-igure ..
The second type of guided+bend test 2ig, shown in $igure %-, is similar to the standard
guided+bend 2ig, e!cept it is e#uipped with rollers instead of hardened shoulders on the
die portion. This reduces the friction against the specimen allowing for lower loads to
achieve the bending. The last common type of guided+bend 2ig bend is referred to as the
wraparound 2ig. It bears this name because the specimen is bent by being wrapped
around a stationary pin, as shown in $igure %.. This type is especially useful for
bending specimens having different strengths of base and weld metal. If there is a great
imbalance, there will be a tendency to bend preferentially in the weaker metal, resulting
in a kink or a bend away from the area of interest.
--
-igure ./
-igure .1
-.
Most !ualification tests for mild steel re!uire that the specimen e ent around a
mandrel having a diameter four times the thickness of the specimen. Therefore, a
.8B inch specimen would be bent around a % K inch diameter mandrel. This results in a
/4E elongation of the outer surface of the end specimen. If a smaller bend mandrel
is used, the necessary amount of elongation would increase.
In !ualification test coupons where the weld metal is much stronger than the ase
metal, there may e a tendency for the specimen to kink in the ase metal next to
the weld rather than forming smoothly around the mandrel. If a wraparound @ig is
not availale, it may e eneficial to select longitudinal end specimens rather
than the standard transverse types. The weld lies in line with the longitudinal a!is of
the longitudinal bend specimen. This is shown in $igure %1 .
-igure .8
,ith any of these bend tests, the specimens must be carefully prepared to prevent test
inaccuracies. "ny grinding or sanding marks on the tension surface should e
running transverse to the direction of ending so they don%t provide stress raisers
which could cause the specimen to fail prematurely. The corners of the specimen
should e radiused to relieve that stress concentration as well. $or specimens
removed from pipe coupons, the side of the bend specimen against the ram may be
re#uired to be ground flat to eliminate the bending in the direction transverse to the
bending direction.
The acceptability of bend test specimens is normally 2udged based on the size and8or
number of defects which appear on the tension surface. The governing code or
specification will dictate the e!act acceptance8re2ection criteria.
-1
The next type of soundness test to e discussed is the nick9reak test. This test is
used almost exclusively y the pipeline industry as descried in ")I ..48. This
method 2udges the soundness of the weld by fracturing the specimen through the weld
so that the fracture surface can be analyzed. typical nick+break specimen is shown in
$igure %5.
-igure .5
Once the specimen has been saw cut, it is then broken by pulling in a tensile machine,
hitting the center with a hammer while supporting the ends or hitting one end with a
hammer while the other end is held in a vise. The method of reaking is not
significant ecause you are not interested in how much effort is re!uired to fail the
specimen. (ou are simply trying to fail the specimen through the weld zone so you can
determine if any imperfections are present. (ou then investigate the fracture surface for
any areas of slag inclusions, porosity or incomplete fusion. If present, they are measured
and accepted or re2ected based on the code limitations. The re#uirements for /I %%01
are shown graphically in $igure %:.
-5

-igure .:
The last soundness test to e mentioned here is the fillet weld reak test. =ike the
other two types, this soundness test is used primarily in the #ualification of welders.
This is the only test re!uired for the !ualification of tackers in accordance with
"WS D....
Once welded, the specimen is broken by loading it against the face of the fillet weld.
With this test, the inspector is looking for a weld having satisfactory surface
appearance. -urther, he inspects the fracture surface to assure that the weld has
evidence of fusion to the root and that there are no areas of incomplete fusion to
the ase metal or porosity larger than 1=1/ inch in their greatest dimension.
These soundness tests are used routinely in many different industries. Their application
and evaluation appear to be #uite straight+forward. "owever, the welding inspector
should be aware that the evaluation of these tests may not be as simple as the various
codes and specifications might imply. $or this reason, it is important that the welding
inspector actually spend some time performing these tests to become familiar with their
performance and interpretation.
-atigue Testing
The last mechanical test method is to be discussed is fatigue testing. This is a type of
test which enales us to determine the fatigue strength of a metal. -atigue, as you
recall, is the cyclic loading of a memer. $atigue tests help designers determine how
well a metal will resist failure when loaded in fatigue. Aormally a series of fatigue tests
are run to arrive at the endurance limit for that metal. So tests are run at various stress
-:
levels until some maximum stress is found elow which the metal should exhiit
infinite fatigue life.

Since fatigue is so strongly influenced y surface finish and configuration, the
preparation of fatigue specimens is extremely critical. 2nly minor lemishes are
necessary to cause drastic changes in the results. *o, if sufficient care is not taken at
this stage, all final results could be invalid.
$atigue tests can be performed in a number of different ways. The specific test to be
used depends upon the e!pected loading of the metal in service. That loading could
e& planar ending, rotational ending, torsion, axial tension, axial compression,
or cominations of these. ,hen loading in the a!ial, or longitudinal direction, the
cycles could be such that the specimen is loaded in alternating tension and compression.
This is the most severe case.
Destructive Tests for 3hemical )roperties
The tests, which have been previously discussed, are used to determine the mechanical
properties of a metal. s you found out in Module 5, there are also important chemical
properties of a metal. In fact, the chemical makeup of a metal determines to a great
degree the mechanical properties of that metal. Therefore, we often have the need
to determine the chemical composition of a metal. The three most common
methods are& spectrographic, comustion, and wet chemical analyses.
The welding inspector will rarely be re#uired to actually perform chemical analyses.
"owever, he may have to review the results of these analyses to determine if a metal
complies with a particular specification. $or more information regarding chemical
analysis of metals, refer to the *TM specifications in which cover this sub2ect. The
particular methods utilized for steel are listed in *TM <5%.
"nother group of tests which can generally e classified as chemical tests are
corrosion tests. These are specific tests designed to determine the corrosion resistance
of a metal or combination of metals. =osses from corrosion resistance of metals cost
industry billions of dollars annually. Therefore, we are very concerned about how a
metal will behave in a particular corrosive environment. The tests used to determine the
degree of corrosion resistance are designed to simulate as closely as possible the actual
conditions which the metal will encounter during its service. Some of the
considerations which must e addressed when setting up a corrosion test are&
chemical composition, corrosive environment, temperature, presence of moisture,
-<
presence of oxygen, presence of other metals, and amount of stress. If any of these
features are ignored, the corrosion test may yield invalid results.
Metallographic Testing
nother way in which we can learn about the characteristics of a metal or a weld is
through the use of various metallographic tests. These tests asically consist of
removing a section of a metal or a weld and polishing it to some degree. Once
prepared, the specimen can then be evaluated with the unaide eye or with the use of
magnification.
Metallographic testing is generally classified as either macroscopic or microscopic.
They differ in the amount of magnification that is utilized. Macro tests are generally
performed using magnifications of .4F or lower. Micro tests, on the other hand,
utili0e magnifications greater than .4F.
" numer of different features can e oserved on a typical macro specimen. ,e
often cross section a weld to provide us with a macro specimen from which we can
determine such things as> depth of fusion, depth of penetration, effective throat,
weld soundness, degree of fusion, presence of weld discontinuities, weld
configuration, numer of weld passes, etc. " picture of a macro specimen is
referred to as a photomacrograph. *ome typical photomacrographs are shown in
$igure %<.
-igure .A
-B
Micro tests can e used to determine various features as well. Included are&
microstructural constituents, presence of inclusions, presence of microscopic
defects, nature of cracking, etc. Similarly pictures of micro specimens are called
photomicrograhs. $igure %B shows some typical photomicrographs.
-igure .6
These various metallographic tests can be very helpful in such matters as failure
analysis, weld procedure and welder #ualification, and process control testing.
The two types of specimens also differ in the amount of preparation re#uired. Some
macro specimens need only e rough ground, whereas micro specimens re!uire
fine grinding and even polishing to produce a mirror finish.
-;
There is so much information which can be gained about the properties of a metal by
making simple macro and micro evaluations that metallographic testing is an important
tool for both the welding inspector and the engineer.
Metric 3onversions
In recent years, there has been an effort put froth to convert our merican sLstem of
measurement to an international system referred to as the International System of
'nits &in $rench, 4=e *ysteme Interantionale d)Enites' or SI. Ionse#uently, the
welding inspector may have occasion to inspect pats or interpret specifications having
these *I units. "e may then be asked to convert those units into more familiar units.
This is part of the 2ob math which the welding inspector may be re#uired to perform. s
a minimum, he will be asked to perform. s a minimum, he will be asked to perform
some of these conversions on the ,* I,I e!amination. The following discussion is
meant to provide a basis for the welding inspector to begin to understand how these
conversions are done.
3efore actually looking at the conversion process, it is important to understand some of
the notations which will be used to e!press different numeric values. One of the
techni#ues used to e!press numbers which may be very large or very small, is called
scientific notation. ,hat this method does is reduce some number to an e!pression
which includes a number multiplied by some power of ten. $or e!ample, the number
-.1,5:< would be e!pressed as -..15:< 9 %0
5
in scientific notation.
(ou probably noticed that the same digits are used but the decimal place has been
moved. The decimal place is always moved to where there is only one number
appearing to its left. The number of spaces which the decimal place was moved
becomes the power of ten in the scientific notation e!pression. If the decimal point was
moved in the opposite direction, as would be the case for a number less than one, then
the power of ten becomes a negative number. The e!amples below show how scientific
notation is used.
Scientific 7otation *xamples
-.1 7 -..1 ! %0
-
0.0-.1 7 -..1 ! %0
+-
5.:<B ! %0
.
7 5,:<B
5.:< ! %0
+1
7 0.0005:<
.0
$rom this e!ercise, it is evident that movements of the decimal point one space is
e!uivalent to multiplying or dividing y ten, depending on the direction in which it
is moved. )owers of ten that are positive result in moving the decimal point to the
right. )owers of ten that are negative result in moving the decimal point to the
left.
nother type of notation with which you should become familiar are the various
prefi!es which are used to indicate powers of ten. These are shown in $igure %;.
-igure .D
These are simply abbreviations to reduce the number of digits re#uired. s an e!ample,
4kilo5 means %000, so a kilometer is %000 meters. *imilarly, 4milli5 means one+
thousandth, so a millimeter is one+thousandth of a meter or there are %000 millimeters
in one meter. *ome e!amples of the use of these prefi!es appear below.
*xamples of )refixes for )owers of Ten
15:,000,000 /a 7 15: Mpa
5: km 7 5:,000 m
-.1,000 mm 7 -.1 m
15: g 7 0.15: kg
,ith this background, we can now begin our discussion of how to actually perform
numeric conversions from E* to *I and *I to E*. The initial point to understand is that
you are not intended to memorize all of the conversion factors. @very factor which will
be needed for the ,* I,I e!amination will be provided to you. (ou must simply be
capable of manipulating the numbers to arrive at some solution. $igure -0 lists some of
the important conversion factors for units relating to welding.

.%
.-
..
-igure /4
=ooking at this table, there are a few important features. (ou see that the table is
arranged in four columns, entitled> 4/roperty,5 4To convert from,5 4To,5 and 4Multiply
by.5 (ou will use these columns in the same order as they are listed.
$or any conversion e!ercise, the first step is to decide what particular property is
described by the units which are to be converted. Once the proper category has been
chosen from the 4/roperty5 column, you now look at the second column &4To convert
from5' and locate the line which contains the unit that is given in the e!ercise. That is,
this is the unit which you are asked to convert from. Moving straight across this same
line to the right, look for the unit that matches the unit to which you are asked to
convert. ,hen you locate the line which contains both the know and desired units, the
value found in the last column &4Multiply by5' is the appropriate conversion factor. t
this point, you simply multiply the number of the known units by this conversion factor.
The result is the number of desired units. *everal e!amples appear below to show how
to use this table to perform typical conversions.
3onversion *xample .& n o!ygen gage shows a pressure of 10 psi. "ow
many pascals is thatM
%' /roperty 7 pressure &gas and li#uid'
-' Cnown unit 7 10 psi
.' ?esired unit 7 pascals
1' Ionversion factor 7 :B;5 &or :.B;5 ! %0
.
'
*olution> 10 psi ! :B;5 7 M pascals
10 psi 7 -<5,B00 /a &or 0.-<5B Mpa'
3onversion *xample /& tensile specimen was pulled and displayed and
ultimate tensile strength of :5.5 M/a. This corresponds to how many psiM
%' /roperty 7 tensile strength
-' Cnown unit 7 :55 M/a &:.55 ! %0
B
/a'
.' ?esired unit 7 psi
1' Ionversion factor 7 %.150 ! %0
+1
*olution> :50,000,000 /a ! 0.000%150 7 M psi
:50,000,000 /a 7 ;1,;<5 psi
.1
3onversion *xample 1& ,hat is the diameter in millimeters of a 58.- inch
electrodeM
%' /roperty 7 linear measurements
-' Cnown unit 7 58.- inch &0.%5: in'
.' ?esired unit 7 mm
1' Ionversion factor 7 -5.1
*olution> 0.%5: in ! -5.1 7 M mm
0.%5: in 7 ..;: mm
3onversion *xample 8& ,elding parameters were ad2usted to produce a weld
metal deposition rate of <.. kg8h. ,hat is that deposition rate in terms of lb8hM
%' /roperty 7 deposition rate
-' Cnown unit 7 <.. kg8h
.' ?esired unit 7 lb8h
1' Ionversion factor 7 -.-05
*olution> <.. kg8h ! -.-05 7 M lb8h
%.. kg8h 7 %:.% lb8h
The e!amples above are typical of problems which will appear on the ,* I,I e!amination.
Ao matter how large or small the numbers may be, the problems are all dealt with in the same
manner. *imply go through the various steps and utilize the table to provide you with some
multiplier. Then, all that is left to do is perform the arithmetic.
.5