Professional Documents
Culture Documents
PRODUCTION NOTE
University of Illinois at
Urbana-Champaign Library
Large-scale Digitization Project, 2007.
ABSTRACT
I. INTRODUCTION . . . . . . . . . . . . . 1
A. Use of High - Strength Bolts 1
B. Object and Scope of Investigation 2
C . Acknowledgments 3
II . DESCRIPTION OF TESTS 5
A . Relaxation Tests 5
B. Specimens for Tests with Sloping Surfaces 7
C . Fatigue Test Specimens 7
V. CONCLUSIONS 31
VI . REFERENCES CITED 34
FIGURES
21 . Bolt Te nsion at " Snug " of 5000 lb . Plus One- Half Turn for
A354 BD Rel a x ation Tests
A. USE OF HIGH-STRENGTH BOLT ASSEMBLIES and began using methods of tightening which
For a decade after its adoption, the are independent of any load-torque relation-
high-st rength structural bo lt assembly included ship. The reason for the change was the
a bolt with a finished hexagon head, a heavy desire for a simp ler, faster, and possibly
hexagon nut, and two hardened washers. Prima- more accura te means of tightening the bolt
ri ly, the washers served three purposes: assemblies. The objections sometimes offered
(a) to protect the outer surfaces of the con- to the use of tor~ue as a basis for bolt load
nected material from damage or galling as the control are: (a) a lack of dependability
bol t or nut was tor yued or turned; (b) to as- (because of differences in thread fit, thread
sist in maintaining the high clamping force in clean I iness, rust, etc.), (b) the possible
the bolt assembly; and (c) to provide surfaces resulting variations in bolt tension through-
of consistent hardness so that fairly depend- out a joint, and (c) inefficient use of pneu-
could be established for the various bolts and A preliminary study at the University
3
bolt diameters . of lllinois( ) and an investigation conducted
bolt assemblies made use of a relationship be- tightening high-strength bolt assemblies.
tween applied torque and bolt tension . How- Thi s method has provided a means of obtaining
ever, recently erectors and engineers developed at least the required minimum bolt tension
This report combines most of the data pre- nut method the bolt is tightened to, or above,
sented in tw9 previously unpublished progress
reports(1,2)T and other unpublished data the required minimum tension by rotating the
availab le on research with high-strength bolts.
nut or bolt a prescribed number of turns from
t Superscript numbers in parentheses refer to
references cited. either a finger-tight or a snug position.
2
The "finger-tight" position is that to which Three of the questions that must be
the nut can be tightened when turned by hand answered to determine whether one or both wash-
after the joint has been drawn tight with ers can be omitted from a bo lt assembly with-
fitting-up bolts; usually this has been fol - out any significant effect on its ability to
lowed by one complete turn (360 degrees) of form a sound and effective connection are as
the nut or bolt head. The "snug" position is follows: (1) Can the required clamping force
the point in the tightening at which a pneu- be developed and maintained without the hard-
matic impact wrench ceases to r un freely and ened washers when a high- strength bolt assembly
the impact mechanism begins to operate, or the connects several plies of structural materials?
tightness attained by the full effort of a man (2) Is the relaxation of tension in an assembly
using a n ordinary spud wrench. From the snug without washers the same as, or comparable to,
position the bolt or nut is usually given that for an assembly with washers? And (3) will
approximately a one-half turn (180 degrees) to the deformation and galling of the connected
achieve the fully tightened position. However, material which results from the tightening of
joints with wa rped, heavy, or many plies of the nut directly against that material cause
material may require that the bolts be touched the connection to be any less satisfactory,
bolt load in a ll the fasteners. A more de- To answer the first two questions
tailed discussion o f the turn-of-nut method as most of the types of nuts commonly used in
used by some f ab ricators was presented by Ba l l structural work were included in this study of
B. OBJECT AND SCOPE OF INVESTIGAT ION heavy semi-finished hexagon nut, the only type
Since the turn- of-nut method does permitted prior to 1960 for structural use;
not depend on torque to control the bolt ten- (b) a heavy- thin nut, which is the same dimen-
sion, the question was raised by structural sion across the flats as the heavy nut, but is
fabricators and erectors as to whether the only as thick as the finished-series nut;
hardened washers could be eliminated from the (c) a finished-series nut which was permitted
bolt assemb ly. By the same token, would only as an alternate* in the 1960 Specifications of
"k
one washer under the turned element be adequate Because of the development of the heavy hex-
agon head, structural A325 bolt and the result-
when the torque control or the calibrated ing lack of interest in finished nuts among
structural us~r~, the 1962 Research Counci I
wrench method is used? Specification~ 6 J mentions specifically only the
heavy hexagon nut.
3
the Research Council on Riveted and Bolted percent slope relative to the plane perpen-
Structural Joints; (d) a finished-thick nut dicular to the bolt axis. Variables in the
which has the same across-fl ats dimensions as tests included the use of washers, relative
a finished nut, but is as thick as a heavy nut, directions of sloping surfaces, position of
and (e) a flanged nut which has the same bolt, bolt hardness, method of tightening, and
a l so has an integral washer on one face. Also The third question to be answered
included in the tests were two types of bolt concerning elimination of washers involved
heads: the regular semi-finished hexagon head twenty-one fatigue tests. The variables em-
which has been specified in the past and a ployed were type of nut, type of bolt head,
heavy semi-finished hexagon head which was hole size, location of the critical plate
approved as an alternate in the 1960 Specifica- (inner or outer plates of a three plate spe-
tions of the Research Counci I and became the cimen), element tightened (the bolt or the
standard bolt in the 1962 Council Specifica- nut), bolt tension, and bolt diameter . These
tions. ( 6) fatigue specimens were four- bolt, double- lap,
The program included a large number shear- type joints similar to the one shown in
determination was made of the variation in bolt recently performed at the University of l lli-
tension with time using a strain gage load cell nois, as well as at other schools, the Speci -
within the grip as shown in Figure I. The fications{6) of the Research Council were
lubrication . In some of the relaxation tests, The tests described in this report
the assemblies were tightened by turn ing the are part of an investigation being conduct ed
bolt rather than the nut. as the result of a cooperative agreement be-
Other tests were performed with the tween the Engineering Experiment Station of
strain gage load cell and with a commercial the University of I llinois, the I llinois Divi-
hydraulic load cell to determine the turns sion of Highways, the Department of Commerce--
necessary to properly tighten the assembl ies Bureau of Public Roads, the Research Council
when one or both outer surfaces have a five on Riveted and Bolted Structural Joints, the
4
American Institute of Steel Construction, and Committee B of the Research Council on Riveted
the Industrial Fasteners Institute. The and Bolted Structural Joints. The chairman
investigation is a part of the Structural of the Committee was Mr. A. H. Smith, and its
Research Program of the Department of Civil members included E. R. Estes, Jr., N. G. Hansen,
we re performed under the general supervision T. W. Spi !man, 0. L. Tarlton, and G. S. Vincent.
Associate Professor of Civi 1 Engineering. and T. W. Spilman . The authors wish to thank
A number of the tests described in them for the ir careful analysis and criticisms.
;·~
A. RELAXATION TESTS i n ha rdne ss from a Rockwe ll B of 80 for t he
specimens cons i sted of two 4 x 4 x 3/4 -inch 95 for the heavy-thin nuts (Tab 1e 1, Co 1umn 6) .
steel pl ates (with a 13/16- inch diameter hole The ASTM- A354 bolts were specified to be grade
in the center, except in the ca se of the over- sot and were near Rockwell C = 37, the maximum
sized hole specimens), a 3/4 x 6-inch high- ha rdness for that grade (Ta ble 2) .
strength bo l t, and a load cell. The steel The load ce ll s (Figure 1) were
p la tes were ASTM-A7 structura l grade stee l with 1 11/16 inches in diameter and 3 1/8 inches
surfaces le ft in the a s-rolled cond i tion; how- long. These cells were designed to provide
ever, i n some ca ses the plates were rusted from average axi a l strains of 800 to 900 micro-
having been stored outside . Before the pl ates inches when subjected to the maxi mum load a nd
we re used ,the y were wi re- b rushed to remove any were made from AISI 434.) stee l, he at-treated to
loose sca le or rust. provide a high yiel d point to reduce the possi-
Da t a on the bolt and nut ha rdnesses bility of local yield ing. Ea ch load cell had
a re in Tables a nd 2. It is import ant to note eight stra i n gages (four placed axially and
tha t there were significant differences in the four ci r cumferent ia lly) arra nged in a fou r- a rm
ha rdnesse s and ult imat e strengths of the regu- br id ge for high sensitivity.
Ja r- and heavy- head bolts used in the relax- In the relaxation tests, the output
a tion tests for A325 bolts. The heavy-head
bolts (Rockwe ll C = 24) were on the low sides These nuts were obtained before the 19 58
changes in ASTM-A325.
of the ha rdness and strength ranges a ll~~ed by
5
6
of the load cell was recorded while the bolt the method of making the hole (either punched
was being tightened with a manual torque wrench or drilled), and the effect of lubricating the
and for five additional minutes to provide a contact surface between the face of the nut
continuous graphic load- time plot. For tests and the plate.
that were run more than 5 minutes, periodic The bolt-head tests included over-
strain readings were taken during the remai nder sized holes in some of the specimens, torquing
of the test. Measurements were also made of the head rather than the nut in four of the
the change in length of most bolts as a result tests, and regular as well as heavy heads.
of tightening. The effects of removing the Two specimens were tightened to a specific
washers from under the bolt heads and from under bolt tension and all other bolts were tightened
the nuts were studied separately. The members using a turn-of-nut method. For the latter
were designated as 11 nut study" or "bolt-head method, each bolt was tightened to a tension
study" specimens depending on which part of of 5 ,000 pounds to simulate the load in the
the bolt assembly was in contact with the mild bolt when it is snugged up; then the nut or
steel plates . The load cell was placed on the bolt head was given a one-half turn (180 de-
opposite side of the plates from the nut or gree rotation).
head and was always faced with two conventional In order to identify the variables
hardened washers. The one exception to the studied in the different relaxation tests,
assembly just described was the one used in the letters and numbers have been used to specify
relaxation test of the specimen with punched the test conditions for each of the specimens.
holes. In this case, the load cell was placed The letter "N" or "H" before the specimen
between the 3/4-inch plates to permit the bolt number indicates whether the test was a nut
head to bear on one lip and the nut to rotate test or bolt-head test, respectively. A
against the other lip of the punched hole; specimen number which includes the letter "S"
neither the bolt head nor the nut had a washer. identifies a short- time test (usually lasting
The variation in grip in the A354 bolt tests only 5 minutes after the bolt was tightened)
required some changes in the plate thickness as opposed to those tests which lasted from 3
The variables in the . nut tests in- specimen number indicates the use of over-sized
cluded, besides nut type, the hole size (the holes- - 11811 for the 27/32-inch diameter holes
customary 13/16- inch ho le and oversized holes a nd "C" for the 7/8-inch diameter holes. ln
of 27/32 and 7/8 inches to simu late reaming), those cases where a specimen was tightened to
7
a specific bolt load, that load in kips * appears would involve considerable time and expense
as the last figure in the specimen number. The and is much more extensive than could be in-
four specimens that were head torqued (tightened eluded in this investigation. However, a
by turning the bolt head} are identified by "HT" number of these variables have been explored
in the specimen number. One specimen was tested and are summarized in Table 3.
type between the nut face and the surface of the of a bolt and a suitable nut tightened against
plate; this specimen is identified by the letter sloping mild steel plate surfaces (mi II scale
"L" in the specimen number. The letter "X" surface, as-received) with some type of load
(Table 2) identifies the A354 BD bolts used in indicating device included in the grip -- either
the relaxation studies and the letters "D", "E", a steel load cell (cylindrical dynamometer as
and "F" indicate a variation in the number of in Figure I) or a commercial hydraulic load
A thorough study of the behavior of which provided a record of load, turns, and
assemblies tightened on sloping surfaces would app lied torque and also permitted an evaluation
number of factors. The following are of primary Accuracy of the Illinois load cells
importance: bolt head size and nut size was verified by recalibrations under axial
(distance across flats); grip and number of loadings, with eccentric loadings, and with a
threads in the grip; diameter of bolt and hole; check of the effects of residual torque. The
hardness or strength of bolts , nuts, plates and Skidmore-Wilhelm device was also calibrated
washers; position of the bolt in the hole -- under axial and eccentric loadings. Both kinds
centered, on high side of slope (maximum slid- of equipment showed excellent response under
washer-faced nuts, multiple plies, flats or The fatigue test specimens were
corners on the high side, and torquing head or designed to supplement the relaxation tests on
nut. The systematic study of all these factors A325 bolt assemblies and to determine the
Kip = 1000 pounds (kilo-pound) effects of galling when washers were omitted
8
from the fastener assembly. Specimens that and 5. Hole sizes and fastener diameters are
were designed with the center plates critical given in Figure 3. All the specimens were
(center plate more highly stressed than the double-lap, butt-type joints fastened with high
outside plates) served primarily to indicate strength bolts; all holes were match-dri ll ed.
whether the clamping force developed in the The bolt assemblies were tightened
bolts had been reduced when the washers were by turning the nut, except for Specimens 8 and
omitted; previous fatigue programs have shown 9 (Table 4) . The bolts in Fatigue Specimens
that the fatigue lives of bolted structural through 4, 6 through 8, and Series A were
joints decrease with a reduction in the initial installed by the turn-of-nut method using a
3
tension in the bolt.( ) Therefore, if there is pre-load or snug position of 5,000 pounds, the
any tendency for a high-strength bolt without same value as that used in the relaxation
washers to work its way into the surface of the tests. In Specimen 5 the assemblies were
connected material and thus lose its clamping tightened to a tension of 29,000 pounds, and in
force, a reduction of the fatigue resistance Series B and 0 a bolt load of approximately
could be expected to appear in these fatigue 28,400 pounds was used, the latter correspond-
tests. In those specimens where the outside ing to the minimum permitted by Specifi-
plates were designed to be critical (the outside cations.(G) In the case of the l-inch bolts
plates having the same or higher nominal stress of Specimen 9, the bolts were tightened to
than the center p late), the objective was to 10,000 pounds before adding the final one- half
determine whether the galling on the outer plates turn. This change was based on tests which
resulting from the tightening of the assemb ly, indicated that the snug position for the l-inch
combined with the high plate stress and possible bolt was more closely represented by 10,000
loss in bolt tension, would cause a reduction in pounds than by 5,000 pounds. For Specimen C- 1
the fatigue resistance. snug was obtained using 100 foot-pounds torque,
To provide a correlation with the data after which an additional 180 degree turn-of-
the specimens were similar to those used for The fatigue tests were carried out in
·
spec1mens . prev1ous
1n . s t ud.1es. (J) Tens1on:s
. hear the 200,000 pound University of Illinois fatigue
ratios* for all specimens are shown in Tables 4 machines which are ful l y described in Reference
* The "tension:shear ratio" is the ratio of the 8. The load was applied at the rate of approxi -
net section tensile stress to the shear st ress on
the c ross-sectional area of the bolts. The ten- mately 200 cycles per minute to all specimens
sile area used in the calculations for a "nomi-
nal" T:S ratio was computed using the nominal except Specimens 03, 04, and 05 . These speci-
fastener diameter plus 1/8 inch regardless of
actual hole size. The shear a rea assumed was mens were tested at a loading rate of 100 cycles
calcul ated using nominal bolt diameter.
9
per minute to reduce the tendency for slip and what were considered to be the most severe
pounding which had developed in the second conditions possible - - oversized holes, minimum
All plates for the fat igue specimens normal bearing stress under the bolt head and
were of ASTM-A7 structural steel. The plate nut. To accomp lish this, a ll bolt holes
surfaces were left in the as-rolled condition; (except for Specimen Cl) were d ri lied 1/8 inch
however, in some cases these surfaces were larger in diameter than the nominal diameter
rusted and slightly pitted from having been of the bolt to simulate excessive field ream-
stored outside. The plates were lightly wire ing, and regular series hex bolts with finished
brushed before use to remove any loos e material, hex nuts were used to give an unusually high
and the contact surfaces of the joint were bearing stress at the plate surfaces under-
cleaned with a solvent to remove any cutting- neath the bolt heads and nuts. Specimens with
oil remaining from the hole-drilling operation. less severe geometries would have been tested,
The bolts for the f atigue specimens if necessary, to find the conditions required
met the hardness and ultimate strength require- to produce fatigue lives comparable to those
ments of ASTM-A325. Bolt and nut hardnesses obtained with joints having two hardened
are shown in Table 4 for fatigue test Specimens washers for each bolt; however, the severe
I through 9 where the hardness tended to be conditions noted above did not appear to
low, but within the allowable range. In general, produce any significant reduction in the fatigue
for specimen types A, B, and D, the hardness of strength of the joints and no further tests were
the high side of the a llowable Rockwell C range First row slip was measured a t each
(RC = 30); the hardness of the finished nuts edge of all the fatigue specimens by means of
was Rc = 26 (ASTM-AI94, grade 2H). The regular mechanical dials (Figure 2). ''First row slip 11
hex bolts of Series C hao d hardness of RC = 32 is the relative movement which takes place
and the hex nuts in Series C had a ha rdness of between the edges of the center plate and the
. .r:
R
6
= 82 to 84. outside p lates and includes both elastic and
mens of Series A, B, C, and D were designed for Where bolt tensions in the fatigue
to those used in the fatigue specimens and with made possible the construction of a typical
the same grip was tightened in a Skidmore- load-elongation curve that was used to es-
Wilhelm Calibrator and its elongation measured tablish the loads or tensions of the bolts in
A. RELAXATION TESTS OF A325 BOLT ASSEMBLIES upon completion of the torquing there was a
In examining the Table I relaxation small drop in load. The load for each bolt,
data for A325 bolt assemblies it is necessary recorded in Column g of Table I, is the peak
to keep in mind the wide range in bolt and nut load on the trace. However, because the maxi-
hardnesses which existed in the specimens test- mum load in the bolt is only an instantaneous
ed. Because of this range, comparisons should load and rarely measured, even in the labora-
be made only between assemblies with comparable tory, t h is value has not been used as a basis
hardnesses and with the same nut and bolt head. of reference. Instead, the load in the bolt
Furthermore, since each type of nut and bolt one minute after reaching the peak load has
head was tested bot h with and without a washer, been used as the base from which to determine
the contribution of the washer to each assembly the percent loss in bolt tension with time
can be studied without considering hardness as {Column 10 , Table I). The drop in load from
a variable. In order to provide additional the maximum load to the one-minute load varied
information, one assembly of each type was from 2 to 11 percent, with an average of about
the bolt elongation was measured at various thought to be a result of an elastic recovery
increments of load. These measurements were which takes place when the wrench is removed
compared with elongation readings taken on the and possibly a result of creep or yield in the
relaxation specimens as a check on the bolt bolt, produced by high stresses a t the root of
tension recorded by the load cell. the threads. In addition, some flow may occur
In the relaxation tests a continuous in the steel plates under the head and nut.
record of bolt tension versus time was obtained Simil ar tests made on a hydraulic
from the load cell with a Sandborn Recorder; a load analyzer by the Lamson and Sessions Compa-
typical curve is shown in Figure 4. It was ny(]) show similar percentages of bolt tension
evident from these records that immediately losses immediatel y after tightening and over a
11
12
period of days. Comparable results were also shown in Figure 7. The flanged nut produced
reported in Table 6 of Reference 9 for 7/8- inch the greatest bolt load, yet the torq ue required
ticular tests. and the size of the pneumatic wrench and the
The effect of the washer on the maxi- air compressor capacity are factors.
mum torque required to tighten an A325 bolt In Figure 6 it is also important to
assemb ly is evident upon examination of Figure note that: (1) With washers under the nuts,
6. The higher torque required to tighten nuts the torques for 5,000 pounds plus 1/2 turn
without washers is the result of galling be- were relatively consistent regardless of the
tween the nut and the soft plate material. nut types, and (2) there was considerab le
One of the more severe cases of galling is variation in torque between the different types
omitted under them while the bolt head s pro- 0.380 square inches). This may account for
duced depressions in the plates if their washers the fact that the heavy- head bolt, when
were omitted. A typical bolt- head depression tightened in a 7/8-i nch diameter hole, he ld
is shown In Figure 8. The results of the bolt- its load in much the same manner as did the
head tests, both with and without washers, are regular- head bolt when tightened in the
compared in Figure 9; in all cases the bolt 13/16-inch hole. However , even the regular-
heads were in contact with the plates or washers head bolt, when used in a 7/8- inch hole and
and the bolt assemblies were tightened by turn- without a washer, developed nearly the same
ing the nuts. Also included for comparison are tension as did the bolt in a 13/16- inch hole
the tests in which the holes were either with a washer, when tightened to snug plus l/2
1/32- inch or l/16-inch oversize. I n the case of turn.
the regular semi-f inished hex-head bolts {A325 3. Bolt Head Torque Tests . Several
type) and the 13/16-inch diameter holes, there relaxation tests were conducted after turning
was no difference in bolt load whether or not the bolt head rather than the nut. The results
there was a washer under the bolt head. When of these tests, conducted with the regular
the hole diameter was increased and no washer semi - finished hex head A325 type bolts in
was used, a small dec rease in load occurred, various hole sizes, are shown in Figure 10.
although the final load was sti 11 well above the The oversized hole produced a small decrease
minimum tension now specified. For the softer, in the bolt load, but the maximum torque was
heavy sem i- finished hex-head bo 1ts {A325 type) approximately the same. It is apparent that in
there was no apparent loss in bolt load when either case, omitting the washer and applying
the washer was removed or as a result of the 5,000 pounds plus 1/2 turn produced more than
~~
oversized holes. The bearing area of the bolt proof load in the bolts, but the required
head around the perimeter of a 7/8-inch diameter torques were considerably larger than would
hole when a heavy semi - finished hex-head 3/4- inch have been required for proof load only, if
bolt was used {approximately 0.402 square inches) washers had been used. Where the regular semi-
was slightly greater than the bear ing area of a finished hex bolt head and the heavy semi -
regular semi-finished hex-head 3/4- inch bolt finished hex bol t head were torqued to approxi-
a round a 13/16- inch diameter hole {app roximately mately the same load (Figure 11), a somewhat
head type. Note that the heavy- head bolt was tightened in a 7/8-inch diameter hole. The
ins t al led by 5,000 pounds plus 1/2 turn while flanged nuts (Figure 12c) developed con-
the regular- head bolt was tightened to approxi- slstently high bolt loads with both hole sizes.
mately that same load, rather than by amount The difference in behavior between
of turn. the finished and the heavy nuts in the case
A comparison of the depths of the of the 7/8-inch diameter hole may be the
depressions in specimens tightened by tu rning result of the differences in the load- turns
the heads to approximately the same bolt load, relationships for the two types of nuts and
such as H- 2S- 34 (.0089 inch, regular head) and may be caused by differences in geometry and
H-4 (.0028 inch , heavy head), indicates the hardness. In order to explore this problem
marked reduction in the depth of dep ression further, the load-turns relationships for
resulting from the increased bearing area bolts with either the heavy or t he finished
provided by the heavy- head bolt. Study of the nut were determined and are plotted in figure
data on depth of depression and nominal bearing 13. Note that for the softer finished nuts,
pressures showed that, at a given nominal bear- the bolt load at 5,000 pounds + 1/2 turn was
ing pressure, specimens tightened by turning considerably less than in the case of the
the bo lt heads had deeper head depressions than heavy nut. It can also be seen that the maxi -
similar specimens tightened by turning the mum tension attainable wi th the finished nuts
nuts.(!) I t appears that as the heads were and washers was only 34,000 pounds, a value
turned mechanica l wear of the plates produced only slightly above tha t produced by 5,000
the greater depressions. However, there pounds+ 1/2 turn. Harder finished nuts would
seemed to be little difference in the load have produced higher bolt tensions and might
produced whether the head or nut was turned. have produced a load-turns relationship compa-
4. Effect of Hole Size. The results rable to that of the heavy nuts.
of tests ru n with several types of nuts and It is interesting to note the wide
regular bolts in different size holes are shown range in bolt loads obtained between bolts on
in figure 12 for easy comparison. for t he t he high and low side of the specifications.
finished nuts (Figure 12a) there was little Comparison of the test results in Table I for
variation in load as a result of removing the SPecimens N- 5 and N- 5S (40,450 pounds) with
washer or as a result of turning the nut over the results of N- 17S (33,700 pounds) emphasizes
an oversized hole. However, in the case of the this difference. It is a lso worthy of note
heavy nuts (figure 12b) there appeared to be a that the flanged nut cons istentl y developed a
small decrease in load when the bolt was load equa l to or greater t han the load
Bol r H£AO Sruoy
5PECI M £N
F I GURE 2 . TYPICAL FATIGUE SPECI/'1.:N St-()WJNG MECHANICAL DIALS FOR " SLIP" t-EASURE/'1.:NTS
h......._ Actual Hole Dlam. • d
+ + + + a~ 1 + + + +
+ + + +
9
I ','
a ..J--
h /~
d
-- i-
.1 + .j. + +
e p e
I thru 5 3/4 Center I 3/8 3/8 5/8 13/16 I 1/4 6 3 1/4 0 3 1/4
6 3/4 Outside 1 5/8 3/8 3/4 13/16 1 1/4 6 1/2 2 1/4 1/4 3 1/4
7 3/4 Ouuide 1 1/4 3/8 3/4 7/8 I 1/4 5 3/4 2 3/4 1/4 3 1/4
8 3/4 Outside I 1/4 3/8 3/4 13/16 1 1/4 5 3/4 2 3/4 1/4 3 1/4
Speci111en
types A, 3/4 Outside I 1/4 3/8 3/4 7/8 I 1/4 6 1/4 2 3/4 1/4 3 1/4
II, C,
and 0
C1 5/8 Outside 1 1/4 3/8 3/4 11/16 I 1/4 6 1/4 2 3/4 1/4 3 1/4
40
I' One Klnute? L
I \ Load
-•.
~
Q. 30
J I'Haxl. . Load
~
20 I I
~ Spec. No. N-145
FIGURE 4. RECORD FROM SANBORN RECORDER -•...
0 10 J Heavy Nut Thin
llo Washer
I
Regular ltud lolt
0 I
o o.2 o. 4 o. & o.a 1.0
Tl•• Klnutes fro. Maxi- Load
40
30 30
...
~
'V
0
...J
.Q
~
0
20 20
m
"c:
~
::J
co
N
-
%: 'V
Ill
0
0"c: ....110
~
10 10
0
0
~
Q..
0
Heavy Heavy Nut-Thin Finished Finished Flanged
Nut Nut Nut-Thick Nut
3/4 x 6 Regular sen~i-finished hex head bolts, Tightened to 5000 lb. + 1/2 Turn
11
200
100
0
200
100
0
200
100
0
200
100
0
II
Heavy Heavy Nut-Thin Finished Nut Finished Flanged
Nut Nut-Thick Nut
* Sued on T • 0.2 PO
3/4 x 6" Regular semi-finished hex head bolts, Tightened to 5000 lb . + 1/2 Turn .
Plots are average values where duplicate tests were conducted.
FIGURE 6 . NUT TESTS - EFFECT OF WASHERS ON TORQUE OF A325 BOLTS
SPE --\ M EN I v. 7 s
\1 AV'< \'l T
0 WA H&::-R
SPEC\MEN No. 2
REGUL.Aa BoLT HEAD
No WAsHER
FIGURE 8. DEPRESSION OF PLATE SURFACE
RESULTir-.K; FROM BOLT HEAD
'V
•.,
z
:30 •"'
'V
.
.Jt
s J"'
.!"'
••
1_,
. ..•
:a
:a
J
0 20
..•I
•
X
J 10
:30
~
~
It
I....!
roof Lo.d (28.400 lb.)
.,.
.Jt to pr<*tu Proof Lo.d
., N when washers arc used • when washers arc • • ·
]
..
0 20
::. 600
.
~
3
20
•
..•
~
•
0
..•
:I
c .
z:
-...
.a
.....
z:
•c
0 10 IS
. g• 10 ...
::i IN ; 200 H.O--l~H
~ I~
~
., I~
:I
f
~ 100 H~--l_.,H -
......"
:I
c:r
0
!::
0 ~ ~ 0 '--"a.D..---LU.......J 0
13/ 16 7/8 Hole S"lze 13/16 7/8 Reg. Heavy Reg. Heavy
3/ 4 x &• Reg. semi-fin i shed hex head bolts 3/ 4 x 6" Bolt - 13/ 16 l n. dl a. hole
(no washers) (Spec l-.ns torqued to appr~l~a t ely
salle load - no washer)
FIGURE 10 . A325 BOLT HEAD TESTS - HEAD
OR NUT TORQUED; EFFECT OF HOLE DIAMETER FIGURE 11. A325 BOLT HE.AD TESTS - HEAD
~ BOLT LOAD MD TORQJE TORQUED; EFFECT OF HEAD SIZE ~ TORQUE
.. 30
..,
z
..
Q. ...
....•
:5 ...
,;
... ..•
1
3
.
~
..•
.! Jl
-
0
;•
20 Jl
0
'"'
c
&
•c
0
10
II
Kole ot-ters
(a) fini shed Series Nuts (b) Keavy S.rlea Mvta (c) Flanged Mvts
4Q
~
.J'l""""~
\lJvy
(
Nut
• 9 1)
/
v v_-
•
A.
..........._
.
Jll.
,
.!
.r01.
..•
5
c
....
20
f/ I I
I I
Specl•n
~ ) I I
•0
~
10 I I
I Motu :
Wuher under llut
13/1& ' " · 01.-. Hole
Regular Head lolt
0 I I
0 1/4 1/2 3/4
Turns fro. "Snug" of S kips
FIGURE 13. LOAD-TURNS RELATIONSHIP FOR HEAVY AKJ FINISHED NUTS USED IN THESE TESTS
ApprOKIMate torque
req 1 d to produ~e
Proof Load when
washers are used •
.
..0
.
.......• 400
"~
fT
.....~ 300
200
100
0
13 13 27 7 13 13 27 7 13 7
T6 T6 3i 8 T6 T6 3i 8 16 8
Hole Diameters
{a) Finished Series Nut (b) Heavy Series Nut (~) Flanged Nut
FIGURE 14. EFFECT OF HOLE SIZE ON TORQUE FOR 5000 LB. PLUS ONE-HALF TURN
70
I
. ......---'- -- - 1-------- _- f-.': 7/8"4- A3S4 BO
60
~roof Load 7/8"¢ ' -rl!---+---~
.. I A3S4
..,__...,._ 80- - - - - - - - - --
1 , 7, 8··¢
~ so
1--i- . . P:-L: 7/811~ "! A3S4 BC
......
.>t ----- - --
"'0 t/
~ 3/4 ¢> A32S l 11
------ - I
. 40
; : 7-- p. L. 3/4"¢> (Hl gh Hardness 32 Rc) _ .{ 314,~
~
.g
c:
..."
30
,IL
IV'. _ A3S4 80 - - - - - - - - - - , - - - -
!1 --;~:.-3~4"¢>
A3S4 8C
P.L. 3/4'~ A32S
113/41 ~ A32S
- A3S4 BC
( Low Hardness 23 R)
c
-I- - 1 --
---~--
I
I
~
0
liD 20 t-1--t- - +- - + -- + -- - + - - · - __ j__-
All bo l t s had g rips ~f
2 1/411 and were torqo, ad
10 In a Skldnlore-W I Ihl!lr- -
Calibrator
70
60
..
~
50
A3S4 BO Bolt
2 1/4 ln. Grip 7 ·/ A3S4 BD Bolt, 1 3/4 ln. Grip
------ ------r--
-------
.>t ____ _ J
~
...3 ,, ;
-Proof Loed
---- r---- -- ---
.. 40
T A3S4 BO /
A32S Bo l t
I 3/4 ln. Cirlp
...
.. ~~:
~
g
c
..."
30
I A325
----- -z---- -
A32S Bolt
2 1/4 ln. grip
-.::::...--- ....
-
~ All bolts were 3/4 l n. dla~. by
I
•0 20 3 l n. long wi th suitable nu t s .
I The I 3/4 ln. grip Incl uded ap -
proxiMately 7 th reads and 2 1/ 4
I ln. grip approximate l y 12 threads
10
i/ Torqued In a
Ca II bra tor .
Skl~re - WI lhelQ
0 I I I
0 0. 02 0.04 0.06 o.oa 0.10 0. 12 0.14 1) .16
Ave rage Over-all Elongation , Inches
-- -·- ---
I I I I I I I I I I I
I I
/ ;;;.._/7/8''~ A3~ 80
60
Proof Load 7/811~
- A3~ BC t' ~~
. --
I --- 1 - - - ec
~:l· 7/8"~
A3~ eel
l
~"' /:,_3/4'~ A3~ 80
-----1.
f_11&·~ A3~
1
',...,
P.L. 3/4"~ / ~ ::::::-- --~
-- -
~'
/" 3/4"~
~
A325 High Hardnen
A3~
--- ----- -r---
'~
...0 BOj _:::1-~
-- :---
f----
--
~
,_2.. L.
A3~
3/4"~
BC I 1,' ~A325./ 3/4"~ A3~ BC---"- -..... r--- 1---
......
.. 30 P. L. 3/4"~
c:
..
...."
A32s
,rt
,g 20
II I
,_1
10
jl All bolts had grips (head to nut)
of 2 1/4" and were torqued In a -
v' SkidMore Wilhel• Calibrator .
0
J/ I I I I I 1 I I I I I I I I
0 0. 5 1.0 1.5 2.0 2.5
Average NURiber of Turns fro. "Snug'' of 5 kips
FIGURE 17. COMPARISON OF LOAD-TURNS DATA FOR VARIOUS A325 AND A354 BOLTS AT IDENTICAL GRIP
70
I I I I I I I I I I I
60
. so
,.,.-· --- r--~ A354 80 Bolt, I 3/4" 6irlp
-~ . ----
·----L -J
G.
~
--8 £0 ~-r:.3-'T
- A32S Bolt, 2 1/4" 6irlp --
r
~-
r---
.
c
30 Proof L011d
AJ2S
-~
:!
:::.0
20
I
OD
All bolts were 3/4" dl-. by
3" long, Torqued In a SkidMore -
Wllt~l• C.llbrator
10
~/
0
~,
I I I I I I I I I I
0 o.s 1.0 1. 5 2. 0 2.5
Average Nuilber of Tums fro. "Snug'' of 5 kips
FIGURE 18. C(M)ARISON OF LOAD-TURNS DATA FOR DI FFERENT GRIPS A325 MD A354 BOLTS
~
3
... ·-
GJ.J:!
~
~
-I-
IU '- 2
....0 "c:
~
~-
.... ~
M
"'E 8 ~
~ ...
67l~~l
.... 'to-
0
Threads ll2 3 3 3 ~ 4J.2
In Grip 2 2 ...
2 , c.:l
'-- '----..,-------'
Grips 1.35" I. 5011 ).sou
Used
(a) Bolts Tightened to Failure In Solid Blocks
~--------------------~~----_, :~
rom
II
.J::
~
N
M
<t
0
Threads sJ.2 7 ,oJ.2 11 12 12
In Grip
Grips 1. 7511 2.25"
Used
{b) Bolts Tightened to Failure in SkidMore-
Wilhel• Calibrator
(All Bolts were 3/4 in. diaMeter)
FI GURE 19. CCM'ARISON OF TURNS TO FAI LURE FOR VARIOUS BOLT TYPES, GR IPS, An) CONDITIONS
70
where~ •
I I I 1
Approxl.ately T • 0.2PO TorqJe, ln.- lb.
P • Bo I t Load , I b.
0 • Bolt Olaa., ln.
r·
60 T • 0.2(30,000)(3/4)
• 4SOO ln. lb • 37S
ft.
...0.. so
..- A354 80 Bolt
""' /'
.......
"0 Proof Load A354 BO ,/
~40 A354 BC Bolt
....
.. I I
.:~ 1-
~
A32S Bolt
~
0 ~oof Load, A354 BC
I I
~
g
..
c
"
1-
0
30
20
~oof Load, A32S
/
v .J
~
CD
All bolts were 3/4"~ by 311 long and
v
10 Hardened washers were used under nut~.
't'
0 0 100 200 300 400 500 600 700 800 900 1000
Average Torque, Foot - Pounds
FIGURE 20. LOAD-TORQUE RELAT I CNSH IPS FOR VARI OJS BOLT TYPES
;-;( I)
so I ... - ~ (1)
I
11 Proof
J!L
I U I I
I
l---- =! !.-
\Load
4)
I
I
.
I
1
~ I
(2)-- "'
-'~
I( 1)-
Q..l -
1
I
I wt
~
1--
-u..
( 2):;
-Q,.
:J
I--
....~ .
-u... -... -...
1-- Q, ~ lE:
~-
Q,
Q, Q,
~
~
.... ..
u u
] o..-
011-
1--
-..c: ~ ~
1"-
.... ... ... ....
"0 vc:
.c-
-
1-- ...
v "0
.. _
v
"...
1- J:
1-
J:
1- ......
U'-
J:J:
0
1---
..____
J:
1-
~
.,.. - -
N N
:a.o
L--
FIGURE 21. BOLT TENSION AT " SNJG" OF 5000 LB. PLUS ONE-HALF TURN FOR A354 BD RELAXATION TESTS
Note: All fasteners are 3/4 ln. nominal dla~ter.
60~--------~------~--------~--------~--------~---------,
A354 BD Bolt, Tl plates ,
--- Shear P 1ane I
354 BD Bolt, A242 plates,
in Shear Plane
50
I
A325 High Hardness Bolt,
A7 plates, Shank In
Shear Plane
-• 40
Q.
.,.
.Jtl.
~
•
•uc ? ()
....
0 10 20 30 40 50 60
Shear Load to.ponent, kIps
FIGURE 22. INTERACTION CURVES OF A325 AND A354 BD BOLTS
UI\OER COMB INED TENSION AND SHEAR SHO.<I ING FAILURE LOADS
52
48
Average Curve -
(II Tests)
44
40
36
.
-..
Q.
~
32
g
..
c
u
28
1-
.... 24
0 Average Curve - Is
IQ
(27 Tests)
20
16
12
Flat Surface
8 Tests
Approx. 1/2 Turns
(Extremes) T~o 5% Converging or
4 Crossed Bevels Tests
0
0 1/4 1/2 3/4 ) 1/4 ) 1/2
Turns frOftl "Snug'' of 5 kips
FIGURE 23 . EFFECT OF BEVELED SURFACES ON BOLT TENSION VS . TURNS, 3/4 IN. DIAM. A325 BOLTS
72 ~--~---.----.----,---.----,----.----.----r----.---,----,
68 ~--+---~--~----+----r--~r---+----+--~----~---r--~
64 ~--4----+----+---~--~r----r--_,----~---+----r---~--~
56
Average Curve -
(8 Tests)
52
..
~ 40 t--- - + - - 4-+--14- -l
..
..g 3& ~--+---~~~~~~~~
c Average Curve - Bevels
...." (8 Tests)
~ 32 ~--+--~4-~-~~~~~---t----
0
•
16
Approx. 3/8+ Turns
- (Extrea.es)
12 ---+-·-
D Flat Surface
Tests
~ Two 5% Converging
~ Bevels Tests
0 L---~--~----L---~--~----~--~--~----~--~--~--~
0 l/4 1/2 3/4 1 1/4 l l/2
Turns fran~ "Snug" of S kips
FIGURE 24. EFFECT OF BEVELED SURFACES ON BOLT TENSION VS. TURNS 1-JN. DIAM. A325 BOLTS
(a) Locatio n of Fatigue Crack in Specimen B l
SPECIMEN-BJ
's"~HOlf- 29,4oo•soLr TENS.
SR : 0+30 KSI N:: 11 883,100
••
. -I....
.... -
l it
• ~
lit~
....0
J I
I 40 N
Deal .. StreaMs
AlfA-MSIIO Speclflcetl . . s, A7 St"l
llote:
llvet date
fro. Ref. 15
-40 -30 -20 -10 0 10 20 30 40 50 60
Klnl~ Stress, ksl (eo.pressl ..) KIRI~ Stress, ksl (Tension)
(All StrMMS ere lased on let SectiOft)
FIGURE 27. MODIFIED GOODMAN DIAGRAM FOR 21 000 1 000 CYCLES
100 0
-I
c
\ -•i&
.:
g
-..
60
r--.
- ~'* A fasteners In lpec:l . .ns that
failed at Rc»t A
40
-•..
....
0
c
...." 40 II II I
!
60
c
-• II.
•
....•0
~
0
l r---1~Rc»t II
Center
~
20 f+·--1 ~; A··Crltlc:al 80
f4'.:::..~ sec:tlon for out•
side pletes (See
_( ~lg. 3)
~ Outside Plates
0 100
0 5 10 100 1000 10,000 100,000 1,000,000 10,000,000
Applied Cyc~s
FIGURE 28 . AVERAGE LOSS IN BOLT TENSION VS. APPLIED CYCLES JN FATIGUE TESTS
TABLE 1. RESULTS OF BOLT- TENSION RELAXATION TESTS FOR A325 BOLTS
( 1) ( 2) (3) (4) (5) (6) (7) (8) (9) ( 10) ( 11) ( 13) ( 14) ( 15) ( 16) ( 17) ( 18)
( 12) ( 19)
Specimen Bolt Nut Washer Ha r dne ss Torque Elong. Max. 1 Min. Percent Loss of Load f rom 1 Minute Load Depth of
Nlllllber* Head Bolt Nut ft. lb. in. Load Load 5 Da s Depression
Rc RB Min . 3 4 5 6 7 21 in.
Nut Tests
N- 5 Regular Flanged No 32 90 580 .03204 40 , 230 39,150 0.4 3.9 3.9 3.8
N-55 Regular Flanged No 32 90 530 42,800 41,740 0.8
N- 5s- c Regular Flanged No 32 90 520 42,880 41,800 1. 3
N- 6 Regula r Heavy Nut 32 90- 92 .02577 40,930 39,430 0.1 3.2 3.8 4.1
N-6S Regular Heavy Nut 32 90- g2 460 .02268 41,040 39,340 0. 4
N- 8 Reg u lar Finished Nut 32 84- 86 .01356 33,640 30,020 0.5 1.6 1.6 1.6
N- BS Regula r Finished Nut 32 84-86 420 . 01545 36,700 33,240 0.3
N-9 Regular Finished No 32 84- 86 490 .01518 35,900 33,100 0 1.8 1.8 1.8
N- 9S Regular Finished No 32 84-86 500 . 01520 33,960 31,680 0.4
N- gs-B Regular Finished No 32 84- 86 620 .01805 36,100 33,960 1.2
N- 95- C Reg ular Finished No 32 84- 86 725 .01458 33,660 31,760 0.3
N- 10 Re gula r Fin.- Thick Nut 32 79-82 480 . 01791 35,970 34,350 0. 1 3.2 3.4
N- IOS Reg ular Fin.- Thick Nut 32 79-82 440 . 01914 37 . 740 35,960 0.3
N- 11 Regular Fin.- Thick No 32 79- 82 750 . 01534 34,440 33,720 0. 3 2.0 3.0
N- IlS Regula r Fin.-Thick No 32 79- 82 600 .01664 35,700 34,700 0.9
1 32 92- 95 450 . 01947 40,650 38,060 0 2.2 2.2 2.2
N- 12 Re gu Jar Heavy- Thin No
N-12S Regula r Heavy-Thin! No 32 92- 95 540 . 02186 41 , 140 39,240 0.3
N- 13 Regular Heavy- Thin Nut 32 92-95 420 .02075 40,560 37,280 0.4 2.6 2.4
N-13S Regular Heavy- Thin Nut 32 92-95 460 .02144 40,780 38,680 0.2
N- 145 Regular Heavy- Thin 2 No 32 92- 95 610 . 01893 38,880 37,500 0.3
( 1) (2) (3) (4) (5) (6) ( 7) (8) (9) ( 10) (11) ( 12) ( 13) ( 14) ( 15) ( 16) ( 17) ( 18) ( 19)
Specimin Bolt Nut Washer Hardness To rque Elong. Hax. 1 Hin. Percent Loss of Load from 1 Hinute Load Depth of
Number Head Bolt Nut ft . lb. in. Load Load 5 Da s Depression
Rc R8 Hin . 3 4 5 6 7 21 in.
Head Tests
H·l Regular Heavy Head 32 90- 92 40,200 38,670 1.0 4.7 4.8 4.8 5. 1
H-IS Regu lar Heavy Head 32 90- 92 480 39,800 38,620 0.8
H- 2 Regular Heavy No 32 90-92 .01940 39,960 37.980 0.6 3.4 3.5 3.6 .0087
H- 2S Regular Heavy No 32 90- 92 480 41,140 39,120 0.4 .o 112
H· 2S·B Regula r Heavy No 32 90- 92 420 .01627 35,800 34,260 0.5 .0105
H-2S-C Regular Heavy No 32 90- 92 420 .01682 37,960 36,560 1.0 .0143
H- 2S- 34 Regula r Heavy No 32 90- 92 440 .01620 35,340 34,840 0.2 .0089
H-3 Heavy Heavy Head 24 90- 92 34,080 33,500 1.1 2.2 2.2 2.2
H-3S Heavy Heavy Head 24 90.92 380 .03547 31,000 30,420 0.5
H-4 Heavy Heavy No 24 90- 92 .03178 34,300 33,700 1.2 3.0 3, I 3.2 .0028
H-4S Heavy Heavy No 24 90- 92 420 .03195 34,400 33 , 620 0. 4 .0031
H-4S-C Heavy Heavy No 24 90- 92 440 . 03200 33,800 33,180 0.5 .0069
H-20S- 31 -HT Regula r Heavy No 32 90- 92 460 .01502 32,920 31 , 960 0.8 .0082
H-21S- HT Regu l ar Heavy No 32 90- 92 480 .01868 39,340 38,480 0.7 .0132
H·22S· C- HT Regula r Heavy No 32 90- 92 500 .01610 36,520 35,220 0.3 . 0166
Lubrication Tests
N-23S Regula r Heavy No 32 90-92 620 .02626 41,100 3g,560 0.6
N-235-L Regula r Heavy No 32 90- 92 340 .02291 41,400 39,640 0.6
H - Bolt Head Test; N - Nut Test; p - Punched Hole; S - Short Time Test (5 minutes); 8 - Oversize Hol e- 27/32 in. dia.; C - Over size Hole- 7/8 in. dia.
HT - Head torqued to tighten bolt; L - Lubrication between nut face and plate.
Proof Load : 28,400 lb. Hinlmum Bolt Tension : 25,600 lb. required by 1954 Specification; 28,400 lb. by 1960 and 1962 Specification of Research Counci 1.
I. Washer Face Down Hardness range by specs. on 3/411 '1' bo Its: { A325- 58T Rc = 23- 32
2. Chamfered Face Down A325- 61T Rc • 23- 35
TABLE 2. RESULTS OF BOLT-TENSia-l RELAXATION TESTS CONEXJCTED WITH A354- BD BOLTS
(1) (2) (3) (4) (S) (6) ( 7) (8) (9) ( 10) ( 11) ( 12) ( 13)
Spec. Bolt Nut Washer Hardness Torque Hax. Hin. Percent Loss of Load From
No. Head Bolt Nut Load, lb Load, lb l Min . Load
Rc Rc Days
S Min. 2 3 90
NX-245-0- l Regular Heavy- 2H No 37 3S soo 40,600 39 , 210 0. 2
NX-24-0- 2 Regular Heavy-2H No 37 3S S40 37,200 3S,900 O. B 1.3
NX- 24-0- 3 Regular Heavy- 2H No 37 3S soo 36,8SO 34,8SO 0 .9 1. 0
NX-24S-0-4 Regular Heavy- 2H No 37 3S 610 40,SOO 40,300 0. 4
NX-24S-E- l Regular Heavy-2H No 37 3S sao 4l , lSO 40,000 0.6
NX- 245- E-2 Regular Heavy- 2H No 37 3S S30 39,860 38, 1SO 0.4
NX-24S-E-3 Regular Heavy- 2H Nut 37 3S S20 47, ISO 46,840 0.7
NX-24S- F-l* Regular Heavy- 2H No 37 3S 4SO 35,3SO 34,2SO 0.5
NX- 24S- F-i* Regular Heavy-2H No 37 3S 450 34 ,600 33,300 1.1
NX- 24-F-3** Regular Heavy- 2H No 37 3S 37,BSO 36, 7SO 1.1 4. 2
HX- 255- E- 1 Regular Finished- 2H No 37 26 480 36,230 34,700 0
H • Nut Test ; X; A354-BO Bolt; S • Short Time Test (S Hinutes); 0 =3 Threads in Grip; E • 6 Threads in Grip;
F • 12 Threads in Grip; A7 Plate used under head and under nut .
Proof Load- 40,100 lb. 5 Washers In Grip Hardness by Spec. - A354 BO Bolts 32- 38
* S/8 Plate in Grip Al94 2H Nuts 24- 37
**
TABLE 3. SlM-1ARY OF VAR IABLES FOR TESTS OF BOLTS TIGHTENED ON BEVELED SURFACES
SPECIAL TESTS
Position of Bolt: Maximum Sliding (4); Minimum Sl iding (4)
Bolt Strength: High Ha rdness (2); L~ Ha rdness (2)
Minimum Slide bolt placed against I~ side of bevel t o minimize bolt movement
d~n s lope .
Maximum Sl ide = bolt placed against high side of bevel to permit maximum possible
movement d~n be veled slope during tightening.
Conve rge a Beve ls placed with sloping p lanes conve rging (or dive r ging) .
Pa ra llel • Bevels placed with sloping planes para llel but with 5 percent slope
at head and nut .
Crossed .. Bevels pl aced with sloping planes "crossing" or t urned at go deg.
rel ati ve to each other.
TABLE 4. RESULTS OF FATIGUE TESTS, SPECIMENS 1 THROUGH 9
(I) (2) (3) (4) (5) (6) (7) (8) (9) (10) (II) ( 12)
Spec. Fastener Type Hardness Stress Cyc Ie Applied Kethod of Critical Remarks Average % Loss in Nom.
Bolt Nut Net Section Cycles Tightening Plate Bolt Load Bolt Load Coef.
Rc Ra Stress, ksl Before Test During Test of Frlc.
5 Heavy Head Bolt, 3/411 24 0 to +30 2,257,900+ Bolt Load • Center During first cycle, specimen slipped 30.3 19 0.32
Heavy Nut with no Washer 90-92 29,000 lb. Into bearing when net stress was
approximately 29,000 psi .
T:S • 1.0:0.75
6 Heavy Head Bolt, 3/4" 24 0 to +30 2,416, 900+ 5 kips* Outside Failed in outside plate on bolt head 33.5 25(a) 0.35
Heavy Nut with no Washer 90-92 + 1/2T side. During first cycle, specimen
experienced large slip at 27,000 psi.
T:S • 1.0:0.90
7 Regular Head Bolt, 3/4" 32 0 to +30 3,073,900+ 5 kips * Outside Hole- 7/8 in. india. Slipped 32.1 24 0.28
Heavy Nut with no Washer 90-92 + 1/2 T during first cycle at 26,200 psi.
T:S • 1.0:0.75
8 Heavy Head Bolt, 3/4u 24 o to +30 5,148,600+ 5 kips* Outside Slipped during first cycle at 35.2 29 0. 20
Heavy Nut with no Washer 90- 92 + 1/2 T 20,400 psi. Specimen bolts were
tightened by turning nuts on 2 bolts
and bolt heads on the other 2 bolts.
T:S • 1.0:0.75
9 Regular Head Bolt, 111 24 0 to +30 4,892,400+ 10 kips* Outside Slipped during first cycle at 62.8 3 0.24
Heavy Nut with no Washer B8 + 1/2 T 24,800 psi. Specimen bolts were
tightened by turning nuts on 2 bolts
and bolt heads on the other 2 bolts.
T:S • 1,0:0.75
+ Specimen did not fall. * 5 or 10 kip bolt load was developed by use of torque-tension relationship; nut was then given an additional 1/2 turn.
(a) Since Specimen No.6 failed, the loads of the 2 bolts on the line of fracture are not Included.
TABLE 5. RESULTS OF FATIGUE TESTS, SERIES A1 B1 C1 AND D
Spec. Stress Cycle App I ied Bolt Tension, Net Tensile Area, T:S Ratio Nom. Stress Nom. Coeff.
Net Section Stress, ksl Cycles sq. in. at 1st cycle of Friction
Nominal Actual Average % Loss Nom. Actual Calc.* Actua l slip, ksl based on 1st
Ini tial, During Fatigue Meas. cycle slip
kips Test
+ 34.8 14 2.84 2.74 I :0.80 I :0.77
AI 0 to +28.8 0 to +29.8 2,435,400
A2 0 to +28.6 0 to +29.6 2, 110, 100+ 36.5 17 2.86 2.75 1:0.80 1:0.78 25.5 0.25
A3 0 to +28.2 0 to +29.2 2,657,600+ 33.2 28 2.89 2.79 1:0.82 1:0. 79 26.9 0.28
15.7(a) 0.19(a)
81
82
0 to +30.0 0 to +31.1
0 to +30.0 0 to +31.1
I • 883 • 100(+ I)
2,067,200
29.4
29.3
-
21
2.88
2.87
2.78
2.77
I :0.81
1:0.81
1:0.78
1:0.78
19.9(b)
20.9
0.24(b)
0.26
83 0 to +30.0 0 to +31.1 2,455,7oo+ 29.2 21 2.87 2. 77 I :0.81 I :0.78 18.6 o. 23
Cl 0 to +30.0 0 to +30.0 2,390.5oo(l) 28.7 33(c) 3.07 3.07 1: 1.25 I: 1.25 14.8 0.20
Dl
02
±20 .0
±20.0
±20. 8
±20.8
1,860,200~2)
20,800 3)
28.4
28.4
-- 2.83
2.87
2.73
2. 77
1:0.80
I :0.81
I :0.77
1:0.78
18.4
8.1
0.23
o. 10
03
04
±18.0
±20.0
±18. 7
±20.8
Jo,ooo{J)
2,178,000(4)
28.4
29.0
-
8
2.85
2.84
2.75
2.74
I :0.80
1:0.80
1:0.78
1:0.77
14.4 0.18
Static Specimen: Ultimate strength • 60.8 ksl based on nom. tensile area
+ Indicates no failure (a) Minor slip (b) Major slip (c) At unfractured portion
* Calculated using actual hole diameter, Nominal T:S ratios were 1.0:0.75 for A, 8, and D with 1. 0:1 . 14 for Cl.
(I) Failure In outside plate (bolt-head side) at critical section. Fracture s t a rted at inside faying area--
possibly a result of fretting.
(2) Failure In outside and center plates at critical sections. Fracture In outside plate (nut side) appeared to
initiate at faylng surface; center plate fracture (bolt-head side) possibly a result of fretting.
(3) Test stopped because of excessive slip, pounding and high temperature (see text) .
(4) Specimen stress Increased from ±16 ksi to ±20 ksi afte r 253,000 cycles. Thus total specimen life was
2,431,000 cycles without fail ure.
(5) Specimen stress Increased from ±16 ksi to ±20 ksi after II , 100 cycles. Thus total specimen life was
2,278,100 cycles. Failure occurred in outside plate (bolt-head side) slightly beyond critical section.
Fracture appea r ed to Initiate by fretting between edge of bolt head and plate.
TABLE 6. RESULTS OF PREVIOUS FATIGUE TESTS
BOs-1 Regular Head Bolt, 7/8", with Washer !22 .8 1,206 , 000 Bolt Tension • Center Failed in center plate at critical section
Heavy Nut with Washer 56,500 lbs. T:S • 1.0:0. 75
BOS-3 Regular Head Bolt, 7/8" , with Washer ±20 4,405,000 Bolt Tension • Center Failed in center plate at critical section
Heavy Nut with Washer 56,500 lbs . T:S = 1. 0:0. 75
BOS-4 Regular Head Bolt, 7/&' with Washer ±20 2,083,000+ Bolt Tension • Center T:S = 1.0:0 . 75
Heavy Nut with Washe r 56,500 lbs.
BOS-7 Regular Head Bolt, 7/8", with Washer 0 to +30 1,973,000+ Bolt Tension • Center T:S 1.0:0 . 75
Heavy Nut with Washer 56 , 500 lbs .
( 18}*
Northwestern Unive rsity Tests
K(A7}-1 Regular Head Bolt, 3/4", with Washer 0 to +37. 382,000 Bolt Tension Center All specimens failed. T:S • 1.0:0.75
K(A 7}-2 Heavy Nut with Washer 312,000 26,500 lbs.
K(A 7) -3 225,000
K(A7)-4 Regular Head Bolt, 3/4", with Washer 0 to +35 I ,337,000 Bolt Tension • Center All specimens failed. T:S • 1.0:0.75
K(A7)-5 Heavy Nut with Washer 1,244,000 26,500 lbs.
K(A 7}-6 Regular Head Bolt, 3/4", with Washer 0 to 30 3,000,000+ Bolt Tension • Center T: S ~ 1. 0:0. 75
Heavy Nut with Washer 26,500 lbs.
+ Specimen did not fai 1. * Numbers In parentheses indicate references from which test results were obtained.
15
developed when a hardened washer was used with other than the following were to be justified:
a heavy nut. (l) The turn-of-nut method (snug+ l/ 2
The torques required to tighten the turn) without washe rs a nd in holes
bol ts i n holes of dif ferent diameters and using up to l/8 inch larger than the
the f in ished, heavy, and flanged nuts are nominal diameter will produce bolt
p lotted in Figure 14. The torque required to loads above the minimum for A325
tighten the finished nut increased with the bo lts as now specified, especia ll y
removal of the washer and also with the increase when heavy hex heads and nuts are
in hole diameter. The heavy nut had a higher used.
torque without a washer, and as the hole size (2) Omi ssion of the washer under t he
increased , the torque appeared to decrease. In turned element may produce torque-
the case of the f l anged nut the torque was load relationships somewhat more
fairly constant despite the differences in ho le erratic than usual; thus, torque
size , but was somewhat greater than that re- methods shoul d not be used for the
quired for the finished or conventional heavy installation or checki ng of bolts
nuts with hardened washers. having no washers under the parts
When the washer was omitted, a t orque turned during tightening or in-
of ap prox imately 700 foot pounds was recorded spection, except for bolts installed
for the heavy nut Specimen N- 7S; however, this by tur n- of- nut method in accordance
high torque was probably a result of the severe with the 1964 edition of Reference 6.
galling shown in Figure 7. Specimen N- 7 did 5. Lubrication. Two specimens
not have s uch severe gal l ing and a somewhat (N-23S and N-23S-L) were t est ed to study the
lower torque was obtained. The lower torque in effects of lubrication. One had lubrication
the case of t he 7/8- inch diameter hole is re- (molybdenium disulfide) only between the nut
lated in part to t he smaller load developed in fac.e and t he plate face, and the other was with-
differ ences in behavior shown between the torque requi red to tighten the lubricated
various types of nuts may be, in part , experi- specimen was just slightly more than one-half
ment a l scatter. Additiona l testing to give that required for the non-l ubricated specimen.
average values would be needed if obse rvations Tests run a decade ago with case- hardened
16
washers and employing a lubricant called exponent i al manner typical of creep data;
"No-Oxide" showed the following comparative whether there were wa s hers under the nuts or
results:(IO) under the bolt heads did not seem to influence
These results are of importance to erectors Those results agree well with the results of
using the turn- of-nut method and may warrant the tests reported herein and were summarized
impact wrenches and compressor capa cities may '~he tensile load which the bolt carries
is transferred through the washer to a
be used to develop the same bolt load. Since narrow, annular r i ng surrounding the bolt
hole. When the stresses in th i s ring
an advantage can be expected with appropriate reach or exceed the compressive yield
point stress of the pl a tes, the ring
lubrication of threads as well as the nut face, yields. As the ring yields, the annular
area becomes larger and the grip smaller;
this might be achieved commercially by treat- thus the unit stresses are reduced. The
yielding of this area is a lmost instan-
ing only the nut. taneous. If an elongation of a certain
amount is desired, and if the bolt load
6. Long-Time Relaxation Tests. A for this elongation is high enough to
cause yielding of the plate material (as
number of the A325 relaxation specimens were would be the case with nea r ly all con-
struction jobs where the bolt tension is
left tightened from 3 to 21 days to determine at least 100 percent of the proof load),
and if a torque is applied to the bolt
if there was any significant difference in the until this elongation is reached and then
the torque is released, there will follow
load-time relationship when the washer had an immediate dropoff in the bolt tension
(within 15- 20 seconds) of a magnitude in
been omitted (Table I, Columns 12 to 18). the order of 5 percent of the total load.
I f after having reached the desi red
Specimen H-1, which remained tightened for 21 elongation , the torque is maintained for
a short period (a minute or so) and then
days, showed a 5.1 percent loss from the released, there wi II still be an immediate
dropoff of the bolt tension, not to exceed
!-minute load. Ninety percent of this small about 5 pe r cent; the amount of loss will
depend upon how long the torque was
loss, however, occurred in the tirst day. maintained . If the torque is maintained
unti l there is no further yielding, we
During the remaining twenty days the rate of assume that there would be no loss of
bolt tension when the torque was released.
change in load continued to decrease in an However, the time e lement involved in such
a procedu re makes it impractical.
*
Insofar as is known, no tests or field "Following the immediate dropoff, there is
studies have shown whether such lubrication a further, gradual loss of bolt tension
will result in the nuts backing off in service , lasting for one to two hours, and amount-
especial l y under vibration or fatigue loadings. ing to another 5 percent of the load.
17
This loss may be attributable to the cations for high strength bo l ting and included
heating of the bolt during torquing and
its subsequent cooling, changes in room two grades of bolts-- the A325 fastener
temperature, heating of the annular ring
during yielding and subsequent cooling, (similar to SAE Grade 5) which has become so
and other similar effects .
familiar to the structural profession over the
·~his yielding of the plate should not
have any great effect upon the recorded past decade and the less familiar (and less
bolt tensions. The initial, large loss
is over , normally, before the bolt common) A354 bolt. Initial consideration by
tensions are measured, and the subsequent,
long time loss is small enough to be with- AISC appeared to favor the A354 Grade BD bolt
in the calibration errors.
(comparable to SAE Grade 8), but the ASTM
"The times and percentages indicated
above are merely to give an indication specification limit on size a nd the economics
of the order of magnitude. Their absolute
value will depend upon the yield point involved in producing the larger sizes led to
of the plate materi a l, and the bolt
tension and properties of bolt and nut. selection of the A354 Grade BC bolt for the
Also such local factors as the rate of
application of the torque and plate AISC 1961 specification and for the 1963
surface conditions may affect this -:(
behavior." revision.
It is interesting to note that the Since little information was avai !able on
loss in bolt tension after one minute, observed the characteristics of either grade of A354
in the recent University of I llinois tests, bolt then under consideration, an exploratory
confirmed the tests that Stevens made nearly program was initiated at the University of
a decade earlier. This same percentage of loss Illinois to provide information on such bolts
in bolt tension, about 5 percent, agrees in structural sizes. Initially, the Grade BD
closely with the loss in bolt load with time bolt was compared in behavior with the A325
reported by Sanks in a recent series of tests bolt; subsequently, when the AISC specification
on specimens which had mi llscale contact was in its final form, some additional tests
surfaces as well as a variety of other surface were made on the A354 Grade BC bolt.
9
conditions.( ) It is also confirmed by the The results of these preliminary tests
tests on A325 bolts conducted over a period of are included herein as a means of comparing
days in the Lamson and Sessions load the relative behavior of A325 and A354 bolts.
ana lyzer. ( 7) However, the tests are neither extensive nor
exhaustive.
Because Grade BC bolts were not averaged 47 . 0 kips. The A354 BD 3/4-inch
commercially available in the sizes and lengths diameter bolts had strengths in direct tension
desired for these tests, Grade BD bolts were averaging 56.9 kips. The A354 BD bo l ts of
annealed and re- heat-t reated to produce the 7/8-inch diameter averaged 73.2 kips.
are permitted variations in hardness ranges to the maximum bolt load obtained when bolts
(based on ASTM A325- 63T) depending on bolt are torqued to failure. The torqued strength
diameter, but the A354 bolts (ASTM A354- 58T) was 82 percent of the tensile strength for the
have a single hardness range for each grade few tests available for A354 BD bolts. This
which must be maintained for all sizes. The is 4uite similar to the 84 percent reported
largest diameters of A325 correspond in hard- by the University of I llinois for A325 bolts(ll)
ness, but not in elongation or reduction of and close to the 85 percent value commonly
area, to the A354 Grade BB bolt; the smaller assumed for A325 bolts.
diameters of A325 have hardnesses not greatly Load-elongation behavior at a given grip
different from those for the A354 Grade BC provides another convenient means of comparing
fasteners; and A325 bolts, l-inch diameter or the behavior of the A325, A354 BC, and A354 BD
l ess, of near maximum hardness would meet most bolts. Such data are shown in Figure IS where
requirements of A354 Grade BD. For these the four lower curves are for 3/4- inch diameter
reasons, low hardness and high hardness A325 bolts. Here we see that the behavior of the
bolts have been included in the following A325 bolts may be similar to the behavior of
comparisons between the behavior of the the A354 bolts, depending on their hardnesses.
familiar A325 bolts and the A354 fasteners. As might be expected from our knowledge
No attempt was made to select the hardnesses of the A325 bolts, reducing the number of
of the A354 bolts used. Al l bolts were pro- threads in the grip of the A354 fastene rs also
vided with appropriate nuts, either A325 or tends to increase the maximum bolt tension
Al94 Grade 2 for the A325 and A354 BC bolts, obtained. This is shown in Figure 16 where
and Al94 Grade 2H for the A354 BD bolts. average load-elongation data have been plotted
The low hardness A325 3/4-inch for two grips and the A325 and A354 BD bolts.
diameter bolts had an average tensile s trength The A325 bolts used in all these series of
of 40.1 kips (the minimum permitted), while tests had the old* thread lengths and thus
the same diameter high hardness A325 bolts * The length of thread for these bolts was
nominally two diameters plus 1/4 inch.
19
were simi Jar in this respect to the A354 bolts It has been shown in studies at the
tested. University of Illinois, Lehigh University , and
Possibly of greater value to those elsewhere that, although the apparent strength
who may use these fasteners are the comparisons of a high-strength bolt may be increased by
of A325 and A354 bolts in the load- turns plots reducing the number of threads in the grip,
of F igure 17 and Figure 18. As befor e, the the turns- to-failure is simultaneously de-
load versus turn-of-nut data have been adjusted creased. This condition might then become
17 shows that at snug plus 1/2 turn the A325 of A354 bolts, particularly if thread lengths
bolts easily provided a prestress or bolt comparable to those on the new heavy head
tension equal to the proof load; in fact, the structural bolts (A325} were to be provided
proof load prestress is reached i~ 1/4 to 3/8 for the higher strength A354 or A490 bolts .
turn of the nut from snug. Because of their The decrease in turns-to-failure
higher proof loads, the A354 bolts of both with a decreasing number of threads in the
grades require more than 3/8 turn and up to grip, is shown in Figure 19a. At a grip of
1/2 turn of the nut beyond snug to achieve the 1. 35 inches the A325 bolt had about I 1/2 full
desired preload. An average for six 3/4-inch threads included and the A354 bolts had about
diameter A354 BD bolts showed snug plus 150 3 full threads (the number of threads in the
degrees to be required for proof load. This grip are shown for al l cases}. Note that the
suggests that more nut rotation may have to be turns shown in Figure 19a are from finger-tight
specified for the A354 bolts than is the case or about 1/4 turn greater than from snug. The
for A325 bolts in order to insure that during maximum torque required in these solid block
field tightening at least the minimum required tests was often reached at about one turn from
bolt tension is achieved. In the case of the finger tight. Figure 19 illustrates again the
A354 Grade BD bolt the factor of safety against probability of a smaller number of turns-to-
over- tightening and against resulting bolt failure with the A354 BD bolt than with the
breakage is reduced accordingly. Fractures A325 bolt for bolts with the same nominal
at about 1/2 turns from snug and at about The difference in turns-to-failure
I 3/4 to 7/8 turns from snug for the 7/8- inch between bolts torqued in the Skidmore-Wilhelm
having the same grip appeared to be small as manne r, the relaxation behavior of A354 SO
shown by comparing Figures 19a and 19b . How- bolts was evaluated by means of a limited
ever, other studies at the University of number of tests. These incl uded various
I llinois and elsewhere indicate that there may thread lengths within the grips; two nut sizes;
be substantial differences in turns required washers and no washers; and varying lengths of
to reach a given load , especially in the elastic time, up to three months , for the tension to
range of the bolt and in the range of normal relax. A354 BO bolts with regular finished
installation. Some differences in stiffnesses hexagon head s and A7 plates were used to create
may occur from calibrator to calibrator depend- a n opportunity for creep and load relaxation
ing on the condition of the unit, the oil in because of the very high clamping forces possi-
the cylinder, etc. A Skidmore-Wilhelm type ble. The results of these tests have been
calibrator may indicate from 25 percent to as summarized in Figure 21 and in Table 2.
much as 75 percent more turns to reach a given From Figure 21 it is quickly apparent
bolt load or bolt tension than are required for that, especially without washers at the plate
an actua l steel assembly, depending on the su rface where turning occurs, the pre- load in
number , t hicknesses, and types of plies included the bolts may not reach the appropriate proof
in the gr ips. load after snug p l us 1/2 turn. This may be
The initial portions of the load- the result of three or more effects: (1) The
torque relationships for the A325, A354 SC, and high level of proof load in A354 bolts requires
A354 SO bolts are all similar as shown in more bo l t deforma tion and produces greater
Figure 20. The differences in behavior appear plate deformation than does the load in A325
in the loads for which the relationship is no bolts; (2) the pl a t e surface galls or brinel l s
longer linea r . With the famili ar formula at while some plate material i s actually worn
the top of Figure 20 an estimate can be made of out from under the nut at the very high loads;
the t orque required t o reach a given pre-load and (3) the load-turns relationship is still
(below, say, the bolt proof load) for any A325 quite steep near the snug plus 1/2 t u rn level
or A354 (A490) bolt, provided a hardened washer and thus is sensitive to any errors, especially
is used under the nut and the threads are clean those errors which may result in Jess than one-
and as recei ved . half turn from snug (Figures 17 and 18).
2. Relaxation. The relaxation be- Two of these bolts were subsequently
havior of A325 bolts has been described in torqued to failure, and the resulting maximum
earlier sections of this report. In a similar loads are also depicted i n Figure 21. One
21
specimen, NX-24-D-3, reached a maximum load of one minute load ranged from less than one
52.5 kips after an additional 225 degrees beyond percent to more than five percent with an
the snug plus 1/2 turn level and a maximum average value of just over three percent . The
torque of 840 foot-pounds at 315 degrees beyond relaxation in the first five minutes was about
the 1/2 turn stage. The other bolt reached one percent of the one minute load. After
' beyond the
48.5 kips at about 300 degrees 90 days the bolt load was sti 11 over 95 per-
initial snug plus 1/2 turn. The maximum torque cent as great as that recorded a t the end of
later reached 1450 foot-pounds and when fracture t he first minute . Thus, relaxation does not
of the bolt occurred, the threaded end of the appear to be a major problem with the A354
bolt was locked in place because of the large (A490) bolt even when used with ASTM-A7
able change in microstructure of the grade 2H Shear. At the time that tests were being made
nut, but the cold worked A7 plate was found to in another study on A325 bolts under combined
have doubled in hardness in the 1/64-inch layer tension and shear, tv10 similar series of tests
below the surface under the nut. The deforma- ;~ith A354 BD bolts were included. In the
tion and galling would certainly have been interest of making possible a comparison of
marked ly reduced if a higher strength plate A325 and A354 bolts in this report, some of
material had been used. the data are included . The complete report of
Use of a washer under the Grade 2H the tests can be found in Reference 13. All
nut considerably reduces the wear or galling fasteners were tightened to approximately the
and increases the probability of reaching at bolt proof loads. The average results are
least proof load with snug plus 1/2 turn for presented in Figure 22 in terms of the bolt
the A354 type fastener. However, only one test tensile and shear strengths a nd the a ctual
under the head and nut. It shou 1d a 1so be noted In general, the A325 bolt seemed t o
that in Figures 15 through 18 and 20 the have a s l ightly greater shear capacity rel a tive
calibration curves also had washers at the nut. to its tensile strength than did the A354 BD
considered (Table 2) there is little difference tion of cross sectional areas, either type of
in the behavior of A354 and A325 bolts. The bolt is only about 80 percent as strong in
losses in bolt tension from maximum load to a shear through the threads as through the shank.
22
These data and those in the more complete study is interesting to note that the behavior of the
of combined tension and shear(l 3) suggest that maximum hardness A325 bolt would probably be
similar relationships exist for the usual quite similar to that of a low hardness A354
structural sizes of bolts . For large grips one BD bolt .
might expect a somewhat different behavior.
and the 1961 and 1963 American Institute of require an additional 1/4 turn for each outside
Steel Construction specificat ions. The A354 face in the grip with a 1 to 20 s lope.
BO bolt had a thread shear strength comparable In late 1962 it was reported that a
to the shank shear strength of low hardness number of new field tests did not support fully
A325 bolts and a tensile strength about one- the requirement for additional turns on a
third greater than that of those A325 bolts . I to 20 slope. Simultaneously, the University
For shear through the shanks in both cases, the of Illinois had undertaken a pilot study to
A354 BD bol t was about one-th ird stronger than explore the relaxation behav i or and fatigue
the A325 low ha rdness bolt. These comparative behavior of bolts on 1 to 20 slopes. The
strengths are generally in agreement with the L & S (Lamson and Sessions} load analyzer(?}
comparative proof load requirements for the two was used to assist in these studies . Because
types of bolts, although it must be remembered of the Importance of determining accurately
that the A354 bolts tested in this study were the extra turns required with beveled surfaces;
actually about 14 percent st ronger than the and in order to provide maximum protection
mi nimum required. With these facts i n mind it against twist-off with the new, shorter t hread
23
lengths, the study was expanded to include a load. Furthermore, for the 3/4-inch A325
number of calibrations on sloping surfaces. bolts, the proof load was reached between 1/4
The tests were performed in the same steel and 3/8 turn from snug when flat surfaces were
load cells (cylindrical dynamometers) used in used, but more than 1/2 turn was needed for
the relaxation tests (Figure 1), in a Skidmore- proof load when two I to 20 sloping surfaces
Wilhelm device, and in the L & S load analyzer. were in the grip.
One of the problems encountered in In Figure 24 one can see that with
such tests is that of controlling the accuracy two I to 20 bevels in the grip, 1/8 to 3/8
of measuring turns from snug and the decision turn additional is required to meet proof
of what bolt tension should be selected for load for the l-inch bolts , depending on the
snug. For a number of years, a 5,000 pound extremes considered. On the average, 1/4 turn
axial load has been used a t the University of additional is needed to insure proof load.
Illinois. This snug load seems to be a reason- Thus, the results of tests with l- inch diameter
able lower level for 3/4- a nd 7/8- inch bolts. bolts agree with data shown in Figure 23 for
Therefore, al l data have been adjusted to this 3/4-i nch diameter A325 fasteners.
level. In the elastic or linear portion of Through study of the detailed data
the load- turns relationship, a difference of from these tests, it is possible to present
10 degrees in turning will produce a difference the following observations:
of approximately 2,000 pounds in load. However, (I) Pneumatic tightening seemed to
as the bolt yields, a variation in turns has a produce faster seating of the head and nut on
much smal ler effect. the bevels . For example, in the tests of
Table 3 is a summary of a l l test data l-inch diameter bolts, nut seating occurred
which were reviewed in this study on effects of near ISO degrees in manual tightening and at
beve l ed surfaces. With the exception of tests about 75 degrees for pneumatic tightening;
involving one bevel or parallel bevels, these head seating occurred near 270 degrees in
data are represented in Figures 23 and 24. manual tightening and near 225 degrees in
Thus a wide variety of factors and a large pneumatic tightening. Pneumatic tight ening
number of tests are portrayed in the curves. tended to reduce slightly the number of turns
In Figure 23 it may be noted that at required to reach a given bolt tension .
proof load and with two beveled surfaces, on (2) Hardened washers tended to reduce
the average 1/4 additional turn is needed to slightly the turns necessary to obtain a given
reach the same bolt tension at and above proof bo lt load, possibly because of less wearing at
24
the contact between the nut and the surface. lit tl e or no effect; a t longer grips the effect
in the grip reduced the number of turns to (7) It has a l so been found f rom the
(4) Tests in load cells, Skidmore- of A325 bolts on the 1 to 20 sloping surfaces
Wilhelm calibrators, and the L & S load ana lyzer is essentially of the same magnitude as t ha t
al l show this consis t ent pattern: addi tional found wi th non- sloping surfaces. An initial
turns are needed to insure proof load when loss from the peak bolt tension o f about 5 per-
(5) The effect of one bevel in the by a loss of about 5 percent from the one
grip was generally intermediate between the minute load over a period of two days.
results fo r no bevel and two bevels. Usually, A convenient method of achieving the
proof load was attained in snug plus 1/2 turn. desired additional turns in t he field would be
(6) Relative position of t he bevels- to specify 1/4 additional turn for two bevels
difference Jn behavior. For converging bevels, (1 to 20) bevel since proof load on an A325
positioning the bolt toward the top or bottom bolt is usually reached in less than 1/2 turn
of the bevel made a difference in behavior. of the nut . This ~1ould provide at least proof
Placing the bevels so t hey were crossed (as load in almost every case presented in t hese
may occur in fas tening crossing beams) produced data, and in addition would provide for a
a more consistent and median pattern. When t he slightly greater reserve agains t twist- off
beve l s were paral lel, the effect appeared than the requirement stipulated in the 1962
tests were made. The first of these in- using the tu rn-of-nut method, to serve as a
eluded Specimens I through 9 and the second basis of comparison between the resu l ts of
included Specimen Types A, B, C, and D as the present program of t ests and the results
shown in Figure 3 and Tables 4 and 5 . The of similar tests conducted in 1957. Despite
second set of specimens was designed with this reduction in clamping force to a minimum
extreme conditions; oversized holes, finished level, the specimen did not fait.
rather than heavy hex-head bolts and nuts, Four specimens with the outer plates
no washers, and in most cases minimum clamp- critical were tested in the first series
ing were employed. The fatigue tests may be (Table 4) and only one failed ; al l specimens
sub-divided into two groups -- those with the of the second series (Table 5) had higher
center plate critical (i.e., having the stresses in the outer plates.
higher nominal stresses) and those with the If we look first at Table 4 we find
tests (Figure 3 and Table 4) we find that ksi stress cycle. And, a l thoug h failure oc-
several different combinations of nuts, bolts, curred in the outside plate, a close examina-
and washers were used to determine whether tion of the fracture indicated that the point
the omission of washers would reduce the fa- of initiation of the crack was on the inside
tigue life of the members. None of these face of the plate and that failure resulted
specimens failed within 2,000,000 cycles even from fretting rather than from bolt head de-
though they were tested at a stress cycle of pression or any other stress concentration.
diameter), slipped into bearing during the specimens were generally two million cycles
first cycle of loading and ran 3,073,900 or more, despite the severe conditions of
cycles at 0 to 30 ksi without failure. To oversized holes, high pressures under the
determine if tightening a bolt by turning head and nut, etc. Failure occurred in only
the bolt head, rather than the nut, might four specimens, disregarding the tests of
produce a difference in fatigue life, Speci- Specimens 02 and 03 which were discontinued
men 8 (without washers) was assembled with because of excessive heat a nd pounding. The
two bolt heads and two nuts torqued on the failures occurred in the series B, C, and 0
same side of the specimen. Neither the specimens with minimum specified bolt tension,
omission of washers nor the torquing of the and then only two of these failures initiated
heads contributed to an early fatigue failure. at less than two million applications of
stress of 20.4 ksi, a value just above the The failures in Specimens Bland
current AISC design load for A7 structural Cl appear to have started on the inside face
stee l, a fatigue failure did not occur. The of an outer plate and not at the depressions
test was discontinued after 5,148,000 appli- caused by the bolt heads (Figure 25). The
cations of a 0 to 30 ksi stress cycle. Speci- appearance of the fractures suggests that
men 9 was assembled in a manner similar to fretting rather than the bolt head depressions
that used for Specimen 8, except that l-inch or other stress concentrations at the holes
diameter bolts were used. Slip occurred in initiated the failures. Figure 25 also pre-
the first cycle at 24.8 ksi . This test was sents a typical fatigue failure for specimens
discontinued without failure after 4,892,000 with the outer plates critical and indicates
cycles of loading at 0 to 30 ksi . Thus, even th~ type of seating action that occurred under
without washers, the joints sti 11 have a large the bolt head. Galling of the plate surfaces
fatigue resistance, providing the bolts a re similar to that shown in Figure 7 occurred
tigue spec imens of the second series (all with in both an outside plate and the center plate
outside plates critical) were tested at no- at their respective critical sections but may
minal stress levels of either 0 to 30 ksi (or have been i nfluenced by fretting. Although
slightly less) or -20 ksi to+ 20 ksi. At Specimen 01 slipped into bearing dur•ng the
either stress level the fatigue lives of these first loading cycle, it did not continue to
27
slip to any significant extent when subjected B. ANALYSIS OF FATIGUE TEST RESULTS
wer e discontinued because of the excessive them with those of s i milar tests conducted
slip and pounding that developed during the at the University of Illinois and Northwestern
early part of the tests. Examination of these University (Table 6). The zero-to-tension
specimens indicated it was probable that use tests as well as the full reversal tests of
of lubricating oil rather than water-soluble this and the other programs compare very fa-
cutting oil during the drilling operat i on was vorably as far as specimen life is concerned.
responsible for the low frictional resistance, Specimen Cl (Table 5) is unique be-
Because of the difficulty encountered tensile area than shear area and can be com-
in testing Specimens 02 and 03, it was decided pared with the performance of Specimen 6 in
to gradually increase the stress on the nominal Table 4. I t can be seen that for a stress
net sections of Specimens 04 and 05 from+ 16 range of 0 to 30 ksi, the fatigue lives of
ksi to+ 20.0 ksi. Failure occurred only in both specimens were essentially the same,
Specimen 05 and then only after 2,267,000 cy- even though Specimen 6 was designed with a
cles. The fracture initiated in an outside tension:shear · ratio of 1.0 to 0.90 and
plate slightly below the critical section at a Specimen Cl with a tension:shear ratio of
point which suggested that fretting between the 1.0: 1.25. It can also be seen that these
edge of the bolt head and the surface of the higher shear stresses (nominal shear stresses)
outside plate was responsible fo r the failure . did not materially affect the fatigue life un-
This unusual failure can be seen in Figure 26 . der zero-to-tension loadings when high clamp-
In Figure 26a the extent and shape of the fa- ing forces existed in the bolts . This was
tigue crack is shown. The probable locations true despite the greater probability of slip
of the initiation points for the fretting fail- at the higher shear stresses.
ure have been marked by arrows in Figure 26b. Although the full reversal tests of
Once again , the connection withstood more than Specimens 02 and 03 were discontinued after
2, 000,000 cycles of reversal at stresses which relatively few cycles of loading, it is in-
were well above current design levels. teresting to note that their behavior was
28
which was also tested in full reversal. The I. Clamping Force. In order to
fatigue life of this latter specimen, which obtain an indication of the clamping force
was tightened to only 83 percent of the mini- in the bolts of the fa t igue specimens, the
mum bolt tension now required by the specifi - elongation of each bolt of certain specimens
6
cations, ( ) was well below 2,000,000 cycles. was measured periodically during the tests
This low fatigue life, although partly a re- by means of an extensometer. This e longa-
suit of the low clamping force in the bolts, tion measurement made it possible to estimate
also can be attributed to the presence of oil the tension in the bolts from the load-
on the faying surfaces of that join t. elongation curves for similar bolts. The
The average results of the fatigue initial bolt tension obtained from t hese
studies reported herein are presented in the measurements has been indicated in Column 10
Modified Goodman Diagram in Figure 27. This of Table 4 and in Column 4 of Table 5.
diagram was constructed on the basis of stress After completion of the tests,
on the nominal net section of the specimen and measurements were again made of the change
a fatigue life of 2,000 , 000 cycles. Since the in length of the bolts as they were removed.
life of most bolted joints was more than With this information the tension remaining
2,000,000 cycles, the plot is conservative. In in the bolts at intervals during the tests
comparing the test results with current bridge and at the end of the tests could be esti-
design specifications for A7 steel, it can be mated . Figure 28 gives a graphical represen-
seen that the bolted specimens exhibited a tation of the average loss in bolt tensions
fatigue life of 2, 000,000 cycles or more at during these fatigue tests. The loss in t en-
stresses approximately 60 percent greater than sion for bolts in the fatigue specimens that
those permitted by the specifications. did not fa i 1 was on the order of IS percent,
Figure 27 also serves to compare the the major portion of which occurred duri ng
fatigue behavior of bolted and riveted the first few cycles of loading. This beha-
( 14)
joints. The test results curve for the vior agrees well with the results reported
test results curve for the riveted joints, Reference 8 losses in pre-load are repor ted
as fatigue strength is concerned, in the use of 375,000 cycles and about 20 to 35 percent in
high-strength bolts for joints subjected to 540,000 cycles with further rapid decreases
29
as joint slip increased greatly . The bolt critical sections which eventually failed in
pre-stress in Reference 8 was about one- fatigue were found to be reduced considerably
third of that which would be used today, and in the fi rs t few cycles of loading and to
substantial slipping occurred . In Reference about half the initial bolt load when failure
17 with bolt tensions about equal to those occurred. This suggests that there may be
presently specified, the losses in pre-load meri t in periodic checks and re-tightening
varied from about 7 Lo 35 pe rcent depending at the critical section of bolted joints sub-
bo It t ens i on reported in Reference ~~~' were In t he study by Baron and L~rson the
and no joint fa ilure s deve loped . (I 3/ 16-in ch grip) varied from 20,000 to
The results reported herein and in 40 , 000 pounds before the tests and had de-
Reference 16 indicate that approx imately a creased approximately 27 percent by the end of
3 to 5 percent loss in bolt te nsion may be the tests. (IS) These specimens were as-
attributed to plastic flow in the bolts and sembled with washers under the nuts and under
plate materials caused by the high bolt ten- the bolt heads , and the bolts were tightened
sion . Much of the additional loss in bolt to a predetermined torque. I n most cases
load appears to be proportional to the ap- Baron's specimens slipped into bea r ing during
plied stress and attributable to the applica- the first loading cycle . Thus, the loss in
tion of static loading and , among other factors, bolt load during test ing fo r specimens with-
6 out hardened washers was found to be compara-
to any slipp ing which may have occurred. (! )
I t should be noted that many of the fatigue ble to the loss of bolt tension in joints as-
specimens s lipped into bearing during the semb led with washers.
first loading cycle and that the applied 2. Nominal Coefficient of Friction
stresses at first slip were generally welt The nominal coefficient of friction between
above design levels . Bolt tensions at t he the specimen plates is a significant factor
*Two fatigue tests of Reference 18 were since it, along wi th the clamping force, de-
made with bolts torqued into the yield
range . These tests appear to be the earliest termines whether the joint wi II slip or behave
in which it was shown that such high torquing
did not impair the structural joint behav ior . as a friction joint. The value of the nominal
The same study also sho~~d that with washers
under the nut, there was no more loss in clamp- coefficient of friction for each fatigue speci-
ing for bolts having regular finished hex nuts
than for those tightened with heavy hex nu ts . men which slipped wi II be found in Tables 4
30
and 5. These values are based on load-slip The friction values shown in Tables
readings taken on each specimen during the 4 and 5 are generally smaller than the values
first cycle of loading. The nominal coef- of 0.30 to 0.35 frequently reported for mi 11
ficient of friction was computed on the basis seale surfaces. However, the va I ues of the
of the load on the net area and the bolt present tests are based on r elatively small,
clamping force as determi ned f rom the bolt four-bolt, double- lap, shea r-type joints as
Because this report is primarily nuts are used, but the clamping
a resume of a number of studies, much of the force will sti 11 be in excess of t he
preceeding text consists of summary state- A325 proof load . There is no signi-
ments and conclusions. Only the principal ficant difference in the amount of
results and conclusions are repeated in this bolt tension lost with time between
section. The reader is reminded also that specimens with washers and those
these studies generally have been made on without washers.
limited numbers of laboratory specimens. Based on the tests repo r ted herein
On the basis of the tests report ed It appears that A325 and A354 BC and BD (or
and reviewed herein it may be concluded that: the A490) bolts may be compared as follows:
1. Joints properly assembled with I. Both types of bolts have a torqued
high-strength bolts and no washers strength of about 85 percent of the
can be expected to perform in about tensile strength. The same approxi-
the same manner as those assembled mate torque- tension relationship,
with washers. T = 0 .2PD, holds for both types
2. Bolt assemblies without washers may when hardened washers are provided
be expected to develop very nearly at the turning surface.
the same tensions as assemblies 2. Reducing the number of threads in
with washers when tightened by the t he grip (going from old thread
turn- of- nut me thod. In the case of Iengths to new thread Iengths) in-
oversized holes , up to 1/8 inch creases maximum bolt load at fail-
greater in diameter than the bolt, ure, but decreases the number of
there may be some red uction in bolt turns to failure in both A325 and
tension when washers are omitted A354 bolts.
and when f inished hex head bolts and 3. The A354 bolt (particula r ly grade
31
32
2. The major portion of the loss in into bearing, as well as those that did not;
bolt tension during the fatigue on joints with extremely severe combinations
life of a specimen tends to occur of hole size, fastener head size, and nut
in the first few cycles of loading. size; and in many cases on joints assembled
This loss in bolt tension appears with minimum allowable clamping in the high-
It should be noted that these last connections with less severe conditions.
34
IS. F. Baron and E. w. Larson, Jr., "Re- 19. W. H. Hunse, "An Eva luation of the Be-
sults of Static and Fatigue Tests of havior of Structural Connections As-
Riveted and Bolted Joints Having sembled with Hook-Knur l Bolts,"
Different Lengths of Grip," An un- Unpublished Private Report. Urbana,
published report of an investigation Illinois, September 1957.
conducted in the Department of Civi I
Engineering at Northwestern Un iver- 20. W. H. Hunse , 0. T. Wright, and N.H.
sity, January 1952. Also summarized Newnark, "Laboratory Tests of Bolted
in ASCE Transactions, Vol. 120 J oints, " ASCE Transactions, Vol . 120
(1955), Paper No. 2778, Pp. 1322-1334 . (1955). Paper No. 2778, Pp. 1299-1 321.