Research Bulletin 364
September, 1949
Farm Fence End and
Corner Design
By
HENRY GIESE AND
S.
MILTON HENDERSON
AGRICULTURAL EXPERIMENT STATION
IOWA STATE COLLEGE OF AGRICULTURE
AND MECHANIC ARTS
AGRICULTURAL ENGINEERING SECTION
AMES, IOWA
CONTENTS
P age
Summa ry .................................................. ........................... ............... 47
Sta temen t of problem... ... ....... .. .... ... .... ... ................. ...... ..... ......... ..... 49
Design load ................................. ............... ! ... . ............. . .......... . .......... 57
Structural analysis ..............: ............................................................. 65
T ests on models................................ .................................................. 71
Field tests- Series L ........................................................................ 78
Field tests- Series IL .................... .................................................. 87
Time test ...................................................................... ...................... 109
R eferences .......................................................................................... 11 9
47
SUMMARY
Field surveys of farm fences show that many are in poor condi·
tion. The end construction is a critical factor in the suuccessful per·
formance of a fence. The study herein described was undertaken to
find the causes for failure, to appraise the relative value of common
construction methods and to attempt to devise better ones. Particu·
lar attention has be.e n given to labor saving in the hope that knowl·
edge of improved methods would result in more satisfactory construction on the farm.
A field study disclosed the factors responsible for failure, gave
some information on loadings to be expected and some suggestions
for assemblies which might be built with a minimum of materials and
labor and still be expected to give satisfactory service. The experi.
mental work included observations of loads which might be imposed
on the fence end or corner by wire fencing, and a study of forces
necessary to destroy ends and corners fabricated in a variety of
ways. A test was made to show the effe.ct of time and temperature
on the tension in wire fencing and the ability of two types of fence
ends to resist such factors. Field tests were made in one soil type
only and under approximately similar soil moisture conditions. In
most cases, replications have not been possible.
The tension curve in woven wire fencing is beneficial in helping
to maintain a taut condition but is not entirely effective because the
elastic range is small. The manufacturer's recommendation to half
remove the tension curve is not sufficiently specific, is not equally
applicable to summer and winter stretching and is likely to give vari·
able results because of differences in shape and size of the tension
curve. The height of the tension curve., as well as the quality of the
wire involved, definitely affects the tension necessary to remove half
of the height.
Tension CUrves in barbwire would be beneficial in maintaining a
taut fence and also in reducing the total load on the fence end. Ten·
sion springs in the barbwire may provide a simpler means of ac·
complishing the same result.
The load on a fence end will vary with the combinations of woven
wire and barbwire used, but for 832·6·11 woven wire and three
strands of barbwire, the load should be approximately 3,000 pounds.
A drop in temperature may cause the load on the fence to increase
as. much as 50 percent. This increase is caused largely by the barbed
wues.
At least two double strands of No. 9 wire should be used for
tension members in a fence end.
Lengthening the span decreases the vertical force on the end post
of an unanchored assembly.
A horizontal brace located at the top of the posts proved superior
to other types of single span bracillg.
The double span end assemblies all displayed more favorable
48
characteristics with respect to vertical and horizontal movement than
did the single span assemblies. The single spans failed
through a combination of vertical and horizontal movement, while
the double spans failed as a result of the buckling of the long col·
umns formed by the horizontal braces. The single span end post roo
tated about its base, while the double span end post remained more
nearly vertical as it moved through the soil.
Increasing the depth of set of the double span with horizontal compression braces from 2'-6" to 3' -6" nearly doubled its holding
power . .
The 16'-6" horizontal compression brace double span end held
214 percent of the load with 54 percent of the horizontal and 43
percent of the vertical movement of the 8'-6" single span end.
The single span corners failed because of vertical movement of
the corner post. The double span corners could not be pulled to rupture with equipment available (fig. 25), but the indications were
that failure would have resulted in a considerable horizontal movement before a corner would rise out of the soil.
The operating characteristics of a single span were affected adversely when subjected to corner conditions, while those of a double
span were definitely improve.d. This is doubtless explained by the
difference in type of failure of the single and dOllble spans. When
used in a corner, one double span assembly tends to stabilize the
other.
A 16'-6" double span corner holds approximately 230 percent
of the resultant rupture load held by an 8'-6" single span corner
without failing. The horizontal and vertical movements of the 16'6" double span corner were 12 percent and 6 percent, respectively,
of that of the 8'-6" corner at loads equal to the rupture load of the
latter.
In both the single and double span corners, an increase in the span
length gave an increase in the holding power, but the increase for
the single span was not as great as was indicated by the tests on end
structures.
Taut bracing decreased the horizontal and vertical movement of
the corner post considerably, thereby increasing the holding power.
Improved performance can be secured by:
(1) Improving the resistance of end construction.
(2) Decreasing the load.
The load can be decreased by:
(1) Improving the elastic properties of the tension curves III
woven wire.
(2) Constructing tension curves in barbed wires or accomplishing the same results with tension springs.
Tests on small-scale models indicated the possibility of still further increasing the holding capacity of the double span arrangement
by using cross braces in both spans and applying the load to the
center post.
Farm Fence End and Corner Design
By
HE NRY GIESE AND
S.
MILTON HENDERSON 2
Fence construction principles have received little experimental attention. Many opinions have been offered but facts concerning tested
structural designs are still inexistent. An editorial published in Agricultural Engineering (3) expressed the status of farm fencing thus:
Taking the country as a whole, it seems sure that no other part of the farm
plant is so far gone into disrepair and disorganization. It seems equally ob·
vious, in general, that few structural improvements are so quickly self-liquidating.
The farmer's present-day investment in fencing is difficult to estimate because of the variation between communities in types of
farming and size of farm. Miller (7) estimates the amount of fence
and investment as follows:
The fence must be of some service to farmers, or the average farm in the
United States would not have between bOO and 700 rods of fencing, nor would
the average midwestern farm have 10 rods of fence per acre, for the 160·acre
farm.
The cost for a woven wire fence in place is about $1 per rod for a 4·fool
high fence. About 80 cen ts of the cost of this fence is for material and 20 cents
for labor.
Aitkenhead (1) makes the following statement with regard to
fence costs:
A survey of 30 farm s averaging 160 acres showed a fenc e investment of
$1,500.00 each or 10 percent of value of the farm. Costs of upkeep were 18
cen ts per rod per annum.
The end construction is a basic factor in the effective life of a
fence. Failure of line posts usually causes only local disturbance
and posts are easily re.p laced. Failure of the end, however, necessitates complete rebuilding of the fence. A survey was made of end
and corner units on farms to determine those factors which may
cause failure and also construction methods which have given satis factory performance. Particular attention has been given throughout
this study to the development of techniques by means of which a
successful fence can be built with a minimum of labor. Farmers
generally appear to have been unwilling to put forth the effort necessary to construct anchors which have demonstrated their ability
to hold a fence in place. Vertical movement of the end post and/ or
IProjec t 6]8 of th e Towa Agricidtural Experiment Statio n, in cooperation with the Farm Fenc·
jng Association.
2The authors wish to acknow le dg e the contriblltions of two Icscarch fellow s, G. L. Hazen and
M. D. Stron!!_
50
Fig. 1.
Typical failure of e nd post and compression brace.
horizontal movement of the entire end or corner assembl y may result
from one or more of the followin g construction faults:
1.
2.
3.
4.
5.
Inadequate tension member.
Inadequate compression brace.
Inad equate methods of fastening the component parts.
Infer ior materials.
Inferior workmanship.
Perhaps the most evident type of failure in end and corner constructions on farms was the vertical movement of the corner or end
post. In all but a few of the constructions observed there was some
vertical movement. It was not possibl e to determine in all cases
Fi~.
2. Failure resultin g from insufficient depth of se t.
51
Fi g. 3.
A large end post may not be adequate.
whether the construction was anchored, or how deep the end or cor·
ner post was set in the ground, nor to correlate one factor with any
other.
Figure 1 shows a typical failure resulting from vertical movement
of the end post. The length of span was approximately 9 feet and
the brace height was 3% feet. There were five No. 12% gauge
barbwires attached to the corner post. The tension member comprised
a double strand of No.9 galvanized wire.
The length of span in fig. 2 was approximately 11 feet. There
were only three barbwires fastened to the end post. The end post
Fig. 4.
Failure of a short spa n.
52
in this particular structure was completely out of the ground showing
a depth of set of approximately 2 feet.
The corner post of the. construction shown in fig. 3 was ap·
proximately 10 inches in diameter and the span was approximately
8 feet in length. An 832·6-1P woven wire fence and three barbwires were attached to the corner post.
The vertical movement in this case was not as pronounced as in the
constructions shown in figs. 1 and 2.
A short span of approximately 5 feet is shown in fig. 4. This
construction had two diagonal braces each of a doubled strand of
barbed wire. As can be seen the construction was intact structurally,
but it was not holding the fence.
The horizontal movement of end posts was difficult to observe and
analyze in the field because the soil tends to fill in around the base
of the posts. However, in some cases where the corner construction
was intact the fence was loose, indicating either that the wire had
stretched or that the corner post had moved laterally. Wood compression braces less than 4 inches in the narrow dimension usually
showed signs of buckling. That in fig. 2 was made of a 4"x4" wood
-member about 11 feet long. It showed no sign of buckling.
Failure.s in the tension member were rather difficult to isolate in
the field because a slack tension member could be caused by other
3This number indicates the st yl e of woven wire. Tn this case th ere arc 8 line wires, the
height is 32 inches. The stay wi res orc 6 inches apart, and the filler wires are 11 gauge.
Fig.
5.
Faih! re
due
to
inadequat e
compression
brace
fastening.
53
Fig. 6.
Failure of tension member.
types of failures without any evidence concerning the .condition of
the tension member before failure occurred. In most cases the tension member was made from a double strand of No.9 gauge smooth
wire, or a double strand of barbwire. The member was tightened
by inserting a rod through the strands and twisting until the entire
structure had the desired rigidity. Failure of the tension member
could result from stretching of the wire by overload, by untwisting
of the wire, or by loosening of the wire at the points of fastening
and subsequent slipping on the post. In fig. 6 the end post was set
sufficiently deep in the ground to prevent rotation of the base of
the post, but the load imposed by the fence bent the post above the
ground line. The fence was still in fair condition notwithstanding
an ll-inch lateral movement of the top of the post. The rod used
to twist the tension member wire together slipped from the loop and
allowed the wire to untwist.
Most compression members were toenailed to the end and brac!'
posts. Figure 5 shows a typical connection between the end post
and the compression brace. The ends of the compression member
were weathered more than the remainder of the member, and the
nails no longer held the brace in place. A slight load applied perpendicular to the axis of the brace was sufficient to dislodge the
brace. The support furnished the brace by the wire tension members was almost as effective as were the nails. The large number of
total failures caused partially or wholly by the decay of the materials used in the construction emphasized the importance of adequate structural materials.
Figure 7 illustrates two unique methods employed by farmers to
carry the heavy load imposed by a fence . They are, however, not to
be recommended. The stone is unsightly. The stump, in this case
54
Fig. 7.
Unique methods of llnchoring are some times employed.
conveniently located is likely to decay soon, giving only short-lived
performance_ Adequate construction is possible without resorting
to such expediencies_
The survey disclosed several successful end constructions which
having demonstrated satisfactory performance, provided stimulus
for some of the research he.rein reported_ Most of these employed
what might be termed a double span arrangement
In a few cases an anchor of some type was used, but very few
farmers seeme.d willing to spend the extra effort necessary for a construction of this kind_
The double span end shown in fig_ 8 had been in service in a livestock enclosure for 4 years and was still in excellent condition_ The
arrangement was 1 rod long and use.d the first line post for the
second brace post, thereby reducing the cost of the end structure_
The compression braces were run from a point approximately three-
Fig. 8.
Thi s double 8pau has given good service for 4 years.
55
Fig. 9.
This doubl e span has held th e fen ce well.
fourths of the above-ground height of the post to the mid-point on
the brace post.
The arrangement shown in fig_ 9 is in use around the highway
maintenance grounds 2 miles east of Atlantic, Iowa. The fence is
composed of two 726-6-9 strands of woven wire and one piece of
double strand No. 12% gauge barbwire, making a total height of 54
inches. The load resulting from this type of fencing is as severe as
will be found in almost any farm fence, ye.t the corner construction
was in perfect condition with the exception of the second compression
brace which was bending slightly. There was no e.vidence around the
post of any movement, either horitontal or vertical.
Fi g:
)0 .
A doubl e span. usin g horizontal bra ces.
56
Fig , 11. Equipment u sed in testing tension curves used in woven wire. Top : Th e ex tensometer measu lc changes in l ength. Bottom: Loading the wire sa mpl e in tes ting ma chin e.
The horizontally braced double span, shown in fig. 11, maintained a tight pasture fence after years of service. The one double
strand of wire had stretched, allowing the posts to lean.
The principal problem in fence end construction arises from the
fact that the loads applied to end posts by tight fence wires are
above ground whereas the resistance is supplied by the soil below
the ground level. This eccentric loading causes a tendency toward
rotation of the fence end and exerts a vertical force which fre-
57
quently lifts the end post out of the ground. In numerous fence
constructions, this te.ndency of the end posts to rotate and lift has
been met satisfactorily by nailing lugs to the bottom of the end post
or by anchoring the tension member, which is fastened to the top
of the brace post, to a "dead man" in the ground beyond the end
post. This construction re.lieves the end post of the vertical force
applied by the tension member.
DESIGN LOAD
The proper initial tension for stretching Vari01,IS wire arrange·
ments at the time a fence is erected has not be.en definitely deter·
mined. Reynolds (9) gives the proper initial tension for summer
stretching in a standard 832·6·11 woven wire fence as 1,600 pounds.
This pulled the tension ctl'rve to one-half its tensionless size, which
is the recommendation commonly given by manufacturers. The
load for a No. 12% gauge barbwire was given as 250 pounds. Hazen
(5) placed a dynamometer in a standard 832·6·9 woven wire fence
at the time of stretching, and the loads for two separate stretches
were 2,300 pounds and 2,600 pounds. A dynamomete.r placed in a
stretch of 726·6·11 style woven wire on a farm showe.d that it was
under only 800 pounds of tension.
Several specimens of wire tension curves were teste.d (a) to dete.rmine the load imposed upon an end construction by the initial
stretching of various types of wire, (b) to compare the. elasticity of
the crimped wire with the straight specimens and (c) to predict the
effectiveness of the curve in relieving the stress in the fence caused
by temperature changes.
Both straight and crimped specimens were chosen from gauges No.
9, 10, 11 and 12% which represent the most common sizes used in
fencing.
A conventional laboratory testing machine was utilized, in loading
the specimens.
An extensometer was constructed to determine the deformation in
the tension curve during the loading period. This device with a
specimen of wire in place ready for testing is shown in fig. 12. It
consisted of an Ame.s dial gauge and two clamps pivoted on an arm
opposite the dial gauge. The two clamps, containing a 4-inch length
of the specimen, were. made of pointed thumb screws which fastened
on either side of the tension curve.
A piece of metal was clamped to the wire across the tension curve
so that a micromete.r could be used to determine the amount of curve
removed for a given ' load.
The procedure followed throughout the tests was as follows: The
wire on either side of the curve was straightened to fit in the jaws of
the. machine, the piece of metal used in taking the micrometer read-
<00
T<;:ST
'5
50
"''3\NII2E . FACTOR.Y
TE "-l 5\Ot-J CURVE
CURVE ~ALF
RE.MOV ED -------.....
4 0 ::>
I
ELOt--lC1A.T\Ot-J
~
.Ji
d)30 :J
--- ~-; rj
..J
D
~
.:( Zo o
o
-1
100
o
--1 / ./-
L--C
~
~
[ZV
~v
o
2""'"
0 .1
0.2
~
V
V!
/{/
Loss ItJ Ht:.IC;HT
R
OF CURVE
•
I
O:~
0;4
0.5
0 .<0
L
jV
O.T
DEFORMATI O N
Fig . 12A.
~
O.B
-
0 ,9
Q I
(I N 'S .)
Charac teristic loading curves
for
fence wir
59
N0 . 12 C.A.
BARB
\NI~E
YIELD
POIN
-f
6U~----+-~~------~ ----~--~--------~-,~--~
/"THEO RET\C.AL
I ~TRE"5'S-~TRA'~
CUR-VE. F O g.
'::>TEEL '
I
I
200
I
I
1
IOo~--~--~~~~~+-~~~+-~~---+------~
\0
'2.0
o
ELONGATION
fig. 12D.
It-.}
\0
400FT.
~o
00
-(l ~C.\-H::CS)
Characteristic 10ading curves for fe nce wire.
60
tJo. c;, ~Jrt... <:;ALV, WIR~ WITH TE..NSlor.J
C.UR-V£:.,& IQri 0 C .
,OO'b---_+--~~----r_--_+----~
yl~POaJT
G09-~~~--_+----+_----~--_+----+_--~--__~+__++r--_1
5.~--T-~----+_--_1----~----+_--_1~~~--T_t__i_F--~
iii
~
, 40. 1-~--~---+----~----~--~~~~~-rr-~++--r+f----1
T~EORE.TICAL "'S-n~E"S~-STEA."J
o
S
I
I
CURVE. Fe>", 'STI!.E:.L
~~-+--~----+----4~~/_/~---+++--~~~'
aD
Fi g. 12C.
'90
{DO
Characteristic loading curves for fence wire.
ings clamped on the specimen, the wire placed in the extensometer,
the specimen clamped in the testing machine and initial readings
taken. The load application consisted of applying 40-pound increments and observing the stretch in the wire at each increment. The
load was completely released at 120-pound intervals to observe any
permanent se.t. After each release, loads were applied as before. Observations of elongation and decrease in depth of tension curve were
made at the various loads. The results are presented in table 1.
These data will not apply to all brands of fencing. Elastic properties of fence wire vary widely with the result that some would
permit more stretching load than others. Manufacturers could
profitably study this problem further toward the end that tension
curves would be more effective in maintaining proper fence tautness.
In general the higher tension curves require less load to produce
half reduction in height than the shallower curves. A considerable
variation in the initial curve heights made it difficult to establish
an accurate relationship between curve deformation and load. The
fact that woven wire is usually made of two wire gauges further complicates the problem. It is, however, permissible to assume average
conditions applicable to a line. of fence since a great number of
tension curves are represented. In a cattle fence with 1047-6-11
woven wire topped with one strand of barbwire, the top and bottom
61
TABLE 1.
TESTS ' OF WOVEN WIRE TENSION CURVES.
(ALL DIME NSIONS IN IN CHES; LOADS IN POUNDS.)
Conditions wh en cu rve wa.s reduced half th e
in iti al height
Initia l
curve
height
Loap
Ib s.
.266
.324
.360
.394
.336
520
400
440
324
421
E l ongation
of curve
No. 9 Gauge
~
Av.
I
I
1
Ult.
Per man ent
se t
strength
.012
.0 11
.012
.010
.0 11
.058
.088
.1I 2
.090
.089
1,440
1,320
1.360
1,240
1,353
.012
.0 15
.014
.043
.11 8
.080
1.360
1,3 20
1,340
.012
.0lD
.009
.010
.042
.034
.032
.036
1,040
1.040
920
1.000
E l astic
range
Wire
.070
.099
.124
.105
.100
No. 10 Gauge Wi re
-1I
.430
.750
.590
No. 11 Gauge Wire
Av.
.200
.210
.236
.215
Av.
.235
.252
.255
.247
1
400
400
320
- -373- -
.054
.044
.041
.046
No. 12\!, Gauge Wire
]40
280
180
200
.067
.067
.060
.067
I -'~I
.017
.018
.0 19
.01 8
.049
.047
.049
760
880
720
787
No.9 Gauge Wire (Lab. Mad e Curve)
Av.
I-~I
.260
.270
---.226
760
520
600
627
.081
. 081
.094
.085
1
I _.070
'O~I _ ~~O_
1,260 .
.014
.Oll
.01r,
--.014
.078
.071
1.500
1.373
------
Ten sil e Strength of Plain \Vi re
No.9
No. 10
No. II
No. 12t;,
(2 tests)
(I tes t)
(2 tests)
(I test)
_' ~1,400
1 1,100
800
wiTes of the woven fence aTe No.9, each of which would require 421
pounds to reduce the height of th e tension curve to one-half its value
under zero stress.
The filler wires are No. 11, each requiring 3.73 pounds. The total
necessary load for two No.9 and nine No. 11 wires would thus be
4_199 Ibs. To this add 250 Ibs. , the recommended tension for a barbwire, and the total initial load would be 4,449 Ibs. A hog fence with
832-6-12% woven wire and three barbwires would require an initial
end load of 2,793 Ibs.
The character and significance of the elastic ran ge are demonstrated in fig _ 12 which was plotted from a test chosen at random.
62
\
Further tests we.re made to determine and to compare ·the elastic
properties of plain No.9 fence wire, No.9 fence wire with tension
curves and barbwire, and also to relate or use the findings to estimate the. effect of temperature upon fence tension.
No.9 galvanized fence wi.re and No. 12 harbwire were used in test.
The barbwire was American Glidden Cattle fence made by the
American Steel and Wire Company. The two-point barbs were
spaced at approximately 5 inches.
The wire was twiste.d 1% turns between barbs. The cross sectional
area of the No. 9 wire, 0.0173
square inches, was nearly equal
to the area of the two· 12-gauge
wires used for the. barbwire which
was 0.0176 square inches. Since
no single-strand wire with tension
curves was available, a length of
the No. 9 wire was crimped with a
shop-constructed tool designed to
duplicate the tension curve in the
top and bottom No. 9 wires of
woven wire fencing. The curves
were spaced 6 inches on centers as
is customary in woven wire.
One end of a 400-foot length
was fastene.d securely to a 6-inch
steel pipe set in concrete, the other
to .a tractor which was fitted with
a scale for measuring the load
(fig_ 13) _ At the 400-foot point,
just behind the tractor, a board
was placed under the wire to provide. a reference point for measuring the elongation_ A steel rule,
graduated in hundredths of an
inch, was used to measure the
elongation, pencil marks on the
board and wire being used for
reference. points.
The load was applied by putting
the tractor in low gear and turning the crank. Observations of
elongation were made for every
Fig. 13. Appara tu s used for de terminin g
50-pound increment of load. For
th e e l as ti c prope rti es o f fe nce wir e. Th e l oad
each 100·pound increment the. load
was appli e d by movin g th e trac tor ahead
by hand c rankin g. Th e pull e y ove r wh ich
was removed and the amount of
th e ca bl e passed was fitt ed with ant i- fri cpermanent set observed. The friction . bea rin gs to minimize drag.
63
tion of the wire on the grass and of the pulley arrangement accounted for about 15 pounds for the No_ 9 wire, both plain and
with tension curves, and 70 pounds for the barbwire. These values
were determined by detaching the wire from the steel post and noting the force necessary to slide the wire. through the grass.
These tests were made in a level, close-cropped pasture.
The results are reported graphically in fig. 12-B. The theoretical
stress-strain curves are based on data for structural steel and are
incorporated in the chart as reference standards. The resistance of
the No. 9 wire upon the ground was so small it was not considered
in plotting the data. A 35-pound compensation, half the total resistance value, was applied to all the barbwire computations.
The wire used in these tests contained many bends of various
shapes and sizes, this being a condition which is always found in
coiled wire or any wire which has been handled. These bends
have a small but definite elastic quality which affected the plotted
results. It is this condition which gives the graph for the No.9 plain
wire (fig. 12-B) a curved characteristic rather than straight below
the yield point. This condition is present in the other graphs but
is covered up by the elasticity of the tension curves.
The unloading curves are of much importance since they indicate by their slope and position the elastic properties of the wire_
The slope of the unloading curves gives the elasticity of the wire.
The elongation at zero load for an unloading curve is the permanent
set or stretch resulting from a particular load.
The value of design as it affects the elasticity may be evaluated
by comparing the elasticity of the wire to the material from which
it is made; viz., the slope of the unloading curve for a 365-pound
load on barbwire is 2.18 inches per 100 pounds. The slope of the
curve of the material, 0.94 inches per 100 pounds, from which the
wire was made is 2.32 times that of the wire_ Stating this relationship in another manne.r, the wire is 2.32 times as elastic as the material from which it is made. This value may be used as an index
of elasticity. Similar values for the wires tested at various loads
are tabulated below.
ELASTI C INDICES OF WIRES TESTED
Rati o of Elast icity o f Wi re to Wire Mate ri al.
Load
(lb •. )
No. 9 gal v.
pl ain
No. 12
barbed
No. 9 ga l v. with
tension curve
200
1.38
3.92"
400
1.32
2.21"
2.88
600
I.l 2
1.63"
1.81
800
1.]]
1.38"
1.34
*De te rm ine d by int e rpol ati on.
5.19
64
The coefficient of thermal expansion for steel (annealed) is
6.1x10-6F. The decrease in length of a 400-foot length of wire,
plain, barb, or with tension curves, for a drop in temperature from
80 degrees (F.) to -20 degrees would be 0.30 feet or 3.12 inches. If
the ends of the wire are fixed so no movement can take place, the
decrease in temperature will cause an increase in load. For example,
let us assume that when the temperature is 80°F., a 400-foot length
of No.9 wire is stretched to 380 pounds, the amount Strong found
necessary to half remove the tension curve. A drop in temperature
to -20°F. would increase the tension by the same amount as stretching it 3.12 inches, or to approximately 530 pounds. This would
result in some permanent stretch in the wire and a consequent drop
in load to 260 pounds when the temperature again reached 80° F.
The barbwire under the recommended load of 250 pounds under
the. same temperature conditions would increase to 330 pounds,
dropping to approximately 185 pounds. The No. 9 wire with the
tension curves (fig. 12-C) under a 380-pound load would increase
to 395 pounds, dropping to approximately 285 pounds. The significant observations to be made here are that, under the selected
temperature condition which is typical, the increase in tension of a
plain No.9 wire is 150 pounds; a barbwire, 80 pounds; and a No.
9 wire with tension curves, only 15 pounds. Barbwires twisted
tighter initially than the one tested would approach the plain No.9
wire in performance, a condition which is not desired.
The results of a field study of the effect of temperature and time
upon fence tension was made in connection with a time study of
two end constructions and will be found reported elsewhere in this
bulletin.
The loading in the fence wire may be affected by one or a combination of factors such as initial stretching of the wire, contraction
of the wire caused by a change of temperature and impact or transverse loading caused by animals leaning on and running into the
fence. The magnitude of the fence loads may be quite severe as is
evidenced by the fact that the wire is often given a permanent set. .
A good fence end construction will resist displacement when the
loading caused by the fence wire is placed upon it. The demand
upon the fence end, however, could be made less severe by improvement of tension curves and proper selection of metal in woven wire.
The barbwire is, however, the principal offender. Not only does it
possess elastic properties unfavorable to the fence end., but its location at the top of the post maximizes the moment around the
point of rotation of the post.
An attempt should be made to impart similar properties to barbwire making it retain a large measure of tautness in spite of some
movement of the end post or changes in wire length due to temperature variations. This might be accomplished by tension curves
65
in the barbwire or perhaps might be obtained more easily by in·
serting tension springs in the barbwire line. In such a case it would
be imperative to so drive the staples that the wire would be free
to move within the confines of the staple. The importance of this
can be visualized by considering the potential change in length of
the wire in a long fence due to changes in temperature. Many 160·
rod stretches of fence and those approaching 320 rods are not un·
common. Unless restrained by the ends, a fence line 160 rods in
length will shrink 0.2 inch for each degree (F.) temperature
drop. A 50° temperature drop from, let us say 80 0 to 30° , would
cause a shrinkage of 10 inches or a marked increase in the load on
the fence end.
STRUCTURAL ANALYSIS
The object of this analysis ~as to determine (a) the effect of arrangement of the structural members of an end construction on the
load distribution by graphical analysis, and (b) the actual and
allowable load on some of the braces found in the field.
The end construction shown in part A of fig. 14 has an arbi·
trary load on the end post which is transmitted by simple beam
action to the panel point C and the ground line AB. The magnitudes
of the portions of load transmitted to C and AB are represented by
loads P 1 and P 2 , respectively. The load P 2 is taken by the support
which is the earth around the base of the end post. This is obvi·
ously at some variance with the actual situation as the soil is unable
to provide positive support at its surface. Load P 1 is transmitted
through the braces to the end and brace posts, and finally to the
ground. Since load P 1 causes the vertical force on the end post, the
stress diagrams for this part of the study were drawn using the
load P 1 applied at point C. Space and stress diagrams for a change
in length of span are given for A. Note that the longer the span
the smaller the vertical component. For a given load and span no
change in vertical component could be obtained by changing the
arrangement of the bracing as shown in fig. 14 part B.
Lack of information regal'ding the support given by the soil is a
limiting factor in an attempt to make a structural analysis.
From diagram D in fig. 14 it will be observed that inclining the
brace post offers a possible method of reducing the vertical force.
An angle of 45,0 seems to be the most satisfactory from the stand·
point of reducing the vertical force and at the same time keeping
the brace post and compression braces reasonably short.
The characteristics of the soil pressures fO'r unbraced posts were
studied in an effort to predict the possible action of the soil about
the bases of the end and corner assembly posts. An analysis of
66
.J
I
4:fV<fJ
-0
STR ESS D IAGR A M ,
K
P_Bm;.jll~
1F_I2H3H4H5H"H7__t8-~~~gl~~
P-;',._-_g
I
B
LOAD
DIAGRAM
A
F
!5" W~
gw
1~~~=17e
"
DIAGRAM
SPACE
I,
I
A
~3
4@C1
A t:~i
z
3
6
t- 0
3
E
'i.
I~~
• :> 8
2.
c
il
SPACE
F
DIAGRAM
B
Po~
:jzf
~Z
jS~EF~·
'y- C
~r
uz
i=W
o: z
~~
tA- WO
.4
-
~
LOAD Po Z U
STRE.55 DIAGRAM ,
DIAGRAM
o
UNIT 5PAN LENGiH
LOAD Po
STRESS DIAGRAM .
B
~f-~
d
Z 0.
0~
• LOAD P.
P.~k.. D
A
~
u. ffiZ
00
Z
MEMBERS
p
A
~o.
o
CURVE SHOWING RELATI ONSHIP
BETWEEN SPAN LENGTH &VERTICAL
COMPONENT FOR TYPICAL END
CONSTRUCTION .
E
Fig. 14.
Structural analy sis of end con struction s.
stabil ity studies of transmission poles has been made by Seiler (8).
Quoting from his paper:
[Figure 15] represents the principal forces acting on a pole when set in
compressible or granular soil and subjected to a horizontal load W acting at
a distance of Z from the top of the pole. The pressures developed are com·
monly . considered as the ordinates to a parabola whose position is such that
the pressure area on one side of the pole bears the same relation to that on
the other side as R does to P , these being the butt reactions.
P
=
.
y
W-and R
x
=
P
+W
In general, the ratio of R to P is not far from 1.1 and the pressure areas
very closely satisfy th is relation when the neutral axis of the figure occurs at
a point distance b
O.324d from the butt of the pole.
=
Apparently the diameter of the post has little effect upon the sta·
bility. This fact is brought out by Seiler (9) in the following
statement :
67
[Figure 15] indicates approximately the shape of the prism of earth which
must be ruptured loose when a pole falls down. The point b determines the
altitude of the prism, and its distance from the pole depends on the depth
of setting. The shearing areas, then, are proportional to the following:
Area A (two sides) = cd 2 sec
Area B =
(D
e
+ kd + D) cd sec e
2
2D)cd sec e
2
kcd 2 sec e
Dcd sec
2
(kd
+
+
A+B
cd 2 sec
e
d 2 (c sec
+
e
e
cd 2 k sec e
2
ck sec e
+
-2-
+ cdD
+ dcD
sec
sec
e
e
Since the angle e is constant for any soil, the shearing area varies as the
square of the depth of setting; hence the resistance of a pole to overturning in
a given soil would vary approximately in the same manner, and can then be
represented by the ordinates to a parabola. Hence,
Resistance to overturning = Md 2 -r Nd
The diameter of the pole D is reflected in the constant N but since in practical
cases D varies approximately as the depth of setting d, we can write:
+
Resistance to overturning = Md 2
N1 d 2
= (M
N 1 )d 2
Of course the weight of the prism of earth to be moved affects the equation,
but a careful analysis of the problem indicates that this would be reflected
in the increased shearing resistance of the soil, which would tend to raise slight·
.
ly the exponent of the variabl'e d.
Obviously, even considerable variations in the diameter of the pole have
little if any effect in increasing the shearing areas, and therefore resistance
to overturning, because D affects only the first power term of the equation.
Moreover, the soil would rupture along surfaces of weakness, more or less independent of the size of the pole butt, and this is substantiated by actual
tests.
+
This analysis of the unbraced post provided a background for a
study of the single span horizontal brace end construction. The
function of the braces is to keep the end post vertical at all times.
The brace post theoretically serves the one purpose of kee.ping
braces in position, but actually it takes part of the load. This is
because immediately after the structure is loaded the soil yields, the
end post moves, and the brace post then takes part of the load. The
brace post acts in a manner typical of the unbraced post as illustrat·
ed by fig. 16. The end post, on the other hand, acts somewhat differently. The bracing tends to keep the post approximately in a vertical position, and the entire structure has a tendency to rotate about
68
the point on the brace post where the tension and compression braces
meet. Rotation of the structure results from a combination of vertical and horizontal movement_ The vertical movement is produced
by the vertical component of the load carried by the tension member. From these observations it is only reasonable to deduce that
the rotation of the end post in itself must be about the butt which
moves not only horizontally but vertically.
The double span arrangement would act similarly to the single
span with the exception that the vertical and horizontal movement
for a given load would be reduced.
The work of Seiler shows that the curve of earth pressures between the ground line and rotation point is a parabola. The single
Cl
Section G-G-G
A
A
ox.
-,<lI'0'
(L'
~I
-0
.:.:
0
0
~!
J
cd
'I b
Area A
Elevati on
0
cd
,I
P ·lan
C
Fig. 15. Earth pressures resu ltin g from load.ing an unbraced post. A shows th e force distri·
bution on a post l oaded at W. Vi ews B. C, D define approx im ate ly the mass of earth rlIDturcd
loose when a po le or post is overturned.
b9
a
P/z..
R
~
0
III
~
-rJ
-
cJ
J
'1
R.
A... sutTIe Pl'"e:.-
p
sure Cl..!rve
p '0,_'~~~~~
0'
I
PJ"
A'!.... um
eo
pilI
Fres"",-,re Curve
~
'R~
_
Y,.~
(IJ
p'"
(b) 1NIER'I'-A~Pl,1>.\ E(c) E't-.\.'V 'D>::'!:...CE
e>RACE PO~\
'PO'S\'
fic:o 16.
Theoretical ana l ysis
or
sin~l e
and double span end.
70
span end with posts showing these coresponding earth pressures
is shown in fig. 16.
In an unanchored construction the diagonal braces carry the load
which causes the vertical force on the end post. These loads may be
divided into a horizontal and a vertical component. The vertical
component of the load on the end post in combination with the
heaving of frost action constitutes a major cause for failures of an
end construction by the vertical movement of the end post. The
vertical component of the load must be taken either by an anchor,
the weight of the post, or by the friction of the end post against the
soil, or by all thre.e.
The curve shown in fig. 14E gives the relationship between the
length of span and the vertical force on the end post for the given
loading and height of brace. The equation of this curve is xy = c,
where x and yare the vertical for<;e and length of span respectively,
and c is a constant. This curve shows that a short span must resist a relatively large vertical force as compared to a longer span.
The length of span is the most important of the factors affecting the
vertical force.
In the case of the arrangement shown in fig. 16, all the horizontal component of the brace load is taken by the end post. Crossed
braced arrangements as the one shown in fig. 14C distribute the horizontal load between the end post and the brace post. If the compression brace is attached to the brace post near the ground line and
the soil is sufficiently strong to vrevent rotation, then the tension
member would not carry any load. In this case all the horizontal
component of the compression brace would be taken by the brace
post. However, since the soil in most cases is not sufficiently strong
to prevent rotation, there will be a load in the tension member, and
as a result, a portion of the horizontal component of the brace
load will be transmitted to the end post.
The inclined brace arrangement shown in fig. 14, part D, reduoes
the horizontal component of the brace load as the angle between the
compression brace and the brace post is increased. If it were possi·
ble to increase the angle between the brace post and the compression brace to 180 degrees. the horizontal component would be taken
entirely by the compression brace and the brace post. and no tension member would be necessary.
In many of the constructions observed the compression member
had buckled. This failure could be caused by the use of too small
a cross section or by the deterioration of the materials from which
the brace was made.
Most such failures are evidently caused by the use Qf poor materials and not because the section of the member was too small. The
use of initially warped members or members of poor durability, and
in the case of steel, the use of discarded boiler tubes which have
71
rusted through, all lead to premature failures of the end construction_
Giese (4) recommends three duuble strands of No_ 9 wire for a
tension member for a 9-foot crossed braced arrangement. On basis
of these calculations, three double strands would carry a maximum
safe load of 2,700 lbs., which is ample for this type of arrangement because the tension member carries a load equal to or less
than the brace load. This will devend upon the size of posts, depth
of set and condition of the soil in which the end construction is
set. Many of the end constructions in the field are failing because
the one double strand of No. 9 wire, which is commonly used, is
not sufficient.
TESTS ON MODELS
The results of the structural analysis of the fence end construction indicated that the length of span and arrangement of the structural members had an important bt::aring on .the vertical force on the
end post. Preliminary tests were made on scale models in the lab·
oratory to determine the general characteristics of various arrangements and lengths of span on the holding power of the fence end
assembly.
The dimensions of the structural members were chosen from the
results of Allbaugh's (2 ) survey and the recommendations by
Giese (4 ) . The assemblies were chosen as a re.sult of the field study
and were representative of those found in practice.
The same end and brace posts were used throughout the tests,
except that the end post had to be replaced once during the tests because of fracture.
Fi g.
17.
Pos t co nnector.
72
Fig. 18. Loading a model fence- end.
The height of fastening the compression member to the end post
was taken as 4 feet full size, or 2 feet to the scale used.
The post connector shown in fig. 17 was used in making all com·
pression member fastenings.
Earth was eliminated because of difficulty in securing uniformit~ ·
from one test to anothe.r. Dry sand confined against flowing by
partitions or bearing plates was tried, but the re~ults were indifferent. The sand would pack and slip at irregular intervals, and
the end post would not pull out uniformly. Also the sand was hard
to handle and pack around the base of the posts.
Sand moistened to varying degrees was found to be the most satisfactory testing medium. The moisture content of the sand finally
used was approximately 3% percent, or just enough to leave a
slight trace of moisture when pressed firmly in the hand.
In the box (fig. 18) in which the specimens were set, the sand
was confined against failure in lateral shear by means of two partitions placed one on either side of the end post and parallel to
the specimen. These partitions served as bearing plates, transmitting the force. equally over the entire cross-section of the sand.
The load was applied by means of a Buffalo testing scale. A
3fs-inch wire rope cable was attached to the machine and run over
two ball bearing pulleys to give the proper height of load application. The pulley on the bottom of the testing machine was anchored
to the floor.
The vertical movement of the top of the end post was observed
by means of a Dumpy level and °a scale. The load was measured
by the testing machine. and the horizontal movement of the end post
73
L
'·1
'\,{PE 1
c.poo
L::: -o'-c."
o
o
o
A
H
~ 500
..J
[
~
o)
\}SOO f--4'-----+--~~--_+_---C>jk>-f__-_011!1~--___li_-___l
Co
0
I
e.
'Dt:.FOR\..AA'\ON- \ N5.
Fig, 19 .
Performane~
of small-scale end construction. Type 1.
74
L
,---
J_
<II
III
! /lS h!'
I
V
:!
P=>,7-
~~III
'//.= FI"/
~
III
~ \
'0
.'
(\J
TYPE
2
3 ,000
.
•
~
0
2 ,500
0
.v
:/
2 ,000
~
"
0
~: " /"" 7""
!f "/v"
0
00
1,500
I
~
1,000
.Ii
Ql
.J
I'
500
I--H
L = 4'-,,"
H ~ 2'- 6"
[1
0
<!
0
.J
~"J tzL=
a
1,500
Vo'riO
""
o
o
0
"~ ~v
"
M
/"
t1 v" ZH
"
'7
6>
Nty
' - 6"
'-6"
L H =
"
0
0
0
v.
"/
I~ oVo
I R?" of 1/KH
11 V
/" V.
V
f..- V
0
~"
0
A
~
0
~
0
~
p....v
"
~ H
0
a
~ v"
""
T""
00
0>.
0
500
1<.4
0
0
1.000
"
"
"lJ6
o •
<R
~
v
~
~
[?
0
I
Fi g . 20 .
2
L
""
L = 4'-6"
H =3'-0"
V
0
I
3
DEFORMATION
2
-
=:
!SI-b ll
He 3'-0"
:3
5
INS ,
P e dorman ce of small- sca l e end co nstru c ti on. T y pe 2.
was assured by means of th e strain gauge built into the machine. The
sand was tempered and mixed thoroughly at the beginning of each
day, and from time to time additional moisture was added to compensate for the drying caused by handling after each test. Each
6-inch layer of sand was tamped as firmly as was possible with
hand tampers. The slack was taken out of the cable by applying an
initial load of 50 Ibs. The load was applied as uniformly as possible, and the read ' ags for load, horizontal movement and vertical
movement were t8' en when the balance beam on the testing machine
75
first raised for the desired increment of load. No attempt was made
to keep the beam balanced during a set of readings.
Failure was considered as that point at which the specimen continued to move without an addition of load. This point was well
defined for most of the arrangements.
The data for these tests are presented in figs. 19 to 24, inclusive.
Heavy arrows indicate points of applications of loads. Each curve
represents the average of several tests, H being horizontal and V the
vertical movement of the end post.
These te.sts show rather conclusively the effect of the arrangement of the structural members on the vertical force on the end post.
The conditions as they were set up in the laboratory allowed failure
to occur only as a result of the vertical force on the end post. Be·
/ /_-
;::.
I
a 'j
.1
[\J
~~oo.----.-----'-----'-----r-----r----~~--~----~--~~
6
:2.F->OO f-----j-9~<>_+-----jHl/al~--+--__,~--+-__l>_<~--+_-__l
If)
cO
.J
I
cpoof-~~---4~~-+--+~~~--~~-&-+_--+_-__l
o
«
Sl,~O OI-<!r.----j--9l~---+--_+<rIs>_-~---<>_IV_+~.:..-.+_--+_-__l
OL-__
~
__
o
~
____
~
_____ L_ _ _ _
~
_ _ _ _~_ _ _ _~_ _ _ _L __ _~
o
Dc.F"ORM"'-"TION- INS . .
{
Fig. 21.
J.
P e rformance of sma ll- sca le end constructi c -.
Type 3.
76
cause of the holding medium used and the confinement against
failure in bearing and lateral shear, these same results would not
be expected in tests under actual field conditions.
In every case, the crossed brace arrangements held the load well
up to a certain point, and then would jump out of the sand.
The arrangements with braces fastened at a greater height on the
brace post than on the end post all twisted considerably more than
did the other arrangements as is shown by their relatively large
horizontal movements.
The ends with point of load application above. the j unction be·
tween the brace post and the compression member (type 1 ) tend
to rotate clockwise about the. junction. This provides a lifting force
2,000
1,500
0°; X ~
""
500
o
r
>0
1,000
.-t'
11.~ •
"It>
fT
.
•
0
~-
M
(IJ
I--H
w
'0
L
I
-///~~
II,
.'
~
~=III
'_ b
L~
1/
o~o
I
o
o
.J
0
500
0
of
(
tF.
Q
0
.J.
V
;
It>
V. ~
~
VA
"
IbP
III
' /
',j" "
_ ~"
L=4
0
J;
.~
I?
•
lY.
•
I llH
t v/!
••
~H
~
0;10
•
0
r-V
0
:(0 7
of
0
V
0
~
TYPE 4
~
0/
Cl
4:2.500
1,000
~
"
Ifi 3
ro
,ooo
1,500
/II=IH"'(
fir
OJ
.J
2.000
I
'-6"
L=
y
I
2
3
a
I
3
DEFORMATION - IN S.
Fig. 22 .
Performance of small-scale e nd co ns tru ction. Type 4.
4
5
77
~,ooo r----'----.---'---'----r---'----.----r----'
c,50or---r~~~--+--M-~--+---~-~---+--~
!Ii
~epoO+--4~~~~~~~~~--~-~+--~~~~--~
..J
I
~\)500t--""'-+----"-¥--6----1---+---+<Ii"--+_~*---f---l
o
..J
,ENo;:,\ON
M'\':.~'OE.~
1>-.\
'&
:,
40
\
C.
OET"ORMP--,\ON- INo;:, .
Fi g. 23 .
P e rforman ce of small -scale end con stru c tion. T ype 5.
on the end post which probably caused the. posts to jump from the
sand. The ends with the load applied below the junction (types
2, 3, 4 and 6) would tend to rotate in the other direction, holding
the end post down.
Span length was the most important factor affecting the. holding
power.
The cross-braced assemblies moved less horizontally than Other
types, but for a given span length carried less load. Inclining the
brace post of a horizontally braced assembly apparently increased
its holding power.
78
o
.'rJ
%:
:;;:/,I,-~
"'0
--0':
\\1
~
TY P E <0
c.,oOO
:c:p0 ol---+--...;....q....-=--I>---<lr7-+--+-----+--+--I---l
III
cP
.J
J \o004----~-----+~4-~----4-----+-----+_~-~~~~~~
D'
''E.NS\ON
J.
lv\'E.M'C t:.'R.
A., A.
o \,ooohl----1----,~-+----l---+_----1--_I__I_-_J.l--+--__l
.J
~
4
~
0
DEFO R lv\J:>..T \ON - \ N5.
Fig. 24.
P e rf o rm an('~
of sma ll· sca l e end construction. Type 6.
FIELD TESTS- SERIES 1
Tests we.re continued in the field under conditions comparable to
those experienced in actual service. Full size spe.cimens were used
and designed to secure further information on combinations which
had shown up well in the sand tests using scale models. Two series ,
of tests, spaced about a year apart, were made by different operators.
The plot of ground used was selected primarily with uniformity
of soil texture and moisture in mind. Tests were timed for operation
under as nearly uniform conditiuns as possible. Soil texture and
moisture content are. important factors affecting the behavior of
fence ends. Thus far the tests have been limited to only one soil
type and to one moisture content in an attempt to get comparable
tests of a large. number of assemblies. Soil moisture presents a serio
ous problem. ReferenGe to Seiler's analysis (fig. 15) and to photographs to be discussed (fig. 27 parts A, B, C, D) later shows that the
amount of soil move.d by an overturning post depends to a very
large degree upon the. moisture in the soil. The moisture content
usually varies from the surface to that at a level with the bottom
of the posts. Disturbing changes can take place in the midst of a se-
79
0" PIPE FILLED WITH
CONCRETE
L AYOUT FOR
T EST ING COR N ERS
TRACTOR AND WEIGHTS TOTALED
APPI20XIMATELY BOOO POUNDS
I-BEAM EVENER WA5 TWO'" 16.5
BE.AMS WE.LDED TOGETHER
SIMPLEX
JACK
CATERP1LLAJ;2
TRACTO'i2
WITH WEIGHTS
FOR ANCHOR
LAYOUT FOR TESTING E ND:::
Fi g. 25.
A ppara tu s and l ayou t used in testin g end and co rn e r cons truc ti o ns.
80
ries of tests. On a hot day, the surface soil lose.s moisture rapidly.
A series of tests may be completely upset by an unexpecled rain.
The test plot consisted of a clay loam soil to a depth of about 29
inches. Below this was a thin laye.r of clay changing at a depth
of 34 inches to 36 inches to a gravelly, sandy clay. Analysis of the
clay)oam showed hygroscopic moisture content 7.8 percent, liquid
limit 52.8 percent, plastic limit 31.9 percent and plasticity in·
dex 20.9. Hogentogler (6) makes the following statement about 'this
type of soil: "A clay loam is a fine·textured soil which breaks into
clods or lumps which are hard when dry. When the moist soil is
pinched between the thumb and finger, it will form a thin ribbon
which will break readily, barely sustaining its own weight. The
moist soil is plastic and will form a cast which will bear much
handling. When kneaded in the hand, it does not crumble readily
but tends to work into a heavy, compact mass." The moisture con·
tent of the soil was checked frequently throughout the tests in an
e.ffort to detect any change which might affect the results of the
tests.
A sketch of the apparatus used is shown in fig. 25. The various
assemblies to be tested were set in a circle the center of which was
a large anchor post (fig. 26 ) used in the pulling. The. anchor post
. was made from a heavy 6·inch steel pipe 9 feet in length. It was
set 6 feet in the ground in concrete 18 inches in diameter and fille.d
with concrete. A %.inch wire rope cable was use.d between the
I·beam e.vener and the dynamometer and between the dynamometer
and the anchor post. The cable was fastened to the anchor post
with a heavy log chain.
The objectives of this series were to determine the relationship of
diame.ter of end and brace posts and length of span to strength of
various ends. The effect of an anchor and of the depth of set of the
end post was also studied.
Unless otherwise stated, the conditions were as follows: The posts
were drive.n in holes bored to size, to reduce the complications resulting from variable tamping necessary when large holes are
bored. The posts were tapered approximately 1;4 inch from the
groundline. to the bottom to provide a snug fit. All the joints were
made with the special connector shown in fig. 17. The tension
member was either 2 itT -inch steel rods or a 4-inch diameter post
fastened with the connectors, both of which had ample. strength.
The moisture content of the soil ranged from 24.26 percent to
28.23 percent during the testing period. This was below the plastic
limit of the soil.
Failure was considered as that point after which no additional
load could be applied. Figure 27 shows several end posts raised out
of the ground after failure.
81
48'- 0"
Fi g . 26.
Arrangement of end s around anchor pOSI.
The specimens were loaded by means of Simplex push-pull jacks
as shown in the ske.tch. An initial load of approximately 300 pounds
was used to take up the slack in the testing apparatus and the
specimen. The increment of load was taken as 100 pounds dynamometer reading, the actual load depending upon the constant of the.
dynamometer mechanism. The horizontal movement of the top and
bottom of the end post was observed by counting the number of
turns of the. jack in each line. A plumb bob was used to keep the
I-beam evener in a vertical position at all times. Thus if the top
of the post moved further than did the bottom, a greater nUlI1ber
of turns of the top jack would be required to keep the e.vener in a
vertical position. The vertical movements of the top and bottom of
the end post were observed by means of a Dumpy le.vel and scale
reading .2 inch. The movement of the post after failure. started
was not recorded. The earth was remove.d from around the base of
the end post to determine the extent and type of failure below the
surface of the ground. In most cases the. soil was moved horizon-
82
Fi g. 27.
T ypi ca l soil fai lu res show ing varyi ng ang l es o f rupt ure.
tally in front of the. end post over the entire depth of set, the move·
ment being the greatest at the groundline. A half·conical se.ction of
soil in front of the end post, varying from 6 inches deep in some
cases to as deep as 18 inches in others, was lifted when the. post
was pulled beyond the point of failure. The diameter of this section
of soil at the groundline varied from 12 to 18 inches.
The mass of e.arth ruptured loose as cited by Seiler shows clearly
in fig. 27. In A, the arc between the sides of the ruptured mass is
nearly 180° . In B, C and D the arc becomes progressively less.
In D, the sides of the arc are. parallel (zero degrees). It appears
evident that in A, the diameter of the post, within reasonable limits,
would have little or no effect upon its overturning resistance. In D,
on the other hand, the diameter of the post is the. most important
factor. In the problem of fence end design, this means that post
diameter is of little importance in dry soil but becomes increasingly
important as the soil takes on moisture.
83
In figs. 28-35 inclusive and all similar graphs, reference is made
to the vertical movement V, of the end post and to the horizontal
movement HT and HB, at top and bottom (ground level) of the post,
respectively. The difference between HT and HB is the amount of tip
or rotation which the post experienced. If HT and HB are parallel,
I"
L
?)A..t:0-a
-0
b
.J
~
'if
/.'/.:::1.
-
\9
ih
-";;:;,,'
,"
T,{PE 1
~
'5,000
4,000 1---~--~--I__-=~V::-+----jI_----j--___4---__1
~.000~--~~~-,4-----~--~~~~-+-----+----~----~
L'" 11'- 0"
D,'" ""'''
'Gpo 0
D~-::4Z;i'
1----48----c?-----~---I__M--!l-I__-~I_--=~--___4--__1
L",e<-o"
D == S"
D~~4Y~'
c.-c"x4-·x4" LUGS
FAsTENED TO '&01'
,OM 01'" eND pO'ST
OL-__~____~____L -__- J____~____~__~~__~
o
"C..
0
I
c:..
~
DE.f"OR.M»"'"T\ON -IN5.
Fig. 28.
Perfo rm ance of cro ssed braced arrangement. Type 1.
4
84
the post moves laterally through the soil. If the difference between
HT and HB gradually increases with an increase in load, the. post is
rotating. Characteristics to be desired in a fence end are little or no
vertical movement, parallel movement of HT and HB (both of which
should be small) and high strength at failure. The deformation at a
load of from 2,500 to 3,000 pouuds, the duty of an average fence
end, should be considered as an iwportant index of performance.
The holding power of the crossed brace arrangements (type 1,
figs. 28-29 inclusive) varied from 3,400 pounds to 6,000 pounds;
in most cases the longer spans with larger posts gave superior per·
formance. The addition of lugs to the end post increased its holding
power but did not improve the deformation characteristic.
The. holding power of the horizontal brace arrangements (type 2,
figs. 30·33, inclusive) with a span of 9'·0" and posts set 3'·6" var·
ied from 3,500 pounds for the end with a 9·inch end post which
was set with an oversize auger and tamped, to 6,300 pounds for the
one with 5· and 6·inch posts set in bored holes.
4,000.-----.-=--r:::;...,----,r=--
I'
L
'I
4,oool---,R-+--,ff---l---\,J.---h..lm:-lt-..,.--+---l---1
Fig. 29.
Performance of c rossed bra ced arrange ment. Type 1.
85
Note that for comparable conditions the horizontal brace arrangements had less vertical and more horizontal movement than
the crossed brace arrangements, although the holding power was
superior.
The end with the wire tension member (fig. 30) had an excessive
amount of horizontal movement, the wire failing at 5,000 pounds.
'I
<0,000
5.0001---+l---+-~~----,,f!---~--+----:""''-----+~--I
.2.,00 0
I--l--d-__--__-__..jl---__..j+-~~--_+
---l----f-----l
If)
,/l
.l
I
~
3
o~---+----~--~--_44---+_--_+~--~~~--~
~pool--..J-I-~~-I---I-----..j~-4-. -l.........,£~=----+---+--
L~ ~'-O"
D,=
<;;"
.,
p ...= 4-'/....
"TEM5tON
C. '5"\R.I'\ND'5
Ol____
~
_____L_ _ _ _L__ _
o
~
~
_ _ _ __L_ _ _ _
~~~
0
\
C.
t>\:.'FO'R. ........ """\ION- INS.
Fig. 3lJ,
MEM\!J~R
NO.~W\RE.
____
~
Performance of horizontal braced arrangement.
~
__
~
86
r..-__.-=L'------_~
b
r-
~
II -
If/ ..c//I
~
I~
/I
~
7,000
Iii
--#l~
jl
"T,{PE G
I
;n
'----
c,POO
'3,000
4,000
-:',000
2,000
c!l
~
1,000
1
0
<[
0
:1
o
t-----
jV
o
0
=;
~
K~
O
J:
/;V
~
t~
5,000
;7
L= ,'--0..
D,= c,"
D-a=S"
}
It.
~k,
II
-
L=9'-- 0"
<:0"
D~"S"
0,=
'(u
'0,-=
o
~
H,
v~
L'=' \ \'-0"
1,000
/
V
\I\RRANCiL\,,\I::N,
NO, PU\..U:'D
"\0 RUP"\URI:.)
<:'..0 00
~
~
H
-:';'=)00
.
,: I
{/
- k,
I j:
1)
IL
JJ
4iJOO
~
~,,-
"
~~
/f j I
vy
6
V
0
.J
~
0
0
o
.sDI "-""Dc.
DI A coO,
..s_
J
V
Go
D2.= S"
V-
fI
~6
~
/
-
'-LKT
--L =
0'-0"
D,= Sl"
D~= 4- 'Ii.'
P 05,
s,,::r
VV\\H
OVERSIZE A.U<5E.R
t\:
o
~
Fig . 3 1.
,,,-\,APED
40
\
C.
Dl:.rORI--AI>.' \ON- INS.
Perfor manc e of horizont a l braced arra ngement.
87
A deep set end post (fig. 32) permitted a greater load but did
not exhibit superior performance at lower loads.
An anchor or dead man (fig. 33) increased the strength and de·
creased the deformatior,.
Lowering the brace (fig. 33) changed the performance character·
istics but did not improve them.
The incline.d brace post arrangements (type 3, fig. 34) showed
no general superiority.
Neither of the double span arrangements (type 2·1 and 2·2, fig.
35) could be pulled to failure. However, type 2·2 deformed much
more than type 2-1, thus indicating inferior performance.
FIELD TESTS-SERIES II
The object of this series was to study the holding power of the
double fence end, to determine the loads carried by the bracing
members, and to study the performance of arrangements when set
up as corners.
The method of end construction was simplified somewhat to provide designs which could be constructed more easily on the farm.
All the posts were set in oversize holes and tamped unless otherwise stated.
The compression members were fastened to the posts with Teco
toothed ring connectors (fig. 36). Unless othe.rwise stated, all tension members were two double strands (4 wires total) of No.9
wire twisted. The rod used for tWIsting the wire was left in the wire
and hooked over the compression brace to keep the wire from untwisting. The horizontal movement was measured from a reference
point established behind the end post.
The method for testing the corner constructions is shown in fig.
25. Measurements were made as for ends. The load on the tension
and compression members was measured by a dynamometer and
two calibrated compression springs (fig. 37). The moisture content
of the soil varied from 13 to 28 percent during the series of tests.
The se.ries of tests with single ends and corners (figs. 38-41) was
made to observe the performance of the single arrangement when
used on a corner, to check the effect of method of set upon strength,
and to determine the load on the individual brace members. The superior performance of the end set in holes bored to size when tested
at once is shown conclusively in fig. 38. Note that the arrangement
had comparable performance chamcteristics when used as a corner
(fig. 39). The load on the individual members is shown graphically in fig . 39. Additional member load observations were made
on a longer span (fig. 40). The longer span, tested in a corner
assembly (fig. 41), was found to be better than the shorter span.
88
Tests of double span ends and corners (figs. 42·48) were made
to check the, results of similar ends tested in series 1 and to secure
data on the effect of depth of set. The center post twisting out of
line was the major cause of failure. The end set 2% feet deep
(fig. 43) was definitely inferior to the comparable end with 3%·foot
set (fig. 51). The results of three tests of a short double span cor·
ner are shown in figs. 44, 45 and 46. Some difficulty was encountered
L
~I
\9
D\A.~Dt.
.1
r<>
111=
'0
,=
,II
4 -
-
,I
I
/I/= lfl':
\'{PE "2..
tfl
{l
,It
1 33
",
8.000
7.000
G,OOO
ui
'=>,000
cD
..J
1
0
4.
4 pOO
0
..J
~,oOO
AR'?/>..'N.G.EM~NT
N
\"
'?UU.. E D \" 0 RUPT
RE
c.,OOO
L= £>'- 0"
D\::: c;:,"
"
D'2.= 4 1/'2.
I ,00 O~~-f----+----+-----+---=--t---t---i
OL-__~____~__~____-L__~L-__-+__~
c...
a
~
4-
'::>
D'C.rORMA\"\ON-IN'5
Fig . 32.
P e rformance of horizontal braced arrangement.
<Q
('
89
due to equipment failure. The variation in these results is believed
due in the main to variation in soil structure and soil moisture.
Figure 47 which has no tension member in the second span is lit·
tle bette.r than a single end. That portion of the load transmitted
by the second compression member must be taken entirely by the end
brace post which, having no bracing, can withstand little load.
When the first tension member is absent (fig. 48), part of the load
is taken by the post bearing on the soil, the other part being trans·
mitted to the top of the second span. Consequently, the load on the
se.cond span is less than on a single similar span. The results demo
onstrate the importance of the tension members, especially in the
second span.
The comparative performance of tension members made of two
L
<D,OOOI----t---J'+-_Y---+---I----t--_t_--__-___l
,jJ
'5,000
to
.J
I
o
4,OOOI--~~--+--_r--+_~-~-~--_t_--_I_-___l
~
g~,000
L"" 8>'-0"
'0,:. S"
'Ii.'
Dz.= 4
G,OOOJ-r----t--=:....s..---r'--=-_r---I:--a-,f___.t.+----t--=--+--_1_-___l
CO""PR~SS\ON Mt:'Iv\~t:.'R
LOWL~tD
, 0 'C.
\ ,0 00 H'---t--_t_--_I_--ff'----t 'NO
p., NC 't-\ O~,
O L __ _~_ _~_ _ _ _~_ _~_ _ _ _~_ _- L_ _ _ _L __ __ L_ _~
o
G
~
0
\
c:.
'O'C:.'fORMA.' ION. - \ NS.
Fig. 33. Performan ce of hori2.0ntal braced arrangement.
~
<1.
~
90
gr
1/,
1"
~!lI-
L
I
1
~;~A.~D~
c.-';.Z'RO'0S t ":J
D\"~
:;./'
II =
-'''~'~'~
~
"-
~
i.9
ill
,Y ? E "0
0
/
--,...v
4.00 0
.<'-po 0
<fl
cO
.-1\ 00 0
I
'
2
o
0
)
0
1 It
'M,
)
:/;
[7
\,00 0
0
7
~
V
iv
-= ~ '
D~ = 4 1/G"
D
./
v
lI:'/
1
~Y
~ ~'r-\'
o
L=- !- o "
D ,= co"
D2.~ "::J'
c.
J
L~ ~ '-O"
D,,::: ~ "
D~=: 41/ Z:'
~
VG
V~.... . /~
II ~
vI
t<-'rI,
~
0
\.. '" e> '-0'
G"
0, =
D ,,=
I
D~F O ">!..\-J\ ,b..\ \ Oh.\-
Fig. 34.
V
/- iLHT
~ ..... ~
~
fll
v
v
./
r;'r-\y,
)
V.;< V
L "' ( ~O"
) ~/
~
Ij
~v
~b
k
0/
3,00 0
c,oo 0
0/
v
I J
spo0
c:.
~
\NS.
Performance of inclined brace post arrangements.
t;:,"
~
91
10.000
"T'<P~ -~-\
COM. PR't.,::>o;:,ION
ML\J\e.t:.R !>.;:x J>...
'3,000
9
7,000
~l
0
I
epOO
L5V
H~
0
.Ii
dl<a,ooo
j
V
~
J
o«'5.000
9
.t
'I
I
f
Of
' ( .....RR"'~«.I;.\;'\tN'T
A "-Kl'T PU\. \... 't. 0
1.000 'T O RUP'TURt:.)
I
1
H~-z.:
)
I
f
0
I
~l--\,
C6
e.pOO
o
[?V
III
4.000
j
1
I
~
Fig. 35.
I
""TYPE - 2,-<':.
COMPRIC'::>'::> ION
ML~'e~R J>...\ 'D
I
/
r
,/
'r-\-I
/
/
6
J
V
(A.RR .....NCO'<- M \:. N"T
NO, PUU... ED
, 0 RU?'TUR"-')
(I
I
\
~
0
"DU"'ORM,b.., 10N- \NS.
Performance of doub l e span arrangement.
4
~
~
92
and four strands of No.9 wire are demonstrated in fig. 49. The
loads on both tension and compression members are shown in figs.
49, 50 and 5l.
A comparison of the average movements of the three most important types of end constructions at a load of 3,000 pounds, which
was the assumed average load on a fence end, gave the followin g
results:
Length of Construction
8' - 6"
10' - 8"
16' - 6"
single
sin gle
doubl e
Horizontal movement
Vertical movement
51 %
32%
100%
100%
23%
17%
The end shown in fig. 38, set in holes bored to size, was omitted
from this comparison_
A comparison of the average results of the various tests made on
horizontal braced arrangements, eliminating the test on the 8-foot
6-inch single span set in holes bored to size of the posts, gave the
following results: '
Total load carried
Horizontal movement
Vertical movement
100%
100%
100%
214%
54%
43%
146%
81 %
37%
These comparisons serve to manifest the superiority of the double span arrangement over the single span end.
In comparing the methods of setting ends, the single horizontal
braced arrangement was set in holes bored to diameters greater than
the post and tamped in place, and was also driven in holes bot:ed to
the size of the post. A comparison of the load and movements gave
the following results when tested shortly after setting :
Tamped Tests
Total load carried
Horizontal movement
Vertical movement
Driven Tests
100%
100%
100%
169%
84%
• 63%
In comparing the double horizontal braced span when set at
2-foot 6-inch and 3-foot 6-inch depths, the followin g results were
obtained:
21,6 ft. Dep ths
Total load carried
Horizontal movement
Vertical movement
31,6 ft. Depths
189%
45%
57%
100%
100%
100%
A comparison of the movement of the corners at a load of 3,300
pounds, which was the rupture load of the single 8%-foot horizontal
braced arrangement, is as follow!>:
Horizontal movement
Vertical movem ent
8' - 6"
Single
10' - 8"
Single
13' - 8"
Double
100%
100%
66%
84%
18%
9%
93
Fi g. 36.
Connector used for fastening wood braces.
The corner arrangement recommended from these tests is composed of two end arrangements at right angles to each othe.r with
members the same size as those recommended for the end, with the
exception of the corner post which should be 6 inches in diameter.
This arrangement required approximately 4% man-hours to erect.
Experienced fence contractors report 8 man-hours required for an
anchored corner construction.
When the double span arrangement was set 2% feet deep, the
Fig. 37.
Fence end ready for testi ng with appara tu s for measuring bra ce l oads in place.
94
1// ~//;.
~
,I
-=1//;:- ",
=~ /il
,
'4
to
(,000
0,00 O I----l------.tt'~+_--+_--_1_--+__--+____"..~
Sp()()~--~~--+_--+_-~~--4_--4_-__4
. :3,0 0
Of--.l---+_--b.-.joo~=--+---~'-::'=---I---+_-----l
if)
ell
..l
I c.,OOO
9
"'-
1I--'-------,jL.,,;C-- - I - -
(0
0pOS\''S SE'T W\\H ,"
&
•
A.UG'C.'i2..
0
~
'TA.v.?'i:.\).
?OS'T~ 'S\:.\ \"i \-\o\"'t..';:,
'OOR\:.\) , 0
o
~\ZE.
..l
~,()OOI----IHN~----+----+---+---+---l
{,Oc)O~~~~--~---I----+----+---+--~
00~---+----~~~--~~~--~4----~o~---*----=
D't:.~ORMf'>..\\ON-\NS.
Fig . 38.
P e rformance of sing l e e n"ds and co rners .
95
/(1""'1'1-= =1115
I/e-I/~
y~
,I
c=:-
IE.
lj
4
\9
[0
7>,000
4)000 I----+---+~~.f_±+=--h,.._t!!=--+_--_+_--___t
!Ii
d1
J~,o 0 OI----+------,~-+u~+_--I____q
--_+__-___4
I
D
4:
OEPOO
PO:::'\":::' 'SE-r IN
.J
HOL'€.S 'CORY-D
"\0 'S\Z~
c.
~
4-
'5
7
D£'~OR'MA.\ION-IN5.
I
a
I
IpOO c..OOO 0,000 4,000 ~OOO ~ooo 7,000
LO}o,.D ON '~D\Y\\)UAL M~'t-A'O'E~- \..'0-'<:..
Fi g. 39.
P e rformanc e of sing l e e nd s an d corners.
structure folded up and twisted out of line at a very small load when
compared with the constructions set 3% feet deep.
The previously established loadin g of 3,000 pounds was used in
calculatin g the end construction member sizes. The. exact earth pres·
sure on the portion of the post below the ground line could not
be determined, but it was known from the structural analysis that
the pressure could be resolved into a resultant at some point below
the ground. To determine the resultant and its line of action, free·
body diagrams of the end posts were drawn, as shown in figs. 53
and 54. The load was applied uniformly and the brace loads de·
termined from the brace load formulas given in fig. 54, which were
taken from pressure data as plotted in the figures as indicated. Only
the horizontal forces were considered. as the soil friction around the
96
•o
I
7, 0 0 0 r---""""'"""lrt----tt----r--.---++----r------,
6,OOOt-----jf---t--+-+-+""""""7.e-+--+---i
5, OOOr--~r-----~--~~L---~----~--~~--~
4 ,OOOI--+--+--+-BL-f---f---+--+--l
(/)
(i) ,
.J
D
<l:
o
.J
::3,0 0 0 I-----,f-+-=--:~+- -J'I-...6-------t------t------t-------j
1,00 0 tt---f-nr+---r-----t-----t-----t-----~-----l
0~-4~~
o
____L __ _ _ _L __ _ _ _L __ _ _ _L __ _ _ _L __ _
7
5
4
2.
3
DEFORMATION - IN'S .
~
I
o
I
LOAD ON INDIVIDUAL
Fi g. 40.
I
1,000 2,000 3,000 4 ,000
I
5 ,000 6,000
MEMBERS
-
Perfor mance o f sing l e end with member l oads indicated .
~OOO
LBS.
97
I~
"I.
o
Dlf>o.-.:=.(Q"
I
~
='
'19
1/1 E
'II-i
iii
'II=III=-"~
~
.1
Ij/;I!I
ill
fl
4,000
3,000 \ ----+-----t----:#"':.,' F - - - - t - - - - - j - - - - j - - - - i
rj)
60
dl
..J c.pO O f----~~--+--_I_--+_--+_----t---l
I
2
o)
Fi g. 41.
Perfo rman ce of s ingl e co rn e r.
base of the pos't offset the vertical component of the tension member
load. The A and B diagrams of fig. 53 apply to the 8·foot 6·inch
single span arrangement, and the C and D diagrams of the same
figure are connected with the 10·foot 8·inch single span structure.
Figure 54 presents the diagrams for the double horizontal brace type
of arrangement. The magnitude of P' was determined by the summa·
tion of horizontal forces, and its. line of action was obtained hy
taking moments about RB.
The force P" for the middle post in the double span was obtained
in much the same manner as was P'. The diagrams for this post are
given in fig. 54. In making computations on the end brace posts of
all the arrangements, the line of action of the two resultant forces,
P'" and R, acting beneath the grDund line, were obtaine.d from the
material presented by Seiler (10) and shown in fig. 15. The magni.
tudes of these same forces were obtained by taking moments about
one and solving for the other.
The shear and moment diagram!;, shown in figs. 53 and 54, were
drawn after the forces which acted on the various posts were de·
termined. The maximum moment was used to determine the post
98
e'-~"
t..
'"I
D\ "'-.. c,"
:>;<f,
11'5:11/'"
\9
,I
(11
8,000
"f...ND
I:.
I:.
co,OO
u5
dl
)0,000
1
2
0
4 ,000
,.l
c)OOC)HL~--+-----4-----~-----4
Fig. 42.
Performance of doub l e span end and ('orner.
99
e '- ~ ·
'5pOO
4,0 CO I--~=---+'-:'<jIC--+-------i
tri
dl
.J ::'0 0 ~--,r--+---,"--l---f----l
[
D
4.
"}
o <..p 0
OH-----,I-t---__-~I__-_;
..J
,po 0
I - + - - t - - -I__-
- I _ _ -_;
Ol..-_ _J.......:~_l.__ _~_~
a
\
Co
~
4-
DE-FOR. MA..T\CN-\ N S.
Fig. 43.
Pe rformance o f dou bl e span end and corner.
diameters of the three types of arrangement. The formula used in
these calculations was S = McII, where S = allowable stress in
pounds/ inch 2, M = maximum moment, and IIc = section modulus
of the post. The size of the compression braces was determined from
the formulas (8 ) :
p
O.274Ad 2£2 for rectangular members
12
and
p
O.274E'lT2r4 for round members
}2
100
when P = allowable load in pounds
A = area of the cross section in inches
d = dimension, in inche.s, of least side of column
E = modulus of elasticity (1,600,000)
I = unsupported length, in inches, of column
r = radius (% diameter) of round column.
These formulas apply for I ratios greater than K where
D
G;,'-\O"
"11"'
1 "'17TC=rr,,="*1I==~~=rm,==~=wl
f9
'IITI/'=:
II/HO
Wt:'K~
\ M
e,ooo
(,000
CO'K NER
<opoe
r!J
I
o
~
4-POO
IJ
..J
-0,000
c.,CCO
1,000
a
/' ~
v/V
ui
J '=',000
\"\""''O~~
CONNE.C\"OR~
;h
~
--
~
~
A.v.
'rl,
{I
to
6
IXl
Ij
I
a
C.
3
4'5
"D'i:. 'rO~ Mp....\ \ON-\ ~'S.
Fig. 44..
Go
PerfOrm'lDCe of double span end and corner.
U'S"t:.D
JOIN.\"'S
101
K
=
O.64 VE
S
For steel compression members, the formula
12
was used.
S = 18,000 +
18,OOOr2
I
I = length of column in inches
r = radius of gyration in inches
S = allowable unit stress
<0 '-\0"
4"
D\~:=
1'/:- 1/1' :
NO I"Iv.er.'R.
II;:
= ""CONN'C.C\"O"R '5
\9
~
W~ 'R. 'C. U OS\:.t:>
\N JOIN\~
8,OOOr-----r----,~----r_--_,----_,----_,----_,----_,
/I
7)OOOt-==-=.:...:::.;~o.=~, '--;~-I---+-C
-'O
-R
----1,,-\ ~
-~
-+---+------1
(ON't P,(MII.MOMn'tR ;A\U.t»
/I
(6,OOOI---..::::I!"....----:F--If----'''---f-----J-----J-+-+-f---+F-- - j
~ S,OO Ol--I.r.-+-I--I-O--+---+----f---+--k--"?'I-~-+--_I
.J
I
04,000
4
o
...l
3,0 00 H--I--t----1I----I-----1I---.f----1~--f---f--_I
?O£:.T-::' SEI" \
N
T-\C'-'E.~ ~OR.'E.D
"TO SIZE.
o~--~----~--~----~----~--~----~--~
o
<::.
~
C
\
C
'3
4Dl:.FOR\J\/>...\"ION- \NS.
Fig. 45.
Performance of double span end and co rn er.
<0'-\0"
102
<0"-\0"
;eNG "\\\.A\':>L)(.
C.O""''Nt.C\O~~
WE.RE USt:.D
\N JOIN'\S
~,ooo.----.-----r----.----.
8,0CX)~----4---~~----_+--~~
7,000 1---~--.fD--+---+F-=----A'---;A.-l
rfl
cD GpCX)~-~rt---t-"'+-t-------1
.J
J
o4.S,OO 0
1--- I'----+-
-----iH-- --+--
---t
9
4pOO~--I---+I---~----_+----_I
~)OOOI-l-__/___if-----+--_+--__l
Fiw;. 40.
Periormance oi double span e nd and corner.
103
(applicable when ~ratio falls between 100 and 200)
r
The following is a result of the calculations for compression
braces:
Type of Brace
Calculated Allowable Load
(1) 2" steel pipe 11' in length ............................................ 7,6oo lbs.
(2) Ph" steel pipe 9' in length ....................................................5,300 lbs.
(3) 4" x 4"{full size) x 11' wood membeL~ ........................6,400 lbs.
(4) 4" x 4" x 11'·0 (actual 3% x 3%) .................................. ..4,350 lbs.
(5) 4" round wood member 11' in length ....................................5,000 lbs.
(6) 3" round wood member 11' in length .................................... l ,600 Ibs.
The breaking load for a single strand of No. 9 gauge standard
smooth galvanized wire is approximately 1,400 pounds. The allowable stress 'on wire is usually taken from 1/3 to 1/5 of the breaking
strength so the allowable load on a double strand of No. 9 wire
would be approximately 900 pounds. For a double strand of No.
12% gauge barbed wire the allowable. load would be 850 pounds.
The following is a result of the calculations on the end member
sizes:
Minimum Size
Type of Member
8' . 6" Single Span
6"
End post (8' - 0")
Brace post (8' . 0")
4V2"
4"
Compression member (8' - 0")
2 double strands of No. 9
Tension member
10' . 8" Single Span
End post (8' . 0")
6"
Brace post (8' - 0")
5"
5"
Compression member (10' - 2")
Tension member
2, double strands of No. 9
16' . 6" Double Span
End post (8' . 0")
5'"
First brace post (8'
0")
4"
End brace post (8' - 0")
3V2"
First compression member (8' - 0")
4"
Second compression member (8' . 0")
3V2"
Tension members
2 double strands of No.9
post
post
post
wire
post
post
post
wire
post
post
post
post
post
wire
The brace load constants will vary with the conditions of the set
and the soil type. In the calculations on the member sizes, ample
factors of safety we.re chosen to provide for any discrepancies in
the brace loads resulting from changed conditions.
'
The double span arrangement required about' 2% man·hours to
dig the holes and set the structure.
The superiority of the double span over the single span has been
shown. Experience in connection with the tests suggested further possible improvements which have not been explored. Possible improvements are:
1. The double span end fails by buckling (fig. 52) rather than by
movement out of the ground. A continuous compression member ex-
104
tending over both spans and capable of carrying considerable bend- .
ing moment might materially improve its load-carrying capacity_
2. Preliminary tests on small-scale models showed a double span
with cross braces as shown in fig. 9 was definitely superior to the
horizontal brace (figs. 43, 49, 50, 51) if the load were applied
to the middle post of the. assembly. With this construction, any tendency of the middle post to lean in the direction of the applied
force is accompanied by a tendency for the other two to lean backward thus offering still gre.ater resistance toward movement out of
the soil.
~
0
I
~
/1=
~
~
I
en
'"
(,000
cD ,CO
°
END
~
,) 'S,OOO
I
D
.J.4,o00
0
.J
0,000
POSTS 5'E\' W\lH
7"AUG't:'R ~ 'TJ>.M'P't.'U
c..POO
Fig. 47.
Pe rformanc e of doubl e e nd as r ela ted to tension me mb e rs.
105
o
DIA.?G·
)
....s
.r
NO '2lWIRE
III~
1==:'/1 -=
\9
~
epoC)
11:/ / £"
rt ~
III
~
7Jo
7,000
v-
G,oao
/-
L-->V
([)5poa
V
I.-"
~A
[s;-HT
I/'
ca
)
.)
I
D4poO
4:
o
.)~poo
epOo
1,000
o
III.?III.E~
/1.=11/ 2 /1/
;
7
I
'POS\S St:.'T
\N 7"
KOL~ ~ \"~P~O.
l{
I
o
c..
~
<\-
S
O'C..'f"ORMAI ION- \NS.
Fig
48.
Performance of double end as Iel ated to tension members .
106
.o
-'v
III~ I I
11 1::III~r
~£
/1/;:
Ii,
" r
.~
-th
o
10,000,---'0---,---0--,.-------,---,-----,------,-----,
o
3,OOOI ---~--_+---~---vL---+---+_--~--_1
o
o
~aoo~---+--~~----+-~--~---+--+--H~--+---
7,0 00 1-----\+----+-------.H------t----+-I-~'+_--hr____1
,"",00 0
1--=--I-F-=------1'----=--F---+-----\t--I'-+--_"'--+---~
til
til
.J
15~()0~~~~-~~~-+----+---~-~~·~~--+---__4
~
o
..J 4. P 0
iENSION
Mt~bE~
O~*'..-.r+--h--.-+_--_+_--_+-l----j'----I-____1 'i' IRST
60
-4-
'i:>OU~U:'
'S? AN
5TRIo.ND
NO.'=> WIR'C. U5ED
f+--I--+r-OR ''E.NSIO,,-\
.. """I:..M'DER (2 iE.STSj
"' '-c. \)OU~\'E.
2iTR.I'JIDS
1H_ _+ No.'2>'NIRE U5ED
FOR ,ENSION
M'i..M'DE.R (. I ,1':S'T)
°O~-~---*G--~3~-~O~-~I,--~G,--~~~-~4
O'\':.'f"ORMA.'\ON-\N"""
I
o
I
I
1,000 2.,000 3.000 4,000
LOAD ON TENSION MEMBER -LB5
Fig. 49.
Pe rforman ce of doubl e end with memb er loa ds indi ca ted.
107
Models were made from steel rod to the scale of 1Yz" = l' - 0".
Vertical posts were y7(i of an inch in diameter and the braces 111 of an
inch in diameter. Loads were applied similar to those in the field
tests. The attempt made first was to determine whether pulling from
the middle post or second brace post would improve the performance
of the double span. It seemed reasonable that the buckling would be
reduced or eliminated by that means. Tests were made in a laboratory box, the posts set to the scale equivalent to 3 feet 6 inches and
the soil tamped thoroughly. When the load was applied to the end
post, failure occurred by the customary buckling experienced in the
8pOO r----,----~----_r----~----r_--~----_r----~--~
IPOO r----;-----+----~----+_----~--~~--~~--~~~
<0.00 0 t----1-----+----+----+-----1---4-~_/i_N__I_J.--_l_--_1
if!
rP..l '::>.ooo r-t--;'I----J~----~~~+_----I+--MrIJ..~_+_----~--___l
J
~ 4,000t-6--,H'-----j-+--+t---+--...,H--J--./;I-----I.----L--I
o..J
~poo ~~~-i-+i~~--1_-+_I~~~~~~-~-~
c.
~
0
\
a.
:,
4-
I
I
s
O'i:.FORlv\A'T\aN-'N~.
o
I
I
I
I
I
I
I
1,000 2,000 3,000
0
1,000 2 ,000 3 ,000 4,000 5,000
LOAD ON INDIVIDUAL
ME.MBE.R - LBS
Fig. 50.
Performance of double end with member loads indicated .
108
'"0
,
"i
/I/E
~
"
(I')
9000
END
8000
7000
"000
If)
oJ
.J 5000
0
<t 4000
0
.J
3000
COMPRESSION
MEMBER,
FIRST SPAN
2000
1000~~~~--~~--~-----4-----4----~
O~L&.--I--~--~--.....l:---~_---J
2
3
4
5
~
o
DEFORMATION - INS.
o
I
I
I
I
\000 2000 3000 4000 5000 ~OOO
LOAD ON INDIVIDUAL MEM BER - LBS.
Fig. 51.
Performance of double end with member l oads indicated.
109
field tests. Loading the middle post or the second brace post (front
post with reference to the pulling mechanism) resulted in a different type of failure but no significant improvement in performance. The entire model moved from the soil as a unit, much like
the behavior of the single span end.
Another double span model was constructed with diagonal braces.
When the load was applied to the
middle post of this assembly the
performance was definitely improved. The laboratory tests made
could 110t be taken as valid proof
of the approximate superiority,
and there has been no opportunity
for making field tests. It would
appear however that the advantage
might be as great as 60 percent.
In practice there would probably
be some objection to stretching the
woven wire to the middle post and
later fastening a short length from
this to the end post. This difficulty
would not present itself with the
barbwire. Perhaps the practical
solution would be to stretch the
woven · wire to the end post as
usual and to stretch the barbwire
to the middle post. Since the barbwire causes the greatest increase
Fi g, 52. Characteri sti c failure of doubl e
sp a n assem bl y.
in load due to tern per a t u r e
changes, the advantages w 0 u I d
still be retained. There is a possible additional advantage due to the
fact that the forces caused by the barb and· woven wire would tend
to oppose each other with reference to the dire.ction of post rotation.
TIME · TEST
The objects of the time test were to compare the structural aspects
of the single and double span under actual fence loads and to set
up criteria by which the action of other end structures might be
predicted. The. assemblies as detailed in figs. 38 and 42 were chosen
because the single span arrangement was a common type observed
in the field while the double span assembly had given the most
satisfactory results in the tests.
The end post of the single span was 6 inches in diameter and the
brace post 5 inches. The end post of the double span was 4 inches
. in diameter and the other two posts only 3 inches.
110
s
~@~
39.
sao. in.-Ibs.
FREE BODY DIAGRAM
SHEAR DIAGRAM
MOMENT DIAGRAM
END POST OF e'-6" SPAN
Rs-Z220'
HT, -1890' -
T.
! .~
20,550 In-It.
Rc1Z07·
P "~874'
0 7 4"
FREE BODY DIAGRAM
SHEAR DIAGRAM MOMENT DIAGRAM
BRACE POST OF 8 '- io' SPAN
Tot-o I
Uniform
LDad <300 0 '
III~
5
46,510 in; lb •.
FREE BODY DIAGRAM
SHEAR DIAGRAM
END POST OF 10'- 8" SPAN
12f1~
R-730"
P"-530·
in.-Ib.
530·
FREE BODY DIAGRAM
SHEAR DIAGRAM MOMENT DIAGRAM
BRACE POST OF 10'- 8' S PAN
Fi g. 53 .
Load . she ar and mom e nt diag rams fo r posts for singl e span end s.
The fence comprised two IS-rod sections of 832-6-11 woven wire
with four strands of barbwire above. The barbwire consisted of two
strands of No. 12% gauge wire and was spaced at 3, 4, 5 and 7-inch
interval s. Line posts consisted of sawed halves made from round
posts 4% to S% inches in diameter and 7 fee t in length.
An anchor post was set halfway between the two ends to furnish
a place to insert the dynamometer f or reading the loads in the
fence and to serve as a fixed end so the individual characteristics
of the assemblies could be studied.
III
Ze>.400\n:\bs.
FRU. BOOY DIAGlU,M SHEJ>..R
EN'C)
-~ ,., eM- i
\:)IA"'~/'..l.J,
pas,
Re;=\"3'::>ou
p '""loeo"
fR\:'\:'
~OD,( DIAGRJ>..l>\
CE.NI'i:.R
loeo"
'5'r\t./>"R \)I/>..<:.>I./>.."-I MO,,",,'C.\ol.'
'OR"C.'L 'PO';:"
D\/>..G~""l.J,
R:B~ r=-so­
H,€,\c.rO·
\20
P'~c.IG"
ill
,,;
R"c..'2lC'."
<:'Iz'''
N
FR'i:.'E. bOD'" DIAGRAM
'5~E/>"R P\AGRAV.
'i:.ND 'OR .......C.~
Fig. 54.
MOMENi DIAGRAlv\
pas,
Load, shea r and momeilt di ag ram s for posts for 16'·6/1 double span e nd s.
The device for inserting a dynamometer in each line. of fence on
either side of the anchor post is shown in detail by fig. 56. The
apparatus makes use of two Simplex push-pull jacks which were
fastened to I-inch rods, and tighte.ned so that the connecting link
could be removed. The dynamometer was bolted in place and the
jacks released, placing the load 011 the dynamometer. The bolt near
the. anchor post was used to draw the fence to the proper tension on
the initial stretch.
The single span end was set on the north side of the plot near a
grove of trees in Webster clay loam soil. The surface soil to a depth
of 10 to 12 inches consisted of a black, silty clay loam grading into
black, rathe.r compact, fine-grained plastic clay which at a depth of
112
PIc.
\9
~
fio?OM F\<5 . 39
\O~
FROM
"I
FROM
8"
F\G.40
.4'0, \"+C
f"1~S .
49 -5\
Fig. 55 .
Brace load formulas from l oad observations.
20 to 24 inches became lighter in color. At a depth of 26 inches a
layer of fine gravel was encountered.
The double span was set on the higher ground. The soil type at
this end was Clarion loam, with the surface soil being dark grayishbrown friable loam, extending to a depth of approximately 12 inches. The subsurface soil, to a depth of 24 inches, was a medium
113
Fig. 56.
Single span. double span and anchor used in the time tests.
brown loam, changing to a yellow silty clay loam and at 42 inches
tontained quite a bit of gravel.
The ends were allowed to settle for a few days before the wire
was stretched. Care was taken to insure a straight fence. Line posts
were set 2% feet deep. Concrete bench marks were set just back
of the posts, deep enough to insure against any possible upheaval
from frost action. A double jack. stretcher was used on the woven
wire and a block and tackle on the barbwire. Two rolls of woven
114
.12
r>
z:
W'
2r-
wW
>W
all.
Ground
Line .:>,
.08
~ ~Brace
j
Il:Brocle HF i9 7t )
..... /
.30
.20
>LL .10
0
2?
Height
Ground Line ....
.04
~~ .00
f-I
ZI
wr
LW
Ww
-
V- -
Hjiqhi -- -
~
Ground
Line"
I -----
Broce}
V
~ Ground
~ Line
.00
2900
2700
cJ)
r
~ 2500
l
::l
~
z
2300
02100
4:
o
..l1900
I
\
\
/"
V "-V
;--
1700
r
VI
/
}
L
\.
\
/~
f--
J
1500
15
15
July
15 15 15 15
&ept.
Nov.
Aug .
Oct
15 15 15 15 15 15
Ju Iy
&ept.
Nov.
Aug.
Oct.
De c .
DATE OF READING
DOUBLE 5PAN END
SINGLE. SPAN END
Fig. 57.
Comparison of load and movement of the tim e tes t end s.
wire were used, one being fastened to the end construction and the
other to the I·beam at the anchor post. The double jack stretcher
was fastened in the middle and the wire pulled taut. A center splice
was made, and before the jacks were released, a hand tool was used
to crimp the wire between the clamps of the stretcher to remove any
slack.
115
In stretching the wire. an effort was made to follow Reynolds' (9)
suggestion of removing one· half the tension curve. This was not
followed through, however, as the load became so much larger than
the 1,600 pounds which had been recommended. The barbwire was
stretched with an initial tension of 250 pounds per strand. When
the wire was fastened in place, the bolt at the anchor post was
tightened until the load was 2,942 pounds. The staples were. left
partially driven so the wire would have room to slide back and
forth.
The two end structures after the test had been set up are. shown
in fig. 55.
Load observations were made at the time movement readings
were taken and during part of the winter when only temperature
data were taken.
.
During the test, observations of load, movement of end post, elongation of tension curves and air temperature were noted.
Most of the observed change in load and end movement (fig. 57)
took place immediately ·following the loading operation. At the end
of this period both ends had twisted slightly out of line and the memo
bers of the double span were showing signs of overstress. Members
of the single span had not shown signs of overstress due to their
larger size.
A number of factors affected the load readings, such as move·
ment of the ends, temperature, wind and sunlight.
The vertical movement indicates that the double span' is superior
to the single span. The double :;pan moved slightly in a vertical
direction during the stretching operation, but since that time has reo
mained level. The single. span end has displayed quite different
characteristics in this mode of failure. During the first two months
most of the movement took place. The lesser vertical movement in
the double span arrangement is due to the advantageous structural
features displayed by this construction. The second span relieves
part of the load carried by the tension member, and only through
this member can vertical uplift be produce.d on the end post.
Therefore a decrease in the tension load naturally decreases the ver·
tical movement.
The load on both fences dropped approximately 20 percent during
the first 24 hours and almost 40 percent in the first month.
In spite of the fact that the single span end moved the greater dis·
tance of the two end constructions, the load remained slightly highe.r
during the first month. This condition can be explained by pointing
out that the single span end move.d farther during the loading op·
eration. If the differences in movement at the ground line taken
between July 13 and Aug. 19 are considered, the double span moved
farther. After Aug. 19 the load reading on the double span reo
116
mained higher than that of the sil'lgle span. The load reading is dedependent not only upon the movement but upon changes in temperature. Peaks occurring in the load curve can be accounted for
by a drop in temperature.
The greatest factor contributing to permanent drop in load prior
to complete failure is horizontal movement of the end post.
During December, January and February a dynamometer was left
in one or both fences at all times, and readings were taken at various periods throughout the day. Temperature and extensometer read. ings were usually taken simultaneously with the load readings. The
extensometer readings were discontinued during the latter part of
January and February. These readings indicated the elongation of
a wire tension curve chosen at random in the fence. From Dec. 17
to Dec. 27 the extensometer was placed on a tension curve of the top
No.9 gauge wire in the double span end fence, from Dec. 28 to Jan. 1
the instrument was placed on one of the No. 11 gauge intermediate
wires of the same fence, and during the time between Jan. 1 and
Jan. 12 the extensometer was placed on the top No.9 gauge wire of
the single span end fence.
Figure 58 presents the elongation of the wire tension curves in the
fences. The dotted line represents the theoretical elongation of a
straight piece of wire for the given changes in temperature. In plotting the elongation of the various wires, the difference in extensometer readings rather than the actual readings was plotted. The
lower graph presents the change in load with the change in temperature.
In studying the action of the tension curves, the two graphs
were considered together to obtain a comprehensive picture of the
importance of the curves. First of all, one must consider that a larger part of the fence consisted of straight wire than was contained
in the curves, and a greater percentage of the wire clamped in the
extensometer was included in the tension curve than when the total length of wire between curves is considered. With these thoughts
in mind an analysis was made. At the higher temperatures, the wire
in the specimen elongates more for a given change in temperature
than would a straight unloaded wire for the same change in temperature, because the wire in the fence is more or less fixed and a
change in temperature increases the load, thus producing a tendency
for the wire to elongate. The tension curve elongates more easily
than does the straight wire. Therefore, the part of the wire contained within the instrument not only elongates for that wire within
the extensometer, but for the straight part outside. Most of the increase in fence tension due to temperature was 'no doubt due to
the four barbwires. Since they contained no tension curve there was
no way to dissipate the shrinkage due to temperature. If the temperature drops from 70° to 20° F. the wire decreases 0.0585 percent
117
70
",0
....
50
,l1.
W
It
:J
f--
....
~O
O~
v
10
\1\
f,>,<>' ~
a--
~
~
.....
' ~
.
o~K Middle wire on fence
..........
,~wi th q o uble Span End
Top No.9 W,reo
-10 I- on fen ce wdh ~ ~
0
........
SIngle Span End
-20
-30
.th Double
Spon End
"
U.
:2:
Top No 9 Wire on fence
Strolg t wire
30
It
ill
-...... i'-<;---....
...... ,
'
40
f-- 20
<t
lJ.I
---
' ....
o
0005
.0010
EXTENSOMETER
.0015
.00 20
READING IN INCHES
80
70 .......
bO
,tL
ill
a:
:J
f--
~
50
40
..............
..........
30
<t
20
(L
10
0:
ill
~
r-.... I
• t
Sine Ie
W 0
f-- -10
"•• ;;-~
""..
o.
• •
Span - '!.../' '-
j
•
~
.
2000
2 100
L OAD
Fig. 58.
/ D oubl e
,(
0..........
~
-20
1900
"
South
" N or th
"
"
•
-.30
\800
"
Deportment
"
~
~.
•
~
o Chotilion Dynamometer in Souih Fence
North
"
•t:J..
...
2200
IN
"
...:....~
Spoc
b
"'" ~
~
2300
2400
~
I"-....
~
2500
2bOO 2700
POUND5
Load and deformation in tension curve as related to temperature.
of its length. But since it is fixed at both ends the shrinking ten·
dency would cause an increase in tension instead, which, for 8 . 12%
gauge wires, would be 970 pounds. If the temperature drop would
manifest itself in fence shrinkage (barbwires), a quarter mile fence
would decrease in length 0.77 feet. Apparently temperature is one
of the agents which contribute to end failure.
A drop in temperature from 70 ° to 20° F. gave an increase for
the entire fence under test of approximately 900 pounds, or 50
percent in the load.
From these tests it appears in five months of observation on
horizontal movement of both ends, approximately 50 percent in
118
the movement came during the. loading operation or within 24 hours
thereafter. In the double span, 90 percent of the total average hor·
izontal movement for the whole time came during the first Il)onth,
while 83 percent .of the movement for the single span came during
the same period. The average. hurizontal movement of the double
span end was approximately 80 percent that of the single span end.
Horizontal movement of the end post is the largest factor con·
tributing to the reduction in fence loads prior to complete failure.
All of the vertical movement on the double span, thus far, occurred
during. the loading operation and the first night, while 50 percent
of the vertical movement on the single. span end occurred dur·
ing the same period. The average vertical movement of the double
span end was 5.3 percent that of the single. span end. The small
vertical movement in the double span arrangement can be attributed
to the reduction of the load in the tension member of the. first span.
The load on both fences dropped approximately 20 percent during
the first 24 hours and approximately 40 percent during the first
month. Tempe,rature changes may increase the load in a tightly
drawn fence 50 percent or more. The load in an 832·6·11 woven
wire with three strands of barbwire should not exceed 3,000 pounds.
119
REFERENCES
1. Aitkenhead, William. Information on cost of farm fence. Private communication. Lafayette, Indiana. 1936.
2.
Allbaugh, L. G. 1929 farm fence survey report on 146 farms in four
county farm business associations. Farm Fence Institute, Chicago. 1929.
3. Floto, Walter M. Problems in the production of farm fencing. Agr.
Engr., 16:480. 1935.
4. Giese, Henry. Farm fence handbook. Republic Steel Corp., Chicago. 1938.
5. Hazen, G. L. Analysis and test of fence end construction. Unpublished
thesis_ Library, Iowa State College, Ames, Iowa. 1939.
6. Hogentogler, C. A. Engineering properties of soil. 1st ed., p. 37, 44_
McGraw Hill Book Co., 1937.
7. Miller, R. C. An engineering viewpoint on farm fen cing as an investment.
Agr. Engr., 16:479. 1935.
8. National Lumber Manufacturers Ass'n. Wood structural design data, 2d
ed. 2d printing rev., 1:235-2 37A. 1941.
9. Reynolds, F.
J. A demonstration of better fencing. Agr. Engr., 19: 121. 1938.
10. Seiler, J. F. Effect of depth of embedment on pole stability. Wood Preserving News, 10: 152. 1932.
11. Strong, M. D. Wood post end and corner construction for wire fences.
Unpublished thesis. Library, Iowa State College, Ames, Iowa. 1940.