SIMPLE CONNECTION FOR SQUARE TUBULAR SHAPES

.,

by

Leroy Sauerwine

Fritz Engineering Laboratory

Department of Civil Engineering
Lehigh University
Bethlehem, Pa.

June 1969

Fritz Engineering Laboratory Report No. 237.61

 TABLE OF CONTENTS

Page

1. INTRODUCTION 1

2. DISCUSSION 3

3. SPECIMEN DESCRIBED 6

4. TEST PROCEDURES 8

4.i Tensile Coupon Te~ts 8

4.2 Connection Tests 8

5. TEST RESULTS 11

5.1 Tensile Coupon Tests 11

5.2 Connection Tests 11

6. ANALYSIS OF RESULTS 14

7. SUMMARY AND CONCLUSION 17

TABLES AND FIGURES 19

·REFERENCES 36
 LIST OF FIGURES

Figure

1 SIMPLE FRAMING CONNECTION DESIGN 21

2 LOCATION OF ELECTRICAL STRAIN GAGES 22

3 CONNECTING BEAMS 23

4 SKETCH OF ELASTIC LOADING TEST SETUP 24

5 SKETCH OF ULTIMATE CAPACITY GEST SETUP 25

6 TEST SETUP FOR ELASTIC LOADING 26

7 TEST SETUP FOR ULTIMATE CAPACITY TESTING 26

8 MOMENT VERSUS ROTATION FOR ELASTIC LOADING OF 6-2 27

9 MOMENT VERSUS ROTATION FOR ULTIMATE 28

CAPACITY TEST OF 8-1

10 TUBE WALL STRAINS AROUND CONNECTION OF 6-2 29

11 HORIZONTAL WALL STRAINS NEAR TOP OF CONNECTION 30

OF 8-1

12 HORIZONTAL WALL STRAIN NEAR BOTTOM OF 31

CONNECTION OF 8-1

13 VERTICAL WALL STRAIN NEAR BOTTOM OF CONNECTION 32

OF 8-2

14 VERTICAL WALL STRAINS OF CONNECTION WALL ABOVE 33

CONNECTION OF 6-2

15 VERTICAL WALL STRAINS OF CONNECTION WALL

ABOVE CONNECTION OF 8-.1 34

16 VERTICAL WALL STRAINS OF UNLOADED WALL ABOVE 35

CONNECTION OF 8-2
 1. INTRODUCTION

Tubular shaped structural steel has come into use only in

recent years. Tubular shapes, mostly round, have been used

successfully in the formation of trusses. The problems of connections

with.round shapes have been investigated and on trusses the usual

solution has been various types of welded connections. However, little

research has been done on simple beam and column framing connections

for square or rectangular structural steel tubing. R. N. White and

1
P. J. Fang have investigated some framing connections, but according

to the authors none of the connection studies have the combination of

strength and flexibility which is desired for the simple framing

connection.

The purpose of this report is to investigate simple framing

connections for square tubular shapes and develop a theory for the

. ~nalysis of these connections. After studying the performance of the

framing connection design used for tests in this project, the results

wer~ compared with reports of other connection designs previously

investigated. With this information a theory for predicting the

behavior of these connections was developed.

The framing connection investigated for this report is an

8-1/2 inch plate, 3/4 inch thick, which was fillet welded in a slot

-1-
 -2

in the tubular column. Two sizes of tubes were used, 6 x 6 x 3/16 and

8 x 8 x 1/4; and each tube was tested with light and heavy beams o

The tubes and beams were selected so that some comparisons could be
.1
made with connections previously tested by R. N. White and P. J. Fang.

In all four tests only two-way connections with symmetrical cantilever

beams connected to the sides of the square column were studied. The

beams were first loaded at their ends and the stresses were held

below the yield point of the connection. Next the setup wa$ changed

so that the load was applied closer to the column face and the load

was increased until failure of the connections occurred. In these

tests the only load carried by the column was that transmitted to the

column through the connections. '

 -3

2. DISCUSSION

•
Because of its shape the square tubular column has similar

properties in all directions while the wide flange column has a strong

and a weak axis. Although the wide flange shape has a higher moment

of inertia in one direction, the square tube has a high moment of

inertia in all directions which means it has good bending and torsional

resistance. Since the tubular shape also has a high radius of gyration

in both directions, it has good strength in compression. Therefore

in use as a column section the square tubular shape has more favorable

geometric properties than wide flange shapes and could have advantages

in some framing conditions.

The problem for square tubular columns has been to find adequate

simple framing connections. The connections must be strong enough to

carry the beam reactions and yet flexible so only a small bending

moment is transmitted to the column wall. If the connection is too

stiff, the large bending moment can cause harmful distortions in the"

tube wall. The connection should be designed such that it does not

develop s"tresses in the column which would cause a great reduction in

the column strength. Besides the design requirements the connection

must be easy to fabricate and economical.

Because of the tubular shape and ease of fabrication the only

practical method of fastening the connection to the tube wall is

welding. Since the weld transmits all the stresses to "the column, it
 -4

probably is a critical part of the connection. Using a typical fillet

weld the critical stresses in the weld will form in the throat area.

Because the weld is an important part of the connection, the plate

should be welded over the entire length of the slot. A 3/16 inch weld
. . is the most common and economical since it is placed with one pass of

the electrode. It should have sufficient strength for the simple

connection tested, but would probably tear under large shear loading

conditions.

The tubes used in the testing program were rolled into the

square shape and welded with a longitudinal butt weld down the center

of one face. The bending and welding of the tube caused residual

stresses in the wall. Since the tube was cold worked, the steel has

a higher yield in tension but a lower yield in compression. The

residual stress distribution for the welded and rolled tube is not

known and could cause some significant reduction in the column

strength. When the slot is burned in the tube wall, the residual

stress distribution is changed since the stress must be zero at the

edges of the slot. When the connection plate is welded in the slot,

more residual stresses are induced. One question is how detrimental is

the connection to the column strength of the tube. The slot burned

in the wall may be more beneficial than the additional stresses

produced by the weld. It is very difficult to analyze the effect of

the simple connection on the column strength without a large number

of tests.

When load is applied to the beams framing into the column,

the column strength can decrease. As the connection transmits the

 -5

load to the tube wall, stresses are induced in the wall which may reduce

the column strength. With the welded plate the stresses are highly

concentrated. Since the connection is welded near the center of the

tube wall, the concentration of stresses can cause distortions of the

wall. When the load on the beams is increased, high stress concen-

tration form in the wall at the top and bottom of the plate. At these

points confined areas of plastic deformation begin while the surrounding

elastic areas restrain the wall from large deformations. As the loading

increases the plastic areas become larger and the wall distortions

increase. When the distortions are clearly visible, the effective

area of the tube at the connection has decreased which means a

reduction in column strength. The amount of reduction in column

strength due to a rotation of the connection would depend on the

width to thickness ratio of the column and the ratio of connection

&

length to column size. A complete analysis of this problem is beyond

the scope of this report and would require a larger assortment of

specimens.
 -6

3. SPECIMEN DESCRIBED

..
For this report tests were performed for simple framing

connections to square tubular columns. Two tube sizes of ASTM A36

steel, 6 x 6 x 3/16 and 8 x 8 x 1/4 were used for the tests. Each

tube was 4 feet long and had a weld running along its length near the

center on one of the walls. The slots for the connections were

burned in the tube wall on sides which did not have the weld. This

removed one possibJe problem in analyzing the performance of the tube

near the connection. The slots were 'located 5/16 inch off the

centerline at midheight of the tube as shown in Fig. 1. This located

the beam at the center of the column face.

The connection plate was cut from 3/8 inch ASTM A36 steel.

The same size plate, 8-1/2 inches by 4 inches, was used for all

connections. Three 15/16 inch holes were punched in each plate with

the line of holes 1-1/4 inches from the plate edge and the holes

spaced 3 inches apart as shown in Fig. 1. The slots in the tube were

filed and ground until the plate could fit into the slot. The plate

was positioned such that the center of the holes were 2 inches from

the tube face. The ends of the plate were tack welded to hold it in

position as a 3/16 inch weld was placed along the length on both sides

of the plate.

Electrical resistance strain gages were placed on the specimen,'

Three gages were placed on the connection plate on the side opposite
 -7

the centerline of the tube wall. The gages ak the edge were placed as

close to the edge as possible. The remaining gages were placed as

. shown in Fig. 2. There were gages on the inside and outside at the

ends of the connection plate in both horizontal and vertical directions.

A pair of gages were put on the centerline of the connection wall and

unloaded wall. Other instrumentation during testing included exten-

someters to measure deflection of the loading points on the beam and

a rotation gage with a 6 inch gage length on the centerline of the

beam above the connection.

Each tube size was tested with heavy and light beams of ASTM

A36 steel. The 6 x 6 x 3/16 tube was first tested with lOB15 beams.

The l2W3l beams were used for the heavy beams with the 6 x 6 x 3/16

tube and as light beams for the 8 x 8 x 1/4 tube. The heavy beams for

the 8 x 8 x 1/4 tube were l8W55. All the beams had three holes

punched for the connection to the tube. The holes were centered on the

beams 1-3/4 inches from the edge except for the l8W55 beam which had

the holes 1-1/2 inches from the edges as shown on Fig. 3. Stiffener

.. plates of 3/8 inch ASTM A36 steel were placed on the beams 6 inches

from the holes to give the beam added strength at the loading point

for the ultimate load.

The beam was connected to the tube by 7/8 inch ASTM A325 high

strength bolts which were 1-3/4 inches long. The beams were aligned

perpendicular to the column face and the bolts were tightened by the

turn of the nut method. After the bolts were snug, a quarter turn was

added to the nuts. This was sufficient to hold the beams securely

while the specimen was loaded.

 -8

,
I
I
\.

I
i
I 4. TEST PROCEDURES
I
4.1 Tensile Coupon Tests

I
! Six coupon specimens were made for the tensile coupon tests.

Two coupons of ASTM A36 steel were made from each of the following:

6 x 6x 3/16 tube, 8 x 8 x 1/4 tube, and 3/8 inch plate material.

The four coupons from the tubes curved considerably indicating high

residual stresses in the tubular structural steel. The coupons were

the standard shape as specified by ASTM A370.

k
The tests were performed on the Tinius Olsen l20 machine.

The automatic recorder was used to graph the strain from the ext en-

someter with an 8 inch gage length and the machine load. The machine

was operated with a crosshead speed of 0.025 in./min. and after the

steel was stressed beyond the elastic range, the machine was stopped

for five minute intervals to obtain static yield. When the strain had

reached 0.03 in./min., the extensometer was removed and the machine

speed was increased to 0.5 in./min. This speed was constant until

failure of the tensile coupon occurred.

I
i 4.2 Connection Tests
!

For the first part of the connection testing the specimen was
k
placed in the Baldwin 300 machine. An 8 inch spreader beam approxi-

mately 6 feet long was used so the load could be applied at the ends
 -9

of the beam. Mechanical jacks were placed at the ends of the spreader

beam and dynamometers were positioned on the jacks as shown in Figs.

4 and 6. With this setup the load on the beams was applied approxi-

mately 28 inches from the face of the tube wall. The readings from the

dynamometers were used to check for symmetrical loading conditions.

The· actual load reading was taken from the machine load applied to

the column by the spherical head. The specimen was inverted compared to

the usual building conditions so that the upper part of the column

against the spherical head would be the base in normal column loading.

The loading of the specimen was done in 200 pound increments.

Each load was held while strain gage and ~ther instrumentation readings

were taken. When one of the strain readings reached the plastic range

the testing was stopped and the specimen was saved for testing of the

ultimate load capacity of the connection.

To test the ultimate capacity the specimen was placed i.n the
k
Riehle 800 machine. Again the column was loaded by a spherical head

but the supporting apparatus for testing in this machine was two

2-foot pedestals with 4 inch rollers on which the beams rested. These

rollers were centered 8 inches from the column face as shown in Figs.

5 and 7. With this loading setup the load was increased in 2 kip

increments until failure of the connection occurred.

For the ultimate capacity testing, the first two specimens

were whitewashed. Because of the cold rolled condition of the column

section the whitewash did not aid observations and was eliminated for

the remaining specimens. On the first test it was noticed that the

outer edges of the beam flanges were bearing against the tube wall at
 -10

relatively low loads. To allow more freedom of rotation to the con-

nection and remove the effect of the flanges at low loading

conditions, the outer edges of the beam flanges of the remaining

specimens were cut so that the flange could not bear against the

corners of the tubular section.

 -11

5. TEST RESULTS

5.1 Tensile Coupon Tests

Since the tubular columns ha~ high residual stresses from

cold working, the yield point of the steel was higher than expected.

The residual stresses also cause a smooth transition from the elastic

to plastic range on the stress-strain curve. A 0.2 percent offset

was used to find the static yield point. The yield point for the

6 x 6 x 3/16 tube was 57 ksi and for the 8 x 8 x 1/4 tube it was

49.5 ksi. The 3/8 inch plate used for the connection did not have

any noticeable residual stresses and produced a more ideal stress-

strain curve. The change from the elastic range to the plastic range

was obvious. The static yield point was 30 ksi for this steel.

Using the data from the tensile coupon tests, as shown on

Table 1, the strain corresponding to the yield point can be determined.

This was used as the upper limit in the elastic testing of the

connections.

5.2 Connection Tests

The first part of the test was basically a moment loading of

the framing connection. When a strain reached the plastic range the

load was removed but the strain gages did not return to their original

readings. During the testing the only visible movement or rotation

 -12

was the slight deflection of the ends of the beams, but the rotation

gage on the flange above the connection did show that the connection

was rotating, as shown in Fig. 8. After the load was removed the

beams returned to their original positions and no signs of damage to

the specimen were observed.

The next step in the testing was the ultimate capacity of the

connection with a larger shear load as compared to moment under usual

. loading conditions. For each specimen the rotation of the connection

was visually observed after a few increments of loading. With the outer

edge of the flange cut, the remaining flange bore against the tube

wall after approximately six loading increments. As shown in Fig. 9,

the rotation of the connection was only slightly constrained by the

flange bearing on the wall since the flange was not bearing 'on the

corners of the tubular section.

Besides the visual observation of the bending of the connection

wall due to loading, the unloaded wall was also buckled. The square

tube is a rigid structural member and the bending of the connection

~all caused a corresponding buckling in the unloaded wall. From the

loading condition the wall above the connection was forced in and the

wall below the connection was pulled out from the column while the

unloaded wall of the tube was buckled out above the connection and

buckled in below it.

The ultimate failure of the connection was a tension tear at

the weld at the bottom of the connection plate. In all tests cracking

sounds were heard at high loads before the cracks at the weld were

visible. Figure 7 is a picture of a specimen in the machine after

 -13

failure which illustrates the obvious rotation of the connected beams.

For the 6 x 6 x 3/16 specimens the failure appeared to be in the column

material while in the 8 x 8 x 1/4 specimens the weld material failed.

The 6 x 6 x 3/16 specimens failed at machine loads of 28 kips and 47

kips while the 8 x 8 x 1/4 specimens failed at machine loads of 49 and

46 kips.

On the tube wall the strain gages showed an average of inside

and outside readings to be tension. . This tendency for the average

strain to be tension was the effect of the bending of the wall caused

by the rotation of the connection. An example of the strains on the

wall around the connection is given in Fig. 10.

The strain readings on the connection plate did not show any

consistent pattern. During the elastic testing the lower part of the

plate was in tension with the center in slight compression and the

upper edge was always compression. In the ultimate capacity tests the

pattern of the strains were not consistent in the specimens but the

lower part of the plate was always in high tension.

 -14

6. ANALYSIS OF RESULTS

The strains measured in the 3/8 inch connecting plates were

generally lower than strains in the tube wall. The pattern of strains

in the connecting plates varied from test to test as the size of beam

was changed. In the final test on each specimen the pattern of strains

was also affected by the bearing of the flange of the beam against the

tube wall near the end of the test. The strain measurements on the

connecting plates were not considered to be significant because failure

of the connection would not involve failure of the plate for any of the

combinations of beams and columns included in this program.

The pattern of strains in the wall of the tube near the

connection produced an indication of the behavior of the tube due to

bending and shear stresses produced by loading the beams. The pattern

of strains for anyone specimen was similar for the loading in the

moment-rotation test and the ultimate load test. Because of this, only

the strain data from the ultimate load test have been plotted for each

specimen.

The transverse strain in the tube wall at the top and bottom

. of the conne~ting plate resulted in a very consistent pattern. This

may be seen by comparison of Figs. 11 and 12. The gages on the outside

and inside surfaces at both locations gave relatively large strain

 -15

readings with the average strain being in tension for the upper side of

the plate and approximately zero for the other side. For some tests

the average strain was compression for the lower gages.

The strain data for the vertical gages at the same locations

was .not as consistent as the data from the horizontal gages. The

average strain above the beam was tension, but the pattern of strains

for the inside and outside gages varied from test to test.

The vertical strains measured on the wall below the connection

were more consistent. Generally the strain on the outside wall was

tension from the start changing toward compression for higher loads.

The strain on the inside wall was compression at the start changing to

tension at higher loads. A typical pattern of strains is shown in

Fig. 13. The average strain was tension at this location throughout

the test. Failure of the connections resulted from tearing of the welds

at this location. This would seem to correlate with the high average

tensile strains read on the vertical gages.

The strain gages centered on the tube were located 2 inches

above the connection plate and served to indicate the extent of the

local effect resulting from the loads on the connection. The length

of the column which is affected by the connection loading is a function

of the column size, the beam size, and the magnitude of the applied

load. Figures 14 and 15 show that there is an apprecialbe bending

strain present at the location of these gages. Since the average

strain is tension at this location, the effect of the connection is to

cause other parts of the cross section to carry the compression load.
 -16

There is a tendency for the loaded sides of the column to be in

compression ~t very high loads.

Figure 16 shows that the unloaded sides of the column tend

to have a net tension stress also. This indicates that the compressive

stress is carried at the four corners of the column near the

connection. The unloaded walls exhibited a tendency to deflect outward

while the loaded walls deflected inward. Since the residual stress at

the corners of the column is apt to be tension, the redistribution of

stress in the cross section at the connection does not tend to reduce

the column strength even though the bending in the walls is

considerable.

If one considers the results of the column tests reported in

Ref. 1, it will be noted that the results of these tests are not

unlike the results to be expected if only the bare column were

loaded. In other words, the connection is not more detrimental to the

column strength than is the residual stress pattern in the original

section. In fact the fabrication procedure used in making this

connection may produce a more favorable distribution of residual stress

in the vicinity of the connection. This would help to explain why the

connection did not produce lower than normal results in the column

tests of Ref. 1. This aspect of the problem deserves further study.

 -17

7. SUMMARY AND CONCLUSION

The 8-1/2 inch plate used in these tests seems to be the

proper size for connecting the beams involved with the tube sections

included in the program. This size of plate is a little large for the

10B15 shape. In testing the ultimate capacity of the connection the

shear load did not reach half of the allowable shear for the

standard 8-1/2 inch connection which consisted of two 8-1/2 inch

angles as given in the AISC Manual. The welded joint does not have

the necessary strength to carry a force equal to the capacity of the

. 1
beams. In tests performed by White and Fang the connection tested

was able to sustain much larger shear loads, but it should be noted

that their tests were made with a pure shear loading.

Another important characteristic of the simple connection is

the flexibility as determined by the moment-rotation characteristics

of the connection. In the tests performed the average moment required

for a 0.01 radian rotation was 14 kip-inches for the 6 x 6 x 3/16 tubes

and 18 kip-inches for ·the 8 x 8 x 1/4 tubes. As shown in Table 2,

these values are lower than any connection tested by White and Fang.

Therefore this connection is more flexible than previous framing

connections tested for square tubular columns. Greater flexibility is

generally considered to be a desirable characteristic for simple

connections.
 -18

The weld is the critical area in the simple fr~ing connection.

The shear capacity of the weld was never reached because the applied

moment produced a tearing of the weld at the bottom of the plate.

Fo+an adequate connection to be found, ~ weld size which allows the

rig~t combination of strength and flexibility must be selected. The

fabrication process of this connection is an important element for

this connection. If the slot is not properly prepared, the full

capabilities of the weld can not be developed. The most important

part of the weld is at the end of the plate which is also the most

difficult place tO,achieve the quality needed in the fabrication. If

some fabrication method could be found which would develop a satisfactory

weld at the ends where it is needed, the capacity of the connection

would be greatly increased. From these tests it appears that a larger

weld could satisfactorily"be used for, these connections, and a smaller

clearance in the slot is desirable.

The type of connection tested is adequate for light loads,

but should be used with caution where heavy loads are encountered.

More tests are required with different size connection plates and varied

column and beam combinations before a satisfactory theory for the

connection behavior can be formed. With further studies this type of

connection may prove to be a good solution for simple framing connections

to square tubular columns.

 -19

TABLE 1 TENSILE COUPON TEST RESULTS

Coupon Static Tensile Reduction

No. Source of Coupon Yield Stress Strength of Area a Elongation b
(psi) (psi) (%) (%)

A-I 6 x 6 x 3/16 tube 57,100 71,600 41.9 17.6

A-2 6 x 6 x 3/16 tube 56,500 72 ,300 45.7 23.8

B-1 8 x 8 x 1/4 tube 50,200 65,700 47.2 21. 6

B-2 8 x 8 x 1/4 tube 48,900 63,300 46.8 24.8

C-l 3/8 inch plate 29,600 53,500 53.4 30.5

C-2 3/8 inch plate 30,200 53,600 52.5 29.7

aReduction of area at location of necking

bElongation over the 8 inch gage length

 -20

TABLE 2 CONNECTION TEST RESULTS

Beam Shear
Specimen Tube Size a Size Moir b Test Load
(kip- inches) (kips)
D_1 c 6 x 6 x 3/16 12W27 140

E-1 c 6 x 6 x 3/16 12W27 135

6-1 6 x 6 x 3/16 10B15 13 28.0

6-2 6 x 6 x 3/16 12\f31 14 27.3

A-5 c 8 x 8 x 1/4 12W31 22 108

8-1 8 x 8 x 1/4 12W31 19 48.0

8-2 8 x 8 x 1/4 18W55 18 46.0

aA11 connection plates are 8-1/2 inch (3/8 inch ASTM A36)
with the 7/8 inch H. S. bolts.

bConnection moment for rotation of 0.01 radians.

cSpecimens from White and Fang tests.

 -21

5 ..
-L~6 =r<) -~
~ -I

t5 ..
•
. ~ -~
1'16
.. C\I C\I

-+ -+I I
<.0 CO

. I I
I I
I
I
I --~
I rt)
I II I
I~4 2" I
2" I ~4" -- ¢ -en
• ,----
.~ ~I
I
r1
I
r- ~
TYP "-
V
3/ ..
16
,
-- -0-

~
-0-
I
jl
II
II
II
.11
I
I I
I :
I I
, I
1I
-0-

-0-

-G-
I
I ~
=
~
<X>
I !J 1L_I I
I I I
I +
I
I
I
=v
~
I
I
I
~
- rt)

en
I
I
I
,
I I
I
1
All Holes 15/16"
...
,'l~1 ~ ASTM A325 Bolts

SPEC. TUBE PLATE BEAMS

6xl 6x6 3/ 16 10815
6x2 6x6 3/16 4 3/axalfa I 2 \!IF 3
axl 8x8~ 12\~3
8x2 4 3/ax8lfa 18W55

Fig. 1 SIMPLE FRAMING CONNECTION DESIGN

 -22

14(15)
12 (13)
I-I--.- 11
I
--+-- <==>1 1(2) 2

o
lOll czl°Q

~
II
o

~----
Gages in parentheses are
on inside of tube wall

"

Fig. 2 LOCATION OF ELECTRICAL STRAIN GAGES

 -23

BEAMS
I

I
I
10 B 15

1%1J~ j
II 6"
•-
----=.3---=O=--- II

~ .

I 12VF31
-rt'>

11
6
II I
- II

-to j
~ ~
_.

I 18VF55
-rt'>
A ~

3/1 41J1-- Gil

II

3_-_0
I II
~
- .
All Holes 15'JS II

Fig. 3 CONNECTING BEAMS

 300 K Macnine

,. I
iP 1
I

"'- 2 in. Bearing Plate

...

......L
0 0
0 0
0 0
---.-

Dynomometer Approx. 28"

L I""..,J

IT Mechanical
I
l--l

14"
Jack

I I

,.i:d I
N
+'
Fig. 4 SKETCH OF ELASTIC LOADING TEST SETUP
 TEST SET-UP
" 800 K Machine
,P
~ Bearing Plate

o o
o o
o o

Pedestal

I
N
V1
Fig. 5 SKETCH OF ULTIMATE CAPACITY TEST SETUP
 -26

Fig. 6

TEST SETUP FOR ELASTIC

LOADING

Fig. 7
I

TEST SETUP FOR ULTIMATE

CAPACITY TESTING
 .'

-27

LOAD MOMENT
(kips) (kip-in.)
30

2.0

.~

-(/)
Q.
.-
-«
~

0
1.5

0
...J

W
-:r.:
Z
(.)
15 '
« 10
2

0.5
I .

I
.01 .02 .03 .
. ROTATION (radians)

Fig. 8 MOMENT VERSUS ROTATION FOR ELASTIC LOADING OF 6-2

 -28

LOAD MOMENT
(kips) (kip - in.)

100

20

50

10

.~
\

."

o .01 .02 .03 .04 .05

ROTATION (radians)
Fig. 9 MOMENT VERSUS ROTATION FOR ULTIMATE CAPACITY TEST OF 8-1
 -29

LOAD MOMENT
(kips) (kip-in.)
I 3 2~

40
150

30

100

20

o Outside
o Inside

Compression Tension
---.J I
10000 5000 o 5000' 10000
STRAIN (fL in·/in.)

Fig. 10 TUBE WALL STRAINS AROUND CONNECTION OF 6-2

 -30

LOAD MOMENT
(kips) (kip-in.)
Avg. 4

. .

150

30

100

20

50
o

o Outside
o Inside

I • Compression Tension
I ~ . ~1_ _---::~'~_ _---1e-_ _- L _

2000 1000 0 1000 2000-

-' STRAIN AT GAGES 3 ¢ 4 (fL in./in.)
Fig. 11 HORIZONTAL WALL STRAINS NEAR TOP OF CONNECTION OF 8-1
 -31

LOAD MOMENT
(ki ps) (kip - in.)
6 5
Avg.~

,i
:
,
.

150.

30

c Outside
o Inside

10000 5000 0 5000 10000

STRAIN AT GAGES 546 (fL in./in.)
Fig. 12 HORIZONTAL WALL STRAIN NEAR BOTTOM OF CONNECTION OF 8-1
 -32

LOAD MOMENT
(kips) (kip- in.)

150 7 8

30

100

20

o Outside
o Inside

Tension

1000 0 2000 4000 6000 8000

STRAIN AT GAGES 7 ¢8 (fL in./inJ

Fig. i3 VERTICAL WALL STRAIN NEAR BOTTOM OF CONNECTION OF 8-2

 -33

LOAD MOMENT
(ki ps) (kip - in.)

12 Avg. 13
~,
~
,
~ .
,
40~ ,
~
, 150
~
\
t,
t,
30'

'.

o Outside
o Inside

.Compression Tension

2000 1000 o 1000 2000

STRAIN AT GAGES 12 $ 13 (,u in.lin.)

Fig. 14 VERTICAL WALL STRAINS OF CONNECTION WALL ABOVE CONNECTION OF 6-2
 -34

LOAD MOMENT
( kips) Odp- in.)

12 13

o Outside
o Inside

Compression Tension

1000 500 a 500 1000

STRA'IN AT STRAIN I? ¢ 13 (fL in./i~.)

Fig. 15 VERTICAL WALL STRAINS OF CONNECTION WALL ABOVE CONNECTION OF 8-1
 -35

. LOAD MOMENT
(ki ps) (kip-in.)

o Outside
o Inside

.-
.; Compression Tension

800 400 400 800

STRAIN AT GAGES 4 ¢ 5 (p.. in./ii1.)

Fig. 16 VERTICAL WALL STRAINS OF UNLOADED HALL ABOVE CONNECTION OF 8-2

l
 ~36

REFERENCES
•

1. White, Richard N. and P. J. Fang

FRAMING CONNECTIONS FOR SQUARE STRUCTURAL TUBING,
Proc. Journal ASCE, Vol. 92, No. ST2, April 1, 1966,
p. 175

2. American Institute of Steel Construction

MANUAL OF STEEL CONSTRUCTION,
American Institute of Steel Construction, 1966

3. Tall, L., et al
STRUCTURAL STEEL DESIGN,
The Ronald Press Company, New York, 1964

:
•