ILLINOI S

UNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN

PRODUCTION NOTE
University of Illinois at
Urbana-Champaign Library
Large-scale Digitization Project, 2007.
 ABSTRACT

EXTENS IVE TESTS OF INDIVIDUAL A325 BOLTS

~/ERE rvADE TO EVALUATE THE EFFECT OF WASHERS ON

THE CLAMPING FORCE DEVELOPED IN HIC~-S TRENGTH

BOLT ASSEMBLIES TIGHTENED BY A TURN-OF-!IUT

METHOD. THE RESULTS OF THESE TESTS INDICATE

THAT THE~ IS NO SIGNIFICANT DIFFERENCE IN THE

CLAMPING FORCE OF STRUCTURAL BOLTS, EI THER

WITH OR WI THOUT WASHERS. SIMILAR RESULTS WERE

OBTAINED WITH A354 GRADE BD BOLTS .

A STUDY OF THE D-l,A./\K;ES 1'-ECESSARY IN THE TURN-

OF-NJT METHOD WHEN THE SURFACES AT THE HEAD AND

NJT OF THE BOLT WERE ON FIVE PER CENT SLOPES

INDICATES THAT ADDITIONAL TURNING IS REQUIRED

TO INSURE PROPER TIGHTENING.

FATIGLE TESTS ~!ERE CONDUCTED TO DETERMINE THE

EFFECT UPON JOINT LIFE OF OMITT1NG THE WASHERS

FRCJ-1 THE BOLT ASSEMBLY . ALL SPEC IMENS HERE

FOUR-BOLT, DOUBLE-LAP, SHEAR-TYPE JOINTS AND

INCLUDED CONt-ECTIGJS DESIGNED WITH EITHER THE

INSIDE OR THE OUTSIDE PLATES CRITICAL. SPECI -

MENS WERE GENERALLY ASSEr13LED WITH 3/4-INCH

D!Atv'CTER FASTEt-ERS. ON THE BASIS OF THESE AND

OTHER TEST RESULTS, IT APPEARS THAT THE FATIGUE

LIVES OF PROPERLY BOLTED (A325) JOINTS ASSEMBLED

WITHOOT HARDENED WASHERS ARE GENERALLY AS Let.G

AS lHOSE OF JOINTS ASSEMBLED WITH WASHERS,

This page is intentionally blank.
 CONTENTS

I. INTRODUCTION . . . . . . . . . . . . . 1
A. Use of High - Strength Bolts 1
B. Object and Scope of Investigation 2
C . Acknowledgments 3

II . DESCRIPTION OF TESTS 5
A . Relaxation Tests 5
B. Specimens for Tests with Sloping Surfaces 7
C . Fatigue Test Specimens 7

III . TEST RESULTS . . . . . . 11

A. Relaxation Tests of A325 Bolt Assemblies 11
B . Relaxation Tests of A354 Bolt Assemblies
and Comparison of Behavior with A325 Bolt
Assemblies . . . . . . . . . . . . . 17
C . Effects of 1 t o 20 Bevels on Bolt Tension 22

IV . RESULTS AND ANAL YSIS OF FATIGVE TESTS 25

A. Res u lts of Tests .... 25
B. Analys i s of Fatig u e Test Results 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 o f 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 Was h er

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 Relationsh i p 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 Bolt
Types at Identical Grip

16 . Comparison of Load -El ongation Data for Different Grips

17 . Comparison of Load - Turns Data for Various A325 and A354

Bolts at Identical Grip

18 . Comparison of Load - Turns Data for Different Grips A325

and A354 Bolts

19 . Comparison of Turns to Failure for Various Bolt Types ,

Grips , and Cond itions

20 . Load -Torque Relat ion ships for Various Bolt Types

21 . Bolt Te nsion at " Snug " of 5000 lb . Plus One- Half Turn for
A354 BD Rel a x ation Tests

22 . Interaction Curves of A325 and A354 BD Bolts Un der Combined

Ten sion and Shear Showing Fs ilure Loads

23 . Effect of Beveled Surfaces on Bolt Tension vs. Turns , 3/4

in . Diam . A325 Bolts

24 . Effect of Beveled Surfaces on Bolt Tension vs . Turns l - in .

Diam . A325 Bolts

25 . Typical Fatigue Failure

26 . Views of Fatigue Fai lure of Specimen OS

27 . Modified Goodman Diagram for 2 , 000 , 000 Cycles

28 . Average Loss in Bolt Tension vs . Applied Cycles in Fatigue

Tests
This page is intentionally blank.
TABLES

l . Results of Bolt-Tension Relaxation Tests for A325 Bolts

2. Results of Bolt - Tension Relaxation Tests Conducted with

A354-BD Bolts

3. Summary of Variables for Tests of Bolts Tightened on

Beveled Surfaces

4. Results of Fatigue Tests , Specimens l Through 9

5. Results of Fatigue Tests , Series A' B ' c' and D

6. Results of Previous Fatigue Tests

This page is intentionally blank.
I. INTRO!XJCTION"

A. USE OF HIGH-STRENGTH BOLT ASSEMBLIES and began using methods of tightening which

For a decade after its adoption, the are independent of any load-torque relation-

high-st rength structural bo lt assembly included ship. The reason for the change was the

a bolt with a finished hexagon head, a heavy desire for a simp ler, faster, and possibly

hexagon nut, and two hardened washers. Prima- more accura te means of tightening the bolt

ri ly, the washers served three purposes: assemblies. The objections sometimes offered

(a) to protect the outer surfaces of the con- to the use of tor~ue as a basis for bolt load

nected material from damage or galling as the control are: (a) a lack of dependability

bol t or nut was tor yued or turned; (b) to as- (because of differences in thread fit, thread

sist in maintaining the high clamping force in clean I iness, rust, etc.), (b) the possible

the bolt assembly; and (c) to provide surfaces resulting variations in bolt tension through-

of consistent hardness so that fairly depend- out a joint, and (c) inefficient use of pneu-

able relationships between torque and tension matic impact wrenches.

could be established for the various bolts and A preliminary study at the University
3
bolt diameters . of lllinois( ) and an investigation conducted

Most of t he ea rly methods employed by F. P. Drew for the Association of American

4
for the development of tension in high-strength Rai lroads( ) led to a turn-of-nut method for

bolt assemblies made use of a relationship be- tightening high-strength bolt assemblies.

tween applied torque and bolt tension . How- Thi s method has provided a means of obtaining

ever, recently erectors and engineers developed at least the required minimum bolt tension

without reI i ance upon toryue. In the turn-of-

This report combines most of the data pre- nut method the bolt is tightened to, or above,
sented in tw9 previously unpublished progress
reports(1,2)T and other unpublished data the required minimum tension by rotating the
availab le on research with high-strength bolts.
nut or bolt a prescribed number of turns from
t Superscript numbers in parentheses refer to
references cited. either a finger-tight or a snug position.
2
The "finger-tight" position is that to which
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 r un freely and
the impact mechanism begins to operate, or the
tightness attained by the full effort of a man
using a n 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 wa rped, heavy, or many plies of
material may require that the bolts be touched
up by addi tional tightening to insure fu ll
bolt load in a ll the fasteners. A more de-
tailed discussion o f the turn-of-nut method as
used by some f ab ricators was presented by Ba l l
and Higgins.(S)
B. OBJECT AND SCOPE OF INVESTIGAT ION
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 assemb ly. 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 bo lt 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 clampi~g force
in the assemb 1 i es. 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
"k
Because of the development of the heavy hex-
agon head, structural A325 bolt and the result-
ing lack of interest in finished nuts among
structural us~r~, the 1962 Research Counci I
Specification~ 6 J mentions specifically only the
heavy hexagon nut.
 3

the Research Council on Riveted and Bolted percent slope relative to the plane perpen-

Structural Joints; (d) a finished-thick nut dicular to the bolt axis. Variables in the

which has the same across-fl ats dimensions as tests included the use of washers, relative

a finished nut, but is as thick as a heavy nut, directions of sloping surfaces, position of

and (e) a flanged nut which has the same bolt, bolt hardness, method of tightening, and

across- flats dimensions as the heavy nut, but bolt size.

a l so has an integral washer on one face. Also The third question to be answered

included in the tests were two types of bolt concerning elimination of washers involved

heads: the regular semi-finished hexagon head twenty-one fatigue tests. The variables em-

which has been specified in the past and a ployed were type of nut, type of bolt head,

heavy semi-finished hexagon head which was hole size, location of the critical plate

approved as an alternate in the 1960 Specifica- (inner or outer plates of a three plate spe-

tions of the Research Counci I and became the cimen), element tightened (the bolt or the

standard bolt in the 1962 Council Specifica- nut), bolt tension, and bolt diameter . These
tions. ( 6) fatigue specimens were four- bolt, double- lap,

The program included a large number shear- type joints similar to the one shown in

of tests of single assemblies which are re- Figure 2.

ferred to as relaxation tests, wherein a Based on these and other studies

determination was made of the variation in bolt recently performed at the University of l lli-

tension with time using a strain gage load cell nois, as well as at other schools, the Speci -

within the grip as shown in Figure I. The fications{6) of the Research Council were

variables in the relaxation tests included the revised early in 1964.

type of nut, bolt head size, bo lt strength,

hole size, hole preparation , and the effect of C. ACKNOWLEDGMENTS

lubrication . In some of the relaxation tests, The tests described in this report

the assemblies were tightened by turn ing the are part of an investigation being conduct ed

bolt rather than the nut. as the result of a cooperative agreement be-

Other tests were performed with the tween the Engineering Experiment Station of

strain gage load cell and with a commercial the University of I llinois, the I llinois Divi-

hydraulic load cell to determine the turns sion of Highways, the Department of Commerce--

necessary to properly tighten the assembl ies Bureau of Public Roads, the Research Council

when one or both outer surfaces have a five on Riveted and Bolted Structural Joints, the
4

American Institute of Steel Construction, and Committee B of the Research Council on Riveted
the Industrial Fasteners Institute. The and Bolted Structural Joints. The chairman
investigation is a part of the Structural of the Committee was Mr. A. H. Smith, and its
Research Program of the Department of Civil members included E. R. Estes, Jr., N. G. Hansen,

Engineering of which Professor N. M. Newmark T. R. Higgins, N. P. Maycock, A. K. Robertson,

is head. J. L. Rumpf, and W. H. Hunse. Additional

The work was carried out by the testing was coordinated with Committee 15 of
following research assistants in Civi 1 the Research Council chairmaned by
Engineering, whose efforts are gratefully Mr. T. R. Higgins and having the following
acknowledged: H. D'Arcy, P. C. Birkemoe, members: L. s. Beedle, R. B. Belford,
N. Faustino, A. c. Knoell , c. w. Lewitt, F. H. Dill, E. L. Erickson, F. E. Graves,
J . F. Parola, and E. W. J. Troup. The studies w. H. Munse, E. J. Ruble, A. Schwartz, Jr.,

we re performed under the general supervision T. W. Spi !man, 0. L. Tarlton, and G. S. Vincent.

of W. H. Hunse, Professor of Civil Engineering, This manuscript was reviewed by

and under direct supervision by E. Chesson, Jr., w. G. Waltermire, R. B. Belford, E. H. Gaylord,

Associate Professor of Civi 1 Engineering. and T. W. Spilman . The authors wish to thank

A number of the tests described in them for the ir careful analysis and criticisms.

this report were planned in cooperation with

II. DESCRIPTION OF TESTS

;·~
A. RELAXATION TESTS i n ha rdne ss from a Rockwe ll B of 80 for t he

As shown in Figure I, the rel axa tion f inished-thick nu t s to a Rockwell B of 9 2 to

specimens cons i sted of two 4 x 4 x 3/4 -inch 95 for the heavy-thin nuts (Tab 1e 1, Co 1umn 6) .

steel pl ates (with a 13/16- inch diameter hole The ASTM- A354 bolts were specified to be grade

in the center, except in the ca se of the over- sot and were near Rockwell C = 37, the maximum

sized hole specimens), a 3/4 x 6-inch high- ha rdness for that grade (Ta ble 2) .

strength bo l t, and a load cell. The steel The load ce ll s (Figure 1) were

p la tes were ASTM-A7 structura l grade stee l with 1 11/16 inches in diameter and 3 1/8 inches

surfaces le ft in the a s-rolled cond i tion; how- long. These cells were designed to provide

ever, i n some ca ses the plates were rusted from average axi a l strains of 800 to 900 micro-

having been stored outside . Before the pl ates inches when subjected to the maxi mum load a nd

we re used ,the y were wi re- b rushed to remove any were made from AISI 434.) stee l, he at-treated to

loose sca le or rust. provide a high yiel d point to reduce the possi-

Da t a on the bolt and nut ha rdnesses bility of local yield ing. Ea ch load cell had

a re in Tables a nd 2. It is import ant to note eight stra i n gages (four placed axially and

tha t there were significant differences in the four ci r cumferent ia lly) arra nged in a fou r- a rm

ha rdnesse s and ult imat e strengths of the regu- br id ge for high sensitivity.

Ja r- and heavy- head bolts used in the relax- In the relaxation tests, the output
a tion tests for A325 bolts. The heavy-head

bolts (Rockwe ll C = 24) were on the low sides These nuts were obtained before the 19 58
changes in ASTM-A325.
of the ha rdness and strength ranges a ll~~ed by

the ASTM Specifica tions whi l e t he regul a r-head

t It should be noted that t he A354 BD bolts
have mechanical properties simi Jar to those
bolts (Rockwell C = 32) were on the h igh side prescribed for the new ASTM-A490 structural
bo lt s now incorporated in the 1964 edition of
of the Specification values. The nuts varied Reference 6.

5
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 remai nder
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 11 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 ho le and oversized holes
of 27/32 and 7/8 inches to simu late 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 8 11 or "C" in the
11
specimen number indicates the use of over-sized
holes- - 11811 for the 27/32-inch diameter holes
a nd "C" for the 7/8-inch diameter holes. ln
those cases where a specimen was tightened to
 7

a specific bolt load, that load in kips * appears would involve considerable time and expense

as the last figure in the specimen number. The and is much more extensive than could be in-

four specimens that were head torqued (tightened eluded in this investigation. However, a

by turning the bolt head} are identified by "HT" number of these variables have been explored

in the specimen number. One specimen was tested and are summarized in Table 3.

with a lubricant of the molybdenum disulfide In general the specimens consisted

type between the nut face and the surface of the of a bolt and a suitable nut tightened against

plate; this specimen is identified by the letter sloping mild steel plate surfaces (mi II scale

"L" in the specimen number. The letter "X" surface, as-received) with some type of load

(Table 2) identifies the A354 BD bolts used in indicating device included in the grip -- either

the relaxation studies and the letters "D", "E", a steel load cell (cylindrical dynamometer as

and "F" indicate a variation in the number of in Figure I) or a commercial hydraulic load

threads included in the grip. cell (Skidmore-Wi I helm calibrator). A second

group of tests was conducted at the Lamson

7
B. SPECIMENS FOR TESTS WITH SLOPING SURFACES and Sessions Company in a load analyzer, ( )

A thorough study of the behavior of which provided a record of load, turns, and

assemblies tightened on sloping surfaces would app lied torque and also permitted an evaluation

require a systematic consideration of a large of the relaxation of the fastener.

number of factors. The following are of primary Accuracy of the Illinois load cells

importance: bolt head size and nut size was verified by recalibrations under axial

(distance across flats); grip and number of loadings, with eccentric loadings, and with a

threads in the grip; diameter of bolt and hole; check of the effects of residual torque. The

hardness or strength of bolts , nuts, plates and Skidmore-Wilhelm device was also calibrated

washers; position of the bolt in the hole -- under axial and eccentric loadings. Both kinds

centered, on high side of slope (maximum slid- of equipment showed excellent response under

ing} , or on low side of slope (minimum sliding); all conditions.

and method of torquing-- pneumatic or manual.

Secondary factors might include chamfered or C. FATIGUE TEST SPECIMENS

washer-faced nuts, multiple plies, flats or The fatigue test specimens were

corners on the high side, and torquing head or designed to supplement the relaxation tests on

nut. The systematic study of all these factors A325 bolt assemblies and to determine the

Kip = 1000 pounds (kilo-pound) effects of galling when washers were omitted
8
from the fastener assembly. Specimens that
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
3
tension in the bolt.( ) 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 p late), the objective was to
determine whether the galling on the outer plates
resulting from the tightening of the assemb ly,
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
·
spec1mens . prev1ous
1n . s t ud.1es. (J) Tens1on:s
. hear
ratios* for all specimens are shown in Tables 4
* The "tension:shear ratio" is the ratio of the
net section tensile stress to the shear st ress on
the c ross-sectional area of the bolts. The ten-
sile area used in the calculations for a "nomi-
nal" T:S ratio was computed using the nominal
fastener diameter plus 1/8 inch regardless of
actual hole size. The shear a rea assumed was
calcul ated 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-dri ll ed.
The bolt assemblies were tightened
by turning the nut, except for Specimens 8 and
9 (Table 4) . The bolts in Fatigue Specimens
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 0 a bolt load of approximately
28,400 pounds was used, the latter correspond-
ing to the minimum permitted by Specifi-
cations.(G) In the case of the l-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 l-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 ful l y described in Reference
8. The load was applied at the rate of approxi -
mately 200 cycles per minute to all specimens
except Specimens 03, 04, and 05 . These speci-
mens were tested at a loading rate of 100 cycles
 9

per minute to reduce the tendency for slip and what were considered to be the most severe

pounding which had developed in the second conditions possible - - oversized holes, minimum

Series D specimen. clamping force in the bolts, and higher than

All plates for the fat igue specimens normal bearing stress under the bolt head and

were of ASTM-A7 structural steel. The plate nut. To accomp lish this, a ll bolt holes

surfaces were left in the as-rolled condition; (except for Specimen Cl) were d ri lied 1/8 inch

however, in some cases these surfaces were larger in diameter than the nominal diameter

rusted and slightly pitted from having been of the bolt to simulate excessive field ream-

stored outside. The plates were lightly wire ing, and regular series hex bolts with finished

brushed before use to remove any loos e material, hex nuts were used to give an unusually high

and the contact surfaces of the joint were bearing stress at the plate surfaces under-

cleaned with a solvent to remove any cutting- neath the bolt heads and nuts. Specimens with

oil remaining from the hole-drilling operation. less severe geometries would have been tested,

The bolts for the f atigue specimens if necessary, to find the conditions required

met the hardness and ultimate strength require- to produce fatigue lives comparable to those

ments of ASTM-A325. Bolt and nut hardnesses obtained with joints having two hardened

are shown in Table 4 for fatigue test Specimens washers for each bolt; however, the severe

I through 9 where the hardness tended to be conditions noted above did not appear to

low, but within the allowable range. In general, produce any significant reduction in the fatigue

for specimen types A, B, and D, the hardness of strength of the joints and no further tests were

the regular hex (hexagon) bolts tended to be on considered necessary.

the high side of the a llowable Rockwell C range First row slip was measured a t each

(RC = 30); the hardness of the finished nuts edge of all the fatigue specimens by means of

was Rc = 26 (ASTM-AI94, grade 2H). The regular mechanical dials (Figure 2). ''First row slip 11

hex bolts of Series C hao d hardness of RC = 32 is the relative movement which takes place

and the hex nuts in Series C had a ha rdness of between the edges of the center plate and the
. .r:
R
6
= 82 to 84. outside p lates and includes both elastic and

In order to minimize the number of plastic deformation as well as slipping of the

tests in the second stage of the program, speci- plates.

mens of Series A, B, C, and D were designed for Where bolt tensions in the fatigue

specimens were reported, they were determined

These nuts were from stock ava i !able at the
laboratory and procured prior to the 1958 by means of bolt elongation. A bolt similar
changes in A325 nut proof load.
10

to those used in the fatigue specimens and with made possible the construction of a typical

the same grip was tightened in a Skidmore- load-elongation curve that was used to es-

Wilhelm Calibrator and its elongation measured tablish the loads or tensions of the bolts in

a t various increments of load. This information the fatigue specimens.

III. TEST RESULTS

A. RELAXATION TESTS OF A325 BOLT ASSEMBLIES upon completion of the torquing there was a

In examining the Table I relaxation small drop in load. The load for each bolt,

data for A325 bolt assemblies it is necessary recorded in Column g of Table I, is the peak

to keep in mind the wide range in bolt and nut load on the trace. However, because the maxi-

hardnesses which existed in the specimens test- mum load in the bolt is only an instantaneous

ed. Because of this range, comparisons should load and rarely measured, even in the labora-

be made only between assemblies with comparable tory, t h is value has not been used as a basis

hardnesses and with the same nut and bolt head. of reference. Instead, the load in the bolt

Furthermore, since each type of nut and bolt one minute after reaching the peak load has

head was tested bot h with and without a washer, been used as the base from which to determine

the contribution of the washer to each assembly the percent loss in bolt tension with time

can be studied without considering hardness as {Column 10 , Table I). The drop in load from

a variable. In order to provide additional the maximum load to the one-minute load varied

information, one assembly of each type was from 2 to 11 percent, with an average of about

tightened in a Skidmore-Wilhelm Calibrator and 5 percent. This drop in bolt tension is

the bolt elongation was measured at various thought to be a result of an elastic recovery

increments of load. These measurements were which takes place when the wrench is removed

compared with elongation readings taken on the and possibly a result of creep or yield in the

relaxation specimens as a check on the bolt bolt, produced by high stresses a t the root of

tension recorded by the load cell. the threads. In addition, some flow may occur

In the relaxation tests a continuous in the steel plates under the head and nut.

record of bolt tension versus time was obtained Simil ar tests made on a hydraulic

from the load cell with a Sandborn Recorder; a load analyzer by the Lamson and Sessions Compa-

typical curve is shown in Figure 4. It was ny(]) show similar percentages of bolt tension

evident from these records that immediately losses immediatel y after tightening and over a
11
12
period of days. Comparable results were also
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 va r i-
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 wi 11
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
assemb ly 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 bolts
must withst and wi t hout 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. The
1960, 1962, and 1964 specifications of the
Research Counci I require a t least a tension
equal to the proof load in the bolts as
installed.
shown in Figure 7. The flanged nut produced
the greatest bolt load, yet the torq ue 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
t ighten 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 ga l ling
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 considerab le
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 ha rdened washers
are omitted.
2. Bolt Head Tests. The nuts, when
torqued, galled the plates if the washers were

omitted under them while the bolt head s 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 l/16-inch oversize. I n the case of
the regular semi-f inished 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 dec rease in load occurred,
although the final load was sti 11 well above the
minimum tension now specified. For the softer,
heavy sem i- finished hex-head bo 1ts {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 bear ing area of a
regular semi-finished hex-head 3/4- inch bolt
a round a 13/16- inch diameter hole {app roximately
•k
"Bearing a rea" as used here is computed by
subtracting the hole area from the area of the
washer-faced portion of the bolt head, based on
measured dimensions for these spec imens .
13
0.380 square inches). This may account for
the fact that the heavy- head bolt, when
tightened in a 7/8-i nch diameter hole, he ld
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 l/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 bol t 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
14
head type. Note that the heavy- head bolt was
ins t al led 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 tu rning
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 dep ression
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 bo lt heads had deeper head depressions than
similar specimens tightened by turning the
nuts.(!) I t appears that as the heads were
turned mechanica l 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 ru n with several types of nuts and
regular bolts in different size holes are shown
in figure 12 for easy comparison. for t he
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-
slstently 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 t he 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 wi th the finished nuts
and washers was only 34,000 pounds, a value
only slightly above tha t 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
t he 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 a lso worthy of note
that the flanged nut cons istentl y developed a
load equa l to or greater t han the load
 Bol r H£AO Sruoy
5PECI M £N

Fic;tJRE 1. TYPICAL ASSEMBLIES FOR RELAXATION SPECI/'1.:NS

F I GURE 2 . TYPICAL FATIGUE SPECI/'1.:N St-()WJNG MECHANICAL DIALS FOR " SLIP" t-EASURE/'1.:NTS
 h......._ Actual Hole Dlam. • d
+ + + + a~ 1 + + + +

+ + + +
9
I ','
a ..J--
h /~
d
-- i-
.1 + .j. + +

e p e

Critical Section for Critical Section for

~ Outside Plate) .Center Plate
r ~: I I! i ~: :v:

SPEC I HEN BOLT CRITICAL

a b c d e f 9 h p
OIA. PLATE

I thru 5 3/4 Center I 3/8 3/8 5/8 13/16 I 1/4 6 3 1/4 0 3 1/4

6 3/4 Outside 1 5/8 3/8 3/4 13/16 1 1/4 6 1/2 2 1/4 1/4 3 1/4

7 3/4 Ouuide 1 1/4 3/8 3/4 7/8 I 1/4 5 3/4 2 3/4 1/4 3 1/4
8 3/4 Outside I 1/4 3/8 3/4 13/16 1 1/4 5 3/4 2 3/4 1/4 3 1/4

9 I Outside 2 3/8 3/4 I 1/16 2 1/8 9 4 1/2 1/4 3 1/4

Speci111en
types A, 3/4 Outside I 1/4 3/8 3/4 7/8 I 1/4 6 1/4 2 3/4 1/4 3 1/4
II, C,
and 0
C1 5/8 Outside 1 1/4 3/8 3/4 11/16 I 1/4 6 1/4 2 3/4 1/4 3 1/4

FIGURE 3. DETAILS OF FAT IGUE TEST SPECIMENS

40
I' One Klnute? L
I \ Load
-•.
~
Q. 30
J I'Haxl. . Load
~
20 I I
~ Spec. No. N-145
FIGURE 4. RECORD FROM SANBORN RECORDER -•...
0 10 J Heavy Nut Thin
llo Washer
I
Regular ltud lolt
0 I
o o.2 o. 4 o. & o.a 1.0
Tl•• Klnutes fro. Maxi- Load
 40

30 30

...
~

'V
0
...J
.Q
~

0
20 20
m

"c:
~
::J
co
N

-
%: 'V
Ill
0
0"c: ....110
~
10 10
0
0
~
Q..

0
Heavy Heavy Nut-Thin Finished Finished Flanged
Nut Nut Nut-Thick Nut
3/4 x 6 Regular sen~i-finished hex head bolts, Tightened to 5000 lb. + 1/2 Turn
11

Plots are average values where duplicate tests were conducted.

FIGURE 5. NUT TESTS - EFFECT OF \>/ASt-ERS ON CLAMPING FORCE OF A325 BOLTS

800 800 800.------.

roxl . .te torque required _ _~

700 700 to produce Proof Loecl when
w•shers are used.*
.Q • 600 600
.
....... ..
500 500 1
-" ..
"'
:a
~ 400
~
400 400 0
0 z
1-
300 300 300 300

200

100

0
200

100

0
200

100

0
200

100

0
II
Heavy Heavy Nut-Thin Finished Nut Finished Flanged
Nut Nut-Thick Nut
* Sued on T • 0.2 PO
3/4 x 6" Regular semi-finished hex head bolts, Tightened to 5000 lb . + 1/2 Turn .
Plots are average values where duplicate tests were conducted.
FIGURE 6 . NUT TESTS - EFFECT OF WASHERS ON TORQUE OF A325 BOLTS
 SPE --\ M EN I v. 7 s
\1 AV'< \'l T
0 WA H&::-R

FIGURE 7. AN EXAMPLE OF SEVERE GALLI~~(; THAT RESULTED FROt-1

TURNING THE NJT AGAINST THE PLATE WITHOUT A WASHER

SPEC\MEN No. 2
REGUL.Aa BoLT HEAD

No WAsHER
FIGURE 8. DEPRESSION OF PLATE SURFACE
RESULTir-.K; FROM BOLT HEAD
 'V
•.,
z
:30 •"'
'V

.
.Jt
s J"'
.!"'
••
1_,
. ..•
:a
:a
J
0 20
..•I
•
X

J 10

13/16 13/16 27/32 7/8 13/16 13/16 1/8

Hole Dl..eter , Inches Hole 01.-eter, Inches
3/4 x &• bolt, Tightened to 3/4 x &• Bolt, Tightened to
5000 lb. + 1/2 Tum 5000 lb. + 1/2 Turn
Regular S..I·Finlshed Hex Head Bolt Heavy S•I-Finlshed Hex Head Bolt
(HI9h Hardness Bolt) (Low-Hardness Bolt)
FIGURE 9. BOLT HEAD TESTS - EFFECT OF WASHER AND HOLE SIZE
 ---Torque~~ Head
I ll llillllllllllllllllhegulu Head
V././U//.ATorqued ....vy l§#ig§g§3He.vyNeM
111ft

:30
~
~

It
I....!
roof Lo.d (28.400 lb.)

ApprOK. torque req 1 4

to produce Proof Loed
.30
a. 1-- ~--·
v roof Loed (28.400 lb.)

Approx. torque req 1 4

.,.
.Jt to pr<*tu Proof Lo.d
., N when washers arc used • when washers arc • • ·
]
..
0 20
::. 600
.
~
3
20
•
..•
~
•
0

..•
:I
c .
z:
-...
.a

.....
z:
•c
0 10 IS
. g• 10 ...
::i IN ; 200 H.O--l~H
~ I~
~
., I~
:I
f
~ 100 H~--l_.,H -
......"
:I
c:r
0
!::
0 ~ ~ 0 '--"a.D..---LU.......J 0
13/ 16 7/8 Hole S"lze 13/16 7/8 Reg. Heavy Reg. Heavy
3/ 4 x &• Reg. semi-fin i shed hex head bolts 3/ 4 x 6" Bolt - 13/ 16 l n. dl a. hole
(no washers) (Spec l-.ns torqued to appr~l~a t ely
salle load - no washer)
FIGURE 10 . A325 BOLT HEAD TESTS - HEAD
OR NUT TORQUED; EFFECT OF HOLE DIAMETER FIGURE 11. A325 BOLT HE.AD TESTS - HEAD
~ BOLT LOAD MD TORQJE TORQUED; EFFECT OF HEAD SIZE ~ TORQUE

.. 30
..,
z
..
Q. ...
....•
:5 ...
,;
... ..•
1
3
.
~
..•
.! Jl

-
0

;•
20 Jl
0
'"'
c
&
•c
0
10
II
Kole ot-ters
(a) fini shed Series Nuts (b) Keavy S.rlea Mvta (c) Flanged Mvts

FIGURE 12 . EFFECT OF HOLE SIZE ~ BOLT LOAD

 ..J..,..

4Q
~
.J'l""""~
\lJvy
(
Nut
• 9 1)
/
v v_-
•
A.
..........._
.
Jll.

;;- 30 .& ~ Finished llut

<a. • as)
3
..
.....
L ,r
I_
..... c;:wr

,
.!
.r01.
..•
5
c
....
20
f/ I I
I I
Specl•n
~ ) I I

•0
~
10 I I
I Motu :
Wuher under llut
13/1& ' " · 01.-. Hole
Regular Head lolt

0 I I
0 1/4 1/2 3/4
Turns fro. "Snug" of S kips
FIGURE 13. LOAD-TURNS RELATIONSHIP FOR HEAVY AKJ FINISHED NUTS USED IN THESE TESTS

3/4 x &• Regular hex head bolts

Plots are average values where duplicate tests were ~ondu~ted

ApprOKIMate torque
req 1 d to produ~e
Proof Load when
washers are used •

.
..0

.
.......• 400
"~
fT

.....~ 300
200

100

0
13 13 27 7 13 13 27 7 13 7
T6 T6 3i 8 T6 T6 3i 8 16 8
Hole Diameters
{a) Finished Series Nut (b) Heavy Series Nut (~) Flanged Nut
FIGURE 14. EFFECT OF HOLE SIZE ON TORQUE FOR 5000 LB. PLUS ONE-HALF TURN
 70
I
. ......---'- -- - 1-------- _- f-.': 7/8"4- A3S4 BO
60
~roof Load 7/8"¢ ' -rl!---+---~

.. I A3S4
..,__...,._ 80- - - - - - - - - --
1 , 7, 8··¢
~ so
1--i- . . P:-L: 7/811~ "! A3S4 BC
......
.>t ----- - --

1.; A3S4 BC I _r 3/411 ¢> A3S4 80 f.

1/- - -- -- - r- - ri -__~
1 --
1 ... ...

"'0 t/
~ 3/4 ¢> A32S l 11
------ - I
. 40
; : 7-- p. L. 3/4"¢> (Hl gh Hardness 32 Rc) _ .{ 314,~
~

.g
c:
..."
30
,IL
IV'. _ A3S4 80 - - - - - - - - - - , - - - -
!1 --;~:.-3~4"¢>
A3S4 8C
P.L. 3/4'~ A32S
113/41 ~ A32S
- A3S4 BC

( Low Hardness 23 R)
c
-I- - 1 --
---~--
I
I
~
0
liD 20 t-1--t- - +- - + -- + -- - + - - · - __ j__-
All bo l t s had g rips ~f
2 1/411 and were torqo, ad
10 In a Skldnlore-W I Ihl!lr- -
Calibrator

0 0 . 02 0 .04 0,06 o.oa 0.10 0.12

l
0 ~----~----~------~----~------~-----' ~-----L-
o. 14
__ ____
0 .16
Average Over-all Elongation of Bolt, Inches

FI GURE 15. C<M>ARISON OF LOAD-EL{)ti(;ATI{)II BEHAVIOR

OF VARWJS BOLT TYPES AT I DENT I CAL GRI P

70

60

..
~
50
A3S4 BO Bolt
2 1/4 ln. Grip 7 ·/ A3S4 BD Bolt, 1 3/4 ln. Grip

------ ------r--
-------
.>t ____ _ J

~
...3 ,, ;
-Proof Loed
---- r---- -- ---
.. 40
T A3S4 BO /
A32S Bo l t
I 3/4 ln. Cirlp
...

.. ~~:
~

g
c
..."
30
I A325
----- -z---- -
A32S Bolt
2 1/4 ln. grip
-.::::...--- ....
-
~ All bolts were 3/4 l n. dla~. by
I
•0 20 3 l n. long wi th suitable nu t s .
I The I 3/4 ln. grip Incl uded ap -
proxiMately 7 th reads and 2 1/ 4
I ln. grip approximate l y 12 threads
10
i/ Torqued In a
Ca II bra tor .
Skl~re - WI lhelQ

0 I I I
0 0. 02 0.04 0.06 o.oa 0.10 0. 12 0.14 1) .16
Ave rage Over-all Elongation , Inches

FIGURE 16. COMPARISON OF LOAD-ELONGATION DATA FOR DI FFERENT GRI PS

 10

-- -·- ---
I I I I I I I I I I I
I I
/ ;;;.._/7/8''~ A3~ 80
60
Proof Load 7/811~

- A3~ BC t' ~~

. --
I --- 1 - - - ec
~:l· 7/8"~
A3~ eel
l
~"' /:,_3/4'~ A3~ 80
-----1.
f_11&·~ A3~
1

',...,
P.L. 3/4"~ / ~ ::::::-- --~
-- -
~'
/" 3/4"~
~
A325 High Hardnen
A3~
--- ----- -r---
'~
...0 BOj _:::1-~
-- :---
f----

--
~
,_2.. L.
A3~
3/4"~
BC I 1,' ~A325./ 3/4"~ A3~ BC---"- -..... r--- 1---
......

.. 30 P. L. 3/4"~
c:

..
...."
A32s
,rt
,g 20
II I

,_1
10
jl All bolts had grips (head to nut)
of 2 1/4" and were torqued In a -
v' SkidMore Wilhel• Calibrator .

0
J/ I I I I I 1 I I I I I I I I
0 0. 5 1.0 1.5 2.0 2.5
Average NURiber of Turns fro. "Snug'' of 5 kips

FIGURE 17. COMPARISON OF LOAD-TURNS DATA FOR VARIOUS A325 AND A354 BOLTS AT IDENTICAL GRIP

70
I I I I I I I I I I I

60

. so
,.,.-· --- r--~ A354 80 Bolt, I 3/4" 6irlp

-~ . ----
·----L -J
G.
~

'V #~- ·- --<!..354 80 •

Proof L011d - lolt, 2 1/4;~.
....8 40 ~ ....
...0 A3~ 10 j ,/ "'i.A32S lolt, I 3/4" 6irlp ---...... "'t 6irlp

--8 £0 ~-r:.3-'T
- A32S Bolt, 2 1/4" 6irlp --
r
~-

r---
.
c
30 Proof L011d
AJ2S

-~
:!
:::.0
20

I
OD
All bolts were 3/4" dl-. by
3" long, Torqued In a SkidMore -
Wllt~l• C.llbrator
10
~/

0
~,
I I I I I I I I I I
0 o.s 1.0 1. 5 2. 0 2.5
Average Nuilber of Tums fro. "Snug'' of 5 kips
FIGURE 18. C(M)ARISON OF LOAD-TURNS DATA FOR DI FFERENT GRIPS A325 MD A354 BOLTS
 ~
3
... ·-
GJ.J:!
~
~

-I-
IU '- 2
....0 "c:
~

~-
.... ~
M
"'E 8 ~

~ ...

67l~~l
.... 'to-
0
Threads ll2 3 3 3 ~ 4J.2
In Grip 2 2 ...
2 , c.:l
'-- '----..,-------'
Grips 1.35" I. 5011 ).sou
Used
(a) Bolts Tightened to Failure In Solid Blocks

~--------------------~~----_, :~
rom
II
.J::
~

N
M
<t

0
Threads sJ.2 7 ,oJ.2 11 12 12
In Grip
Grips 1. 7511 2.25"
Used
{b) Bolts Tightened to Failure in SkidMore-
Wilhel• Calibrator
(All Bolts were 3/4 in. diaMeter)

FI GURE 19. CCM'ARISON OF TURNS TO FAI LURE FOR VARIOUS BOLT TYPES, GR IPS, An) CONDITIONS
 70
where~ •
I I I 1
Approxl.ately T • 0.2PO TorqJe, ln.- lb.
P • Bo I t Load , I b.
0 • Bolt Olaa., ln.

r·
60 T • 0.2(30,000)(3/4)
• 4SOO ln. lb • 37S
ft.
...0.. so
..- A354 80 Bolt
""' /'
.......
"0 Proof Load A354 BO ,/
~40 A354 BC Bolt
....
.. I I
.:~ 1-
~
A32S Bolt
~
0 ~oof Load, A354 BC
I I

~
g
..
c
"
1-

0
30

20
~oof Load, A32S

/
v .J

~
CD
All bolts were 3/4"~ by 311 long and

/ tested at a grip of I 3/411 In a

Skidmore - Wilhelm C<!llbrator.

v
10 Hardened washers were used under nut~.
't'

0 0 100 200 300 400 500 600 700 800 900 1000
Average Torque, Foot - Pounds
FIGURE 20. LOAD-TORQUE RELAT I CNSH IPS FOR VARI OJS BOLT TYPES

All bolts were A354 BD 3/4 ln. dl~. x 6 ln.

long. and tightened against A7 plates with
13/16 ln. diameter holes. NUMbers In paren-
theses Indicate no. of tests represent ed.
Add it iona l Load
No When Torqued With
Washers to Fai lure Washers
r·-- -- --i
1111111111011111 --------4
60

;-;( I)
so I ... - ~ (1)
I
11 Proof
J!L
I U I I
I
l---- =! !.-
\Load
4)
I
I
.
I
1
~ I

(2)-- "'
-'~
I( 1)-
Q..l -
1
I
I wt
~

1--

-u..
( 2):;
-Q,.

:J
I--
....~ .
-u... -... -...
1-- Q, ~ lE:
~-
Q,
Q, Q,
~
~
.... ..
u u
] o..-
011-

1--
-..c: ~ ~
1"-
.... ... ... ....
"0 vc:
.c-
-
1-- ...
v "0

" ... .. ..::-

"0 "0
.c 4"0
10 1--- 1- v u
J:

.. _
v
"...
1- J:
1-
J:
1- ......
U'-
J:J:

0
1---
..____
J:
1-
~
.,.. - -
N N
:a.o
L--

FIGURE 21. BOLT TENSION AT " SNJG" OF 5000 LB. PLUS ONE-HALF TURN FOR A354 BD RELAXATION TESTS
 Note: All fasteners are 3/4 ln. nominal dla~ter.
60~--------~------~--------~--------~--------~---------,
A354 BD Bolt, Tl plates ,
--- Shear P 1ane I
354 BD Bolt, A242 plates,
in Shear Plane
50
I
A325 High Hardness Bolt,
A7 plates, Shank In
Shear Plane
-• 40
Q.

.,.
.Jtl.

i ········ A325 Bolt , Tl plates,

!
"0
30 Shank in Shea
Plane

~
•
•uc ? ()

....

10 t--+-+-~-~ A325 Bo 1t,

A7 plates,
Threads In
Shear P Jan

0 10 20 30 40 50 60
Shear Load to.ponent, kIps
FIGURE 22. INTERACTION CURVES OF A325 AND A354 BD BOLTS
UI\OER COMB INED TENSION AND SHEAR SHO.<I ING FAILURE LOADS
 52

48
Average Curve -
(II Tests)
44

40

36

.
-..
Q.
~
32

g
..
c
u
28
1-
.... 24
0 Average Curve - Is
IQ
(27 Tests)
20

16

12
Flat Surface
8 Tests
Approx. 1/2 Turns
(Extremes) T~o 5% Converging or
4 Crossed Bevels Tests

0
0 1/4 1/2 3/4 ) 1/4 ) 1/2
Turns frOftl "Snug'' of 5 kips
FIGURE 23 . EFFECT OF BEVELED SURFACES ON BOLT TENSION VS . TURNS, 3/4 IN. DIAM. A325 BOLTS
 72 ~--~---.----.----,---.----,----.----.----r----.---,----,

68 ~--+---~--~----+----r--~r---+----+--~----~---r--~

64 ~--4----+----+---~--~r----r--_,----~---+----r---~--~

56
Average Curve -
(8 Tests)
52

..
~ 40 t--- - + - - 4-+--14- -l

..
..g 3& ~--+---~~~~~~~~
c Average Curve - Bevels
...." (8 Tests)
~ 32 ~--+--~4-~-~~~~~---t----
0
•

16
Approx. 3/8+ Turns
- (Extrea.es)
12 ---+-·-

D Flat Surface
Tests

~ Two 5% Converging
~ Bevels Tests

0 L---~--~----L---~--~----~--~--~----~--~--~--~
0 l/4 1/2 3/4 1 1/4 l l/2
Turns fran~ "Snug" of S kips
FIGURE 24. EFFECT OF BEVELED SURFACES ON BOLT TENSION VS. TURNS 1-JN. DIAM. A325 BOLTS
 (a) Locatio n of Fatigue Crack in Specimen B l

SPECIMEN-BJ
's"~HOlf- 29,4oo•soLr TENS.
SR : 0+30 KSI N:: 11 883,100

(b) Close-up of Fatigue Crack in Specimen B l ( Note bolt head depression)

FIGURE 25. TYPICAL FATIGUE FAILURE

 (a) Ge neral Vie w of Outside Plate Showing Bolt
H ead Depressions and Extent of Fatigue Crack

(b) Close-up of Fatigue Crack Surface

F IGURE 26. VI EWS OF FATIGUE FAILURE OF SPECit'EN 05

 60

••
. -I....
.... -
l it
• ~
lit~
....0
J I
I 40 N

Deal .. StreaMs
AlfA-MSIIO Speclflcetl . . s, A7 St"l

llote:
llvet date
fro. Ref. 15
-40 -30 -20 -10 0 10 20 30 40 50 60
Klnl~ Stress, ksl (eo.pressl ..) KIRI~ Stress, ksl (Tension)
(All StrMMS ere lased on let SectiOft)
FIGURE 27. MODIFIED GOODMAN DIAGRAM FOR 21 000 1 000 CYCLES

100 0

~ t-· ·-t-· - W.J • JJ~~.. .l_~}~!L.1 J.!illlllfa,UJ

1 1
~
~x'*--A•f;;t~~;+~;- '~.-;.~~~i#;:itjt -
01
c
80
\
\

- t-· - . .....l....l. .U UJJL

I)\«* 8 fasteners
-~~ .L.I W,IJ- ...&. - ~~~~~J..-~
In spec:l•n• that ailed at '* t-·
20

-I
c
\ -•i&
.:
g
-..
60
r--.
- ~'* A fasteners In lpec:l . .ns that
failed at Rc»t A
40

-•..
....
0
c
...." 40 II II I
!
60
c

-• II.
•
....•0
~
0
l r---1~Rc»t II
Center
~

20 f+·--1 ~; A··Crltlc:al 80
f4'.:::..~ sec:tlon for out•
side pletes (See
_( ~lg. 3)
~ Outside Plates
0 100
0 5 10 100 1000 10,000 100,000 1,000,000 10,000,000
Applied Cyc~s

FIGURE 28 . AVERAGE LOSS IN BOLT TENSION VS. APPLIED CYCLES JN FATIGUE TESTS
 TABLE 1. RESULTS OF BOLT- TENSION RELAXATION TESTS FOR A325 BOLTS

( 1) ( 2) (3) (4) (5) (6) (7) (8) (9) ( 10) ( 11) ( 13) ( 14) ( 15) ( 16) ( 17) ( 18)
( 12) ( 19)
Specimen Bolt Nut Washer Ha r dne ss Torque Elong. Max. 1 Min. Percent Loss of Load f rom 1 Minute Load Depth of
Nlllllber* Head Bolt Nut ft. lb. in. Load Load 5 Da s Depression
Rc RB Min . 3 4 5 6 7 21 in.

Nut Tests
N- 5 Regular Flanged No 32 90 580 .03204 40 , 230 39,150 0.4 3.9 3.9 3.8
N-55 Regular Flanged No 32 90 530 42,800 41,740 0.8
N- 5s- c Regular Flanged No 32 90 520 42,880 41,800 1. 3

N- 6 Regula r Heavy Nut 32 90- 92 .02577 40,930 39,430 0.1 3.2 3.8 4.1
N-6S Regular Heavy Nut 32 90- g2 460 .02268 41,040 39,340 0. 4

N-7 Regular Heavy No 32 90- 92 . 02136 40,580 38,840 0.3 3. 1 3. 1 3.2

N-7S Regular Heavy No 32 90- 92 700 .01937 39,960 38,540 0. 5
N- 7S- B Regular Heavy No 32 30-92 600 .02144 41,160 39,400 0.6
N-7s- c Regular Heavy No 32 90- 92 500 . 01708 37, 000 35,600 o.8

N- 8 Reg u lar Finished Nut 32 84- 86 .01356 33,640 30,020 0.5 1.6 1.6 1.6
N- BS Regula r Finished Nut 32 84-86 420 . 01545 36,700 33,240 0.3

N-9 Regular Finished No 32 84- 86 490 .01518 35,900 33,100 0 1.8 1.8 1.8
N- 9S Regular Finished No 32 84-86 500 . 01520 33,960 31,680 0.4
N- gs-B Regular Finished No 32 84- 86 620 .01805 36,100 33,960 1.2
N- 95- C Reg ular Finished No 32 84- 86 725 .01458 33,660 31,760 0.3

N- 10 Re gula r Fin.- Thick Nut 32 79-82 480 . 01791 35,970 34,350 0. 1 3.2 3.4
N- IOS Reg ular Fin.- Thick Nut 32 79-82 440 . 01914 37 . 740 35,960 0.3

N- 11 Regular Fin.- Thick No 32 79- 82 750 . 01534 34,440 33,720 0. 3 2.0 3.0
N- IlS Regula r Fin.-Thick No 32 79- 82 600 .01664 35,700 34,700 0.9
1 32 92- 95 450 . 01947 40,650 38,060 0 2.2 2.2 2.2
N- 12 Re gu Jar Heavy- Thin No
N-12S Regula r Heavy-Thin! No 32 92- 95 540 . 02186 41 , 140 39,240 0.3

N- 13 Regular Heavy- Thin Nut 32 92-95 420 .02075 40,560 37,280 0.4 2.6 2.4
N-13S Regular Heavy- Thin Nut 32 92-95 460 .02144 40,780 38,680 0.2

N- 145 Regular Heavy- Thin 2 No 32 92- 95 610 . 01893 38,880 37,500 0.3

N-ISS Heavy Heavy No 24 90- 02 500 .02822 33,320 31,920 0

N-16S Heavy Finished No 24 84-86 380 .02343 30,200 28,600 0.1
N-17S Heavy Flanged No 24 90 460 .03625 34,200 33,700 0.3
P-185 Regular Heavy No 32 90- 92 510 . 01824 38,880 37,540 0.3 .0090

For footnotes see end of Table.

 TABLE 1 (CONTINUED)

( 1) (2) (3) (4) (5) (6) ( 7) (8) (9) ( 10) (11) ( 12) ( 13) ( 14) ( 15) ( 16) ( 17) ( 18) ( 19)

Specimin Bolt Nut Washer Hardness To rque Elong. Hax. 1 Hin. Percent Loss of Load from 1 Hinute Load Depth of
Number Head Bolt Nut ft . lb. in. Load Load 5 Da s Depression
Rc R8 Hin . 3 4 5 6 7 21 in.

Head Tests
H·l Regular Heavy Head 32 90- 92 40,200 38,670 1.0 4.7 4.8 4.8 5. 1
H-IS Regu lar Heavy Head 32 90- 92 480 39,800 38,620 0.8

H- 2 Regular Heavy No 32 90-92 .01940 39,960 37.980 0.6 3.4 3.5 3.6 .0087
H- 2S Regular Heavy No 32 90- 92 480 41,140 39,120 0.4 .o 112
H· 2S·B Regula r Heavy No 32 90- 92 420 .01627 35,800 34,260 0.5 .0105
H-2S-C Regular Heavy No 32 90- 92 420 .01682 37,960 36,560 1.0 .0143
H- 2S- 34 Regula r Heavy No 32 90- 92 440 .01620 35,340 34,840 0.2 .0089

H-3 Heavy Heavy Head 24 90- 92 34,080 33,500 1.1 2.2 2.2 2.2
H-3S Heavy Heavy Head 24 90.92 380 .03547 31,000 30,420 0.5

H-4 Heavy Heavy No 24 90- 92 .03178 34,300 33,700 1.2 3.0 3, I 3.2 .0028
H-4S Heavy Heavy No 24 90- 92 420 .03195 34,400 33 , 620 0. 4 .0031
H-4S-C Heavy Heavy No 24 90- 92 440 . 03200 33,800 33,180 0.5 .0069

Head Torgue Tests

H-19S-HT Heavy Heavy No 24 90-92 400 .02844 32,300 31,060 2.6 .0040

H-20S- 31 -HT Regula r Heavy No 32 90- 92 460 .01502 32,920 31 , 960 0.8 .0082

H-21S- HT Regu l ar Heavy No 32 90- 92 480 .01868 39,340 38,480 0.7 .0132

H·22S· C- HT Regula r Heavy No 32 90- 92 500 .01610 36,520 35,220 0.3 . 0166

Lubrication Tests
N-23S Regula r Heavy No 32 90-92 620 .02626 41,100 3g,560 0.6
N-235-L Regula r Heavy No 32 90- 92 340 .02291 41,400 39,640 0.6

* Meaning of Letters in Specimen Number

H - Bolt Head Test; N - Nut Test; p - Punched Hole; S - Short Time Test (5 minutes); 8 - Oversize Hol e- 27/32 in. dia.; C - Over size Hole- 7/8 in. dia.

HT - Head torqued to tighten bolt; L - Lubrication between nut face and plate.

Proof Load : 28,400 lb. Hinlmum Bolt Tension : 25,600 lb. required by 1954 Specification; 28,400 lb. by 1960 and 1962 Specification of Research Counci 1.

I. Washer Face Down Hardness range by specs. on 3/411 '1' bo Its: { A325- 58T Rc = 23- 32
2. Chamfered Face Down A325- 61T Rc • 23- 35
 TABLE 2. RESULTS OF BOLT-TENSia-l RELAXATION TESTS CONEXJCTED WITH A354- BD BOLTS

(1) (2) (3) (4) (S) (6) ( 7) (8) (9) ( 10) ( 11) ( 12) ( 13)
Spec. Bolt Nut Washer Hardness Torque Hax. Hin. Percent Loss of Load From
No. Head Bolt Nut Load, lb Load, lb l Min . Load
Rc Rc Days
S Min. 2 3 90
NX-245-0- l Regular Heavy- 2H No 37 3S soo 40,600 39 , 210 0. 2
NX-24-0- 2 Regular Heavy-2H No 37 3S S40 37,200 3S,900 O. B 1.3
NX- 24-0- 3 Regular Heavy- 2H No 37 3S soo 36,8SO 34,8SO 0 .9 1. 0
NX-24S-0-4 Regular Heavy- 2H No 37 3S 610 40,SOO 40,300 0. 4
NX-24S-E- l Regular Heavy-2H No 37 3S sao 4l , lSO 40,000 0.6
NX- 245- E-2 Regular Heavy- 2H No 37 3S S30 39,860 38, 1SO 0.4
NX-24S-E-3 Regular Heavy- 2H Nut 37 3S S20 47, ISO 46,840 0.7
NX-24S- F-l* Regular Heavy- 2H No 37 3S 4SO 35,3SO 34,2SO 0.5
NX- 24S- F-i* Regular Heavy-2H No 37 3S 450 34 ,600 33,300 1.1
NX- 24-F-3** Regular Heavy- 2H No 37 3S 37,BSO 36, 7SO 1.1 4. 2
HX- 255- E- 1 Regular Finished- 2H No 37 26 480 36,230 34,700 0

H • Nut Test ; X; A354-BO Bolt; S • Short Time Test (S Hinutes); 0 =3 Threads in Grip; E • 6 Threads in Grip;
F • 12 Threads in Grip; A7 Plate used under head and under nut .

Proof Load- 40,100 lb. 5 Washers In Grip Hardness by Spec. - A354 BO Bolts 32- 38
* S/8 Plate in Grip Al94 2H Nuts 24- 37
**
TABLE 3. SlM-1ARY OF VAR IABLES FOR TESTS OF BOLTS TIGHTENED ON BEVELED SURFACES

(Numbers in parentheses indicate number of tests performed.)

Tests Performed at : University of Illinois (54); Lamson a nd Sessions Co. (9)

Period Covered: 1961 - 1963
Bolt Sizes: 3/4 x 3 (2); 3/4 x 4 (12) ; 3/4 x 6 (32); I x 6 (17)
Bolt Types (A l l A325): Heavy Head, Old Thread length s (32); Fini shed Head ,
Old Th read Lengths (12); Heavy Head, New Thread Lengths (19)
Nu t Types: A325 (52); A194 Grade 2 Heavy Hex (II)
Grips (where kn~n): 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 kn~n): 3(5); 4(12); 5(2) ; 6(7); 7(5); 8(15);
9 (4); 10(4)
Type of Calibrator: Load Cel I (16); Load Analyzer (9); Skidmore Wilhelm (38)
Method of Torquing Nut: Manua l (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); Nu t only (4);
Head and Nu t (39)
Position of Tv10 Bevels : Pa rallel (4) ; Converging (33); Crossed (2)

SPECIAL TESTS
Position of Bolt: Maximum Sliding (4); Minimum Sl iding (4)
Bolt Strength: High Ha rdness (2); L~ Ha rdness (2)

Minimum Slide bolt placed against I~ side of bevel t o minimize bolt movement
d~n s lope .
Maximum Sl ide = bolt placed against high side of bevel to permit maximum possible
movement d~n be veled slope during tightening.
Conve rge a Beve ls placed with sloping p lanes conve rging (or dive r ging) .
Pa ra llel • Bevels placed with sloping planes para llel but with 5 percent slope
at head and nut .
Crossed .. Bevels pl aced with sloping planes "crossing" or t urned at go deg.
rel ati ve to each other.
 TABLE 4. RESULTS OF FATIGUE TESTS, SPECIMENS 1 THROUGH 9

(I) (2) (3) (4) (5) (6) (7) (8) (9) (10) (II) ( 12)

Spec. Fastener Type Hardness Stress Cyc Ie Applied Kethod of Critical Remarks Average % Loss in Nom.
Bolt Nut Net Section Cycles Tightening Plate Bolt Load Bolt Load Coef.
Rc Ra Stress, ksl Before Test During Test of Frlc.

Heavy Head Bolt, 3/411 24 0 to +30 2,376,300+ 5 kips * Center T: S • 1.0:0. 75

Heavy Nut with Washer 90-92 + 1/2 T
2 Heavy Head Bolt, 3/411 24 0 to +30 2,965,900+ 5 kips * Center T:S • 1.0:0 .75
Heavy Nut with no Washer 9D-92 + 1/2 T
3 Heavy Head Bo It, 3/4" 24 0 to +30 2,692,goo+ 5 kips * Center T:S • 1.0 :0.75
Flanged Nut 90 + 1/2 T
4 Heavy Head Bolt, 3/411 24 0 to +30 6,300,000+ 5 kips * Center After completion of first 2,500,000
Heavy Nut-Thin with 92-95 + 1/2 T cycles test was discontinued. Two
no Washer months later test was resumed and
continued until specimen felled In
grips of machine. T:S • 1.0:0.75

5 Heavy Head Bolt, 3/411 24 0 to +30 2,257,900+ Bolt Load • Center During first cycle, specimen slipped 30.3 19 0.32
Heavy Nut with no Washer 90-92 29,000 lb. Into bearing when net stress was
approximately 29,000 psi .
T:S • 1.0:0.75

6 Heavy Head Bolt, 3/4" 24 0 to +30 2,416, 900+ 5 kips* Outside Failed in outside plate on bolt head 33.5 25(a) 0.35
Heavy Nut with no Washer 90-92 + 1/2T side. During first cycle, specimen
experienced large slip at 27,000 psi.
T:S • 1.0:0.90

7 Regular Head Bolt, 3/4" 32 0 to +30 3,073,900+ 5 kips * Outside Hole- 7/8 in. india. Slipped 32.1 24 0.28
Heavy Nut with no Washer 90-92 + 1/2 T during first cycle at 26,200 psi.
T:S • 1.0:0.75

8 Heavy Head Bolt, 3/4u 24 o to +30 5,148,600+ 5 kips* Outside Slipped during first cycle at 35.2 29 0. 20
Heavy Nut with no Washer 90- 92 + 1/2 T 20,400 psi. Specimen bolts were
tightened by turning nuts on 2 bolts
and bolt heads on the other 2 bolts.
T:S • 1.0:0.75

9 Regular Head Bolt, 111 24 0 to +30 4,892,400+ 10 kips* Outside Slipped during first cycle at 62.8 3 0.24
Heavy Nut with no Washer B8 + 1/2 T 24,800 psi. Specimen bolts were
tightened by turning nuts on 2 bolts
and bolt heads on the other 2 bolts.
T:S • 1,0:0.75

+ Specimen did not fall. * 5 or 10 kip bolt load was developed by use of torque-tension relationship; nut was then given an additional 1/2 turn.

(a) Since Specimen No.6 failed, the loads of the 2 bolts on the line of fracture are not Included.
 TABLE 5. RESULTS OF FATIGUE TESTS, SERIES A1 B1 C1 AND D

{I) {2) {3) (4) {5) (6) (7) (8)

Spec. Stress Cycle App I ied Bolt Tension, Net Tensile Area, T:S Ratio Nom. Stress Nom. Coeff.
Net Section Stress, ksl Cycles sq. in. at 1st cycle of Friction
Nominal Actual Average % Loss Nom. Actual Calc.* Actua l slip, ksl based on 1st
Ini tial, During Fatigue Meas. cycle slip
kips Test
+ 34.8 14 2.84 2.74 I :0.80 I :0.77
AI 0 to +28.8 0 to +29.8 2,435,400
A2 0 to +28.6 0 to +29.6 2, 110, 100+ 36.5 17 2.86 2.75 1:0.80 1:0.78 25.5 0.25
A3 0 to +28.2 0 to +29.2 2,657,600+ 33.2 28 2.89 2.79 1:0.82 1:0. 79 26.9 0.28
15.7(a) 0.19(a)
81
82
0 to +30.0 0 to +31.1
0 to +30.0 0 to +31.1
I • 883 • 100(+ I)
2,067,200
29.4
29.3
-
21
2.88
2.87
2.78
2.77
I :0.81
1:0.81
1:0.78
1:0.78
19.9(b)
20.9
0.24(b)
0.26
83 0 to +30.0 0 to +31.1 2,455,7oo+ 29.2 21 2.87 2. 77 I :0.81 I :0.78 18.6 o. 23

Cl 0 to +30.0 0 to +30.0 2,390.5oo(l) 28.7 33(c) 3.07 3.07 1: 1.25 I: 1.25 14.8 0.20

Dl
02
±20 .0
±20.0
±20. 8
±20.8
1,860,200~2)
20,800 3)
28.4
28.4
-- 2.83
2.87
2.73
2. 77
1:0.80
I :0.81
I :0.77
1:0.78
18.4
8.1
0.23
o. 10
03
04
±18.0
±20.0
±18. 7
±20.8
Jo,ooo{J)
2,178,000(4)
28.4
29.0
-
8
2.85
2.84
2.75
2.74
I :0.80
1:0.80
1:0.78
1:0.77
14.4 0.18

05 ±20. 0 ±20.8 2,267,ooo(5) 28.4 13 2.82 2.72 I :0.80 I :0.77

Static Specimen: Ultimate strength • 60.8 ksl based on nom. tensile area

+ Indicates no failure (a) Minor slip (b) Major slip (c) At unfractured portion
* Calculated using actual hole diameter, Nominal T:S ratios were 1.0:0.75 for A, 8, and D with 1. 0:1 . 14 for Cl.
(I) Failure In outside plate (bolt-head side) at critical section. Fracture s t a rted at inside faying area--
possibly a result of fretting.
(2) Failure In outside and center plates at critical sections. Fracture In outside plate (nut side) appeared to
initiate at faylng surface; center plate fracture (bolt-head side) possibly a result of fretting.
(3) Test stopped because of excessive slip, pounding and high temperature (see text) .
(4) Specimen stress Increased from ±16 ksi to ±20 ksi afte r 253,000 cycles. Thus total specimen life was
2,431,000 cycles without fail ure.
(5) Specimen stress Increased from ±16 ksi to ±20 ksi after II , 100 cycles. Thus total specimen life was
2,278,100 cycles. Failure occurred in outside plate (bolt-head side) slightly beyond critical section.
Fracture appea r ed to Initiate by fretting between edge of bolt head and plate.
 TABLE 6. RESULTS OF PREVIOUS FATIGUE TESTS

( 1) (2) (3) (4) (5} (6) (7}

Specimen Fastener Type Stress Cycle App 1ied Method of Critical Remarks
Net Section Cycles Tightening Plate
Stress, ksi

1957 Tests( 12)*

H- 1 Regular Head Bolt, 3/4", with Washer ±20 1, 802,400 Bolt Tension • Center Failed in outer pl. at 1,012,200 cycles. New
Heavy Nut with Washer 29,000 lbs . plates used to continue; center pl. failed at
1,802 ,400 cycles. T:S • 1. 0:0.75
H-2 Regular Head Bolt, 3/4", with Washer 0 to +30 454,500 Sol t Tension • Center T:S l . O:ll . 75
Heavy Nut with Washer 29,000 lbs.
H- 3 Regular Head Bolt, 3/4", with Washer 0 to +30 3, 144 ,000+ Bolt Tension • Center T:S 1. 0:0 . 75
Heavy Nut with Washer 29,000 lbs.
H- 4 Regular Head Bolt, 3/4", with Washer !20 3,026,600+ Bolt Tension a Center T:S 1.0:0. 75
Heavy Nut with Washer 29,000 lbs .

1954 Tests( 13)*

C3- l Regular Head Bolt, 3/4", with Washer :!: 18 517,400 Bolt Tension • Center Specimen slipped excessively. Aft er failur e,
Heavy Nut wi th Washer 23,700 lbs. oi 1 was found between plates. T: S • I. 0:0 . 75
C3-2 Regular Head Bolt, 3/4", with Washer ±18 2,108,800 Bolt Tension • Center Failed . T:S = 1.0:0. 75
Heavy Nut with Washer 25,100 lbs.
C3-3 Regular Head Bolt, 3/4", with Washer !18 1. 998,200 Bolt Tension • Center Failed . T:S ~ 1. 0:0 . 75
Heavy Nut with Washer 25,500 lbs.

BOs-1 Regular Head Bolt, 7/8", with Washer !22 .8 1,206 , 000 Bolt Tension • Center Failed in center plate at critical section
Heavy Nut with Washer 56,500 lbs. T:S • 1.0:0. 75
BOS-3 Regular Head Bolt, 7/8" , with Washer ±20 4,405,000 Bolt Tension • Center Failed in center plate at critical section
Heavy Nut with Washer 56,500 lbs . T:S = 1. 0:0. 75
BOS-4 Regular Head Bolt, 7/&' with Washer ±20 2,083,000+ Bolt Tension • Center T:S = 1.0:0 . 75
Heavy Nut with Washe r 56,500 lbs.
BOS-7 Regular Head Bolt, 7/8", with Washer 0 to +30 1,973,000+ Bolt Tension • Center T:S 1.0:0 . 75
Heavy Nut with Washer 56 , 500 lbs .

( 18}*
Northwestern Unive rsity Tests
K(A7}-1 Regular Head Bolt, 3/4", with Washer 0 to +37. 382,000 Bolt Tension Center All specimens failed. T:S • 1.0:0.75
K(A 7}-2 Heavy Nut with Washer 312,000 26,500 lbs.
K(A 7) -3 225,000
K(A7)-4 Regular Head Bolt, 3/4", with Washer 0 to +35 I ,337,000 Bolt Tension • Center All specimens failed. T:S • 1.0:0.75
K(A7)-5 Heavy Nut with Washer 1,244,000 26,500 lbs.
K(A 7}-6 Regular Head Bolt, 3/4", with Washer 0 to 30 3,000,000+ Bolt Tension • Center T: S ~ 1. 0:0. 75
Heavy Nut with Washer 26,500 lbs.

+ Specimen did not fai 1. * Numbers In parentheses indicate references from which test results were obtained.

developed when a hardened washer was used with
a heavy nut.
The torques required to tighten the
bol ts i n holes of dif ferent diameters and using
the f in ished, heavy, and flanged nuts are
p lotted 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.
the case of the f l anged nut the torque was
fairly constant despite the differences in ho le
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 t orque
of ap prox imately 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 s uch severe gal l ing and a somewhat
lower torque was obtained. The lower torque in
the case of t he 7/8- inch diameter hole is re-
lated in part to t he smaller load developed in
this specimen .
Note that these results are from one
or two specimens of each type; therefore, the
differ ences in behavior shown between the
various types of nuts may be, in part , experi-
ment a l scatter. Additiona l testing to give
average values would be needed if obse rvations
15
other than the following were to be justified:
(l) The turn-of-nut method (snug+ l/ 2
turn) without washe rs a nd in holes
up to l/8 inch larger than the
nominal diameter will produce bolt
loads above the minimum for A325
bo lts as now specified, especia ll y
when heavy hex heads and nuts are
used.
(2) Omi ssion of the washer under t he
In turned element may produce torque-
load relationships somewhat more
erratic than usual; thus, torque
methods shoul d not be used for the
installation or checki ng of bolts
having no washers under the parts
turned during tightening or in-
spection, except for bolts installed
by tur n- of- nut method in accordance
with the 1964 edition of Reference 6.
5. Lubrication. Two specimens
(N-23S and N-23S-L) were t est ed to study the
effects of lubrication. One had lubrication
(molybdenium disulfide) only between the nut
fac.e and t he plate face, and the other was with-
out lubrication. These specimens were bot h
tightened by the turn-of-nut method and de-
ve loped approximate ly the same load, but the
torque requi red to tighten the lubricated
specimen was just slightly more than one-half
that required for the non-l ubricated specimen.
Tests run a decade ago with case- hardened
16
washers and employing a lubricant called
"No-Oxide" showed the following comparative
results:(IO)
Condition Relative Torque
No lub r icant 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 capa cities 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 for this elongation is high enough to
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 I, Columns 12 to 18).
Specimen H-1, which remained tightened for 21
days, showed a 5.1 percent loss from the
!-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 field
studies have shown whether such lubrication
will result in the nuts backing off in service ,
especial l y under vibration or fatigue loadings.
exponent i al manner typical of creep data;
whether there were wa s hers under the nuts or
under the bolt heads did not seem to influence
the load-time relationships.
Relaxation of high strength bolt
assemblies wa s studied in a prelimina ry fash ion
at the University of Illinois in early 1952 by
Robert F. Stevens on l-inch di ameter bolts.
Those results agree well with the results of
the tests reported herein and were summarized
as follows: ( II)
'~he tensile load which the bolt carries
is transferred through the washer to a
narrow, annular r i ng surrounding the bolt
hole. When the stresses in th i s ring
reach or exceed the compressive yield
point stress of the pl a tes, the ring
yields. As the ring yields, the annular
area becomes larger and the grip smaller;
thus the unit stresses are reduced. The
yielding of this area is a lmost instan-
taneous. If an elongation of a certain
amount is desired, and if the bolt load
cause yielding of the plate material (as
would be the case with nea r ly all con-
struction jobs where the bolt tension is
at least 100 percent of the proof load),
and if a torque is applied to the bolt
until this elongation is reached and then
the torque is released, there will follow
an immediate dropoff in the bolt tension
(within 15- 20 seconds) of a magnitude in
the order of 5 percent of the total load.
I f after having reached the desi red
elongation , the torque is maintained for
a short period (a minute or so) and then
released, there wi II still be an immediate
dropoff of the bolt tension, not to exceed
about 5 pe r cent; the amount of loss will
depend upon how long the torque was
maintained . If the torque is maintained
unti l there is no further yielding, we
assume that there would be no loss of
bolt tension when the torque was released.
However, the time e lement involved in such
a procedu re makes it impractical.
"Following the immediate dropoff, there is
a further, gradual loss of bolt tension
lasting for one to two hours, and amount-
ing to another 5 percent of the load.

This loss may be attributable to the
heating of the bolt during torquing and
its subsequent cooling, changes in room
temperature, heating of the annular ring
during yielding and subsequent cooling,
and other similar effects .
·~his yielding of the plate should not
have any great effect upon the recorded
bolt tensions. The initial, large loss
is over , normally, before the bolt
tensions are measured, and the subsequent,
long time loss is small enough to be with-
in the calibration errors.
"The times and percentages indicated
above are merely to give an indication
of the order of magnitude. Their absolute
value will depend upon the yield point
of the plate materi a l, and the bolt
tension and properties of bolt and nut.
Also such local factors as the rate of
application of the torque and plate
surface conditions may affect this
behavior."
It is interesting to note that the
loss in bolt tension after one minute, observed
in the recent University of I llinois 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 mi llscale contact
surfaces as well as a variety of other surface
9
conditions.( ) It is also confirmed by the
tests on A325 bolts conducted over a period of
days in the Lamson and Sessions load
ana lyzer. ( 7)
B. RELAXATION TESTS OF A354 BOLT ASSEMBLIES
AND COMPARISON OF BEHAVIOR WITH A325 BOLT
ASSEMBLIES
In late 1961 the American I nstitute
of Steel Construction revised its specifi-
17
cations for high strength bo l ting 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 a nd 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 avai !able 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 BD
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.
··k
ASTM approved in 1964 a specifica tion for a
structural "higher" high- strength bolt, A490,
which has a heavy hexagon head, short thread
length, and mechanical properties comparable
to the A3~4 BD bolt.
18

Because Grade BC bolts were not averaged 47 . 0 kips. The A354 BD 3/4-inch

commercially available in the sizes and lengths diameter bolts had strengths in direct tension

desired for these tests, Grade BD bolts were averaging 56.9 kips. The A354 BD bo l ts of

annealed and re- heat-t reated to produce the 7/8-inch diameter averaged 73.2 kips.

appropriate properties of Grade BC. It may be of interest also to compare

1. Calibrations. The A325 bolts the values of maximum strength in pure tension

are permitted variations in hardness ranges to the maximum bolt load obtained when bolts

(based on ASTM A325- 63T) depending on bolt are torqued to failure. The torqued strength
diameter, but the A354 bolts (ASTM A354- 58T) was 82 percent of the tensile strength for the
have a single hardness range for each grade few tests available for A354 BD bolts. This

which must be maintained for all sizes. The is 4uite similar to the 84 percent reported

largest diameters of A325 correspond in hard- by the University of I llinois for A325 bolts(ll)

ness, but not in elongation or reduction of and close to the 85 percent value commonly
area, to the A354 Grade BB bolt; the smaller assumed for A325 bolts.

diameters of A325 have hardnesses not greatly Load-elongation behavior at a given grip

different from those for the A354 Grade BC provides another convenient means of comparing
fasteners; and A325 bolts, l-inch diameter or the behavior of the A325, A354 BC, and A354 BD

l ess, of near maximum hardness would meet most bolts. Such data are shown in Figure IS where
requirements of A354 Grade BD. For these the four lower curves are for 3/4- inch diameter

reasons, low hardness and high hardness A325 bolts. Here we see that the behavior of the

bolts have been included in the following A325 bolts may be similar to the behavior of
comparisons between the behavior of the the A354 bolts, depending on their hardnesses.

familiar A325 bolts and the A354 fasteners. As might be expected from our knowledge
No attempt was made to select the hardnesses of the A325 bolts, reducing the number of

of the A354 bolts used. Al l bolts were pro- threads in the grip of the A354 fastene rs also
vided with appropriate nuts, either A325 or tends to increase the maximum bolt tension

Al94 Grade 2 for the A325 and A354 BC bolts, obtained. This is shown in Figure 16 where

and Al94 Grade 2H for the A354 BD bolts. average load-elongation data have been plotted

The low hardness A325 3/4-inch for two grips and the A325 and A354 BD bolts.
diameter bolts had an average tensile s trength The A325 bolts used in all these series of
of 40.1 kips (the minimum permitted), while tests had the old* thread lengths and thus

the same diameter high hardness A325 bolts * The length of thread for these bolts was
nominally two diameters plus 1/4 inch.
 19

were simi Jar in this respect to the A354 bolts It has been shown in studies at the
tested. University of Illinois, Lehigh University , and

Possibly of greater value to those elsewhere that, although the apparent strength

who may use these fasteners are the comparisons of a high-strength bolt may be increased by

of A325 and A354 bolts in the load- turns plots reducing the number of threads in the grip,

of F igure 17 and Figure 18. As befor e, the the turns- to-failure is simultaneously de-

load versus turn-of-nut data have been adjusted creased. This condition might then become

to a snug of 5,000 pounds. A study of Figure especially troublesome in fie ld installations

17 shows that at snug plus 1/2 turn the A325 of A354 bolts, particularly if thread lengths

bolts easily provided a prestress or bolt comparable to those on the new heavy head

tension equal to the proof load; in fact, the structural bolts (A325} were to be provided

proof load prestress is reached i~ 1/4 to 3/8 for the higher strength A354 or A490 bolts .

turn of the nut from snug. Because of their The decrease in turns-to-failure

higher proof loads, the A354 bolts of both with a decreasing number of threads in the

grades require more than 3/8 turn and up to grip, is shown in Figure 19a. At a grip of

1/2 turn of the nut beyond snug to achieve the 1. 35 inches the A325 bolt had about I 1/2 full

desired preload. An average for six 3/4-inch threads included and the A354 bolts had about

diameter A354 BD bolts showed snug plus 150 3 full threads (the number of threads in the

degrees to be required for proof load. This grip are shown for al l cases}. Note that the

suggests that more nut rotation may have to be turns shown in Figure 19a are from finger-tight

specified for the A354 bolts than is the case or about 1/4 turn greater than from snug. The

for A325 bolts in order to insure that during maximum torque required in these solid block

field tightening at least the minimum required tests was often reached at about one turn from

bolt tension is achieved. In the case of the finger tight. Figure 19 illustrates again the

A354 Grade BD bolt the factor of safety against probability of a smaller number of turns-to-

over- tightening and against resulting bolt failure with the A354 BD bolt than with the

breakage is reduced accordingly. Fractures A325 bolt for bolts with the same nominal

occurred in the 3/4- inch diameter A354 BD bolts thread lengths.

at about 1/2 turns from snug and at about The difference in turns-to-failure

I 3/4 to 7/8 turns from snug for the 7/8- inch between bolts torqued in the Skidmore-Wilhelm

diameter BC and BD bolts. Ca l ibrator and those torqued in a solid block

20
having the same grip appeared to be small as
shown by comparing Figures 19a and 19b . How-
ever, other studies at the University of
I llinois 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 actua l steel assembly, depending on the
number , t hicknesses, and types of plies included
in the gr ips.
The initial portions of the load-
torque relationships for the A325, A354 SC, and
A354 SO bolts are all similar as shown in
Figure 20. The differences in behavior appear
in the loads for which the relationship is no
longer linea r . With the famili ar formula at
the top of Figure 20 an estimate can be made of
the t orque required t o 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 recei ved .
2. Relaxation. The relaxation be-
havior of A325 bolts has been described in
earlier sections of this report. In a similar
manne r, the relaxation behavior of A354 SO
bolts was evaluated by means of a limited
number of tests. These incl uded 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 BO bolts with regular finished
hexagon head s and A7 plates were used to create
a n 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
su rface where turning occurs, the pre- load in
the bolts may not reach the appropriate proof
load after snug p l us 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 bo l t deforma tion and produces greater
plate deformation than does the load in A325
bolts; (2) the pl a t e surface galls or brinel l s
while some plate material i s 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 t u rn level
and thus is sensitive to any errors, especially
those errors which may result in Jess 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 i n Figure 21. One
 21

specimen, NX-24-D-3, reached a maximum load of one minute load ranged from less than one

52.5 kips after an additional 225 degrees beyond percent to more than five percent with an

the snug plus 1/2 turn level and a maximum average value of just over three percent . The

torque of 840 foot-pounds at 315 degrees beyond relaxation in the first five minutes was about

the 1/2 turn stage. The other bolt reached one percent of the one minute load. After
' beyond the
48.5 kips at about 300 degrees 90 days the bolt load was sti 11 over 95 per-

initial snug plus 1/2 turn. The maximum torque cent as great as that recorded a t the end of

later reached 1450 foot-pounds and when fracture t he first minute . Thus, relaxation does not

of the bolt occurred, the threaded end of the appear to be a major problem with the A354

bolt was locked in place because of the large (A490) bolt even when used with ASTM-A7

amount of plastic flow that deformed the plate. materi a ls.

A metallurgical examination showed no notice- 3. Tests Under Combined Tension and

able change in microstructure of the grade 2H Shear. At the time that tests were being made

nut, but the cold worked A7 plate was found to in another study on A325 bolts under combined

have doubled in hardness in the 1/64-inch layer tension and shear, tv10 similar series of tests

below the surface under the nut. The deforma- ;~ith A354 BD bolts were included. In the

tion and galling would certainly have been interest of making possible a comparison of

marked ly reduced if a higher strength plate A325 and A354 bolts in this report, some of

material had been used. the data are included . The complete report of

Use of a washer under the Grade 2H the tests can be found in Reference 13. All

nut considerably reduces the wear or galling fasteners were tightened to approximately the

and increases the probability of reaching at bolt proof loads. The average results are

least proof load with snug plus 1/2 turn for presented in Figure 22 in terms of the bolt

the A354 type fastener. However, only one test tensile and shear strengths a nd the a ctual

was run in these relaxation studies with washers test loads.

under the head and nut. It shou 1d a 1so be noted In general, the A325 bolt seemed t o

that in Figures 15 through 18 and 20 the have a s l ightly greater shear capacity rel a tive

calibration curves also had washers at the nut. to its tensile strength than did the A354 BD

When the amount of relaxation is bolt. As would be expected from a considera-

considered (Table 2) there is little difference tion of cross sectional areas, either type of

in the behavior of A354 and A325 bolts. The bolt is only about 80 percent as strong in

losses in bolt tension from maximum load to a shear through the threads as through the shank.
22
These data and those in the more complete study
of combined tension and shear(l 3) 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 b locks
or holders were of A7, A440, or Tl 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 i s through the bolt threads, in which case
the shear strengt hs are similar . This result
agrees well with the allowable stresses pro-
vided In the 1962 and 1964 Research Counci1( 6)
and the 1961 and 1963 American Institute of
Steel Construction specificat ions. 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 BD bol t was about one-th ird stronger than
the A325 low ha rdness 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 st ronger than the
mi nimum required. With these facts i n 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 196 1 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 p laced against a slope of to 20. The few
tests performed at that time gave a consistent
pattern which could be confirmed by a ma the-
matica l approximation. Thus the 1962 Research
Counc1
01
specaofoacat1ons
0 (6) were wr1tten
0 to
require an additional 1/4 turn for each outside
face in the grip with a 1 to 20 s lope.
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 behav i or and fatigue
behavior of bolts on 1 to 20 slopes. The
L & S (Lamson and Sessions} load analyzer(?}
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 t hread

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 a t the University of
Illinois. This snug load seems to be a reason-
able lower level for 3/4- a nd 7/8- inch bolts.
Therefore, al l 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 smal ler effect.
Table 3 is a summary of a l l test data
which were reviewed in this study on effects of
beve l ed 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
23
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 l-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 l- inch diameter
bolts agree with data shown in Figure 23 for
3/4-i nch diameter A325 fasteners.
Through study of the detailed data
from these tests, it is possible to present
the following observations:
(I) Pneumatic tightening seemed to
produce faster seating of the head and nut on
the bevels . For example, in the tests of
l-inch diameter bolts, nut seating occurred
near ISO 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 tight ening
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
bo lt load, possibly because of less wearing at
24

the contact between the nut and the surface. lit tl e or no effect; a t longer grips the effect

(3) Reducing the number of threads approacheJ that of a single bevel .

in the grip reduced the number of turns to (7) It has a l so been found f rom the

proof load. study with beveled surfaces that the relaxation

(4) Tests in load cells, Skidmore- of A325 bolts on the 1 to 20 sloping surfaces

Wilhelm calibrators, and the L & S load ana lyzer is essentially of the same magnitude as t ha t

al l show this consis t ent pattern: addi tional found wi th non- sloping surfaces. An initial

turns are needed to insure proof load when loss from the peak bolt tension o f about 5 per-

slopes of I to 20 are used. cent immediately after tightening is followed

(5) The effect of one bevel in the by a loss of about 5 percent from the one

grip was generally intermediate between the minute load over a period of two days.

results fo r no bevel and two bevels. Usually, A convenient method of achieving the

proof load was attained in snug plus 1/2 turn. desired additional turns in t he field would be

(6) Relative position of t he bevels- to specify 1/4 additional turn for two bevels

parallel, converging, or crossed--made a and no additional turns for one 5 percent

difference Jn behavior. For converging bevels, (1 to 20) bevel since proof load on an A325

positioning the bolt toward the top or bottom bolt is usually reached in less than 1/2 turn

of the bevel made a difference in behavior. of the nut . This ~1ould provide at least proof

Placing the bevels so t hey were crossed (as load in almost every case presented in t hese

may occur in fas tening crossing beams) produced data, and in addition would provide for a

a more consistent and median pattern. When t he slightly greater reserve agains t twist- off

beve l s were paral lel, the effect appeared than the requirement stipulated in the 1962

dependent on grip . At short grips there was specifications.(G)

IV, RESULTS AND ANALYSIS OF FATIGUE TESTS
A. RESULTS OF TESTS
Two separate series of fatigue
tests were made. The first of these in-
eluded Specimens I 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 p lates critical.
I f we look at the first ser ies 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 ks i. *
* ksi = kips per square inch or 1000 pounds
pe r square inch.
25
The bolts in Specimen 5 were torqued
to approximately 29 , 000 pounds , rather than
using the tu rn-of-nut method, to serve as a
basis of comparison between the resu l ts of
the present program of t ests and the results
of similar tests conducted in 1957. Despite
this reduction in clamping force to a minimum
level, the specimen did not fait.
Four specimens with the outer plates
critical were tested in the first series
(Table 4) and only one failed ; al l 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 5 and
only after 2,400,000 repetitions of a 0 to 30
ksi stress cycle. And, a l thoug h 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 washe r s, and oversize holes (7/8-inch
26

diameter), slipped into bearing during the specimens were generally two million cycles

first cycle of loading and ran 3,073,900 or more, despite the severe conditions of

cycles at 0 to 30 ksi without failure. To oversized holes, high pressures under the

determine if tightening a bolt by turning head and nut, etc. Failure occurred in only

the bolt head, rather than the nut, might four specimens, disregarding the tests of

produce a difference in fatigue life, Speci- Specimens 02 and 03 which were discontinued

men 8 (without washers) was assembled with because of excessive heat a nd pounding. The

two bolt heads and two nuts torqued on the failures occurred in the series B, C, and 0

same side of the specimen. Neither the specimens with minimum specified bolt tension,

omission of washers nor the torquing of the and then only two of these failures initiated

heads contributed to an early fatigue failure. at less than two million applications of

Although slip occurred in the first cycle at a load.

stress of 20.4 ksi, a value just above the The failures in Specimens Bland

current AISC design load for A7 structural Cl appear to have started on the inside face

stee l, a fatigue failure did not occur. The of an outer plate and not at the depressions

test was discontinued after 5,148,000 appli- caused by the bolt heads (Figure 25). The

cations of a 0 to 30 ksi stress cycle. Speci- appearance of the fractures suggests that

men 9 was assembled in a manner similar to fretting rather than the bolt head depressions

that used for Specimen 8, except that l-inch or other stress concentrations at the holes

diameter bolts were used. Slip occurred in initiated the failures. Figure 25 also pre-

the first cycle at 24.8 ksi . This test was sents a typical fatigue failure for specimens

discontinued without failure after 4,892,000 with the outer plates critical and indicates

cycles of loading at 0 to 30 ksi . Thus, even th~ type of seating action that occurred under

without washers, the joints sti 11 have a large the bolt head. Galling of the plate surfaces

fatigue resistance, providing the bolts a re similar to that shown in Figure 7 occurred

properly t ightened. under the nuts .

As indicated in Table 5, the fa- The failure in . Specimen 01 occurred

tigue spec imens of the second series (all with in both an outside plate and the center plate

outside plates critical) were tested at no- at their respective critical sections but may

minal stress levels of either 0 to 30 ksi (or have been i nfluenced by fretting. Although

slightly less) or -20 ksi to+ 20 ksi. At Specimen 01 slipped into bearing dur•ng the

either stress level the fatigue lives of these first loading cycle, it did not continue to
 27

slip to any significant extent when subjected B. ANALYSIS OF FATIGUE TEST RESULTS

to additional cycles of loading . In examining the results of these

The tests of Specimens 02 and 03 fatigue tests , it is interesting to compare

wer e discontinued because of the excessive them with those of s i milar tests conducted

slip and pounding that developed during the at the University of Illinois and Northwestern

early part of the tests. Examination of these University (Table 6). The zero-to-tension

specimens indicated it was probable that use tests as well as the full reversal tests of

of lubricating oil rather than water-soluble this and the other programs compare very fa-

cutting oil during the drilling operat i on was vorably as far as specimen life is concerned.

responsible for the low frictional resistance, Specimen Cl (Table 5) is unique be-

despite cleaning with a solvent. cause of its proportioning to give a larger

Because of the difficulty encountered tensile area than shear area and can be com-

in testing Specimens 02 and 03, it was decided pared with the performance of Specimen 6 in

to gradually increase the stress on the nominal Table 4. I t can be seen that for a stress

net sections of Specimens 04 and 05 from+ 16 range of 0 to 30 ksi, the fatigue lives of

ksi to+ 20.0 ksi. Failure occurred only in both specimens were essentially the same,

Specimen 05 and then only after 2,267,000 cy- even though Specimen 6 was designed with a

cles. The fracture initiated in an outside tension:shear · ratio of 1.0 to 0.90 and

plate slightly below the critical section at a Specimen Cl with a tension:shear ratio of

point which suggested that fretting between the 1.0: 1.25. It can also be seen that these

edge of the bolt head and the surface of the higher shear stresses (nominal shear stresses)

outside plate was responsible fo r the failure . did not materially affect the fatigue life un-

This unusual failure can be seen in Figure 26 . der zero-to-tension loadings when high clamp-

In Figure 26a the extent and shape of the fa- ing forces existed in the bolts . This was

tigue crack is shown. The probable locations true despite the greater probability of slip

of the initiation points for the fretting fail- at the higher shear stresses.

ure have been marked by arrows in Figure 26b. Although the full reversal tests of

Once again , the connection withstood more than Specimens 02 and 03 were discontinued after

2, 000,000 cycles of reversal at stresses which relatively few cycles of loading, it is in-

were well above current design levels. teresting to note that their behavior was
28
similar to that of Specimen CJ-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 -
6
cations, ( ) was well below 2,000,000 cycles.
This low fatigue life, although partly a re-
suit of the low clamping force in the bolts,
also can be attributed to the presence of oil
on the faying surfaces of that join t.
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.
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
( 14)
joints. 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.
I. Clamping Force. In order to
obtain an indication of the clamping force
in the bolts of the fa t igue specimens, the
elongation of each bolt of certain specimens
was measured periodically during the tests
by means of an extensometer. This e longa-
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 t hese
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 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 t en-
sion for bolts in the fatigue specimens that
did not fa i 1 was on the order of IS percent,
the major portion of which occurred duri ng
the first few cycles of loading. This beha-
vior agrees well with the results reported
in othe r studies. (8 , IS, 16 • 17 • 18 ) In
Reference 8 losses in pre-load are repor ted
to have been sbout 20 to 30 percent in
375,000 cycles and about 20 to 35 percent in
540,000 cycles with further rapid decreases
 29

as joint slip increased greatly . The bolt critical sections which eventually failed in

pre-stress in Reference 8 was about one- fatigue were found to be reduced considerably

third of that which would be used today, and in the fi rs t few cycles of loading and to

substantial slipping occurred . In Reference about half the initial bolt load when failure

17 with bolt tensions about equal to those occurred. This suggests that there may be

presently specified, the losses in pre-load meri t in periodic checks and re-tightening

varied from about 7 Lo 35 pe rcent depending at the critical section of bolted joints sub-

on specimen , applied st ress , etc. Losses of ject to cyclic loading.

bo It t ens i on reported in Reference ~~~' were In t he study by Baron and L~rson the

on the order of a few percent to 18 percent , bolt clamping force in t he specimens

and no joint fa ilure s deve loped . (I 3/ 16-in ch grip) varied from 20,000 to

The results reported herein and in 40 , 000 pounds before the tests and had de-

Reference 16 indicate that approx imately a creased approximately 27 percent by the end of

3 to 5 percent loss in bolt te nsion may be the tests. (IS) These specimens were as-

attributed to plastic flow in the bolts and sembled with washers under the nuts and under

plate materials caused by the high bolt ten- the bolt heads , and the bolts were tightened

sion . Much of the additional loss in bolt to a predetermined torque. I n most cases

load appears to be proportional to the ap- Baron's specimens slipped into bea r ing during

plied stress and attributable to the applica- the first loading cycle . Thus, the loss in

tion of static loading and , among other factors, bolt load during test ing fo r specimens with-
6 out hardened washers was found to be compara-
to any slipp ing which may have occurred. (! )

I t should be noted that many of the fatigue ble to the loss of bolt tension in joints as-

specimens s lipped into bearing during the semb led with washers.

first loading cycle and that the applied 2. Nominal Coefficient of Friction

stresses at first slip were generally welt The nominal coefficient of friction between

above design levels . Bolt tensions at t he the specimen plates is a significant factor

*Two fatigue tests of Reference 18 were since it, along wi th the clamping force, de-
made with bolts torqued into the yield
range . These tests appear to be the earliest termines whether the joint wi II slip or behave
in which it was shown that such high torquing
did not impair the structural joint behav ior . as a friction joint. The value of the nominal
The same study also sho~~d that with washers
under the nut, there was no more loss in clamp- coefficient of friction for each fatigue speci-
ing for bolts having regular finished hex nuts
than for those tightened with heavy hex nu ts . men which slipped wi II be found in Tables 4
30

and 5. These values are based on load-slip The friction values shown in Tables

readings taken on each specimen during the 4 and 5 are generally smaller than the values

first cycle of loading. The nominal coef- of 0.30 to 0.35 frequently reported for mi 11

ficient of friction was computed on the basis seale surfaces. However, the va I ues of the

of the load on the net area and the bolt present tests are based on r elatively small,

clamping force as determi ned f rom the bolt four-bolt, double- lap, shea r-type joints as

elongation measurements . shown in Figur e 3 and may provide a lower

limit for the frictional resistance .

V. CGICLUS I {)liS
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 report ed
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 me thod. In the case of
oversized holes , up to 1/8 inch
greater in diameter than the bolt,
there may be some red uction in bolt
tension when washers are omitted
and when f inished hex head bolts and
31
nuts are used, but the clamping
force will sti 11 be in excess of t he
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 repo r ted herein
It appears that A325 and A354 BC and BD (or
the A490) bolts may be compared as follows:
I. 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
t he grip (going from old thread
Iengths to new thread Iengths) 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 (particula r ly grade
32
BD) requires more turns to reach
its proof load than does the A325
bolt. This then reduces the factor
of safety against twist-off. ~ith
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 mo re than two-thirds turn
beyond proof loa d, even though the
total turns to failure fro m snug
may be approximately one and one-
half for both types of bolts.
4. Hardened washers appear desirab le
under the head and nut of the A354
BD (A493) bolts to reduce contact
pressures and to provide proof load
with slightly fewer turn s , especi ally
when the connected parts are of mild
steel.
5. Relaxation behaviors of the A325
and A354 BD bolts are simi lar and
show losses of about 5 percent from
the one minute load over a period of
9J 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
tha t of the A325 to that of the
A354 bolt .
7, There a re numerous similarities in
the behaviors of A325 bolts of high-
est hardnesses and A354 bol ts.
The inclusion of surfaces with a
1 to 20 slope in the grip requires additional
turning to insure optimum tension with the
turn-of-nut met hod. If snug plus 1/2 turn
is used, no additional ro t ation is necessary
for one beveled surface (although the resul t -
ing load may be near the minimum specified};
however , with beveled surfaces under both the
head and nut, additional rota tion should be
used -- for example , snug p lus three- quarters
turn ( 270 degree) •
On the basis of the tests reported
herein, the following conclusions may be
drawn regarding the fatigue behav io r of shear-
type bolted joints assembled without washers:
1. There was no significant reduction
in the fatigue lives of bo lted joint s
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
 33

lives of the joints. conclusions are based on joints that slipped

2. The major portion of the loss in into bearing, as well as those that did not;

bolt tension during the fatigue on joints with extremely severe combinations

life of a specimen tends to occur of hole size, fastener head size, and nut

in the first few cycles of loading. size; and in many cases on joints assembled

This loss in bolt tension appears with minimum allowable clamping in the high-

to be similar whether hardened wash- strength bolts. Thus, these conclusions

ers are used or omitted. appear to be,~ priori, safely applied to

It should be noted that these last connections with less severe conditions.
34
VI, REFERENCES CITED
J, C. W. Lewitt, E. Chesson, Jr., and W. H.
Munse. "Studies of the Effect of
Washers on the Clamping Force in High
Strength Bolts," Unpublished Struc-
tu r al Research Series No. 191. Univer-
sity of Illinois Department of Civil
Engineering, Urbana, Illinois, March
1960.
2. A. c. Knoell, E. Chesson, Jr . , and w. H.
Munse. "Fatigue Behavior of Bolted
Joints Assembled Without Washers."
Unpublished Structural Research Series
No. 242. University of Illinois Depart-
ment of Civil Engineering, Urbana,
Il linois, February 1962.
3. J, R. Fuller and W. H. Munse, "Laboratory
Tests of Structural Joints with Over-
Stressed High Tensile Steel Bolts,"
Unpublished Structural Research Series
No . 44. University of Illinois Depart-
ment of Civi I Engineering, Urbana,
Illinois, January 1953. Also summa-
rized in ASCE Transactions, Vol. 121
(1956), Paper No. 2B39, Pp. 1255 to
1266.
4. F. P. Drew, "Tightening High- Strength
Bolts," ASCE Proceedings, Vol. 81
(August 1955), Paper No. 786.
5. E. F. Ball and J, J, Higgins, "Installa-
tion and Tightening of Bolts," ASCE
Transact ions, Vol. 126, Part 11--
(1961), Paper No. 3241, Pp. 797-810 .
6. Research Council on Riveted and Bolted
Structural Joints, "Specifications for
Structural Joints Using ASTM A325 Bo l ts,"
ASCE Structural Division Journal, Vol.
88, ST5 (1962), Paper No. 3294. (Pre-
vious editions of this specification
were dated 1960, 1954, and 1951. Re-
vised in March 1964 and entitled
"Specifications for Structural Joints
Using ASTM A325 or A490 Bo Its,")
7. W. G. Waltermire, "Evaluating Fastener
and Joint Performances with a Load
Analyzer," Assembly and Fastener
Engineering, Vol. 5, No. I (1962),
Pp . 32-35.
8. w. M. Wilson and F. P. Thomas. Fatigue
Tests of Riveted Joints
(E~periment Station Bulletin No. 302).
Urbana, Illinois: University of Illi-
nois College of Engineering, May 1938.
9. R. L. Sanks and c. C. Rampton, "Grit and
Shot-Reinforced High Tensile Bolted
Joints," ASCE Structural Division
Journal, Vol. 83, ST 6 (November
1957), Paper No. 1435.
10. D. T. Wright and w. H. Munse, "Calibra-
tion Tests of High- Strength Bolts,"
Unpublished Structural Research Series
No.9, University of Illinois Depart-
ment of Civi I Engineering, Urbana,
Illinois, January 1951.
11. J . G. Viner, E. Chesson, R. L. Dineen,
and w. H. Munse, "A Study of Nuts for
Use with High Strength Bolts," Unpub-
lished Structural Research Series No.
212, University of Illino i s Department
of Civil Engineering, Urbana , Illinois,
March 1960.
12. Robert F. Stevens, Unpublished research
notes, University of Illinois Depart-
ment of Civi I Engineering, Urbana,
Illinois, 1952.
13. E. Chesson, N. Faustino, and w. H. Mun se.
Static Strength of High-Strength Bolts
Under Combined Tension and Shear.
Manuscript in preparation. University
of I llinois Department of Civil Engi-
neering, Urbana, Illinois, 1964.
14 . E. Chesson, J, F. Pa rola, and W. H. Munse
Effect of Bearing on Fatigue Strength

of Riveted Connections (Engineering
Experiment Stat ion Bulletin in pre-
paration). Urbana, Illinois: Uni-
versity of Illinois College of
Engineering, 1964.
IS. F. Baron and E. w. Larson, Jr., "Re-
sults of Static and Fatigue Tests of
Riveted and Bolted Joints Having
Different Lengths of Grip," An un-
published report of an investigation
conducted in the Department of Civi I
Engineering at Northwestern Un iver-
sity, January 1952. Also summarized
in 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 of
Joints Under Static Loads," ASCE
Transactions, Vol. 120 (1955~
Paper No. 277B, Pp. 1335-1352.
17. W. M. Wi !son, "High Strength Bolts as
a Means of Fabricating Steel Struc-
tures," Proceedings of the 30th
Annual Meeting of the Highway Re-
search Board, December 1950, Pp.
136-143.
35
18 . K. H. Lenzen, "The Effect of Various
Fasteners on the Fatigue Strength of
a Structural Joint," AREA Bulletin 481,
Vol. 5 1, (June-July !949), Pp. 1-27.
19. W. H. Hunse, "An Eva luation of the Be-
havior of Structural Connections As-
sembled with Hook-Knur l Bolts,"
Unpublished Private Report. Urbana,
Illinois, September 1957.
20. W. H. Hunse , 0. T. Wright, and N.H.
Newnark, "Laboratory Tests of Bolted
J oints, " ASCE Transactions, Vol . 120
(1955). Paper No. 2778, Pp. 1299-1 321.
21. N. G. Han son , "Fatigue Tests of Joints
of Hi gh-Strength Steels ," ASCE
Transactions, Vol. 126, Pa~l
(1961), Paper No. 3241, Pp. 750-763.
22. w. H. Munse and H. L. Cox, The Static
Stren th of Rivets Sub·ected to Com-
bined Tension and Shear Engineering
Experiment Station Bulletin No . 437).
Urbana, I llinois: University of
I llinois College of Engineering,
December, 1956.
This page is intentionally blank.