Embed Size (px)
Citation preview
I LL INO I SUNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN
PRODUCTION NOTE
University of Illinois atUrbana-Champaign Library
Large-scale Digitization Project, 2007.
ABSTRACT
EXTENSIVE TESTS OF INDIVIDUAL A325 BOLTS
WERE MADE TO EVALUATE THE EFFECT OF WASHERS ON
THE CLAMPING FORCE DEVELOPED IN HIGH-STRENGTH
BOLT ASSEMBLIES TIGHTENED BY A TURN-OF-NUT
METHOD. THE RESULTS OF THESE TESTS INDICATE
THAT THERE IS NO SIGNIFICANT DIFFERENCE IN THE
CLAMPING FORCE OF STRUCTURAL BOLTS, EITHER
WITH OR WITHOUT WASHERS. SIMILAR RESULTS WERE
OBTAINED WITH A354 GRADE BD BOLTS.
A STUDY OF THE CHANGES NECESSARY IN THE TURN-
OF-NUT METHOD WHEN THE SURFACES AT THE HEAD AND
NUT OF THE BOLT WERE ON FIVE PER CENT SLOPES
INDICATES THAT ADDITIONAL TURNING IS REQUIRED
TO INSURE PROPER TIGHTENING.
FATIGUE TESTS WERE CONDUCTED TO DETERMINE THE
EFFECT UPON JOINT LIFE OF OMITTING THE WASHERS
FROM THE BOLT ASSEMBLY. ALL SPECIMENS WERE
FOUR-BOLT, DOUBLE-LAP, SHEAR-TYPE JOINTS AND
INCLUDED CONNECTIONS DESIGNED WITH EITHER THE
INSIDE OR THE OUTSIDE PLATES CRITICAL. SPECI-
MENS WERE GENERALLY ASSEMBLED WITH 3/4-INCH
DIAMETER FASTENERS. ON THE BASIS OF THESE AND
OTHER TEST RESULTS, IT APPEARS THAT THE FATIGUE
LIVES OF PROPERLY BOLTED (A325) JOINTS ASSEMBLED
WITHOUT HARDENED WASHERS ARE GENERALLY AS LONG
AS THOSE OF JOINTS ASSEMBLED WITH WASHERS.
CONTENTS
I. INTRODUCTION . . . . . . . . . . . . . . . . .A. Use of High-Strength Bolts . . . . . . .B. Object and Scope of Investigation . . . .C. Acknowledgments . . . . . . . . . . . . .
II. DESCRIPTION OF TESTS . . . . . . . . . . . . .A. Relaxation Tests . . . . . . . . . . . .B. Specimens for Tests with Sloping SurfacesC. Fatigue Test Specimens . . . . . . . . .
III. TEST RESULTS . . . . . . . . . . . . . . . . .A. Relaxation Tests of A325 Bolt AssembliesB. Relaxation Tests of A354 Bolt Assemblies
and Comparison of Behavior with A325 BoltAssemblies . . . . .. . . . . . . . . . .
C. Effects of 1 to 20 Bevels on Bolt Tension
IV. RESULTS AND ANALYSIS OF FATIGUE TESTS . . . .A. Results of Tests . . . . . . . . . . . .B. Analysis of Fatigue Test Results . . . .
. . . . 11
. . . . 11
. . . . 17
. . . . 22
. . . . 25
. . . . 25
. . . . 27
V. CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . 31
VI. REFERENCES CITED . . . . . . . . . . . . . . . . . . . 34
FIGURES
1. Typical Assemblies for Relaxation Specimens
2. Typical Fatigue Specimen Showing Mechanical Dials for"Slip" Measurements
3. Details of Fatigue Test Specimens
4. Record From Sanborn Recorder
5. Nut Tests - Effect of Washers on Clamping Force of A325
Bolts
6. Nut Tests - Effect of Washers on Torque of A325 Bolts
7. An Example of Severe Galling that Resulted from Turning
the Nut Against the Plate Without a Washer
8. Depression of Plate Surface Resulting from Bolt Head
9. Bolt Head Tests - Effect of Washer and Hole Size
10. A325 Bolt Head Tests - Head or Nut Torqued; Effect of
Hole Diameter on Bolt Load and Torque
11. A325 Bolt Head Tests - Head Torqued; Effect of Head
Size on Torque
12. Effect of Hole Size on Bolt Load
13. Load-Turns Relationship for Heavy and Finished nuts used in
These Tests
14. Effect of Hole Size on Torque for 5000 Lb. Plus One-Half
Turn
15. Comparison of Load-Elongation Behavior of Various BoltTypes at Identical Grip
16. Comparison of Load-Elongation Data for Different Grips
17. Comparison of Load-Turns Data for Various A325 and A354Bolts at Identical Grip
18. Comparison of Load-Turns Data for Different Grips A325and A354 Bolts
19. Comparison of Turns to Failure for Various Bolt Types,Grips, and Conditions
20. Load-Torque Relationships for Various Bolt Types
21. Bolt Tension at "Snug" of 5000 lb. Plus One-Half Turn forA354 BD Relaxation Tests
22. Interaction Curves of A325 and A354 BD Bolts Under CombinedTension and Shear Showing Fsilure Loads
23. Effect of Beveled Surfaces on Bolt Tension vs. Turns, 3/4in. Diam. A325 Bolts
24. Effect of Beveled Surfaces on Bolt Tension vs. Turns 1-in.Diam. A325 Bolts
25. Typical Fatigue Failure
26. Views of Fatigue Failure of Specimen D5
27. Modified Goodman Diagram for 2,000,000 Cycles
28. Average Loss in Bolt Tension vs. Applied Cycles in FatigueTests
TABLES
1. Results
2. Results
A354-BD
3. Summary
Beveled
4. Results
5. Results
6. Results
of Bolt-Tension Relaxation Tests for A325 Bolts
of Bolt-Tension Relaxation Tests Conducted with
Bolts
of Variables for Tests of Bolts Tightened on
Surfaces
of Fatigue Tests, Specimens 1 Through 9
of Fatigue Tests, Series A, B, C, and D
of Previous Fatigue Tests
I. INTRODUCTION
A. USE OF HIGH-STRENGTH BOLT ASSEMBLIES
For a decade after its adoption, the
high-strength structural bolt assembly included
a bolt with a finished hexagon head, a heavy
hexagon nut, and two hardened washers. Prima-
rily, the washers served three purposes:
(a) to protect the outer surfaces of the con-
nected material from damage or galling as the
bolt or nut was torqued or turned; (b) to as-
sist in maintaining the high clamping force in
the bolt assembly; and (c) to provide surfaces
of consistent hardness so that fairly depend-
able relationships between torque and tension
could be established for the various bolts and
bolt diameters.
Most of the early methods employed
for the development of tension in high-strength
bolt assemblies made use of a relationship be-
tween applied torque and bolt tension. How-
ever, recently erectors and engineers developed
This report combines most of the data pre-sented in two previously unpublished progressreports(1,2)T and other unpublished dataavailable on research with high-strength bolts.
Superscript numbers in parentheses refer toreferences cited.
and began using methods of tightening which
are independent of any load-torque relation-
ship. The reason for the change was the
desire for a simpler, faster, and possibly
more accurate means of tightening the bolt
assemblies. The objections sometimes offered
to the use of torque as a basis for bolt load
control are: (a) a lack of dependability
(because of differences in thread fit, thread
cleanliness, rust, etc.), (b) the possible
resulting variations in bolt tension through-
out a joint, and (c) inefficient use of pneu-
matic impact wrenches.
A preliminary study at the University
of Illinois (3 ) and an investigation conducted
by F. P. Drew for the Association of American
Railroads (4 ) led to a turn-of-nut method for
tightening high-strength bolt assemblies.
This method has provided a means of obtaining
at least the required minimum bolt tension
without reliance upon torque. In the turn-of-
nut method the bolt is tightened to, or above,
the required minimum tension by rotating the
nut or bolt a prescribed number of turns from
either a finger-tight or a snug position.
The "finger-tight" position is that to which
the nut can be tightened when turned by hand
after the joint has been drawn tight with
fitting-up bolts; usually this has been fol-
lowed by one complete turn (360 degrees) of
the nut or bolt head. The "snug" position is
the point in the tightening at which a pneu-
matic impact wrench ceases to run freely and
the impact mechanism begins to operate, or the
tightness attained by the full effort of a man
using an ordinary spud wrench. From the snug
position the bolt or nut is usually given
approximately a one-half turn (180 degrees) to
achieve the fully tightened position. However,
joints with warped, heavy, or many plies of
material may require that the bolts be touched
up by additional tightening to insure full
bolt load in all the fasteners. A more de-
tailed discussion of the turn-of-nut method as
used by some fabricators was presented by Ball
and Higgins.(5)
B. OBJECT AND SCOPE OF INVESTIGATION
Since the turn-of-nut method does
not depend on torque to control the bolt ten-
sion, the question was raised by structural
fabricators and erectors as to whether the
hardened washers could be eliminated from the
bolt assembly. By the same token, would only
one washer under the turned element be adequate
when the torque control or the calibrated
wrench method is used?
Three of the questions that must be
answered to determine whether one or both wash-
ers can be omitted from a bolt assembly with-
out any significant effect on its ability to
form a sound and effective connection are as
follows: (1) Can the required clamping force
be developed and maintained without the hard-
ened washers when a high-strength bolt assembly
connects several plies of structural materials?
(2) Is the relaxation of tension in an assembly
without washers the same as, or comparable to,
that for an assembly with washers? And (3) will
the deformation and galling of the connected
material which results from the tightening of
the nut directly against that material cause
the connection to be any less satisfactory,
particularly under fatigue loading?
To answer the first two questions
most of the types of nuts commonly used in
structural work were included in this study of
the effects of washers on the clamping force
in the assemblies. These types were: (a) a
heavy semi-finished hexagon nut, the only type
permitted prior to 1960 for structural use;
(b) a heavy-thin nut, which is the same dimen-
sion across the flats as the heavy nut, but is
only as thick as the finished-series nut;
(c) a finished-series nut which was permitted
as an alternate in the 1960 Specifications of
Because of the development of the heavy hex-agon head, structural A325 bolt and the result-ing lack of interest in finished nuts amongstructural users, the 1962 Research CouncilSpecification 6) mentions specifically only theheavy hexagon nut.
the Research Council on Riveted and Bolted
Structural Joints; (d) a finished-thick nut
which has the same across-flats dimensions as
a finished nut, but is as thick as a heavy nut,
and (e) a flanged nut which has the same
across-flats dimensions as the heavy nut, but
also has an integral washer on one face. Also
included in the tests were two types of bolt
heads: the regular semi-finished hexagon head
which has been specified in the past and a
heavy semi-finished hexagon head which was
approved as an alternate in the 1960 Specifica-
tions of the Research Council and became the
standard bolt in the 1962 Council Specifica-
tions. (6)
The program included a large number
of tests of single assemblies which are re-
ferred to as relaxation tests, wherein a
determination was made of the variation in bolt
tension with time using a strain gage load cell
within the grip as shown in Figure 1. The
variables in the relaxation tests included the
type of nut, bolt head size, bolt strength,
hole size, hole preparation, and the effect of
lubrication. In some of the relaxation tests,
the assemblies were tightened by turning the
bolt rather than the nut.
Other tests were performed with the
strain gage load cell and with a commercial
hydraulic load cell to determine the turns
necessary to properly tighten the assemblies
when one or both outer surfaces have a five
percent slope relative to the plane perpen-
dicular to the bolt axis. Variables in the
tests included the use of washers, relative
directions of sloping surfaces, position of
bolt, bolt hardness, method of tightening, and
bolt size.
The third question to be answered
concerning elimination of washers involved
twenty-one fatigue tests. The variables em-
ployed were type of nut, type of bolt head,
hole size, location of the critical plate
(inner or outer plates of a three plate spe-
cimen), element tightened (the bolt or the
nut), bolt tension, and bolt diameter. These
fatigue specimens were four-bolt, double-lap,
shear-type joints similar to the one shown in
Figure 2.
Based on these and other studies
recently performed at the University of Illi-
nois, as well as at other schools, the Speci-
fications(6) of the Research Council were
revised early in 1964.
C. ACKNOWLEDGMENTS
The tests described in this report
are part of an investigation being conducted
as the result of a cooperative agreement be-
tween the Engineering Experiment Station of
the University of Illinois, the Illinois Divi-
sion of Highways, the Department of Commerce--
Bureau of Public Roads, the Research Council
on Riveted and Bolted Structural Joints, the
American Institute of Steel Construction, and
the Industrial Fasteners Institute. The
investigation is a part of the Structural
Research Program of the Department of Civil
Engineering of which Professor N. M. Newmark
is head.
The work was carried out by the
following research assistants in Civil
Engineering, whose efforts are gratefully
acknowledged: M. D'Arcy, P. C. Birkemoe,
N. Faustino, A. C. Knoell, C. W. Lewitt,
J. F. Parola, and E. W. J. Troup. The studies
were performed under the general supervision
of W. H. Munse, Professor of Civil Engineering,
and under direct supervision by E. Chesson, Jr.
Associate Professor of Civil Engineering.
A number of the tests described in
this report were planned in cooperation with
Committee 8 of the Research Council on Riveted
and Bolted Structural Joints. The chairman
of the Committee was Mr. A. M. Smith, and its
members included E. R. Estes, Jr., N. G. Hansen,
T. R. Higgins, N. P. Maycock, A. K. Robertson,
J. L. Rumpf, and W. H. Munse. Additional
testing was coordinated with Committee 15 of
the Research Council chairmaned by
Mr. T. R. Higgins and having the following
members: L. S. Beedle, R. B. Belford,
F. H. Dill, E. L. Erickson, F. E. Graves,
W. H. Munse, E. J. Ruble, A. Schwartz, Jr.,
T. W. Spilman, D. L. Tarlton, and G. S. Vincent.
This manuscript was reviewed by
W. G. Waltermire, R. B. Belford, E. H. Gaylord,
and T. W. Spilman. The authors wish to thank
them for their careful analysis and criticisms.
II. DESCRIPTION OF TESTS
A. RELAXATION TESTS
As shown in Figure 1, the relaxation
specimens consisted of two 4 x 4 x 3/4-inch
steel plates (with a 13/16-inch diameter hole
in the center, except in the case of the over-
sized hole specimens), a 3/4 x 6-inch high-
strength bolt, and a load cell. The steel
plates were ASTM-A7 structural grade steel with
surfaces left in the as-rolled condition; how-
ever, in some cases the plates were rusted from
having been stored outside. Before the plates
were used,they were wire-brushed to remove any
loose scale or rust.
Data on the bolt and nut hardnesses
are in Tables I and 2. It is important to note
that there were significant differences in the
hardnesses and ultimate strengths of the regu-
lar- and heavy-head bolts used in the relax-
ation tests for A325 bolts. The heavy-head
bolts (Rockwell C = 24) were on the low sides
of the hardness and strength ranges allowed by
the ASTM Specifications while the regular-head
bolts (Rockwell C = 32) were on the high side
of the Specification values. The nuts varied
in hardness from a Rockwell B of 80 for the
finished-thick nuts to a Rockwell B of 92 to
95 for the heavy-thin nuts (Table 1, Column 6).
The ASTM-A354 bolts were specified to be grade
BDI and were near Rockwell C = 37, the maximum
hardness for that grade (Table 2).
The load cells (Figure 1) were
1 11/16 inches in diameter and 3 1/8 inches
long. These cells were designed to provide
average axial strains of 800 to 900 micro-
inches when subjected to the maximum load and
were made from AISI 4340 steel,heat-treated to
provide a high yield point to reduce the possi-
bility of local yielding. Each load cell had
eight strain gages (four placed axially and
four circumferentially) arranged in a four-arm
bridge for high sensitivity.
In the relaxation tests, the output
These nuts were obtained before the 1958changes in ASTM-A325.
It should be noted that the A354 BD boltshave mechanical properties similar to thoseprescribed for the new ASTM-A490 structuralbolts now incorporated in the 1964 edition ofReference 6.
of the load cell was recorded while the bolt
was being tightened with a manual torque wrench
and for five additional minutes to provide a
continuous graphic load-time plot. For tests
that were run more than 5 minutes, periodic
strain readings were taken during the remainder
of the test. Measurements were also made of
the change in length of most bolts as a result
of tightening. The effects of removing the
washers from under the bolt heads and from under
the nuts were studied separately. The members
were designated as "nut study" or "bolt-head
study" specimens depending on which part of
the bolt assembly was in contact with the mild
steel plates. The load cell was placed on the
opposite side of the plates from the nut or
head and was always faced with two conventional
hardened washers. The one exception to the
assembly just described was the one used in the
relaxation test of the specimen with punched
holes. In this case, the load cell was placed
between the 3/4-inch plates to permit the bolt
head to bear on one lip and the nut to rotate
against the other lip of the punched hole;
neither the bolt head nor the nut had a washer.
The variation in grip in the A354 bolt tests
required some changes in the plate thickness
and in the number of washers.
The variables in the nut tests in-
cluded, besides nut type, the hole size (the
customary 13/16-inch hole and oversized holes
of 27/32 and 7/8 inches to simulate reaming),
the method of making the hole (either punched
or drilled), and the effect of lubricating the
contact surface between the face of the nut
and the plate.
The bolt-head tests included over-
sized holes in some of the specimens, torquing
the head rather than the nut in four of the
tests, and regular as well as heavy heads.
Two specimens were tightened to a specific
bolt tension and all other bolts were tightened
using a turn-of-nut method. For the latter
method, each bolt was tightened to a tension
of 5,000 pounds to simulate the load in the
bolt when it is snugged up; then the nut or
bolt head was given a one-half turn (180 de-
gree rotation).
In order to identify the variables
studied in the different relaxation tests,
letters and numbers have been used to specify
the test conditions for each of the specimens.
The letter "N" or "H" before the specimen
number indicates whether the test was a nut
test or bolt-head test, respectively. A
specimen number which includes the letter "S"
identifies a short-time test (usually lasting
only 5 minutes after the bolt was tightened)
as opposed to those tests which lasted from 3
to 90 days. The letter "B" or "C" in the
specimen number indicates the use of over-sized
holes--"B" for the 27/32-inch diameter holes
and "C" for the 7/8-inch diameter holes. In
those cases where a specimen was tightened to
a specific bolt load, that load in kips appears
as the last figure in the specimen number. The
four specimens that were head torqued (tightened
by turning the bolt head) are identified by "HT"
in the specimen number. One specimen was tested
with a lubricant of the molybdenum disulfide
type between the nut face and the surface of the
plate; this specimen is identified by the letter
"L" in the specimen number. The letter "X"
(Table 2) identifies the A354 BD bolts used in
the relaxation studies and the letters "D", "E",
and "F" indicate a variation in the number of
threads included in the grip.
B. SPECIMENS FOR TESTS WITH SLOPING SURFACES
A thorough study of the behavior of
assemblies tightened on sloping surfaces would
require a systematic consideration of a large
number of factors. The following are of primary
importance: bolt head size and nut size
(distance across flats); grip and number of
threads in the grip; diameter of bolt and hole;
hardness or strength of bolts, nuts, plates and
washers; position of the bolt in the hole --
centered, on high side of slope (maximum slid-
ing), or on low side of slope (minimum sliding);
and method of torquing -- pneumatic or manual.
Secondary factors might include chamfered or
washer-faced nuts, multiple plies, flats or
corners on the high side, and torquing head or
nut. The systematic study of all these factors
Kip = 1000 pounds (kilo-pound)
would involve considerable time and expense
and is much more extensive than could be in-
cluded in this investigation. However, a
number of these variables have been explored
and are summarized in Table 3.
In general the specimens consisted
of a bolt and a suitable nut tightened against
sloping mild steel plate surfaces (mill scale
surface, as-received) with some type of load
indicating device included in the grip --either
a steel load cell (cylindrical dynamometer as
in Figure 1) or a commercial hydraulic load
cell (Skidmore-Wilhelm calibrator). A second
group of tests was conducted at the Lamson
and Sessions Company in a load analyzer,(7 )
which provided a record of load, turns, and
applied torque and also permitted an evaluation
of the relaxation of the fastener.
Accuracy of the Illinois load cells
was verified by recalibrations under axial
loadings, with eccentric loadings, and with a
check of the effects of residual torque. The
Skidmore-Wilhelm device was also calibrated
under axial and eccentric loadings. Both kinds
of equipment showed excellent response under
all conditions.
C. FATIGUE TEST SPECIMENS
The fatigue test specimens were
designed to supplement the relaxation tests on
A325 bolt assemblies and to determine the
effects of galling when washers were omitted
from the fastener assembly. Specimens that
were designed with the center plates critical
(center plate more highly stressed than the
outside plates) served primarily to indicate
whether the clamping force developed in the
bolts had been reduced when the washers were
omitted; previous fatigue programs have shown
that the fatigue lives of bolted structural
joints decrease with a reduction in the initial
tension in the bolt.( 3) Therefore, if there is
any tendency for a high-strength bolt without
washers to work its way into the surface of the
connected material and thus lose its clamping
force, a reduction of the fatigue resistance
could be expected to appear in these fatigue
tests. In those specimens where the outside
plates were designed to be critical (the outside
plates having the same or higher nominal stress
than the center plate), the objective was to
determine whether the galling on the outer plates
resulting from the tightening of the assembly,
combined with the high plate stress and possible
loss in bolt tension, would cause a reduction in
the fatigue resistance.
To provide a correlation with the data
from other tests, the dimensions (Figure 3) of
the specimens were similar to those used for
specimens in previous studies. (3 ) Tension:shear
ratios for all specimens are shown in Tables 4
The "tension:shear ratio" is the ratio of thenet section tensile stress to the shear stress onthe cross-sectional area of the bolts. The ten-sile area used in the calculations for a "nomi-nal" T:S ratio was computed using the nominalfastener diameter plus 1/8 inch regardless ofactual hole size. The shear area assumed wascalculated using nominal bolt diameter.
and 5. Hole sizes and fastener diameters are
given in Figure 3. All the specimens were
double-lap, butt-type joints fastened with high
strength bolts; all holes were match-drilled.
The bolt assemblies were tightened
by turning the nut, except for Specimens 8 and
9 (Table 4). The bolts in Fatigue Specimens
I through 4, 6 through 8, and Series A were
installed by the turn-of-nut method using a
pre-load or snug position of 5,000 pounds, the
same value as that used in the relaxation
tests. In Specimen 5 the assemblies were
tightened to a tension of 29,000 pounds, and in
Series B and D a bolt load of approximately
28,400 pounds was used, the latter correspond-
ing to the minimum permitted by Specifi-
cations. (6) In the case of the 1-inch bolts
of Specimen 9, the bolts were tightened to
10,000 pounds before adding the final one-half
turn. This change was based on tests which
indicated that the snug position for the I-inch
bolt was more closely represented by 10,000
pounds than by 5,000 pounds. For Specimen C-1
snug was obtained using 100 foot-pounds torque,
after which an additional 180 degree turn-of-
nut was applied.
The fatigue tests were carried out in
the 200,000 pound University of Illinois fatigue
machines which are fully described in Reference
8. The load was applied at the rate of approxi-
mately 200 cycles per minute to all specimens
except Specimens D3, D4, and D5. These speci-
mens were tested at a loading rate of 100 cycles
per minute to reduce the tendency for slip and
pounding which had developed in the second
Series D specimen.
All plates for the fatigue specimens
were of ASTM-A7 structural steel. The plate
surfaces were left in the as-rolled condition;
however, in some cases these surfaces were
rusted and slightly pitted from having been
stored outside. The plates were lightly wire
brushed before use to remove any loose material,
and the contact surfaces of the joint were
cleaned with a solvent to remove any cutting-
oil remaining from the hole-drilling operation.
The bolts for the fatigue specimens
met the hardness and ultimate strength require-
ments of ASTM-A325. Bolt and nut hardnesses
are shown in Table 4 for fatigue test Specimens
I through 9 where the hardness tended to be
low, but within the allowable range. In general,
for specimen types A, B, and D, the hardness of
the regular hex (hexagon) bolts tended to be on
the high side of the allowable Rockwell C range
(RC = 30); the hardness of the finished nuts
was RC = 26 (ASTM-AI94, grade 2H) . The regular
hex bolts of Series C haa d hardness of RC = 32
and the hex nuts in Series C had a hardness of
RB = 82 to 84.
In order to minimize the number of
tests in the second stage of the program, speci-
mens of Series A, B, C, and D were designed for
These nuts were from stock available at thelaboratory and procured prior to the 1958changes in A325 nut proof load.
what were considered to be the most severe
conditions possible-- oversized holes, minimum
clamping force in the bolts, and higher than
normal bearing stress under the bolt head and
nut. To accomplish this, all bolt holes
(except for Specimen Cl) were drilled 1/8 inch
larger in diameter than the nominal diameter
of the bolt to simulate excessive field ream-
ing, and regular series hex bolts with finished
hex nuts were used to give an unusually high
bearing stress at the plate surfaces under-
neath the bolt heads and nuts. Specimens with
less severe geometries would have been tested,
if necessary, to find the conditions required
to produce fatigue lives comparable to those
obtained with joints having two hardened
washers for each bolt; however, the severe
conditions noted above did not appear to
produce any significant reduction in the fatigue
strength of the joints and no further tests were
considered necessary.
First row slip was measured at each
edge of all the fatigue specimens by means of
mechanical dials (Figure 2). "First row slip"
is the relative movement which takes place
between the edges of the center plate and the
outside plates and includes both elastic and
plastic deformation as well as slipping of the
plates.
Where bolt tensions in the fatigue
specimens were reported, they were determined
by means of bolt elongation. A bolt similar
to those used in the fatigue specimens and with
the same grip was tightened in a Skidmore-
Wilhelm Calibrator and its elongation measured
at various increments of load. This information
made possible the construction of a typical
load-elongation curve that was used to es-
tablish the loads or tensions of the bolts in
the fatigue specimens.
III. TEST RESULTS
A. RELAXATION TESTS OF A325 BOLT ASSEMBLIES
In examining the Table I relaxation
data for A325 bolt assemblies it is necessary
to keep in mind the wide range in bolt and nut
hardnesses which existed in the specimens test-
ed. Because of this range, comparisons should
be made only between assemblies with comparable
hardnesses and with the same nut and bolt head.
Furthermore, since each type of nut and bolt
head was tested both with and without a washer,
the contribution of the washer to each assembly
can be studied without considering hardness as
a variable. In order to provide additional
information, one assembly of each type was
tightened in a Skidmore-Wilhelm Calibrator and
the bolt elongation was measured at various
increments of load. These measurements were
compared with elongation readings taken on the
relaxation specimens as a check on the bolt
tension recorded by the load cell.
In the relaxation tests a continuous
record of bolt tension versus time was obtained
from the load cell with a Sandborn Recorder; a
typical curve is shown in Figure 4. It was
evident from these records that immediately
upon completion of the torquing there was a
small drop in load. The load for each bolt,
recorded in Column 9 of Table 1, is the peak
load on the trace. However, because the maxi-
mum load in the bolt is only an instantaneous
load and rarely measured, even in the labora-
tory, this value has not been used as a basis
of reference. Instead, the load in the bolt
one minute after reaching the peak load has
been used as the base from which to determine
the percent loss in bolt tension with time
(Column 10, Table 1). The drop in load from
the maximum load to the one-minute load varied
from 2 to 11 percent, with an average of about
5 percent. This drop in bolt tension is
thought to be a result of an elastic recovery
which takes place when the wrench is removed
and possibly a result of creep or yield in the
bolt, produced by high stresses at the root of
the threads. In addition, some flow may occur
in the steel plates under the head and nut.
Similar tests made on a hydraulic
load analyzer by the Lamson and Sessions Compa-
ny(7 ) show similar percentages of bolt tension
losses immediately after tightening and over a
period of days. Comparable results were also
reported in Table 6 of Reference 9 for 7/8-inch
diameter bolts.
1. Nut Tests. A comparison of the
clamping forces that were obtained in the vari-
ous nut tests in which the regular hex head
A325 bolts were used is shown in Figure 5. The
bolts were tightened by the turn-of-nut method
(5,000 pounds + 1/2 turn). The figure shows
that there was no significant difference in
the one-minute bolt load whether or not a
washer was used; a variation in the type of nut
and nut hardness had a greater effect than the
washers. Nevertheless, in all cases the bolt
load was in excess of the proof load. The
number of threads included in the grip will
also affect the bolt's load-to-turns relation-
ship, but this was not varied in these par-
ticular tests.
The effect of the washer on the maxi-
mum torque required to tighten an A325 bolt
assembly is evident upon examination of Figure
6. The higher torque required to tighten nuts
without washers is the result of galling be-
tween the nut and the soft plate material.
One of the more severe cases of galling is
The "proof load" is that load which ac-cording to ASTM Designation A325-63 the boltsmust withstand without experiencing any perma-nent set. A tolerance of ± 0.0005 inch is al-lowed as the difference between the measure-ments made before and after the loading. The1960, 1962, and 1964 specifications of theResearch Council require at least a tensionequal to the proof load in the bolts asinstalled.
shown in Figure 7. The flanged nut produced
the greatest bolt load, yet the torque required
to tighten it was less than the torque required
to tighten the heavy and the heavy thin-nuts
(without washers) to slightly smaller loads;
this is probably a result of the larger
contact area of the flanged nut, reduced
contact pressure, reduced galling, and, there-
fore, reduced resistance to turning. The
torque required to turn the flanged nut was
greater, however, than the torque required to
tighten bolts with hardened washers under the
nuts because the flanged nut is turned directly
against the relatively soft plate material.
When a hardened washer is used, the nut is
turned against a hard surface and the galling
of the soft plates is eliminated. This may be
of importance where the speed of installation
and the size of the pneumatic wrench and the
air compressor capacity are factors.
In Figure 6 it is also important to
note that: (1) With washers under the nuts,
the torques for 5,000 pounds plus 1/2 turn
were relatively consistent regardless of the
nut types, and (2) there was considerable
variation in torque between the different types
of nuts when the washers were omitted. Clearly,
any method of installation of bolts must be
independent of torque if the hardened washers
are omitted.
2. Bolt Head Tests. The nuts, when
torqued, galled the plates if the washers were
omitted under them while the bolt heads pro-
duced depressions in the plates if their washers
were omitted. A typical bolt-head depression
is shown in Figure 8. The results of the bolt-
head tests, both with and without washers, are
compared in Figure 9; in all cases the bolt
heads were in contact with the plates or washers
and the bolt assemblies were tightened by turn-
ing the nuts. Also included for comparison are
the tests in which the holes were either
1/32-inch or 1/16-inch oversize. In the case of
the regular semi-finished hex-head bolts (A325
type) and the 13/16-inch diameter holes, there
was no difference in bolt load whether or not
there was a washer under the bolt head. When
the hole diameter was increased and no washer
was used, a small decrease in load occurred,
although the final load was still well above the
minimum tension now specified. For the softer,
heavy semi-finished hex-head bolts (A325 type)
there was no apparent loss in bolt load when
the washer was removed or as a result of the
oversized holes. The bearing area of the bolt
head around the perimeter of a 7/8-inch diameter
hole when a heavy semi-finished hex-head 3/4-inch
bolt was used (approximately 0.402 square inches)
was slightly greater than the bearing area of a
regular semi-finished hex-head 3/4-inch bolt
around a 13/16-inch diameter hole (approximately
"Bearing area" as used here is computed bysubtracting the hole area from the area of the
washer-faced portion of the bolt head, based onmeasured dimensions for these specimens.
0.380 square inches). This may account for
the fact that the heavy-head bolt, when
tightened in a 7/8-inch diameter hole, held
its load in much the same manner as did the
regular-head bolt when tightened in the
13/16-inch hole. However, even the regular-
head bolt, when used in a 7/8-inch hole and
without a washer, developed nearly the same
tension as did the bolt in a 13/16-inch hole
with a washer, when tightened to snug plus 1/2
turn.
3. Bolt Head Torque Tests. Several
relaxation tests were conducted after turning
the bolt head rather than the nut. The results
of these tests, conducted with the regular
semi-finished hex head A325 type bolts in
various hole sizes, are shown in Figure 10.
The oversized hole produced a small decrease
in the bolt load, but the maximum torque was
approximately the same. It is apparent that in
either case, omitting the washer and applying
5,000 pounds plus 1/2 turn produced more than
proof load in the bolts, but the required
torques were considerably larger than would
have been required for proof load only, if
washers had been used. Where the regular semi-
finished hex bolt head and the heavy semi-
finished hex bolt head were torqued to approxi-
mately the same load (Figure 11), a somewhat
smaller torque was required for the heavy-head
bolt, but the difference is no greater than
might develop from test to test with the same
head type. Note that the heavy-head bolt was
installed by 5,000 pounds plus 1/2 turn while
the regular-head bolt was tightened to approxi-
mately that same load, rather than by amount
of turn.
A comparison of the depths of the
depressions in specimens tightened by turning
the heads to approximately the same bolt load,
such as H-2S-34 (.0089 inch, regular head) and
H-4 (.0028 inch, heavy head), indicates the
marked reduction in the depth of depression
resulting from the increased bearing area
provided by the heavy-head bolt. Study of the
data on depth of depression and nominal bearing
pressures showed that, at a given nominal bear-
ing pressure, specimens tightened by turning
the bolt heads had deeper head depressions than
similar specimens tightened by turning the
nuts. ) It appears that as the heads were
turned mechanical wear of the plates produced
the greater depressions. However, there
seemed to be little difference in the load
produced whether the head or nut was turned.
4. Effect of Hole Size. The results
of tests run with several types of nuts and
regular bolts in different size holes are shown
in Figure 12 for easy comparison. For the
finished nuts (Figure 12a) there was little
variation in load as a result of removing the
washer or as a result of turning the nut over
an oversized hole. However, in the case of the
heavy nuts (Figure 12b) there appeared to be a
small decrease in load when the bolt was
tightened in a 7/8-inch diameter hole. The
flanged nuts (Figure 12c) developed con-
sistently high bolt loads with both hole sizes.
The difference in behavior between
the finished and the heavy nuts in the case
of the 7/8-inch diameter hole may be the
result of the differences in the load-turns
relationships for the two types of nuts and
may be caused by differences in geometry and
hardness. In order to explore this problem
further, the load-turns relationships for
bolts with either the heavy or the finished
nut were determined and are plotted in Figure
13. Note that for the softer finished nuts,
the bolt load at 5,000 pounds + 1/2 turn was
considerably less than in the case of the
heavy nut. It can also be seen that the maxi-
mum tension attainable with the finished nuts
and washers was only 34,000 pounds, a value
only slightly above that produced by 5,000
pounds + 1/2 turn. Harder finished nuts would
have produced higher bolt tensions and might
have produced a load-turns relationship compa-
rable to that of the heavy nuts.
It is interesting to note the wide
range in bolt loads obtained between bolts on
the high and low side of the specifications.
Comparison of the test results in Table I for
Specimens N-5 and N-5S (40,450 pounds) with
the results of N-17S (33,700 pounds) emphasizes
this difference. It is also worthy of note
that the flanged nut consistently developed a
load equal to or greater than the load
NUT 3TUDY SPECIMEN SPECIMEN
FIGURE 1. TYPICAL ASSEMBLIES FOR RELAXATION SPECIMENS
FIGURE 2. TYPICAL FATIGUE SPECIMEN SHOWING MECHANICAL DIALS FOR "SLIP" MEASUREMENTS
h- Actual Hole Diam. = d+ 4-+ a + + + +
+ + + + a 4
e p e
Critical Section for
b Outside PlatesI i- %\ .9.
Critical Section forCenter Plate
F b I I T 1I '
SPECIMEN BOLT CRITICALDIA. PLATE a b c d e f g h p
I thru 5 3/4 Center I 3/8 3/8 5/8 13/16 1 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 Outside I 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 1 1/4 3/8 3/4 13/16 1 1/4 5 3/4 2 3/4 1/4 3 1/4
9 1 Outside 2 3/8 3/4 1 1/16 2 1/8 9 4 1/2 1/4 3 1/4
Specimentypes A, 3/4 Outside 1 1/4 3/8 3/4 7/8 1 1/4 6 1/4 2 3/4 1/4 3 1/4B, C,and D
Cl 5/8 Outside 1 1/4 3/8 3/4 11/16 1 1/4 6 1/4 2 3/4 1/4 3 1/4
FIGURE 3. DETAILS OF FATIGUE TEST SPECIMENS
0. 30
S 20
o 10mBFIGURE 4. RECORD FROM SANBORN RECORDER
- --- One Minute -Load
J iaximum Load
I Spec. No. N-14SSHeavy Nut Thin
No Washer
Regular Head Bolt
0 0.2 0.4 0.6 0.8 1.0
Time, Minutes from Maximum Load
I • I I : , I 'l* IJl I , I. .. I c
40
a 30
0
- 20
10
0Heavy Nut-Thin Finished Finished Flanged
Nut Nut-Thick Nut
x 6" Regular semi-finished hex head bolts, Tightened to 5000 lb. + 1/2 TurnPlots are average values where duplicate tests were conducted.
FIGURE 5. NUT TESTS - EFFECT OF WASHERS ON CLAMPING FORCE OF A325 BOLTS
Heavy Nut-Thin
01, SA[
iroximate torque requproduce Proof Load wihers are used.*
0Finished Nut Finished
Nut-Thick
6aA
700
600
50o
200
FlangedNut
4.,
1.
te a* 6
a Q
* 0
I.* Based on T - 0.2 PD
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
HeavyNut
3/4
800
700
600
500
400
300
200
100
HeavyNut
__ m40 .
30
20
" i
10 L
0
J2
E
LI
- 0
II
>
w
FIGURE 7. AN EXAMPLE OF SEVERE GALLING THAT RESULTED FROM
TURNING THE NUT AGAINST THE PLATE WITHOUT A WASHER
SPECIMEN NO. aREGUILAR BOLT HEAD
No WASHERFIGURE 8. DEPRESSION OF PLATE SURFACE
RESULTING FROM BOLT HEAD
13/16 13/16 27/32 7/8
Hole Diameter, Inches
3/4 x 6" bolt, Tightened to5000 lb. + 1/2 Turn
Regular Semi-Finished Hex Head Bolt(High Hardness Bolt)
FIGURE 9. BOLT HEAD TESTS -
U
* 0
ii
* 30
S20
o10
013/16 13/16 7/8
Hole Diameter, Inches
3/4 x 6" Bolt, Tightened to5000 lb. + 1/2 Turn
Heavy Semi-Finished Hex Head Bolt(Low-Hardness Bolt)
EFFECT OF WASHER AND HOLE SIZE
-- -Torqued HeadV/IZerATorqued Heavy
Nut
'roof Load (28,400 lb.)
0 L..U.&.AU..J13/16 7/8 Hole Size 13/16 7/8
3/4 x 6" Reg. semi-finished hex head bolts(no washers)
FIGURE 10. A325 BOLT HEAD TESTS - HEAD
OR NUT TORQUED; EFFECT OF HOLE DIAMETER
ON BOLT LOAD AND TORQUE
Regular Need
-- ýHeavy Head
roof Load (28,400 lb.]
-- Approx. torque req'dto produce Proof Lomwhen washers are use
40
.30
S20i
C
)
dd.
Reg. Heavy Reg. Heavy
3/4 x 6" Bolt - 13/16 in. dia. hole(Specimens torqued to approximately
same load - no washer)
FIGURE 11. A325 BOLT HEAD TESTS - HEAD
TORQUED; EFFECT OF HEAD SIZE ON TORQUE
3/4 x I6 Reg. Semi-fin. Hex Head Bolts. Plots give average values where duplicate tests run.
40
2n
20
10
fn
^=1=-- /V //// //
~y/~y/~-y/-^-_%_^_
*// *//// // **--- ^//--''//--- 3
y/ % x"l^l^ i -
_^_%_ :s 5
/<% /A '- "*-%-%- J!! 5^ ^ S o-^""^- :* x
^y /% _-'y,-^>- * v\/// *// I- 'A - /// - I-1-1-
I-1-1- I_^_%_ |
__%_^_ I'y, ///'// '// __-- %-%---^-%--_^_^__
^ ^13 13 27 7 13 13 27 7 13 76 T6 1 T 32 8 16 16 32 8 1-6
Hole Diameters
(a) Finished Series Nuts (b) Heavy Series Nuts
FIGURE 12. EFFECT OF HOLE SIZE ON
(c) Flanged Nuts
BOLT LOAD
*
'I-0
a
a.I-01
0
I- F + I + -4- ~ I
" f SSpecimen 4 I-7--^ "
/ Notes:Washer under Nut13/16 in. Diam. HoleRegular Head Bolt
I I
I I I
0 1/4 1/2 3/4
Turns from "Snug" of 5 kips
FIGURE 13. LOAD-TURNS RELATIONSHIP FOR HEAVY AND FINISHED NUTS USED IN THESE TESTS
3/4 x 6" Regular hex headPlots are average values where duplicate
13 13 27 7 13 13 27 716 16 32 8 16 1-6 32
boltstests were conducted
proximate torquesq'd to produceoof Load whenshers are used.
-1-1-
-- I-__--^
13 716 8
Hole Diameters
(a) Finished Series Nut (b) Heavy Series Nut (c) Flanged Nut
FIGURE 14. EFFECT OF HOLE SIZE ON TORQUE FOR 5000 LB. PLUS ONE-HALF TURN
40
0
i
I-
0do
800
700
600
- 500
. 400
o 300I.-
200
100
0
----- ----- ----- t y " ----- - v-H avy Nut - ---^ ̂ ^ 9 1)
Finished Nut ------
(% 8s)
I I
c
I
I
0 0.02 0.04 0.06 0.08 0,10 0.12
Average Over-all Elongation of Bolt, Inches
FIGURE 15. COMPARISON OF LOAD-ELONGATION BEHAVIOR
0.14 0.16
OF VARIOUS BOLT TYPES AT IDENTICAL GRIP
0 0.02 0.04 0.06 0.08 0.10 0.12 0.14
Average Over-all Elongation, Inches
FIGURE 16. COMPARISON OF LOAD-ELONGATION DATA FOR DIFFERENT GRIPS
0.16
60
50-• o
A 400
"a 30cI-
10
0
60
&-so.,5
0
- 30
I-.
o 20
10
00 0.5 1.0 1.5 2.0 2.5
Average Number of Turns from "Snug" of 5 kips
FIGURE 17. COMPARISON OF LOAD-TURNS DATA FOR VARIOUS A325 AND A354 BOLTS AT IDENTICAL GRIP
70
60
* 50
30I-
S10
0
Average Number of Turns from "Snug" of S kipsFIGURE 18. COMPARISON OF LOAD-TURNS DATA FOR DIFFERENT GRIPS A325 AND A354 BOLTS
- ..... -A354 3D Bolt, I 3/4" Grip
""*3*riA354 3Proof Load , IBolt, 2 1/4-" -.-A354 BD ) A325 Bolt, I 3/4" Grip & Grip
" A325 Bolt, 2 1/4" Grip "- .-Proof Load
A325
All bolts were 3/4" diam. by3" long, Torqued in a Skidmore -Will1el Calibrator,
/
_ ' I I__1 1 1 1 1 1 i I
--1 3 3 3 412 2 2
1.35" 1.50"
(a) Bolts Tightened to Failure
6 ,6 7 '2 72 2,g,
1.80"
in Solid Blocks*uk
0
ThreadsIn Grip
GripsUsed
5. 7 10- 11 12 122 2
1.75"1 2.25"
(b) Bolts Tightened to Failure in Skidmore-Wilhelm Calibrator
(All Bolts were 3/4 in. diameter)
FIGURE 19. COMPARISON OF TURNS TO FAILURE FOR VARIOUS BOLT TYPES, GRIPS, AND CONDITIONS
4.
1. 0»
L.
qft
ESk
Threadsin Grip
GripsUsed
70
60
« 50
S40-o
30
C
o 20
10
0
NoWashers
illIi llllull
Additional LoadWhen Torquedto FailureF .---
WithWashers
(1)P
P roof
Load d
64)
II3r----- o)_ CA 1) (1)
C:. C . - -( -- -C-
CC C C= -iC
C4 C4 -
* *U - -SID
FIGURE 21. BOLT TENSION AT "SNUG" OF 5000 LB. PLUS ONE-HALF TURN FOR A354 BD RELAXATION TESTS
I I I IApproximately T - 0.2PD where T - Torque, in.-Ib.
P - Bolt Load, lb.D - Bolt Diam., in.
T - 0.2(30,000)(3/4)
- 4500 in.lb - 375 ft. lb.
v,. A354 BD Bolt
Proof Load A354 BD|-- A354 BC Bolt
Proof Load, A354 BC A325 Bolt
Proof Load, A325 --- - -
All bolts were 3/4"0 by 3" long andtested at a grip of 1 3/4" in aSkidmore - Wilhelm Calibrator.
-Hardened washers were used under nuts.
0 100 200 300 400 500 600 700 800 900 1000
Average Torque, Foot - Pounds
FIGURE 20. LOAD-TORQUE RELATIONSHIPS FOR VARIOUS BOLT TYPES
All bolts were A354 BD 3/4 in. diam. x 6 in.long. and tightened against A7 plates with13/16 in. diameter holes. Numbers in paren-theses indicate no. of tests represented.
60
50
n ns 0.c40
8.;> e
C w*30
.-
*J 4 20
0010
0
r•
Ea
-
Note: All fasteners are 3/4 in. nominal diameter.
as r lane I
354 BD Bolt, A242 plates,'hreads in Shear Plane
-A325 High Hardness Bolt,A7 plates, Shank inShear Plane
10 20 30 40
Shear Load Component, kips
FIGURE 22. INTERACTION CURVES OF A325 AND A354 BD BOLTS
UNDER COMBINED TENSION AND SHEAR SHOWING FAILURE LOADS
A354 BD Bolt, TI plates,CL.-.L, 2... eL-.. _ Bl__-.
ahS nk
i n She
I4;
(I*S3
S
I-
60
48
44
40
36
" 32
18- 28
424
20
16
12
8
4
00 1/4 1/2 3/4 1 1 1/4 1 1/2
Turns from "Snug" of 5 kipsFIGURE 23. EFFECT OF BEVELED SURFACES ON BOLT TENSION VS. TURNS, 3/4 IN. DIAM. A325 BOLTS
72
68
64
60
56
52
48
44
- 40
-36
I--
28
24
20
16
12
8
4
n00 1/4 1/2 3/4 1 1 1/4 1 1/2
Turns from "Snug" of 5 kips
FIGURE 24. EFFECT OF BEVELED SURFACES ON BOLT TENSION VS. TURNS 1-IN. DIAM. A325 BOLTS
(a) Location of Fatigue Crack in Specimen B1
SPECIMEN-BI/8 HOLE- 2,4'OOBOT TENS.
SSR: 0 -30o Ks NK1 1,883. 00
(b) Close-up of Fatigue Crack in Specimen Bl (Note bolt head depression)
FIGURE 25. TYPICAL FATIGUE FAILURE
(a) General View of Outside Plate Showing Bolt
Head Depressions and Extent of Fatigue Crack
(b) Close-up of Fatigue Crack Surface
FIGURE 26. VIEWS OF FATIGUE FAILURE OF SPECIMEN D5
-40 -30 -20 -10 0 10 20 30 40Minimum Stress, ksl (Coapression) Minimum Stress, ksl (Tension)
(All Stresses are Based on Net Section)
FIGURE 27. MODIFIED GOODMAN DIAGRAM FOR 2,000,000 CYCLES
I I 1111111 I 111||111 11 11111! I IRow B fasteners in specimens without failureqmme s, wIthot failure.
Row B fasteners in speci mens that failed at s
4f A fasteners In spec i
1000
50 60
10,000
-
Center
Row Bow A--Critical
1- section for out-side plates (SeeFig. 3)Outside Plates
100,000 1,000,000 10,000,000
Applied Cycles
FIGURE 28. AVERAGE LOSS IN BOLT TENSION VS. APPLIED CYCLES IN FATIGUE TESTS
20
8m4 0 S
4.,
C
so
0
S60 "
ae
80
0 5 10
amens tha
t tel let a w
8 -
0.
a
00 0000
co. o 0oC.) C Cm C.) C.) C
00 0000mO -t CO ý0 I 0
00 (00 r 00
99 909090
00 0000
ON -(0 -)
M C.) CO C.)
O0 0000O 0 (0 *0
0n 0 0t 0 M o >coi i n o*
oo - o2T o
00 00 00 00 0n)O oI V4 oo Noo 0
lic rl r-* r«- o C t olo Iin
-* o m-I- oo-a r, co rPc.) m t M Mm c.) cn
oo oo oo oo a0- 00 00 00 0
nr. It 0<- 00 ooco )o.) V om M .3 -ý c.)
- -t I*** r1'D 110 c.)
r)-o) (0(0 0)' 0- oo-- -- -N NN -
00 00 00 00 0
00 ooo000 0 00ooo oo oo oo oo o o o o00 CM (C ) 000 NCV !T 0 CM4ONN (0 C$ 0 (0) CNn O - 0 o 0 (0 -m m) Ln L r't 0omW Ln Lfr LOUfr, -* r'O to * * -* to 0n cn It 1n
N4 toC. wO w0(w(to00) 00)00) 00 00 O G0 00
0 0 C0 a I A 1 1cn0)0 0 0 . 0000 '
0n m D0)c)cn0 co
C.4.4 CM4' CM U LU U LO00 D c00 0 00) 0)0) ) 0)
r0 r ) r0 r N N1 a) ( CNN N
0 Ao00) 00)
NNN NN NNNN NN NNNN NN NN NN NN N M M M N(1) cn). C.)).m C.)C').mCV) m .)m.) V.) () (n V) V.)) C.) nCl) C.)) C.) 01 ) Cl) Nq N4 Nl M
0000 3z zz z Zzz
4,* 4.0000 33z z z zz
0 0 0 0 0z z z z z
*0=0 0O
> >.> >>. U) Wn 'Ainin> > > > > > - -
) D C4 C C4C4C4Cm m m m --.mX =XX = .U = i i Ui U z i
.3.3.3 .33 3.3.33
Un0 n c)o 0 01 U9
U)1 UI U)> U) U) U)>(0))))) (0 10 r-i~p-r-1
zzz zz zzzz
.3.3 .3.33MO) o) 0)M ) )
I u
SInc00 0D0)0)0
CC
I-)-Sm>. >-
.3.34)4)= =
1- ~ ~ ~ ) >. >..UL L . L
> > >(0 ~ ~ ~ ~ 4 0 D0«(0 ( 00)0 0)i > )0) 0)0 O)) 0) .3 ID .3
4)4) 4)4) 4)4) 4)4) 4): 4) 4) 4
tU L0 U U) & ()00 ,-,- NNCM4 C,
zz zz zz zz z
U) U) U) U)I) (0 0
I I S Sz z z 0.
00)
0N
0 0
C.)C M N
0. ). - * 0 (0m C.) C4O C .4 C -
It * * 0 I . 0. 0 * *00 00 0000 00 0-0 00 00 0 o0 0 0 0 0
0
0)
000
'--
0000, C 0
C ONN
0C10
C -0
-J
mm 0O-6
*
S
30-
0) .
CC0
0 cc
co
-0S4)*
CD
*0
I-4a
0 * 0 -0000 3z zz z z
m -o o4)4)40
C.)
N
00 00 3=xz zz =X
4z
0ca =
0 -04 000000
co0() 0ý CN to4C')4 0 o 40000 o 0 - -000 0 0 0 0
4n U0.2-
4) C.Q3 0a
C,
-0o r~--j
U
C 40
n>0Q
to44
0
V-0
c4)0
-J
5
m0
U
ac
CO
0 a
C -0
cc
:-3
* -0
c c
0 v
0).0- 0
0*
0-0I- '4-t
f0 -x0 aZ
4)
-C(A
4-*
to
az:
4- --to
0.03
U)i ^
0 4 N -
NC 0
CN (n
o00 LI )moN -4 Ns4fm 0o 0 oo O ' 4 i 00
-o ooo-o 0 -o C oo o oo0 00
0 0Cer- M
* C.)
4040
400"0* C')'
4)
-000
= 0 014t C',
o o o *oo
c'- V) 4 n c
c.) cl) c.)
0 o C'o
00000
00 0000 CIA O cM 4ON sONO40 It r' -4.
CO 0 C')CO CO)CO CO, CO CO CO)
00 000C) 0 00000 cO ) M0M
4- - 4 C')'rC'),O') CO CO)'
o- oo0 o
co o o» oQ
co 0) 0m; - - NC)O CO CC'0 0009 009
00 (041-* 0
U'-
I-- C-1v
- 0
o 0I- ('
10 NOU
0 0 040 co N
- 0 U)CO) C) CO)
0 0 0N1 It Nl0)> CO) U)
C') CO) CO
NM C 00 0 -
0 0 0
00n0 0 4
(n *0I- c')',
c0
44 0 0
-J
0 ooo0000 0 00 0 0 0 0 00OO 0CN CN 0 C0 4 0 o 0 0 C4 ý*
-t I-t* * It V) ^ ýr -t 't in ko m
C4 C44C4C4 CM C )CM C4C4C- M C M4 4 C4 U) C4NN NNNNN NN NNN N N N N NN0) 0m m 0) 0a0) 0)0)0a 0) 0m 0)* 0) 0) 0)00 00000 00 000 0 0 0 0 0A00) 0D 00)0)) 00a) OM0)0m 0) 0) 0 0) 0 0 m
N4N CNNNC N C 44 C4 4 N- C N C CNNCmCO m M CV)C')Cn 0MM CNCN N N CC ) C.) C.) Cm))
0 0 0 0 00 tz = )
o o1)" = S000z z z
0 0 0 0 00z z z z zz
mot tomoC tom m toto m m to mto to MU to toS (0 ( OI Ol 0 0 (0 10 10 v 0 (0 10 1 10UU VVU UU U U U U U = = =XX =XXXX XX XXX X X X X XX
4-I- 4-4-4-4-4.
>- > > > > >o)0c) m o)0)n0)00 to m M4U)4 VUV4)VU QV UU U 4Vw- Cc cc cc C cc Cc
I I
-- NNNN
u.
U)» U)U)<C')" 444*<*^
IX XXX
U- U U.
2 m-
0 C 4
N N N
S1 S1=
N NN CV)
I II
-a
00
. C1
0
-4u
Ioa
L6
*4-
o!
S4-
- J 1
TABLE 2. RESULTS OF BOLT-TENSION RELAXATION TESTS CONDUCTED WITH A354-BD BOLTS
(1) (2) (3) (4) (5) (6)
Spec. Bolt Nut Washer Hardness Torque Max. 1 Min. Percent Loss of Load FromNo. Head Bolt Nut Load, lb Load, lb 1 Min. Load
RC RC Days5 Min. 2 3 90
NX-24S-D-1 Regular Heavy-2H No 37 35 500 40,600 39,210 0.2
NX-24-D-2 Regular Heavy-2H No 37 35 540 37,200 35,900 0.8 1.3
NX-24-D-3 Regular Heavy-2H No 37 35 500 36,850 34,850 0.9 1.0
NX-24S-D-4 Regular Heavy-2H No 37 35 610 40,500 40,300 0.4
NX-24S-E-1 Regular Heavy-2H No 37 35 580 41,150 40,000 0.6
NX-24S-E-2 Regular Heavy-2H No 37 35 530 39,860 38,150 0.4
NX-24S-E-3 Regular Heavy-2H Nut 37 35 520 47,150 46,840 0.7
NX-24S-F-l* Regular Heavy-2H No 37 35 450 35,350 34,250 0.5
NX-24S-F-2* Regular Heavy-2H No 37 35 450 34,600 33,300 1.1
NX-24-F-3* Regular Heavy-2H No 37 35 37,850 36,750 1.1 4.2
NX-25S-E-1 Regular Finished-2H No 37 26 480 36,230 34,700 0
N = Nut Test; X = A354-BD Bolt; S = Short Time Test (5 Minutes); D = 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 BD Bolts 32-38** 5/8 Plate in Grip A194 2H Nuts 24-37
(7) (8) (9) (10) (11) (12) (13)
TABLE 3. SUMMARY OF VARIABLES FOR TESTS OF BOLTS TIGHTENED ON BEVELED SURFACES
(Numbers in parentheses indicate number of tests performed.)
Tests Performed at: University of Illinois (54); Lamson and Sessions Co. (9)
Period Covered: 1961-1963
Bolt Sizes: 3/4 x 3 (2); 3/4 x 4 (12); 3/4 x 6 (32); 1 x 6 (17)
Bolt Types (All A325): Heavy Head, Old Thread Lengths (32); Finished Head,Old Thread Lengths (12); Heavy Head, New Thread Lengths (19)
Nut Types: A325 (52); A194 Grade 2 Heavy Hex (1l)
Grips (where known): 2 1/8"(2); 3"(12); 4 1/4"(4); 4 3/8"(2); 4 5/8"(17);4 7/8"(5); 5"(8); 5 1/8"(4)
No. of Threads in Grip (where known): 3(5); 4(12); 5(2); 6(7); 7(5); 8(15);
9(4); 10(4)
Type of Calibrator: Load Cell (16); Load Analyzer (9); Skidmore Wilhelm (38)
Method of Torquing Nut: Manual (41); Pneumatic (22)
Location of Hardened Washers: None (24); Head only (1); Nut only (2);
Head and Nut (36)
Location of Beveled (1:20) Washers: None (19); Head only (1); Nut only (4);Head and Nut (39)
Position of Two Bevels: Parallel (4); Converging (33); Crossed (2)
SPECIAL TESTS
Position of Bolt: Maximum Sliding (4); Minimum Sliding (4)
Bolt Strength: High Hardness (2); Low Hardness (2)
Minimum Slide = bolt placed against low side of bevel to minimize bolt movement
down slope.
Maximum Slide = bolt placed against high side of bevel to permit maximum possible
movement down beveled slope during tightening.
Converge = Bevels placed with sloping planes converging (or diverging).
Parallel = Bevels placed with sloping planes parallel but with 5 percent slope
at head and nut.
Crossed = Bevels placed with sloping planes "crossing" or turned at 90 deg.
relative to each other.
4 e) 0 0
0 0 0 0
4-
N
0
C-.)
0
U*)wC-
LIi
0.U.
4-0
C0 _
0)-
0) 00. C
03-
40
*O WJ
>- 0s01
In'A.X
I-a
si4-L
0-
C
'A 0)
-4)
0.0
0. v
.cJ
I4)
4) U) 4)14) X4)/
4) a
i.4I. 4.'
0
0.I-
4)4)4)
4)
-o o - 4ft * 4).*
0 0 C 0. 4)4)0. - 4. -4.oi0-4 0. c 11 * 0- 0-.4- I=c - -- 4 -0 - .00 4).Q00 4 - .0 04) u o o- .a0 4)0 43 -4 0 ft) Q. 4N v 4"t.-0 .. " 00 0. o. N 43 NC*-0 4)-0 CO) .04)r- 0.0 4) C 4) C
u C Q c 0. U - 4) 0 (0-
rU (a a) L4C 04.'.- u CM 04Q .'0 >0.m 3 0* 0.0. 4o' >o.. * 0 C U C
0 0 0 4-04)-- 4). 4)0 B - 4)4-' C 4) C tt»-. .. .- 04)0)- 4o0 - a v 'm)0-c 441o .c0 0 0 4-4)4 -CO 0.4- L) 0 w D w C 4 E ' E EC-
*D *L -O- GL va v &c*- i_. -c a - v u 0 a) v w " 0-.H I Zw 044- >.= -- 4-- o. c0 . o
U) .. ** « c u u 1 O> 0« C c Cm 0N C oto 3-CS w4 ) -. c a) .4 (D >.4 w .
I.. I- 4 -. 4424 HIC O c o - 0 4)0 3~- *.04)0 3. *.04O0.4)4.'f 2 * 3- -4.'. 3- 4 .. 03.. 2-.- C:.. 3.- C:..
E U )0 l-3-4)0 C 200 '..-0 00 0 r0 40 03 44a*- 4- * *-.C * m4- ( 0.4)'.'. 0. 4;, t
S 20 4-. )- '- c-- -5 -D4 n c o). x -g - 3 0) 40 )0 )o0 0
3- 4) -4)- C O H 4)« b.3- C 0.0 *-' 011 0.0 .4.0K I«-*4.440. - 03* o - 4-6 0.-C ! =
Q-40.U) -00.U) -3-l) -.- 0 )I- -' C40 )N C«*->- 0 - 2 C 0.. ** )- X 02. -0*^.o- C. .*-4.- C **
<0 00 U D -4)- 0.4)A Uh- X 0 - (U)NM44)I- tUt (M44*I -.
4)> 4)1 4)1 4) 4)C C C C C
B.0
I- 1- I- I
QCL - QCL C4 M Q- C CM
m+ uo+ a+ i+
0f 01 0 0(0 a CM a
o o ,, oo<*0 0 0 0
r- t m o
C' 0) 0 0
4-1 c4^ 41 4D-
0 0 0 0
0 0 0 0
0 0 0 00 0 0M 0
-0 Q
CI 00
'0
*I.- *I.-i 0.
Len + mr +
+ +0 00 0
(0 C')'- N*4- 0
CM C'
t--4..M0. N,,
2-
n+
+
000Cn,I
LI;
4- -4- C 4NM NM NM
3. :It 310. . S
to, m • ;
4 4) 4 ) « = 4-
) 4)
-3 I V - o ,., o -. " c -c
:aA I I " 2
4-4 4-4 4-* 4* 4-4) 4-4 0.4 4-4) 0'-JC ^J - '.C -X ' X O CA-
a'4 m ) C' *')4 4' : ' '
0 4) w 0 - 9,9,0 0 :0 4 4. 0 woo== = = U . C= C Al - = w -CC' .-.. -S -c - -o
2 0.C -*-. S
1C
00 0 M 0D 41'
3 3A 1032«> 4.4 )4 4.)Z 4)+>I.. 2 4) 2*> -> > *->> >> >C >>.C ZZ '-Z 3> >>
4) 4) 4u) 4) ^ &
-O No r4 co
C') Un - NM
0 C) CN I)C' C) 0) V')
N a
C
4)*
CM
0
-0
cc
c
4)
0
00
4> 4
4)
0
C
04CC
V)0
44
4-
0
5 co
V 10
0 044*
0 C
4) 0
I (0
CL L
W V)
+ -
0* £3- HC0-
.4 040 0
4)I I
4) 0
0. 4)
o 3-t- 4-1
0)
4) -3n -
4) C
-C4.'
- .0
(0*- 4)
5) 0
* 3-
C. - n m ^
4.*Ca4-0-0.4- - -4 )4 C-
o - )j u. Zu -0
4- M
to uJ 40.
4 )en 4a -a -
z ID an
2e.e 00
Goo •0•
r-com o
r-. r- r-.
00 N
000
000
00 000r- r- r-.
000
00 000
000
r- M r O r , r r - 0
CM4 C-4 CM4 CM4 C4 04 V)NyNN NNgN 0m
to > cOiD - r -
CM14 C14 C4 C4 CM m>
- - CNN CM M
40! UI CM V! C! M r
~*r 0 co - 0n0 nc
c )O C tO N CM C-000e NNN N
*o+ob000
c to N
0) 0) 0;
N N N00(0 C
0 0 N
000
CIA V)
000
-0 N 00
'C."
??"
000
- + +
co to L
1+1
0 40 0o
000
000- N 2';
0 ~0 0
- C ND
* * *
000
000'c'J'J
000
000
000???0004 n 4J40* 0
;; CM V -in C 1o
.) 00N - -
000
l e .0
In 0o A to M0
00 0000
U .0
C4 C4 4 - - C4 -
Ir 0o cO r- r-rM rs r- r- r-
00000** ** ** *
00000
ri< r. LO rC-4t*! Oi ý *!
CO I a00 CO 00
* * U) .* N
CM 0M CM C" cM
t- N N 4 NC
NN N
I I I CO
-00400)0
0000 I'-
ooo oo o c
Uo N - I' 804 -M
CD C3, Go 0' 0- NN
00000N N -0 C4 N
+( CM +1 +1 +1-4 C.M MW
ooo oo
0000
*dd *0ww 0C00CM CM* MC
I
140
c4
L.4)
0
412
C
4I
4-4)
41-
*l
4,
041
a
U4=
4J
I
SI4J *
- .
, -
4-
a
0
.-C
4.4
-. m-
OCL
-a --a
Ii .
04J
0.
0
4)C
OA
* 4.• 4-
e-
41
l-
0..0
c
- A
c m
a
04 .
I-
L aU.C I
0 .04a
*-S 0
L- 4
. 4)
- 4J0.
L 0
C 0
4-0t0 L
) 0)*o C
C4
4)-.
-4
)C1L.*-
0
4)I
3
4)
4-
e4E
4-
10
4J
0.
'J
-C
SI-
cc
u
Ow 4
an4)
000
L.
41
a0
4.
a
o
+14
0 ;
-n
-a
&
1)0
4) W
L ID*-0
W 4
S3
0 0+1 CIO 4
u -40 S
a -
- >.4J
0-C 0
O 0
UE C00
*.-MU N
C0m
'A4) 4.J3 u
0
4) >
w V4-
-4)
M
0C *-4 4-
0 0
41
c
Q.-0'C
0*->-
'10 1)
4S-4.0
04- >0
- 4)
L .4 4J
3o -C
Om
.C 0-
L. - 0.
(- --- Q.
a -o
* C
4n m- <
a u- -0 .0
U -C
-4-)
>. a 4)
0 I. 4
040)9
- 4j
= .0
LO &
4-. 0L
4- L -.
*--C
ID 4J)1
4)o
0 O*- *+1 'A.0Q
4-.0 30)
-- C 4J.
a.- u
Q U Q.
S0 4J
n ..
8O 4) 4--L3
e-.0
in- c'4D- 4)
4)-
W 01- UC
UN 4J
- . 0oUl W
a - ID
(U U 0.
0C 0 4)4)... L
--~ 004
Q. * ILU)» NM IL
-C- V q" M c I
O 'C 0 - I
4)0)
a 4 J0 La. 'A..J 4
Lg3
4)-"VI u VI4) 0.
D - -
: . 4J
"0.
Q.U'Q.J,
a. O
a
Sno
)JC
>U ) C
0
4)
in <t
0.4
U- I<n c
11j*E
u/
C r
z
-0.
0 N.
0 4-
U-4)
o- -c -* 4 01*
.UC..
r^co0 U-
0.004)O0 4 O
3 1 ý 1I
0 m >
- 1-*~ 4)-1
4)4)N §S-o o o
, 0 - I. tn
0)4 IUU no-a 3
21^ - -
W4
.2Z
CL
4)
4C
0 4)
.- a.
01
4-- CO
0 -
-0 4)044XC XC
4).-
=-0
*- 0aQ.U
- C-
>..- -^Li 44*
0 -
4)J 4 )
40* Z 40l
I-
4)C
C
0
40
I CIV 4)L Li
II II II
C- C- C-4) )U 4I- 0 t-0 "-0
0 0 0
m0 N. O "J a "J
0 00 0
00 0
0 o
0 0
+ +?
0 0
0 0
4) 4.) 4) M4)
f0 0O 1
:3 3 :3
3 >3 3
4- . 4-S L.. 4- 4 4-
mm mmO CO V
44' 4) 44* TO 44- 4)-:3 -:3 -:3
0 0w 0 v vCc0- 0.- % 0 .-4)3 43 4)3= 44* X 44' = 44
3 3 31- Z L4Z 4- Z
-re
>. -2>. - >.30> 30> 3>
U 4) 0140) 044)4)4 4)4 OC X4)
- NJ N) '3I I I S
= = = =
4) N.u v c
L. o0 0
-0
X 4)
>o 461-0 1 1-)0.
o)C
c- . -
an 0 iL in
-- V 0 01
t)-04)0 II II
C
E 0 -O -
Q.- f0 (40>0 L4- UL
C.) Li Li
It It 11
0 0 * 0 .
C- C- C
1-0 c- 0 1-00 0 0
c- o o mnoi
ON) C 041CM 0340
2)aNJ aN aN4G)41
0 0 0'3- 0 0 0
0n ^- co C- N.» 40 40
- 0 01i40 - 01
4)
C
c
4)
0
0 m
0
I o+1 0
01040
P-0
C c
0 0
0 V4 4n
L 0
0 U0
4) TO
a a
a0 00 0 0O C , 0 0
Ili L4i I- I-
0L 1 0 10 0 Lf0
0 * 0.0.0.*
SO o c 4) O0 0 C 0
"0 -0
U II10 In 4I m
.- D040 010 010
U4 404 404 U04
ý C Ci Ci
+ +
0 01 0 0
N)0 *0 0* 0 0
0.4 - - -C
C- C- C f-
41) 0 40 1, v
0 0 40 0.
NJit 0 001ao~ ~
-» cai -c an C 0 con o
v N)
0 0 0 +
>J 0 > z 0
() v- . - v
4) 4" CU cU
:3 :3
N. 4) L. 4) N 4 ) N
.4 -4) -4) .4
-:3 -:3 -:3 -:
00 0 00 0
3o3 4o3 4)3 o33
C0 C00 C00 0CCQ-4-Z '-Z 4-Z 4-Z
-g - o >- -o ). -o
- N) 3 3.
« (U 40 40V U~
o 0 0 0
I-
*0
S oC
-0
4)
F- 0.
> Qc
C
01
z
o0o
0
II
o°I--
h0
4)
ft)
40
i.n
II II
c 0
o o
4) -
Cf C
C C
4)0 4)10C- Cr - f
<I>-0 1t-0
0 1^ 0444 4440 C
000 00 1000 00 0000 00 0
Cm C l Ili Cý
--- N)
+ + N)0 0 0
0 0 0
4) 4) 4)
3. :2 c75 5
-X3: -m 3
> >t 0'4-* T 4-< O 4-< i
4) * 4) 1 4-o>- -o .- :
30> 30> 3:
4) 4) 4)4) 4)l0£ = OC = OC
I ID
N. N. N. N. be. N'3.4 43' C <
00 40 co
+1 +1 +1
al 4) 4)
:3 :3C :
O TO T
3 3 3
-I -a3 -a
f3 'o f
0.- 0- 0
*^m m" 0, m- 0, L
4) 0 w4) 4):I? It O CO )
4.Z 4n-Z in
4) 1 4) t 4'
- >4 >->4 - >.30> 30> 3>
- NJ N)CO CO C
N) N) N)
10Z
c141X
am -o co c
developed when a hardened washer was used with
a heavy nut.
The torques required to tighten the
bolts in holes of different diameters and using
the finished, heavy, and flanged nuts are
plotted in Figure 14. The torque required to
tighten the finished nut increased with the
removal of the washer and also with the increase
in hole diameter. The heavy nut had a higher
torque without a washer, and as the hole size
increased, the torque appeared to decrease. In
the case of the flanged nut the torque was
fairly constant despite the differences in hole
size, but was somewhat greater than that re-
quired for the finished or conventional heavy
nuts with hardened washers.
When the washer was omitted, a torque
of approximately 700 foot pounds was recorded
for the heavy nut Specimen N-7S; however, this
high torque was probably a result of the severe
galling shown in Figure 7. Specimen N-7 did
not have such severe galling and a somewhat
lower torque was obtained. The lower torque in
the case of the 7/8-inch diameter hole is re-
lated in part to the smaller load developed in
this specimen.
Note that these results are from one
or two specimens of each type; therefore, the
differences in behavior shown between the
various types of nuts may be, in part, experi-
mental scatter. Additional testing to give
average values would be needed if observations
other than the following were to be justified:
(1) The turn-of-nut method (snug + 1/2
turn) without washers and in holes
up to 1/8 inch larger than the
nominal diameter will produce bolt
loads above the minimum for A325
bolts as now specified, especially
when heavy hex heads and nuts are
used.
(2) Omission of the washer under the
turned element may produce torque-
load relationships somewhat more
erratic than usual; thus, torque
methods should not be used for the
installation or checking of bolts
having no washers under the parts
turned during tightening or in-
spection, except for bolts installed
by turn-of-nut method in accordance
with the 1964 edition of Reference 6.
5. Lubrication. Two specimens
(N-23S and N-23S-L) were tested to study the
effects of lubrication. One had lubrication
(molybdenium disulfide) only between the nut
face and the plate face, and the other was with-
out lubrication. These specimens were both
tightened by the turn-of-nut method and de-
veloped approximately the same load, but the
torque required to tighten the lubricated
specimen was just slightly more than one-half
that required for the non-lubricated specimen.
Tests run a decade ago with case-hardened
washers and employing a lubricant called
"No-Oxide" showed the following comparative
results: (10)
Condition Relative Torque
No lubricant 100
Lubricate bolt threads only 83
Lubricate nut face and threads 67
These results are of importance to erectors
using the turn-of-nut method and may warrant
further study; with lubrication, smaller
impact wrenches and compressor capacities may
be used to develop the same bolt load. Since
an advantage can be expected with appropriate
lubrication of threads as well as the nut face,
this might be achieved commercially by treat-
ing only the nut.
6. Long-Time Relaxation Tests. A
number of the A325 relaxation specimens were
left tightened from 3 to 21 days to determine
if there was any significant difference in the
load-time relationship when the washer had
been omitted (Table 1, Columns 12 to 18).
Specimen H-1, which remained tightened for 21
days, showed a 5.1 percent loss from the
I-minute load. Ninety percent of this small
loss, however, occurred in the tirst day.
During the remaining twenty days the rate of
change in load continued to decrease in an
Insofar as is known, no tests or fieldstudies have shown whether such lubricationwill result in the nuts backing off in service,especially under vibration or fatigue loadings.
exponential manner typical of creep data;
whether there were washers under the nuts or
under the bolt heads did not seem to influence
the load-time relationships.
Relaxation of high strength bolt
assemblies was studied in a preliminary fashion
at the University of Illinois in early 1952 by
Robert F. Stevens on 1-inch diameter bolts.
Those results agree well with the results of
the tests reported herein and were summarized
as follows:
"The tensile load which the bolt carriesis transferred through the washer to a
narrow, annular ring surrounding the bolthole. When the stresses in this ringreach or exceed the compressive yieldpoint stress of the plates, the ringyields. As the ring yields, the annulararea becomes larger and the grip smaller;thus the unit stresses are reduced. Theyielding of this area is almost instan-taneous. If an elongation of a certainamount is desired, and if the bolt loadfor this elongation is high enough tocause yielding of the plate material (aswould be the case with nearly all con-struction jobs where the bolt tension isat least 100 percent of the proof load),and if a torque is applied to the boltuntil this elongation is reached and thenthe torque is released, there will followan immediate dropoff in the bolt tension(within 15-20 seconds) of a magnitude inthe order of 5 percent of the total load.If after having reached the desiredelongation, the torque is maintained fora short period (a minute or so) and thenreleased, there will still be an immediatedropoff of the bolt tension, not to exceedabout 5 percent; the amount of loss willdepend upon how long the torque wasmaintained. If the torque is maintaineduntil there is no further yielding, weassume that there would be no loss ofbolt tension when the torque was released.However, the time element involved in sucha procedure makes it impractical.
"Following the immediate dropoff, there isa further, gradual loss of bolt tensionlasting for one to two hours, and amount-ing to another 5 percent of the load.
This loss may be attributable to theheating of the bolt during torquing andits subsequent cooling, changes in roomtemperature, heating of the annular ringduring yielding and subsequent cooling,and other similar effects.
"This yielding of the plate should nothave any great effect upon the recordedbolt tensions. The initial, large lossis over, normally, before the bolttensions are measured, and the subsequent,long time loss is small enough to be with-in the calibration errors.
"The times and percentages indicatedabove are merely to give an indicationof the order of magnitude. Their absolutevalue will depend upon the yield pointof the plate material, and the bolttension and properties of bolt and nut.Also such local factors as the rate ofapplication of the torque and platesurface conditions may affect thisbehavior."
It is interesting to note that the
loss in bolt tension after one minute, observed
in the recent University of Illinois tests,
confirmed the tests that Stevens made nearly
a decade earlier. This same percentage of loss
in bolt tension, about 5 percent, agrees
closely with the loss in bolt load with time
reported by Sanks in a recent series of tests
on specimens which had millscale contact
surfaces as well as a variety of other surface
conditions. It is also confirmed by the
tests on A325 bolts conducted over a period of
days in the Lamson and Sessions load
analyzer.(7)
B. RELAXATION TESTS OF A354 BOLT ASSEMBLIESAND COMPARISON OF BEHAVIOR WITH A325 BOLTASSEMBLIES
In late 1961 the American Institute
of Steel Construction revised its specifi-
cations for high strength bolting and included
two grades of bolts-- the A325 fastener
(similar to SAE Grade 5) which has become so
familiar to the structural profession over the
past decade and the less familiar (and less
common) A354 bolt. Initial consideration by
AISC appeared to favor the A354 Grade BD bolt
(comparable to SAE Grade 8), but the ASTM
specification limit on size and the economics
involved in producing the larger sizes led to
selection of the A354 Grade BC bolt for the
AISC 1961 specification and for the 1963
revision.
Since little information was available on
the characteristics of either grade of A354
bolt then under consideration, an exploratory
program was initiated at the University of
Illinois to provide information on such bolts
in structural sizes. Initially, the Grade BD0
bolt was compared in behavior with the A325
bolt; subsequently, when the AISC specification
was in its final form, some additional tests
were made on the A354 Grade BC bolt.
The results of these preliminary tests
are included herein as a means of comparing
the relative behavior of A325 and A354 bolts.
However, the tests are neither extensive nor
exhaustive.
ASTM approved in 1964 a specification for astructural "higher" high-strength bolt, A490,which has a heavy hexagon head, short threadlength, and mechanical properties comparableto the A3b4 BD bolt.
Because Grade BC bolts were not
commercially available in the sizes and lengths
desired for these tests, Grade BD bolts were
annealed and re-heat-treated to produce the
appropriate properties of Grade BC.
1. Calibrations. The A325 bolts
are permitted variations in hardness ranges
(based on ASTM A325-63T) depending on bolt
diameter, but the A354 bolts (ASTM A354-58T)
have a single hardness range for each grade
which must be maintained for all sizes. The
largest diameters of A325 correspond in hard-
ness, but not in elongation or reduction of
area, to the A354 Grade BB bolt; the smaller
diameters of A325 have hardnesses not greatly
different from those for the A354 Grade BC
fasteners; and A325 bolts, 1-inch diameter or
less, of near maximum hardness would meet most
requirements of A354 Grade BD. For these
reasons, low hardness and high hardness A325
bolts have been included in the following
comparisons between the behavior of the
familiar A325 bolts and the A354 fasteners.
No attempt was made to select the hardnesses
of the A354 bolts used. All bolts were pro-
vided with appropriate nuts, either A325 or
A194 Grade 2 for the A325 and A354 BC bolts,
and A194 Grade 2H for the A354 BD bolts.
The low hardness A325 3/4-inch
diameter bolts had an average tensile strength
of 40.1 kips (the minimum permitted), while
the same diameter high hardness A325 bolts
averaged 47.0 kips. The A354 BD 3/4-inch
diameter bolts had strengths in direct tension
averaging 56.9 kips. The A354 BD bolts of
7/8-inch diameter averaged 73.2 kips.
It may be of interest also to compare
the values of maximum strength in pure tension
to the maximum bolt load obtained when bolts
are torqued to failure. The torqued strength
was 82 percent of the tensile strength for the
few tests available for A354 BD0 bolts. This
is quite similar to the 84 percent reported
by the University of Illinois for A325 bolts(ll)
and close to the 85 percent value commonly
assumed for A325 bolts.
Load-elongation behavior at a given grip
provides another convenient means of comparing
the behavior of the A325, A354 BC, and A354 BD
bolts. Such data are shown in Figure 15 where
the four lower curves are for 3/4-inch diameter
bolts. Here we see that the behavior of the
A325 bolts may be similar to the behavior of
the A354 bolts, depending on their hardnesses.
As might be expected from our knowledge
of the A325 bolts, reducing the number of
threads in the grip of the A354 fasteners also
tends to increase the maximum bolt tension
obtained. This is shown in Figure 16 where
average load-elongation data have been plotted
for two grips and the A325 and A354 BD bolts.
The A325 bolts used in all these series of
tests had the old thread lengths and thus
The length of thread for these bolts wasnominally two diameters plus 1/4 inch.
were similar in this respect to the A354 bolts
tested.
Possibly of greater value to those
who may use these fasteners are the comparisons
of A325 and A354 bolts in the load-turns plots
of Figure 17 and Figure 18. As before, the
load versus turn-of-nut data have been adjusted
to a snug of 5,000 pounds. A study of Figure
17 shows that at snug plus 1/2 turn the A325
bolts easily provided a prestress or bolt
tension equal to the proof load; in fact, the
proof load prestress is reached in 1/4 to 3/8
turn of the nut from snug. Because of their
higher proof loads, the A354 bolts of both
grades require more than 3/8 turn and up to
1/2 turn of the nut beyond snug to achieve the
desired preload. An average for six 3/4-inch
diameter A354 BD bolts showed snug plus 150
degrees to be required for proof load. This
suggests that more nut rotation may have to be
specified for the A354 bolts than is the case
for A325 bolts in order to insure that during
field tightening at least the minimum required
bolt tension is achieved. In the case of the
A354 Grade BD bolt the factor of safety against
over-tightening and against resulting bolt
breakage is reduced accordingly. Fractures
occurred in the 3/4-inch diameter A354 BD bolts
at about I 1/2 turns from snug and at about
1 3/4 to 1 7/8 turns from snug for the 7/8-inch
diameter BC and BD bolts.
It has been shown in studies at the
University of Illinois, Lehigh University, and
elsewhere that, although the apparent strength
of a high-strength bolt may be increased by
reducing the number of threads in the grip,
the turns-to-failure is simultaneously de-
creased. This condition might then become
especially troublesome in field installations
of A354 bolts, particularly if thread lengths
comparable to those on the new heavy head
structural bolts (A325) were to be provided
for the higher strength A354 or A490 bolts.
The decrease in turns-to-failure
with a decreasing number of threads in the
grip, is shown in Figure 19a. At a grip of
1.35 inches the A325 bolt had about 1 1/2 full
threads included and the A354 bolts had about
3 full threads (the number of threads in the
grip are shown for all cases). Note that the
turns shown in Figure 19a are from finger-tight
or about 1/4 turn greater than from snug. The
maximum torque required in these solid block
tests was often reached at about one turn from
finger tight. Figure 19 illustrates again the
probability of a smaller number of turns-to-
failure with the A354 BD bolt than with the
A325 bolt for bolts with the same nominal
thread lengths.
The difference in turns-to-failure
between bolts torqued in the Skidmore-Wilhelm
Calibrator and those torqued in a solid block
having the same grip appeared to be small as
shown by comparing Figures 19a and 19b. How-
ever, other studies at the University of
Illinois and elsewhere indicate that there may
be substantial differences in turns required
to reach a given load, especially in the elastic
range of the bolt and in the range of normal
installation. Some differences in stiffnesses
may occur from calibrator to calibrator depend-
ing on the condition of the unit, the oil in
the cylinder, etc. A Skidmore-Wilhelm type
calibrator may indicate from 25 percent to as
much as 75 percent more turns to reach a given
bolt load or bolt tension than are required for
an actual steel assembly, depending on the
number, thicknesses, and types of plies included
in the grips.
The initial portions of the load-
torque relationships for the A325, A354 BC, and
A354 BD bolts are all similar as shown in
Figure 20. The differences in behavior appear
in the loads for which the relationship is no
longer linear. With the familiar formula at
the top of Figure 20 an estimate can be made of
the torque required to reach a given pre-load
(below, say, the bolt proof load) for any A325
or A354 (A490) bolt, provided a hardened washer
is used under the nut and the threads are clean
and as received.
2. Relaxation. The relaxation be-
havior of A325 bolts has been described in
earlier sections of this report. In a similar
manner, the relaxation behavior of A354 BD
bolts was evaluated by means of a limited
number of tests. These included various
thread lengths within the grips; two nut sizes;
washers and no washers; and varying lengths of
time, up to three months, for the tension to
relax. A354 BD bolts with regular finished
hexagon heads and A7 plates were used to create
an opportunity for creep and load relaxation
because of the very high clamping forces possi-
ble. The results of these tests have been
summarized in Figure 21 and in Table 2.
From Figure 21 it is quickly apparent
that, especially without washers at the plate
surface where turning occurs, the pre-load in
the bolts may not reach the appropriate proof
load after snug plus 1/2 turn. This may be
the result of three or more effects: (1) The
high level of proof load in A354 bolts requires
more bolt deformation and produces greater
plate deformation than does the load in A325
bolts; (2) the plate surface galls or brinells
while some plate material is actually worn
out from under the nut at the very high loads;
and (3) the load-turns relationship is still
quite steep near the snug plus 1/2 turn level
and thus is sensitive to any errors, especially
those errors which may result in less than one-
half turn from snug (Figures 17 and 18).
Two of these bolts were subsequently
torqued to failure, and the resulting maximum
loads are also depicted in Figure 21. One
specimen, NX-24-D-3, reached a maximum load of
52.5 kips after an additional 225 degrees beyond
the snug plus 1/2 turn level and a maximum
torque of 840 foot-pounds at 315 degrees beyond
the 1/2 turn stage. The other bolt reached
48.5 kips at about 300 degrees beyond the
initial snug plus 1/2 turn. The maximum torque
later reached 1450 foot-pounds and when fracture
of the bolt occurred, the threaded end of the
bolt was locked in place because of the large
amount of plastic flow that deformed the plate.
A metallurgical examination showed no notice-
able change in microstructure of the grade 2H
nut, but the cold worked A7 plate was found to
have doubled in hardness in the 1/64-inch layer
below the surface under the nut. The deforma-
tion and galling would certainly have been
markedly reduced if a higher strength plate
material had been used.
Use of a washer under the Grade 2H
nut considerably reduces the wear or galling
and increases the probability of reaching at
least proof load with snug plus 1/2 turn for
the A354 type fastener. However, only one test
was run in these relaxation studies with washers
under the head and nut. It should also be noted
that in Figures 15 through 18 and 20 the
calibration curves also had washers at the nut.
When the amount of relaxation is
considered (Table 2) there is little difference
in the behavior of A354 and A325 bolts. The
losses in bolt tension from maximum load to a
one minute load ranged from less than one
percent to more than five percent with an
average value of just over three percent. The
relaxation in the first five minutes was about
one percent of the one minute load. After
90 days the bolt load was still over 95 per-
cent as great as that recorded at the end of
the first minute. Thus, relaxation does not
appear to be a major problem with the A354
(A490) bolt even when used with ASTM-A7
materials.
3. Tests Under Combined Tension and
Shear. At the time that tests were being made
in another study on A325 bolts under combined
tension and shear, two similar series of tests
with A354 BD bolts were included. In the
interest of making possible a comparison of
A325 and A354 bolts in this report, some of
the data are included. The complete report of
the tests can be found in Reference 13. All
fasteners were tightened to approximately the
bolt proof loads. The average results are
presented in Figure 22 in terms of the bolt
tensile and shear strengths and the actual
test loads.
In general, the A325 bolt seemed to
have a slightly greater shear capacity relative
to its tensile strength than did the A354 BD
bolt. As would be expected from a considera-
tion of cross sectional areas, either type of
bolt is only about 80 percent as strong in
shear through the threads as through the shank.
These data and those in the more complete study
of combined tension and shear( 13) suggest that
similar relationships exist for the usual
structural sizes of bolts. For large grips one
might expect a somewhat different behavior.
There appeared to be little difference in the
behavior of these fasteners when the test blocks
or holders were of A7, A440, or TI steels.
Very pronounced effects of the various
fastener strengths appear in Figure 22. The
strength of a 3/4-inch diameter A141 rivet in
tension or shear, for example, is about three-
fourths that of a low hardness A325 bolt unless
shear is through the bolt threads, in which case
the shear strengths are similar. This result
agrees well with the allowable stresses pro-
vided in the 1962 and 1964 Research Council (6 )
and the 1961 and 1963 American Institute of
Steel Construction specifications. The A354
BO bolt had a thread shear strength comparable
to the shank shear strength of low hardness
A325 bolts and a tensile strength about one-
third greater than that of those A325 bolts.
For shear through the shanks in both cases, the
A354 BO bolt was about one-third stronger than
the A325 low hardness bolt. These comparative
strengths are generally in agreement with the
comparative proof load requirements for the two
types of bolts, although it must be remembered
that the A354 bolts tested in this study were
actually about 14 percent stronger than the
minimum required. With these facts in mind it
is interesting to note that the behavior of the
maximum hardness A325 bolt would probably be
quite similar to that of a low hardness A354
BD bolt.
C. EFFECTS OF 1 TO 20 BEVELS ON BOLT TENSION
Initially the problem of 1 to 20
bevels was explored briefly in 1961 during the
revision of the 1962 edition of the Research
Council specification. A question arose as to
whether additional turns of the nut should be
specified when the head and/or the nut were to
be placed against a slope of I to 20. The few
tests performed at that time gave a consistent
pattern which could be confirmed by a mathe-
matical approximation. Thus the 1962 Research
Council specifications (6 ) were written to
require an additional 1/4 turn for each outside
face in the grip with a 1 to 20 slope.
In late 1962 it was reported that a
number of new field tests did not support fully
the requirement for additional turns on a
I to 20 slope. Simultaneously, the University
of Illinois had undertaken a pilot study to
explore the relaxation behavior and fatigue
behavior of bolts on I to 20 slopes. The
L & S (Lamson and Sessions) load analyzer(7)
was used to assist in these studies. Because
of the importance of determining accurately
the extra turns required with beveled surfaces;
and in order to provide maximum protection
against twist-off with the new, shorter thread
lengths, the study was expanded to include a
number of calibrations on sloping surfaces.
The tests were performed in the same steel
load cells (cylindrical dynamometers) used in
the relaxation tests (Figure 1), in a Skidmore-
Wilhelm device, and in the L & S load analyzer.
One of the problems encountered in
such tests is that of controlling the accuracy
of measuring turns from snug and the decision
of what bolt tension should be selected for
snug. For a number of years, a 5,000 pound
axial load has been used at the University of
Illinois. This snug load seems to be a reason-
able lower level for 3/4- and 7/8-inch bolts.
Therefore, all data have been adjusted to this
level. In the elastic or linear portion of
the load-turns relationship, a difference of
10 degrees in turning will produce a difference
of approximately 2,000 pounds in load. However,
as the bolt yields, a variation in turns has a
much smaller effect.
Table 3 is a summary of all test data
which were reviewed in this study on effects of
beveled surfaces. With the exception of tests
involving one bevel or parallel bevels, these
data are represented in Figures 23 and 24.
Thus a wide variety of factors and a large
number of tests are portrayed in the curves.
In Figure 23 it may be noted that at
proof load and with two beveled surfaces, on
the average 1/4 additional turn is needed to
reach the same bolt tension at and above proof
load. Furthermore, for the 3/4-inch A325
bolts, the proof load was reached between 1/4
and 3/8 turn from snug when flat surfaces were
used, but more than 1/2 turn was needed for
proof load when two I to 20 sloping surfaces
were in the grip.
In Figure 24 one can see that with
two I to 20 bevels in the grip, 1/8 to 3/8
turn additional is required to meet proof
load for the 1-inch bolts, depending on the
extremes considered. On the average, 1/4 turn
additional is needed to insure proof load.
Thus, the results of tests with 1-inch diameter
bolts agree with data shown in Figure 23 for
3/4-inch diameter A325 fasteners.
Through study of the detailed data
from these tests, it is possible to present
the following observations:
(1) Pneumatic tightening seemed to
produce faster seating of the head and nut on
the bevels. For example, in the tests of
1-inch diameter bolts, nut seating occurred
near 150 degrees in manual tightening and at
about 75 degrees for pneumatic tightening;
head seating occurred near 270 degrees in
manual tightening and near 225 degrees in
pneumatic tightening. Pneumatic tightening
tended to reduce slightly the number of turns
required to reach a given bolt tension.
(2) Hardened washers tended to reduce
slightly the turns necessary to obtain a given
bolt load, possibly because of less wearing at
the contact between the nut and the surface.
(3) Reducing the number of threads
in the grip reduced the number of turns to
proof load.
(4) Tests in load cells, Skidmore-
Wilhelm calibrators, and the L & S load analyzer
all show this consistent pattern: additional
turns are needed to insure proof load when
slopes of I to 20 are used.
(5) The effect of one bevel in the
grip was generally intermediate between the
results for no bevel and two bevels. Usually,
proof load was attained in snug plus 1/2 turn.
(6) Relative position of the bevels-
paralle), converging, or crossed--made a
difference In behavior. For converging bevels,
positioning the bolt toward the top or bottom
of the bevel made a difference in behavior.
Placing the bevels so they were crossed (as
may occur in fastening crossing beams) produced
a more consistent and median pattern. When the
bevels were parallel, the effect appeared
dependent on grip. At short grips there was
little or no effect; at longer grips the effect
approached that of a single bevel.
(7) It has also been found from the
study with beveled surfaces that the relaxation
of A325 bolts on the I to 20 sloping surfaces
is essentially of the same magnitude as that
found with non-sloping surfaces. An initial
loss from the peak bolt tension of about 5 per-
cent immediately after tightening is followed
by a loss of about 5 percent from the one
minute load over a period of two days.
A convenient method of achieving the
desired additional turns in the field would be
to specify 1/4 additional turn for two bevels
and no additional turns for one 5 percent
(1 to 20) bevel since proof load on an A325
bolt is usually reached in less than 1/2 turn
of the nut. This would provide at least proof
load in almost every case presented in these
data, and in addition would provide for a
slightly greater reserve against twist-off
than the requirement stipulated in the 1962
specifications. (6 )
IV. RESULTS AND ANALYSIS OF FATIGUE TESTS
A. RESULTS OF TESTS
Two separate series of fatigue
tests were made. The first of these in-
cluded Specimens 1 through 9 and the second
included Specimen Types A, B, C, and D as
shown in Figure 3 and Tables 4 and 5. The
second set of specimens was designed with
extreme conditions; oversized holes, finished
rather than heavy hex-head bolts and nuts,
no washers, and in most cases minimum clamp-
ing were employed. The fatigue tests may be
sub-divided into two groups -- those with the
center plate critical (i.e., having the
higher nominal stresses) and those with the
outside plates critical.
If we look at the first series of
tests (Figure 3 and Table 4) we find that
several different combinations of nuts, bolts,
and washers were used to determine whether
the omission of washers would reduce the fa-
tigue life of the members. None of these
specimens failed within 2,000,000 cycles even
though they were tested at a stress cycle of
0 to 30 ksi.*
* ksi = kips per square inch or 1000 poundsper square inch.
The bolts in Specimen 5 were torqued
to approximately 29,000 pounds, rather than
using the turn-of-nut method, to serve as a
basis of comparison between the results of
the present program of tests and the results
of similar tests conducted in 1957. Despite
this reduction in clamping force to a minimum
level, the specimen did not fail.
Four specimens with the outer plates
critical were tested in the first series
(Table 4) and only one failed; all specimens
of the second series (Table 5) had higher
stresses in the outer plates.
If we look first at Table 4 we find
that failure occurred only in Specimen 6 and
only after 2,400,000 repetitions of a 0 to 30
ksi stress cycle. And, although failure oc-
curred in the outside plate, a close examina-
tion of the fracture indicated that the point
of initiation of the crack was on the inside
face of the plate and that failure resulted
from fretting rather than from bolt head de-
pression or any other stress concentration.
Specimen 7, which combined regular
semi-finished hex-head bolts, heavy hex nuts,
no washers, and oversize holes (7/8-inch
diameter), slipped into bearing during the
first cycle of loading and ran 3,073,900
cycles at 0 to 30 ksi without failure. To
determine if tightening a bolt by turning
the bolt head, rather than the nut, might
produce a difference in fatigue life, Speci-
men 8 (without washers) was assembled with
two bolt heads and two nuts torqued on the
same side of the specimen. Neither the
omission of washers nor the torquing of the
heads contributed to an early fatigue failure.
Although slip occurred in the first cycle at a
stress of 20.4 ksi, a value just above the
current AISC design load for A7 structural
steel, a fatigue failure did not occur. The
test was discontinued after 5,148,000 appli-
cations of a 0 to 30 ksi stress cycle. Speci-
men 9 was assembled in a manner similar to
that used for Specimen 8, except that 1-inch
diameter bolts were used. Slip occurred in
the first cycle at 24.8 ksi. This test was
discontinued without failure after 4,892,000
cycles of loading at 0 to 30 ksi. Thus, even
without washers, the joints still have a large
fatigue resistance, providing the bolts are
properly tightened.
As indicated in Table 5, the fa-
tigue specimens of the second series (all with
outside plates critical) were tested at no-
minal stress levels of either 0 to 30 ksi (or
slightly less) or -20 ksi to + 20 ksi. At
either stress level the fatigue lives of these
specimens were generally two million cycles
or more, despite the severe conditions of
oversized holes, high pressures under the
head and nut, etc. Failure occurred in only
four specimens, disregarding the tests of
Specimens D2 and D3 which were discontinued
because of excessive heat and pounding. The
failures occurred in the series B, C, and D
specimens with minimum specified bolt tension,
and then only two of these failures initiated
at less than two million applications of
load.
The failures in Specimens B1 and
Cl appear to have started on the inside face
of an outer plate and not at the depressions
caused by the bolt heads (Figure 25). The
appearance of the fractures suggests that
fretting rather than the bolt head depressions
or other stress concentrations at the holes
initiated the failures. Figure 25 also pre-
sents a typical fatigue failure for specimens
with the outer plates critical and indicates
the type of seating action that occurred under
the bolt head. Galling of the plate surfaces
similar to that shown in Figure 7 occurred
under the nuts.
The failure in.Specimen DI occurred
in both an outside plate and the center plate
at their respective critical sections but may
have been influenced by fretting. Although
Specimen DI slipped into bearing during the
first loading cycle, it did not continue to
slip to any significant extent when subjected
to additional cycles of loading.
The tests of Specimens D2 and D3
were discontinued because of the excessive
slip and pounding that developed during the
early part of the tests. Examination of these
specimens indicated it was probable that use
of lubricating oil rather than water-soluble
cutting oil during the drilling operation was
responsible for the low frictional resistance,
despite cleaning with a solvent.
Because of the difficulty encountered
in testing Specimens D2 and D3, it was decided
to gradually increase the stress on the nominal
net sections of Specimens D4 and D5 from + 16
ksi to + 20.0 ksi. Failure occurred only in
Specimen D5 and then only after 2,267,000 cy-
cles. The fracture initiated in an outside
plate slightly below the critical section at a
point which suggested that fretting between the
edge of the bolt head and the surface of the
outside plate was responsible for the failure.
This unusual failure can be seen in Figure 26.
In Figure 26a the extent and shape of the fa-
tigue crack is shown. The probable locations
of the initiation points for the fretting fail-
ure have been marked by arrows in Figure 26b.
Once again, the connection withstood more than
2,000,000 cycles of reversal at stresses which
were well above current design levels.
B. ANALYSIS OF FATIGUE TEST RESULTS
In examining the results of these
fatigue tests, it is interesting to compare
them with those of similar tests conducted
at the University of Illinois and Northwestern
University (Table 6). The zero-to-tension
tests as well as the full reversal tests of
this and the other programs compare very fa-
vorably as far as specimen life is concerned.
Specimen Cl (Table 5) is unique be-
cause of its proportioning to give a larger
tensile area than shear area and can be com-
pared with the performance of Specimen 6 in
Table 4. It can be seen that for a stress
range of 0 to 30 ksi, the fatigue lives of
both specimens were essentially the same,
even though Specimen 6 was designed with a
tension:shear ratio of 1.0 to 0.90 and
Specimen Cl with a tension:shear ratio of
1.0:1.25. It can also be seen that these
higher shear stresses (nominal shear stresses)
did not materially affect the fatigue life un-
der zero-to-tension loadings when high clamp-
ing forces existed in the bolts. This was
true despite the greater probability of slip
at the higher shear stresses.
Although the full reversal tests of
Specimens D2 and D3 were discontinued after
relatively few cycles of loading, it is in-
teresting to note that their behavior was
similar to that of Specimen C3-1 (Table 6)
which was also tested in full reversal. The
fatigue life of this latter specimen, which
was tightened to only 83 percent of the mini-
mum bolt tension now required by the specifi-
cations, (6) was well below 2,000,000 cycles.
This low fatigue life, although partly a re-
sult of the low clamping force in the bolts,
also can be attributed to the presence of oil
on the faying surfaces of that joint.
The average results of the fatigue
studies reported herein are presented in the
Modified Goodman Diagram in Figure 27. This
diagram was constructed on the basis of stress
on the nominal net section of the specimen and
a fatigue life of 2,000,000 cycles. Since the
life of most bolted joints was more than
2,000,000 cycles, the plot is conservative. In
comparing the test results with current bridge
design specifications for A7 steel, it can be
seen that the bolted specimens exhibited a
fatigue life of 2,000,000 cycles or more at
stresses approximately 60 percent greater than
those permitted by the specifications.
Figure 27 also serves to compare the
fatigue behavior of bolted and riveted
joints. (14) The test results curve for the
bolted joints is well above the corresponding
test results curve for the riveted joints,
thereby indicating a distinct advantage, as far
as fatigue strength is concerned, in the use of
high-strength bolts for joints subjected to
cyclic loading.
1. Clamping Force. In order to
obtain an indication of the clamping force
in the bolts of the fatigue specimens, the
elongation of each bolt of certain specimens
was measured periodically during the tests
by means of an extensometer. This elonga-
tion measurement made it possible to estimate
the tension in the bolts from the load-
elongation curves for similar bolts. The
initial bolt tension obtained from these
measurements has been indicated in Column 10
of Table 4 and in Column 4 of Table 5.
After completion of the tests,
measurements were again made of the change
in length of the bolts as they were removed.
With this information the tension remaining
in the bolts at intervals during the tests
and at the end of the tests could be esti-
mated. Figure 28 gives a graphical represen-
tation of the average loss in bolt tensions
during these fatigue tests. The loss in ten-
sion for bolts in the fatigue specimens that
did not fail was on the order of 15 percent,
the major portion of which occurred during
the first few cycles of loading. This beha-
vior agrees well with the results reported
in other studies. (8, 15, 16, 17, 18) In
Reference 8 losses in pre-load are reported
to have been about 20 to 30 percent in
375,000 cycles and about 20 to 35 percent in
540,000 cycles with further rapid decreases
as joint slip increased greatly. The bolt
pre-stress in Reference 8 was about one-
third of that which would be used today, and
substantial slipping occurred. In Reference
17 with bolt tensions about equal to those
presently specified, the losses in pre-load
varied from about 7 to 35 percent depending
on specimen, applied stress, etc. Losses of
bolt tension reported in Reference 18* were
on the order of a few percent to 18 percent,
and no joint failures developed.
The results reported herein and in
Reference 16 indicate that approximately a
3 to 5 percent loss in bolt tension may be
attributed to plastic flow in the bolts and
plate materials caused by the high bolt ten-
sion. Much of the additional loss in bolt
load appears to be proportional to the ap-
plied stress and attributable to the applica-
tion of static loading and, among other factors,
to any slipping which may have occurred. (16)
It should be noted that many of the fatigue
specimens slipped into bearing during the
first loading cycle and that the applied
stresses at first slip were generally well
above design levels. Bolt tensions at the
* Two fatigue tests of Reference 18 weremade with bolts torqued into the yieldrange. These tests appear to be the earliestin which it was shown that such high torquingdid not impair the structural joint behavior.The same study also showed that with washersunder the nut, there was no more loss in clamp-ing for bolts having regular finished hex nutsthan for those tightened with heavy hex nuts.
critical sections which eventually failed in
fatigue were found to be reduced considerably
in the first few cycles of loading and to
about half the initial bolt load when failure
occurred. This suggests that there may be
merit in periodic checks and re-tightening
at the critical section of bolted joints sub-
ject to cyclic loading.
In the study by Baron and Larson the
bolt clamping force in the specimens
(I 3/16-inch grip) varied from 20,000 to
40,000 pounds before the tests and had de-
creased approximately 27 percent by the end of
the tests. (15) These specimens were as-
sembled with washers under the nuts and under
the bolt heads, and the bolts were tightened
to a predetermined torque. In most cases
Baron's specimens slipped into bearing during
the first loading cycle. Thus, the loss in
bolt load during testing for specimens with-
out hardened washers was found to be compara-
ble to the loss of bolt tension in joints as-
sembled with washers.
2. Nominal Coefficient of Friction
The nominal coefficient of friction between
the specimen plates is a significant factor
since it, along with the clamping force, de-
termines whether the joint will slip or behave
as a friction joint. The value of the nominal
coefficient of friction for each fatigue speci-
men which slipped will be found in Tables 4
and 5. These values are based on load-slip
readings taken on each specimen during the
first cycle of loading. The nominal coef-
ficient of friction was computed on the basis
of the load on the net area and the bolt
clamping force as determined from the bolt
elongation measurements.
The friction values shown in Tables
4 and 5 are generally smaller than the values
of 0.30 to 0.35 frequently reported for mill
scale surfaces. However, the values of the
present tests are based on relatively small,
four-bolt, double-lap, shear-type joints as
shown in Figure 3 and may provide a lower
limit for the frictional resistance.
V. CONCLUSIONS
Because this report is primarily
a resume of a number of studies, much of the
preceeding text consists of summary state-
ments and conclusions. Only the principal
results and conclusions are repeated in this
section. The reader is reminded also that
these studies generally have been made on
limited numbers of laboratory specimens.
On the basis of the tests reported
and reviewed herein it may be concluded that:
1. Joints properly assembled with
high-strength bolts and no washers
can be expected to perform in about
the same manner as those assembled
with washers.
2. Bolt assemblies without washers may
be expected to develop very nearly
the same tensions as assemblies
with washers when tightened by the
turn-of-nut method. In the case of
oversized holes, up to 1/8 inch
greater in diameter than the bolt,
there may be some reduction in bolt
tension when washers are omitted
and when finished hex head bolts and
nuts are used, but the clamping
force will still be in excess of the
A325 proof load. There is no signi-
ficant difference in the amount of
bolt tension lost with time between
specimens with washers and those
without washers.
Based on the tests reported herein
it appears that A325 and A354 BC and BD (or
the A490) bolts may be compared as follows:
1. Both types of bolts have a torqued
strength of about 85 percent of the
tensile strength. The same approxi-
mate torque-tension relationship,
T = 0.2PD, holds for both types
when hardened washers are provided
at the turning surface.
2. Reducing the number of threads in
the grip (going from old thread
lengths to new thread lengths) in-
creases maximum bolt load at fail-
ure, but decreases the number of
turns to failure in both A325 and
A354 bolts.
3. The A354 bolt (particularly grade
BD) requires more turns to reach
its proof load than does the A325
bolt. This then reduces the factor
of safety against twist-off. With
few threads in the grip the A325
bolt may have a reserve of one full
turn beyond proof load while the
A354 bolt may have a reserve of
little more than two-thirds turn
beyond proof load, even though the
total turns to failure from snug
may be approximately one and one-
half for both types of bolts.
4. Hardened washers appear desirable
under the head and nut of the A354
BD (A490) bolts to reduce contact
pressures and to provide proof load
with slightly fewer turns, especially
when the connected parts are of mild
steel.
5. Relaxation behaviors of the A325
and A354 BD bolts are similar and
show losses of about 5 percent from
the one minute load over a period of
90 days. Much of this loss occurs in
the first few minutes and most of the
balance in a few days.
6. Behaviors of the A325 and A354 bolts
under combined tension and shear are
similar although the shear strength
does not increase quite as rapidly
as the tensile strength when the
bolt strength is increased from
that of the A325 to that of the
A354 bolt.
7. There are numerous similarities in
the behaviors of A325 bolts of high-
est hardnesses and A354 bolts.
The inclusion of surfaces with a
I to 20 slope in the grip requires additional
turning to insure optimum tension with the
turn-of-nut method. If snug plus 1/2 turn
is used, no additional rotation is necessary
for one beveled surface (although the result-
ing load may be near the minimum specified);
however, with beveled surfaces under both the
head and nut, additional rotation should be
used -- for example, snug plus three-quarters
turn (270 degree).
On the basis of the tests reported
herein, the following conclusions may be
drawn regarding the fatigue behavior of shear-
type bolted joints assembled without washers:
1. There was no significant reduction
in the fatigue lives of bolted joints
as a result of omitting the washers
from the bolt assemblies, provided
the fasteners were tightened to at
least the currently specified mini-
mum bolt tensions. Although there
was some galling of the plates when
nuts were turned directly against
them, this galling did not have a
noticeable effect on the fatigue
lives of the joints.
2. The major portion of the loss in
bolt tension during the fatigue
life of a specimen tends to occur
in the first few cycles of loading.
This loss in bolt tension appears
to be similar whether hardened wash-
ers are used or omitted.
It should be noted that these last
conclusions are based on joints that slipped
into bearing, as well as those that did not;
on joints with extremely severe combinations
of hole size, fastener head size, and nut
size; and in many cases on joints assembled
with minimum allowable clamping in the high-
strength bolts. Thus, these conclusions
appear to be, a priori, safely applied to
connections with less severe conditions.
VI. REFERENCES CITED
I. C. W. Lewitt, E. Chesson, Jr., and W. H.Munse. "Studies of the Effect ofWashers on the Clamping Force in HighStrength Bolts," Unpublished Struc-tural Research Series No. 191. Univer-sity of Illinois Department of CivilEngineering, Urbana, Illinois, March1960.
2. A. C. Knoell, E. Chesson, Jr., and W. H.Munse. "Fatigue Behavior of BoltedJoints Assembled Without Washers."Unpublished Structural Research SeriesNo. 242. University of Illinois Depart-ment of Civil Engineering, Urbana,Illinois, February 1962.
3. J. R. Fuller and W. H. Munse, "LaboratoryTests of Structural Joints with Over-Stressed High Tensile Steel Bolts,"Unpublished Structural Research SeriesNo. 44. University of Illinois Depart-ment of Civil Engineering, Urbana,Illinois, January 1953. Also summa-rized in ASCE Transactions, Vol. 121(1956), Paper No. 2839, Pp. 1255 to1266.
4. F. P. Drew, "Tightening High-StrengthBolts," ASCE Proceedings, Vol. 81(August 1955), Paper No. 786.
5. E. F. Ball and J. J. Higgins, "Installa-tion and Tightening of Bolts," ASCETransactions, Vol. 126, Part II(1961), Paper No. 3241, Pp. 797-810.
6. Research Council on Riveted and BoltedStructural Joints, "Specifications forStructural Joints Using ASTM A325 Bolts,"ASCE Structural Division Journal, Vol.88, ST5 (1962), Paper No. 3294. (Pre-vious editions of this specificationwere dated 1960, 1954, and 1951. Re-vised in March 1964 and entitled"Specifications for Structural JointsUsing ASTM A325 or A490 Bolts,")
7. W. G. Waltermire, "Evaluating Fastenerand Joint Performances with a LoadAnalyzer," Assembly and FastenerEngineering, Vol. 5, No. 1 (1962),Pp. 32-35.
8. W. M. Wilson and F. P. Thomas. FatigueTests of Riveted Joints(Experiment Station Bulletin No. 302).Urbana, Illinois: University of 1lli-nois College of Engineering, May 1938.
9. R. L. Sanks and C. C. Rampton, "Grit andShot-Reinforced High Tensile BoltedJoints," ASCE Structural DivisionJournal, Vol. 83, ST 6 (November1957), Paper No. 1435.
10. D. T. Wright and W. H. Munse, "Calibra-tion Tests of High-Strength Bolts,"Unpublished Structural Research SeriesNo. 9, University of Illinois Depart-ment of Civil Engineering, Urbana,Illinois, January 1951.
II. J. G. Viner, E. Chesson, R. L. Dineen,and W. H. Munse, "A Study of Nuts forUse with High Strength Bolts," Unpub-lished Structural Research Series No.212, University of Illinois Departmentof Civil Engineering, Urbana, Illinois,March 1960.
12. Robert F. Stevens, Unpublished researchnotes, University of Illinois Depart-ment of Civil Engineering, Urbana,Illinois, 1952.
13. E. Chesson, N. Faustino, and W. H. Munse.Static Strength of High-Strength BoltsUnder Combined Tension and Shear.Manuscript in preparation. Universityof Illinois Department of Civil Engi-neering, Urbana, Illinois, 1964.
14. E. Chesson, J. F. Parola, and W. H. MunseEffect of Bearing on Fatigue Strength
of Riveted Connections (EngineeringExperiment Station Bulletin in pre-paration). Urbana, Illinois: Uni-versity of Illinois College ofEngineering, 1964.
15. F. Baron and E. W. Larson, Jr., "Re-sults of Static and Fatigue Tests ofRiveted and Bolted Joints HavingDifferent Lengths of Grip," An un-published report of an investigationconducted in the Department of CivilEngineering at Northwestern Univer-sity, January 1952. Also summarizedin ASCE Transactions, Vol. 120(1955), Paper No. 2778, Pp. 1322-1334.
16. R. A. Hechtman, D. R. Young, A. G.Chin, and E. R. Savikko, "Slip ofJoints Under Static Loads," ASCETransactions, Vol. 120 (1955),-Paper No. 2778, Pp. 1335-1352.
17. W. M. Wilson, "High Strength Bolts asa Means of Fabricating Steel Struc-tures," Proceedings of the 30thAnnual Meeting of the Highway Re-search Board, December 1950, Pp.136-143.
18. K. H. Lenzen, "The Effect of VariousFasteners on the Fatigue Strength ofa Structural Joint," AREA Bulletin 481.Vol. 51, (June-July 1949), Pp. 1-27.
19. W. H. Munse, "An Evaluation of the Be-havior of Structural Connections As-sembled with Hook-Knurl Bolts,"Unpublished Private Report. Urbana,Illinois, September 1957.
20. W. H. Munse, D. T. Wright, and N. M.Newmark, "Laboratory Tests of BoltedJoints," ASCE Transactions, Vol. 120(1955), Paper No. 2778, Pp. 1299-1321.
21. N. G. Hanson, "Fatigue Tests of Jointsof High-Strength Steels," ASCETransactions, Vol. 126, Part II(1961), Paper No. 3241, Pp. 750-763.
22. W. H. Munse and H. L. Cox, The StaticStrength of Rivets Subjected to Com-bined Tension and Shear (EngineeringExperiment Station Bulletin No. 437).Urbana, Illinois: University ofIllinois College of Engineering,December, 1956.
