ANALYSIS” AND DENGN OF THIRTY FOOT
SINGLE SLOPE WOOD TRUSSES

Thesis for {In Degree of M. S.
MICHEGM STATE UNWER‘SITY
Philip James Mielock

1959

THESIS

 

Universit

 

ANALYSIS AND DESIGN OF THIRTY FOOT

SINGLE SLOPE WOOD TRUSSES

by
Philip James Mielock

AN ABSTRACT

Submitted to the College of Agriculture
Michigan State University of Agriculture and
Applied Science in partial fulfillment of the

requirements for the degree of

MASTER OF SCIENCE

Agricultural Engigeering Department
199

Approved muting f M

ii

The primary objective of this investigation was to

analyze and design single slope trusses suitable for farm
construction. .

A Review of Literature indicated that h-foot truss
spacings were the most economical from the standpoint of
lumber and materials. This spacing also allows the use of
lighter trusses that do not require Special equipment to lift
the truss in place.

Stress diagrams were drawn for various possible truss
designs. With the aid of the stress diagrams, two truss
designs were selected for further analysis. The designs
selected could be fabricated with either glue-nail or ring-
bolt fasteners. Both designs selected made possible the
construction with 2" x.h" and 2" x 6" lumber.

The theoretical deflections of the panel points of each
truss were calculated by the use of Williot-Mohr Diagrams.
Theoretical deflections were based on the design load of
30 psf of roof with trusses on h-foot centers.

In order to compare the actual trusses with the theo-
retical analysis, three trusses of each type fastening and
design were tested. There were a total of thirteen trusses
tested. These trusses were constructed of commercial Douglas
fir, No. 2 or better. Two and one-half inch split-rings and
5/8" exterior-grade plywood were used for the two types of
fasteners.

The trusses were tested in the horizontal position, as

this allowed for easier means<fifcontrolling lateral movement.

111
The test method consisted of loading the top chord of the
trusses by use of 2-inch hydraulic cylinders placed 2 feet
apart. Simulated wall-reaction plates served to hold the
truss in place while a motor-driven pump applied pressure to
the cylinders. Amos dial gauges were used to measure the
deflection of points on the trusses. Two or three loadings
were made on each truss before it was taken to the failure

point.

The test results indicate a greater difference between
the types of fasteners than between the two designs. The
glue-nail trusses for both designs compared favorably with the
theoretical deflections. The ring-bolt trusses deflected
considerably more than either the glue-nail or the calculated
values. When maximum loads were reduced to simulate a 2-months

loading, all the trusses produced load factors of safety

greater than one. Reduced to 2-months load equivalent, the
ring-bolt trusses had load factors of safety of 1.12 and 1.57.
The glue-nail trusses had load factors of safety of 1.85 and

1.93.

The conclusions drawn from.this investigation were:

1. From.the standpoint of strength and economy when.a
light framing construction grade is used, both designs analyzed
are suitable for farm construction.

2. The increased strength of the glue-nail trusses can

be utilized further by increasing the truss spacing.

iv

3. Both designs are easily 'fabricated and assembly

time is held to a minimum, thereby reducing the need for
highly experienced erectors.

A. These basic truss designs, when employing glue-nail

construction, lend themselves to spans of more than 30 feet,

since stresses are evenly distributed in trusses.

ANALYSIS AND DESIGN OF THIRTY FOOT
SINGLE SLOPE WOOD TRUSSES

by
Philip James Mielock

A THESIS

Submitted to the College of Agriculture
Enchigan State University of Agriculture and
Applied Science in partial fulfillment of the

requirements for the degree of

MASTER OF SCIENCE

Agricultural Ehgigeering Department
19 9

ACKNOWLEDGEMENTS

I would like to express my sincere gratitude to
Dr. Merle L. Esmay of the Agricultural Engineering Department
for his guidance and many helpful suggestions during the

course of this study.

I wish to thank Dr. A.w. Farrell, Head of the Agricultural
Engineering Department, for making possible the research

assistantship for this study.

Thanks are given to the West Coast Lumberman's Association,
which provided the Agricultural Engineering Department with

funds for this study.

I also wish to thank Dr. James S. Boyd, Agricultural
Engineering Department, and Dr. Lawrence E. Malvern, Applied
Mechanics Department, for their help in preparing the

manuscript.

Appreciation is expressed to Delores, my wife, for her

interest and help in the manuscript preparation.

ii

TABLE OF CONTENTS

INTRODUCTION 0 O O O O O O O O O 0 O O 0 O O 0 O O 0
REVIEW OF LITERATURE. . o o 0 o g . a o o o o o o .

Stresses in Wood. . . .
Design Loads for Roofs.
Fastening of Wood . . .
Truss Testing . . . . .

TIE INVESTIGATION Q 0 O O O O O O O O O O O O O O 0
Preliminary Considerations . . . . . . . . . .

Modulus of Elasticity . . . . . . . . . .
Design Stress of Lumber . . . . . . . . .

Theoretical Analij-s . o o c o o o o o o o o 0

Selection of Truss Design. . . .
Roof Design Loads and Loading Points .
Computing Axial Stresses. . . . . . .
Glue Areas for Joints . . . . . . .
Ring-bolt Design and Location .
Calculating Change in Length of Members

Calculating Panel Point Deflection. .
Comparison of the Two Truss Designs .

EXPERIMENTAL INVESTIGATION. . . . . . . . . . . . .
Apparatus O O O O O O O O O O O O O O O O O O 0

Test Floor. . .
Hydraulic Cylinders and Loading Apparatus

Hold- down Brackets and Rollers. . . . . .
Material and Construction of Trusses. . .

Testing MethodS. . . . . . . . . . . . . . . .
Loading Method. . . . . . . . . . . . . .

Measuring Deflections . . . . .
Plotting Deflections of Test Results. . .

iii

TABLE OF CONTENTS--Continued

RESULTS AND DISCUSSION. . . . . . . . . . . .
Number of Trusses Tested . . . . . . . .
Comparison of Trusses. . . . . . . . . .
Method of Failure. . . . . . . . . . . .
Maximum Loads and Load Factor of Safety.

CONCLUSIONS . . . . . . . . . . . . . . . . .

SUMMARY . . . . . . . . . . . . . . . . . . .

SUGGESTIONS FOR FURTHER STUDY . . . . . . . .

BIBLIOGMPWO O O O O O O O O O O O O O O O 0

iv

Page
AS
LLS
L15
Sh
56
58
60
62

63

Figure
l.

3.
1+.
5.
6.
7.
8.

10.
11.
12.
13.
11L.
15.
16.
17.

18.

LIST OF FIGURES

Frequency diagram showing the range in bending
strength values in small specimens of clear, .
green Douglas fir. . . . . . . . . . . . . . .

Relation of basic stress to duration of. . . .
maXimllm 1 0 ad 0 O O O O O O O O O O I I O O O I

Truss "A" - Glue-nail fastening details. . . .
Truss "A" - Ring-bolt fastening details. . . .
Truss "B" - Glue-nail fastening details. . . .
Truss "B" - Ring-bolt fastening details. . . .
Completed trusses of type "A" and "B" designs.

Stress diagram of truss "A" - live plus dead .

load 0 O O O O O O O O O O O O O O C O O O O 0

Stress diagram of truss "B" - live plus dead .

load 0 O O O O O O O O C O O O O O O O O O O 0

Stress diagram of truss "A" - wind load. . . .
Stress diagram of truss "B" - wind load. . . .
Williot-Mohr Diagrams for truss "A". . . . . .
Williot-Mohr Diagrams for truss "B". . . . . .
Typical failures in ring-bolt trusses. . . . .
Apparatus used for loading trusses . . . . . .
Schematic diagram of hydraulic loading system.
Comparison of theoretical and actual . . . . .
deflection of top center joint - glue-nail . .

trusses. O O O O O O O O O O O O O O O O O O 0

Comparison of theoretical and actual . . . . .
deflections of point "D" - glue-nail trusses .

Page

18
19
2O
21
22

25

26

27
28

33

37
38
ho
117

11,8

LIST OF FIGURES-~Continued

 

Figure Page
19. Theoretical deflection pattern of truss "A". . A9
20. Theoretical deflection pattern of truss "B". . h9

21. Comparison of deflection of points on truss. .
No. 7 - glue-nail construction . . . . . . . . 51

22. Deflection of panel point "C" (top center) - .
ring-bolt trusses. . . . . . . . . . . . . . . 52

23, Deflection of panel point "D" - ring-bolt. . .
trusses. O O C O O C O O O 0 O O O O O O O O 0 53

vi

Table
1.

LIST OF TABLES

Page
Minimum.g1ue contact area required between the
gusset plates and the individual members for .
axial stresses . . . . . . . . . . . . . . . . 30

Data used for calculating the change in length
of truss members . . . . . . . . . . . . . . . 31

Maximum loads and method of truss failure. . . A6

Load factor of safety based on 2 months. . . .
duration of load . . . . . . . . . . . . . . . 57

vii

INTRODUCTION

An ever-increasing number of farm buildings are being
constructed without interior pole or post obstructions. This
type of construction is commonly referred to as clear-span

construction.

Clear-span farm buildings offer many advantages over
buildings that require interior support. Perhaps the most
important, long range advantage of this construction is that
the buildings are readily adapted to other uses. With this
flexibility, a farm operator can take advantage of changing
market demands by switching to another enterprise. Such a
switch in enterprises would not result in costly building
alterations, as it has in the past.

In addition to the long range flexibility, there are
many immediate benefits in clear-Span type construction. All
forms of mechanization within the building can be used with
greater ease and at less cost. Barns employing clear-span
iconstruction can be cleaned easier with tractor loaders.
Endless feeders and waterers, as in modern poultry houses, are
installed with greater ease and with possibly greater choice
of location. Gutter cleaners in barns and large dropping
pit cleaners in poultry houses have greater flexibility for

location.

In storage buildings, the usable area and accessibility

of movement of mechanical equipment are increased. Forced

2
air drying is more adaptable since there is no loss of air
around poles or posts. Light and air distribution also are

increased When interior posts are eliminated.

Clear-span farm buildings, wider than 20 feet, require
trusses to support the roof between wall supports. Although
steel is used in prefabricated trusses, commercial lumber is
a popular material for clear-span farm buildings. Wood is
used because of its low cost, availability, ease of fabrication,
and strength characteristics. There are many forms of
fasteners Ibr these trusses. Zfilgeneral, glue-nail, ring-bolt,
and nails are the most common. For 2h- to hO-foot apans,
the majority of the farm trusses use glue-nail or ring-bolt
fasteners. l

Spans of up to ho feet have been found practical for
farm buildings. Spans of greater than ho feet require more
fabrication and erection skill, as well as a higher cost per
square foot of building.

Spacing of trusses varies from.2 to more than 12 feet in
different plans. The larger spacings require considerably

heavier trusses and much larger roof girts. The spacing of
h feet on center allows the use of 2" x h" girts for metal-
covered buildings. Using h-foot spacing, l" x 6" tongue and
groove material also can be used, if shingles or roll roofing
is desired (1h). Spacing of h feet offers a light weight

truss that is easily fabricated and placed in position without

excess labor or special apparatus. This spacing can be
utilized on any type wall construction. It also offers more

economical use of connectors and lumber (9).

3

The interest in clear-span, gable roofed buildings has
brought up the question of eliminating poles or supports from
wider lean-toe and additions to existing buildings. Demand
for complete buildings with single-sloped roofs is greatly
increasing. Single-slope design can provide a greater amount
of light in buildings. Snow and rain water can be diverted
to one side of a building. In dairy barns, this diversion
relieves a sanitation problem. Cows would not have to walk
through mud or snow at the building entrance.

While many tested plans of gable type trusses are avail—
able for use in farm buildings, there are no available plans
for shed or single-slope type trusses. The lack of and the
need for trusses of single slope design for farm buildings

have been the reasons for this investigation.

EVIEW OF LITERATURE

The factor of safety involved in construction is dependent
upon many things. Homes and buildings used by the public
frequently have very high safety factors because of possible
loss of human life. It is the author's opinion that farm
buildings do not require this same high factor of safety.

Previous tests of wooden trusses indicated that theoretical
design loads were generally greatly exceeded. This indicates
that some factor of safety generally is applied to allowable
stresses given for lumber._ The design loads for roofs
frequently contain another factor of safety as they are

unrealistic loads.

The Review of Literature was, therefore, made in four

parts:
1. Stresses in Wood
2. Design Leads for Roofs
3. Fastening of Wood
h. Truss Testing

Stresses in Wood

Allowable unit stresses are available for structural
grade lumber. According to Johnson (27), the following
principles are carried out in stress grading:

1. Every individual timber must be capable of

safely carrying its full design load.

\n

2. The working stresses must be applicable to all
conditions of use.

3. Timbers must be capable of carrying the full
design load for the life of the structure.

A. It is assumed that workmanlike fabrication and

design are reasonably good.

Jehnson (27) states further that, in order to arrive at
an allowable stress for a species, small clear specimens are
tested. ASTM (Dlh3-52) "Standard Methods of Testing Small
Clear Specimens of Timber" (7)gives the methods for selecting
and testing these samples.

ASTM (D2h5-h9T) "Methods for Establishing Structural

Grades of Lumber" (7) states that the basic stresses assume

normal variability of material, competent design, good fabri-

cation, reliable grading, and adequate supervision. A nominal
factor of safety is then provided that permits an occasional
slight temporary overload.

Jehnson (27) says that twelve factors are taken into
account when the "built in" factor of safety is used. The
factors included are variability of wood, indeterminancy of
stress analysis, oversize resulting from use of standard sizes,
depth factor, efficiency of grading rules, efficiency of
inspection, range of defects within grade, offsize, duration
of load, imperfections in fabrication, temperature, expected
versus actual load, and other conditions of service.

Johnson (27) states further that the factor of safety on

the most probable timber under the most probable service

6
conditions, as shown by analysis, is 2-1/2 to 3 in bending
and shear and 2-l/h in compression. One timber in 100 has a
factor of safety as low as l-l/h.

Figure 1 is a diagram presented by Wood (hi) that shows
the range in bending strength values in small, clear, green
Douglas fir.

Basic stresses resulting from these tests of clear
specimens must be reduced in actual pieces of lumber to allow
variations in factors that affect the strength. The factors
that affect strength are given in the Wood Handbook (A) as
slope of grain, knots, shakes and checks, wane, pitchpockets,
holes, and moisture content. Stress grading takes these
factors into account when grading of individual pieces is
accomplished according to use.

Wood (hl) states that the direct effect of drying of
wood is the stiffening and strengthening of the wood fibers.
In larger timbers, however, this may be accompanied by checking
or splitting that largely offsets the gain in strength of the
wood fibers. Some increase in strength in drying is recognized
in smaller sizes of structural lumber subjected to bending
and in compression members of all sizes.

ASTM (D-ZhS-hOT) (7) states that it is common practice
to take advantage of the benefit of drying, in material four
inches or less in thickness, by increasing permissable sizes
of knots or other characteristics.

Wood can carry very high loads for short periods.
Johnson (27) relates that engineers have unknowingly used

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I60 I
Z I -
:1 | AVERAGE—7,4530 1+4
0 :20 ' '
z | I
“ I
a: I
g I RECOMNLI‘EN o |
80» so wwoaKmIa ST RES" I
o; | I350 I
um I |
*5 l I
0
2" 4° I I
‘g’ ' I J— I
m l ' I
E o I I
2 4 6 3 IO l2
SENDING STRENGTH (I,ooo RSI.)
FIGURE l - FREQUENCY DIAGRAM SHOWING THE

RANGE IN BENDING STRENGTH‘VALUES IN SMALL
SPECIMENS OF CLEAR, GREEN DOUGLAS FIR.

8
that characteristic of wood for years by increasing the
stresses they have used in designing for wind, impacts, and
snow loads. The ASTM (D-2h5-h9T) (7) states that advantage
can be taken of this characteristic in many structural designs.
Figure 2 taken from.ASTM (D-2hS-h9T) (7) shows the relation
of basic stress to duration of maximum load. This same

relationship holds in converting short-time test loads to

longer loading periods.

Desigg Loads for Roofs

Barre and Sammet (1) recommend a minimum. of 20 pounds
per square foot for snow load. In northern areas of the
United States, these figures Should be increased according to
figures by the National Bureau of Standards (33% In.Michigan
these figures vary from 20 pounds per square foot in the

southern part of the State to 35 pounds per square foot at

the extreme northern-part of the Upper Penninsula. These

loads are not specified for farm buildings.
Radcliffe and Granum.(35) state that they used live snow

load of 25 pounds per square foot for trusses for house
design because it conformed to standard practice. They state
further that such loads probably will not occur and lead to
..conservative designs.

Wind pressures on roofs vary with wind direction and
orientation of any opening. Many different investigators
have measured these effects and results show that, for slopes
of less than 20°, the pressures are fairly constant over the

surface.. The National Bureau of Standards (33) gives a force

040... 232332 “.0 29.2130 9-. mmmmkm o.m<m mo Scan—mm N mane—u

 

 

 

 

9 min I 93.. sagas. ao 29:22.
08.... oo. _ 66 _oo .o

. cod. 2 _.o 50.0 o .088

/ oo.

// 0:

/// om.

// on.

 

/ o!
/ 00.
/ ow.
/ / ot
1/ 00.

cm.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CON

 

 

d» hm .m». In. 0.2 . >49. .m I _ .z 2 _ .omm _

BOWVA $83819 OISVB :JO 39V1N3033d

lO

coefficient of -O,6 on the windward side and -O.h5 on the
leeward side for slopes of 20° or less.

Giese and Henderson (2h) state that, because live loads

in farm.buildings can be determined rather closely, the high

factor of safety is not required for farm construction.

Fastening_of Wood

While many types of fasteners for wood are known, this

review deals chiefly with split-ring connectors and casein

and resorcinol resin glues.
The Design Manual for TECO Timber Connectors Construc-

tion (39) gives the design material for split-rings. The
manual explains that where the wood, not strength of the
rings, determines the load capacity the design figures can be
increased. The National Design Specifications for Stress-
Grade Lumber and its Fastenings (3h) gives an allowable
increase in stresses for both lumber and the fastenings for
short time loading. These increases are 15 percent for two
months' duration, as for snow; 25 percent for seven days'
duration; 33-1/3 percent for wind and earthquake; and 100

percent for impact.
Markwardt (29) indicates that one of the chief reasons

that wood is being used more for structural material is
because of the developments in glue and gluing techniques.
Giese (23) states that one of the important reasons for
using glue is that the bearing area of joints may be increased
to almost any size, thus making possible a more uniform

stressing throughout the structure.

11

The glues most commonly used in farm construction are
casein and resorcinol resin. The main difference in these
glues is in the ability to withstand moisture.

Resorcinol resin, first introduced in l9h3 for use in
aircraft and naval vessels, is waterproof. The Forest Product
Laboratory Report Number 1336 (20) states that resorcinol
resins are as durable as the wood itself. The report points
out, however, that manufacturers do not recommend curing
resorcinol resins glued below 70°F.

Casein glues have been in existence for many years and
are continually being improved. Traux (37) says casein glue
produces joints in most of the common species of wood that
are equal to or greater than the strength of wood itself.

McLaughlin (30) states that while a casein glue line is
weakened by the presence of moisture, it will regain its
original strength when again dried out. Kaufert (28) points
out, however, that continuous moisture or alternating moist
and dry conditions eventually will weaken the joint.

Dunsan (13) found that casein joints, when exposed to
bacteria capable of digesting casein, were not materially
weakened when the moisture content of the wood was below
approximately 55 percent.

As to the durability of casein glue, McLaughlin (30)
states that this type of glue has been used in the United
States for 30 years and even longer in Europe with satisfying
results. One of the conclusions of Giese and Henderson (2h)

was that casein glue is of ample durability if the joint is

protected from direct action of water.

12

Review indicates a wide variation in shear stress for
glue. Boyd (9) notes that a unit stress of 200 psi is practical
for farm buildings. Giese and Henderson (211,) recommend 1130 psi
be used as a design stress for casein glue joints loaded
parallel to the wood grain for farm construction with Douglas
fir. A stress of 215 psi is recommended (for joints loaded

perpendicular to the grain.

Truss Testipg

The ASTM (13-73-52 ) (7) indicates that the number of
tests of like trusses depends on the desired accuracy and
reliability of the results jand purpose of the test. The
particle recommends a minimum of three tests for reliable and
accurate results.

The article further states that compression chords, of
the truss should be braced laterally if it does not interfere
with the free vertical deflection of the truss. The lateral
bracing should duplicate as far as possible the effect of the
complete structure.

ASTM ( E-73-52 I (7)allows load increments of 25 percent
of design load up to design load. After design load ‘is
reached, additional increments of 50 percent of design load
may be used. Rate of loading should be as uniform as possible
and each increment should remain on about five minutes so that
the truss may come to rest.

Boyd (9) tested a gable truss intensively with strain

gauges while using small aircraft cylinders placed 2 feet on

13
center to apply the load. This truss was tested inahorizontal
position instead of the common standing position.

Boyd (9) found that secondary stresses caused by plywood
gusset plates in a Fink truss could be neglected when using
standard practices, as the factors of safety and design
procedures make adequate allowances.

While the Williot diagram assumes pinned joints, it is
common to use this method for trusses with all types of joints.
Grinter (2) indicates that this method is widely used when
a graphical layout is preferred. In combination with the
Mchr diagram, the Williot diagram gives the deflection of all
panel points of the truss.

From results observed in testing gable trusses, this
author noted that the heel joint strength could be improved
by use of a wedge between the lower and upper chords in the
ring-bolt truss.

Acre (6) found, when comparing several heeljoint desings,
that a birds mouth cut in the upper chord at the reaction
could increase the strength of the joint over the conventional

method.

THE INVESTIGATION
Preliminary Considerations

Modulus of Elasticity

The Review of Literature indicated moisture content
influences the stiffness and strengthcn?wood. Representative
samples of the lumber used in trusses were, therefore, tested
for modulus of elasticity and moisture content. ’Each piece
was selected to be clear grained and knot free.

The samples were tested with simple center loading. The
reactions and loading piece were rounded hardwood blocks, as
specified for this test in ASTM ( E-73-52 I (7). The span
between reactions was 2 feet.“ SR-h gauges placed on the
samples were used to measure strain.

After testing, small pieces of each sample were placed
in an oven to determine moisture content.

The values for modulus of elasticity obtained ranged
from l.h9 x 106 to 2.55 x 106 psi. Most of these figures
were higher than the value of 1.6 x 106 psi given for modulus
of elasticity for Douglas fir in the Wood Handbook (h). Part
Of the high values for moduli of elasticity can be attributed
to the lumber being very dry at the time of the test. Moisture
contents for the moduli of elasticity samples ranged from h.5
to 7.2 percent. Other moisture samples, not checked for

modulus of elasticity had a range in moisture content from

9.2 to 15.h percent.

15
According to the Wood Handbook (h), the modulus of
elasticity given in design informationis evaluated for longer

time intervals then this test used. Wood has the property of

deflecting gradually over a period of time after a load is
applied. This fact coupled with the low moisture content
could appreciably change the value of modulus of elasticity.
Boyd (9), after testing, used 1.96 x 10° psi to determine
theoretical truss deflections. However, the author's test
showed higher values for the small pieces, larger members
containing defects and flaws would reduce the total modulus
of elasticity. From this consideration, 1.6 x 106 psi was

used for theoretical design and deflections.

Design Stress of Lumber

The lumber recommended<n1Michigan State p1ans.fln*trusses
is Douglas fir, Number 2 or better (or equivalent). This
correSponds to the Construction Light Framing Grade in new
grading system. The single slope truss also would carry this
Specification.

Since the Construction Light Framing Grade does not carry
a stress grade, some allowable stress had to be selected for
design. Barre and Sammet (1) give allowable unit stresses
for stress graded lumber. The stress of 2150 psi is given
for select structural joists and planks of Douglas fir in
tension parallel to grain and for extreme fiber in bending.

Allowable unit stress for compression parallel to grain is

1750 p810

16

The 1957 ammendments to National Design Specifications
for Stress-grade Lumber and its Fastenings (3h) gives ranges
of 1200 to 2050 psi for tension parallel to grain or extreme
fiber in bending, depending on grade of joist and planks. The
values for compression parallel to grain range from.lOOO to
1650 psi, again depending on grades. .

According to the National Design Specifications for
Stress-grade Lumber and its Fastenings (3h), these stresses
can be increased by 15 percent for snow loads, based on two
months of loading.

Considering this allowable increase and realizing that
the best yard-grade lumber may not approach stress-grade
lumber, a stress of 2000 psi in extreme fiber in bending was
used for design. The allowable compressive stress used was
1600 psi. This value used in conjunction with the 30 pounds
per square foot roof load gtill should allow a considerable

factor of safety.

Theoretical Analysis

Selection of Truss Design

Many different 30-foot single slope truss designs are
possible. This investigation is not intended to cover all
the variations.' The basis for consideration in design was
balanced stressing between members, ease of construction,
simplicity of design, ability to use available lengths of limber
without waste, and economy. lhiaddition, designs that allowed

the use of both glue-nail and ring-bolt fasteners were used.

17

Because of the intended use of the trusses, two-point-type
support was selected. Indeterminate, three-point support,
would complicate problems of additionsto existing structureS.
Two-point support also gives more head room for storage.
Frequently a single slope roof covers a hay feeding shed
attached to a hay storage barn. This operation requires more

. room than is possible with a three-point supported truss.

Various possible truss designs were checked with stress
diagrams. The load for each member indicated possibilities
of balanced design.

Some possible trusses were eliminated because fabrication
with rings and bolts was not possible without members crossing
planes.

Two 30-foot trusses were selected from the remaining
designs. Each of these trusses had the top chord supported
in three places, in addition to the and supports. With this
span, the stress due to bending plus the axial stress allowed

the use of 2" x 6" lumber in the top chord.

The two basic designs selected are shown in Figures 3 toé
with both types of fasteners. Figure 7 shows the designs

after they were constructed.

Roof Design Loads and Loading Points

Twenty-five pounds of live load per square foot were used
in design. This, in addition to five pounds dead load, gave
a total load of thirty pounds per square foot of horizontal

roof projection. This is the loading given for most areas of

18

mug—ho 02.25.93. ..=<zum:...o I..<.. mmDm... m mane...

 

 

 

 

 

 

 

 

 

 

 

 

 

mme<4a emmmae
mo... 52545 02:50 .

¢—_._

 

 

 

 

v.

\II

.\
\

 

 

 

 

 

 

 

 

 

 

 

l9

 

 

 

mn__<._.mo 92.2953. ._.I_Om 024 02E 1:4: mmzmh v mmnoi

 

 

 

 

 

 

 

 

 

20

 

m..__<._.mo oz_zm._.m<.... J<Zum346 l..m.. mmam... n maze...

 

 

 

 

 

 

 

 

 

 

mmhdqn. hummsw
mOu 24104.0 GZELDU

 

 

._ u .m. Liom

 

 

 

 

 

 

21

 

 

 

mu:<r_.mo OZ_ZM._.m<u .Zom oz< oz_mu..m.. mmnm... w mmswi

 

 

 

 

 

 

 

 

 

 

 

22

 

 

FIGURE 7b - TRUSS ”A" - RING-BOUT CONSTRUCTION

23

I

‘I ‘I ‘ ‘ ' I VI.‘ I: I- H :I‘ I
IIIl IIIIIIIII ‘8 z .. . ' .
I

_ IIII IIII I: . ‘ . ‘l I'W '
“I! I! I I . . ‘ . .‘ I ‘ ' I I‘ll I
‘ ‘I I}, .‘ “ - ~".‘ ‘ , . I '
. ”I ”I“ II ' ' . ‘ I" . . , I I...
I ‘ Il ‘ "w“ ‘ . '. ‘ ' ‘ vl l I ‘I ll II . ‘ ”MI“
I” “NM ' II Insu'a'. I.
H . <v 7| 1‘ ‘ I ‘ . Il‘ ‘ l: .'
. l "I ‘ ' I ' . ‘ . . IP‘I ‘
. I “ ' ' I H I -
I , I I I I I
F

7.7.1,, III II" 1 :HI' ”'I II .“II'IWI” IIIIIIW IIIIIIW 'I p ' f . .. , II
.. .. I. III , ' . I I . I. . _ .III

I -_.._

M I III III‘ I“. .

 

FIGURE 7d - TRUSS "B" - RING-BOLT CONSTRUCTION

2h
Michigan by the National Bureau of Standards (33). These
design loads are given for all types of building uses. If
allowable stresses given in the Wood Handbook (A) are used in
addition to* these roof loads, the result will be a building
that is rather conservative for farm use. While no information
is available, itisapparent that many new buildings<x1Michigan
farms do not meet these standards.

Theoretical design was based on panel point loading. In
actual practice with metal roof construction, the load is
applied every 2 feet of truss. This would correSpond to
2" x u" girts placed on 2-foot centers. For theoretical
purposes, the additional stress due to bending is small, so
that the asSMption of panel point loading is not in geat error.

The stress diagrams for the design load of 30 psf with

trusses h feet on center are given in Figures 8 and 9.
The wind load on the trusses was based on a force coef-
ficient of -0.5. The wind. velocity pressure was taken as

20 psf. The resultant stress diagrams from.this wind load
are shown in Figures 10 and 11.

Computing Axial Stresses

Axial loads were determined from a stress diagram of
each truss. The design load carried by each truss was based
on 30 pounds per square foot of horizontal projection and
truss spacing of )4. feet. The appropriate standard size lumber
I was selected to fit the leads given by the stress diagram.
In the top chord, the additional stresses due to bending also

were considered in selecting lumber size.

25

MEM LOAD-lbs. . 0
ob -4350
be -4080

cd -38Io /

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

scale: 8L: IOO“

   

 

cf +l380 "‘

 

 

 

 

 

 

 

 

 

 

 

 

FIGURE 8 STRESS DIAGRAM OF TRUSS "A"
LIVE PLUS DEAD LOAD

26

 

MEM. LOAD-lbs

 

  
 
 
 

 

 

 

DH -4300 ID
ci -33SO
dk -3050
el -3390

 

 

 

 

  

 

 

 

 

 

 

 

 

.Ig +4000 . ,c
d
i
1 e
K
f
//// I" #
l Scale=-—=l00
£11109.
hg +4050 450*F
hi -l050 5
i j -780 90° 6 f
1k ~760 w

 

 

 

 

 

 

FIGURE 9 STRESS DIAGRAM OF TRUSS "8"
LIVE PLUS DEAD L0AD

27

 

 

 

 

  

 

scale: Ibs

 

 
 

 

. . fl em 44420
0 200 400 mg -I500

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

FIGURE 10 STRESS DIAGRAM OF TRUSS "A"- WIND LOAD

 

28

 

 

 

Ik +400

 

 

kj +275

 

 

 

 

 

 

 

 

 
  

- scale= lbs.

 

0 200 400

 

 

  

LOAD-lbs

 

+I420

 

+ll85

 

+l|85

 

4- I420

 

 

- I500

 

 

 

|58

 

FIGURE II STRESS DIAGRAM OF TRUSS "B"-WIND I_0AD

29

It was apparent that balanced stressing could bepossible

by using non-standard sizes in the glue-nail truss designs.
However, this would lead to. more labor and to the possibility

of the lumber being wasted, regardless.

Glue Areas for Joints

Joint size was determined chiefly by the allowable shear
stresses for lumber. Since it is hoped that these designs are

practical for farm construction, economy of use and simplicity
also were considered. In laying out the gusset plates, it

became apparent that a I4! 1: Il' sheet of plywood would serve

one truss. The shapes of the gusset plates then were made so
as, to minimize the plywood requirement. The outline diagram
for the gusset plates for truss "B" appears in Figure 5.

Truss "A" actually required more plywood than a M x 14'
section. Here again, this additional plywood was layed out

so as to maximize use and minimize waste. Figure 3 shows the
gusset plate layout for truss "A."

A shear stress of 200 psi was used for the glue line.
Although this stress is highly conservative, it allows for

variation in application that might occur with ineXperienced
fabricators. The actual limiting stresses in the glued Joints

is the shear stress of the lumber and the cross-section shear
of the plywood. The design of gussets was, therefore, based
on a 95 psi shear stress for lmnber and a 200 psi cross-sectional
shear of plywood. Table I gives the minimum glue contact area
required between gusset plates and the individual member for

axial stresses.

30

TABLE I

MINIMUM GLUE 00NTA0T AREA REQUIRED BETWEEN
THE GUSSET PLATES AND THE INDIVIDUAL
MEMBERS FOR AXIAL STRESSES

 

 

 

 

 

 

 

 

 

 

TRUSS "A"

Member Area-In.2 Member Area-In.2
AB u5.8 HA h3.8
BC E3.0 BR 8.95
GD no.1 0H 1%.2
DE 36.0 CG 1 .u
EF A3.8 CF lu.5
PG 29.1 DF 8.95
GB 29.5

TMES"E'
AB ES.3 CG 8.2
30 35.3 OF 8.0
CD 32.1 AG #2.?
DE 35.7 GF A1.0
BG 11.1 FE u2.1
DF 11.2

 

Ringybolt Design and Location

The design manual for TECO Timber Connector Construction (39)
was used for design. Advantage was taken of allowable loads
for snow loads of 2 months' duration. This increase amounts

to 15 percent over the values given inthe manual or 25 percent

over permanent loads.

Because of loads involved, 2" x h" members could be used
in the trusses. Two and one-half inch split-rings were used
throughout the truss since a combination of 2-1/2 and h-inch

rings would complicate fabrication.

31

Ring location and number are shown in Figures u and 6.

Calculating_Change in Length of Members

The total strain in a member was determined by the
equation 3 = 2% where

P = load in pounds on the member

1 : length of member in inches

A : cross-sectional area of the member (in.2)

E : modulus of elasticity (1,600,000 psi)

The change in length of each member of both designs is

given in Table 2.

TABLE 2

DATA USED FOR CALCULATING THE CHANGE
IN LENGTH OF TRUSS MEMBERS

TRUSS "A"*

 

 

Member Length (In.) Area (In.2) ,Load (Lbs.) fifi (Inches)

 

AB 95 9.1a -h350 -0.0283
BC 95 9.1a -u080 -0.0265
DE 95 9.1a -3u20 -0.0222
BH 32 5.89 - 850 -0.00289
DF 32 5.89 — 850 -0.00289
CH 100 5.89 1350 0.01u3
CF 100 5.89 1380 0.01u6
00 63 5.89 ~1750 -0.0117
AH 100 5.89 u160 0.0AM1
HG 100 5.89 2800 0.0297
CF 100 5.89 2760 0.0293
FE 100 5.89 Aleo 0.0AA1

 

* For glue-nail truss

32

TABLE 2--Continued
TRUSS "B"

 

Member Length (In.) Area (In.2) Load (Lbs.) fl (Inches)

BC 95 9.1h -3350 -0.0217
CD 95 9.1L -3050 -0.0198
DE 95 9.1u -3390 -0.0220
BG 53 5.89 -1050 -0.0059
DF 53 5.89 -1060 -0.00596
00 77 5.89 - 780 -0.00637
CF 77 5.89 - 760 -0.00622
AG 13A 5.89 A050 0.0576
GP 127 5.89 3900 0.0525
FE 13A 5.89 A000 0.0570

 

Calculating_Bgnel Point Deflection

Inorder to compare the theoretical design to test results,
the Williot-Mohr diagram was used to determine panel point
deflection. The change in length of each member is considered

in the Williot diagram. This graphical presentation gives

the theoretical horizontal and vertical deflection of each of
the panel points. Figures 12 and 13 show the Williot-Mohr
diagrams and resulting deflections for each truss. Deflections;
are determined by scaling the horizontal and vertical distance
between respective points on the Williot-Mohr diagrams. Two
different Williot-Mohr diagrams for each truss served as a

check on the accuracy.

The Williot-Mohr diagram for truss "A" was based on the

glue—nail truss. As can be noted, the center diagonal of the

 

 

  

PEI mama

VI"

m1

 

   

O

 

   

 

  

0.37.

0.007

 

  

0.347

0.030

 

  

0

QIIO

 

  

0.320

0.0.

 

 

0.304

0.073

 

 

 

ZO‘HOO.

   

03‘"

 

0.0“

 

 

nus units-c0

 

 

 

FIGURE I2 - WILLOIT-

MOHR DIAGRAMS 0F TRUSS A

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

d O.‘ I]

 

 

 

 

 

 

 

 

 

 

 

 

 

      
    
           

 

' "‘0 H“.-

“I 4.“
'—~——
0 Q“. Q”

NICO “I..-“
I CALI -*

'——_|
0 m 0.!

 

FIGURE I3 -WILLO|T-MOHR DIAGRAMS 'OF TRUSS B

35
ring-bolt truss has two 2." x LI"'s, whereas the glue-nail truss

has only one 2" x h".

Comparison of the Two Truss Designs

Truss "A" appears to be slightly stronger than truss'fiifl'
At 30 pounds per square foot, truss "A" has a theoretical
deflection of 0.375 inches, whereas truss "B" has O.h26-inch
deflection at the top center joint. Each truss contains the
same number of split-rings. Truss "A" has more board feet of
lumber than truss "B." In the glue-nail type connectors,
truss "A" utilized 5/8 of a sheet of h' x 8' plywood, whereas

truss "B" used 1/2 a sheet.

EXPERIMENTAL INVESTIGATION
Apparatus

Test Floor

The test floor consists of steel inserts placed 2 feet
apart in each. direction in the reinforced concrete floor.
Two rows of double "I" beams, Spaced to allow a bolt to drop
through,. form.the wall reaction supports. A heavy steel

plate is bolted across each of these pairs of beams. The
simulated wall reaction point is then mounted on.the heavy
plate. A hinge between the truss and each reaction point

serves to allow some lateral movement of the truss while it
is under load. Figure lha shows the hinged upper wall reac-
tion.

gydraulic Cylinders and Loading Apparatus

Small, 2-inch diameter hydraulic cylinders were used to
apply the load on each truss. The cylinders were placed 2
feet on center and simulatedziverticle downward force on the
truss. The cylinders were mounted on brackets connected to
movable bars fastened to the floor brackets. Since full
travel of the pistons was only 3 inches, the movable bars
gave a means of taking up the slack before loading. The-
cylinders and their mounting are pictured in Figure 15.

Hydraulic pressure was applied to the cylinders with a
two-way hydraulic pump. This pump was motor-driven and could

maintain high pressures. In earlier tests, hand pumps did

37

 

 

.l‘- .-

lha - Hinged upper wall reaction and failure
in.tension:member

 

lhb - Failure in lower center Joint of a
type "A" truss

     

A g. .. _ _ . :
- «L2. . 2.2.2.- .4177! .._'¢.'g‘,;\ ,
lhc - Failure in a lower Joint of a type
"B" design

FIGURE 1h - TYPICAL FAILURES IN RING-BOLT TRUSSES

38

 

s. r,

VlEa - Hydraulic pump and control apparatus
used to apply pressure to the trusses

I

15b - Truss t pe "A” in lace before cylinders
and hol -down brac et have been acolied

\

-

15c - Truss type "B“ in place and ready for
testing

FIGURE 15 - APPARATUS USED FOR LOADING TRUSSES

Q~

 

 

39'
not maintain a steady pressure due to small leaks in the
system. As the result of pressure leaks, a two-man testing

crew was needed when the hand pump was used. While one man
read the deflection, it was necessary for the other to pump

in order to maintain a steady pressure.

Pressure readings were taken in three positions on the

system. Pressure gauges were placed at each end of the bank
of cylinders and at the center of the system where the pump
was located. With the pump located in the center of the
system, there was no noticable pressure drop between the pump
Hand the ends. Figure 16 shows a schematic diagram. of the
hydraulic system, while Figure 15a shows the pump and control

apparatus.

Hold-down Brackets and Rollers
Because of the high compression loads, the top chords

had a severe tendency to buckle up or down in the test

apparatus. To overcome this problem, rollers were placed at
2 to 3-foot intervals under the truss. These rollers and
some additional blocks served to level the truss and to permit
free deflection. The hold-down brackets placed over the
truss at 2 to 3-foot intervals served to prevent the top
chord from buckling up. Rollers on, the hold-down brackets
also allowed free movement of the truss in the loading plane.
Placement of these supports every 2 to 3 feet simulated the
roof girts of a building. ASTM (7) indicates that total
simulated building effect is preferred as long as free de-

flection is not hampered. Figure lab shows a truss before

I40

. 259m oz_o<o._
032%»: L0 zqmoqa 225:8 9 MERE

maimmmm
_lj .Jl_ ‘ .IJ

/.l| mind

 

 

 

m>4<> umflmm II\ _

WE XIHIIII m>._<> >434
m>._<> MES 1mm

”.65 mmammumak’ meg mmammLE/O
J F I r I r #I _.

Z 2 L E. a l:

muozjbk . T

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

12:.

 

 

 

 

 

 

 

El

hold-down brackets and cylinders have been applied. Figure

lhc shows another truss at the start of the test.

Material and Construction of Trusses
The lumber used was Douglas Fir Number 2 or better and
was obtained from a local supplier. The supplier indicated

that lumber was 80 percent Number 2 and 20 percent Number 1

lumber. This corresponds to the Construction grade in The

Light Framing Grades.

No attempt was made to select the lumber to be used in
the truss. When obvious defects were noticeable, however,
another piece was used. These defects generally were confined

to warped pieces or to where large knots were present in the

2" x h" material. Selection beyond this point probably would
not be made by the average farm.builder.

Material for the gusset plates was 5/8" exterior grade
Douglas Fir plywood. Occasionally 1/2" plywood was used with
no apparent difference in ultimate strength of the truss.
Aircraft-type casein glue was applied to both the lumber and

the gusset plate before the 6d nails were used. The nails
served only to hold the gusset to the members until the glue
had hardened.

Each glued truss was allowed to lie for EB hours or more
before testing was begun.

The ring-bolt trusses were constructed with the use of a
l/2" drill and a grooVe cutter. The groove cutter was a four
blade instrument with an adjustable depth control. The groove,

about 3/8" deep, is cut into the face of both pieces on a two

UZ
member joint. The 2-1/2" Split—ring sets about 3/8f into
each member. Washers l-l/2" in diameter were used on each
side of the joint formed with the ring and bolt.

A birds mouth cut was placed in.the lower end of the
top chord of each ring-bolt truss. This cut resulted in the
wall reaction load being placed on both the t0p and lower
chord. Acre (6) found that this increased the strength of
his test trusses by 55 percent. Without the birds mouth cut
or a block, the load placed on the lower chord tended to
produce a twisting movement on the joint. This difficulty'is
not present in the glued trusses since all the members are in

the same plane.

Testing Methods

Loadinngethod

Some slack always remained after the truss was placed
in position and the cylinders attached the first time. In
order to eliminate as much of the slack as possible, a
small load was applied to each truss. In the ring-bolt
trusses, this load tended to set the rings in position and
to produce a more uniform loading curve. This load was quickly
removed.and the truss was allowed to come to rest.

.After the truss had returned to an unloaded equalibrium
position, the dial gauges were set in location and the test
was begun. Loads were applied to the truss as slowly as

possible. In the first part of the test, increments of from

15 to 25 percent of design load were applied. After design

#3

load had been reached, the increments of loading were increased
in some cases. This method was in accordance with ASTM (7).
Each truss was tested two or three times before it was
taken to the breaking point. The first test was taken to
design load or a small amount beyond before the load was
released. In the remaining tests, the final loads were

gradually increased until failure occurred.

MeasuringDeflections

Seven Ames dial indicators, measuring to the nearest
0.001 inch, were used to measure deflection. The gauges were
placed at panel points on the truss. In addition, several
gauges were used to record deflection between panel points of
the top chord.

Since the range of the gauges was only slightly over one
inch, it was necessary to reset the gauges several thmes in
some tests. The gauges were reset when the truss had come to
equalibrium with the last increment of load. Each recording
of deflection also was made after the truss had reached or
nearly reached an equalibrium.position.

Figure lhc shows the dial gauges in place ready for

testing.

' Plotting_Deflection of Test Results

The deflection curves for each point on the trusses were
drawn from the results of 2 or 3 loadings of each truss. Since
the different runs did not always start at the same place, it

was necessary to plot the runs individually and then shift the

5’

IIlI

curve to start at zero. This practice, in theory, did away
with the variation in slack at the start of the tests.

By using two or three test runs for each truss, it was
possible to Spot any major errors or problems in a particular
test. Test runs that produced radically different results
from other tests on the same or similar trusses were examined
to see if they were caused by the test apparatus. In case
cylinder brackets had slipped or something was holding up the
free movement of the truss, the problem was corrected and the
test was repeated or discarded.

The best curve representing the several tests of a truss

was then drawn to represent the final deflection pattern.

RESUDTS AND DISCUSSION
Number of Trusses Tested

Tests were conducted on thirteen different trusses.
Seven of these trusses were type'TT‘and the rest were type'TJ'
There were four tests of type "B" trusses using ring-bolt and
three tests with glue-nail construction. Truss "A" had three
tests of each type of construction. Table 3 shows the maximum
load and type of failure for each truss. The first two letters
under type of truss refer to the method of fastening, glue-nail
or ring-bolt. The third letter indicates the truss design.
Each truss was loaded two or three times before it was loaded

to failure.

Comparison of Trusses

Comparison of the tested trusses indicates a greater
difference between type of fasteners than type of design. The
test results for the glue-nail trusses compare very favorably
with the theoretical deflections from the Williot -Mohr
diagrams. Figure 17 shows the theoretical and experimental
deflections for the top center joint of each glue-nail truss.
These same deflections for the uppermost joint are shown in
Figure 18.

Figures 19 and 20 show the theoretical deflection pattern
for both truss designs. Theoretical deflection is based on

truss spacing of h feet on center and a load of 30 psf. From

he

TABLE 3

MAXIMUM L0AD AND METHOD OF TRUSS FAILJAE E

 

 

 

 

 

 

Truss Types Maximum. Method of Failure
Number Load
1 RB-A 66.2 Lower joint "G" split out some-
thme after load had been on truss
2 RB-B 58.8 Lower joint "F" split out
A RB-B no.6 Both joints "G" and "F" failed
5 RB-B 37.9 Tension member GF was splitting
out of both joints
6 RB-B h0.6 Joint at "G" failed
7 GN-A 69.8 Gusset plate at "F" failed--grain
of plywood placed in wrong direc-
tion--failure due to rolling shear
. in plywood
8 GN-A 58.8 Failure inmember AH--severe slope
to grain at point of failure
9 GN-B 95.6 Failure in member AG near the
' reaction "A"

10 GN-B 58.8 Failure in member AG near the
reaction "A"

ll GN-B 8h.6 Gusset plate at "E" failed in the
glue line of the interior grade
plywood

12 GN-A 91.9 Did not fail at this point

13 RB-A 55.2 Failure occurred at joint "F"

1h RB-A 66.2 Lower reaction "A" was point of

 

 

failure-~bottom chord broke a

ring connection

 

* GN = Glue-naiI, RB = Ring-bolt construction
B I Truss type "B“

A - Truss type "A"

I-I-7

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

80 A-IZ 3:9
/e-II
'//
70 7A
.01 /
/ s-I
THEDRIITIGAI.
/ [17/ I
,_ 40 /
8
35
2 30F
c':
S
20
I0 /
0 n
0 0.25 0.50 075 I00 I25 I.50

DEFLECTION - inches

FIGURE I7 COMPARISON OF THEORETICAL AND ACTUAL
DEFLECTION OF TOP CENTER JOINT - GLUE’NAIL TRUSSES

 

80

 

 

70

 

 

60

.b
O

LOAD- Ibs/sq.foot
on .
O

20

I0

FOR BOTH /
“A" a “a“, l/

B'IO
THEORETICAL / A'B/ I

 

 
   
    

I/
/

4
7

/
//
4/

/

/

 

 

  

 

 

 

 

 

 

 

F

 

 

 

 

 

0 025 0.50 0.75 I00 I25 L50
DEFLECTI 0N - inches

IGURE I8 COMPARISON OF THEORETICAL AND ACTUAL
DEFLEcTIDN 0F POINT "D" - GLUE-NAIL TRUSSES

 

 

SCALE 8 inches

 

0 OBI

FIGURE I9 THEORETICAL DEFLECTION 0F TRUSS "A""

30 EOUNDS PER SQUARE FOOT LOAD 0N TRUSS“
TRUSS SPACING IS 4 FEET 0N CENTERS.

 

 

 

A v
G.
SCALE: inches
0 05 I

FIGURE 20 THEORETICAL DEFLECTIDN 0F TRUSS "s"-

30 POUNDS PER SQUARE FOOT LOAD 0N TRUSS-
TRUSS SPACING IS 4 FEET 0N CENTER.

50
these figures it can ,be seen that truss "B" .has a. single
smooth curve in the top chord when it is under load. The
theoretical deflection pattern for truss "A" has two curves

in the top chord. That is, the center joint of the top chord

of truss "A" is not deflected as much as truss "B." Comparing
the theoretical deflection of point "0" (top center joint)
and point "D" (uppermost joint on top chord) in Figures 19
and 20, this characteristic also is apparent. This same fact

was present in the actual tests. In the tests, however, the
deflection pattern of the top chords was more pronounced.

Comparison of the two designs, as far as other points on

the trusses are concerned, is not valid since the other panel

points are in different locations relative to one another.

The joint on the lower bottom chord of truss "B" is farther
from the wall reaction than is the corresponding joint of
truss "A." Figure 21 gives an indication of the deflection
of various points of truss Number 7.

Figure 22 shows the results of the deflections of the
top center joint in the ring-bolt trusses. The deflections
for the uppermost joint of the top chord of the same trusses
are shown in. Figure 23. Deflection results with trusses of
ring-bolt fabrication for both designs did not compare well

with the theoretical deflections of the Williot-Mohr diagram.
From.Figures 22 and 23, it also is apparent that there

is not as much consistency in ring-bolt trusses as there is

LOAD- lbs/sq. foot

51

 

80

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

E
| C
B F

70 A L 6 £1. 8/2
60

50 .

40 1’ /

30 l‘

20

I0

0
0 0.2 5 0.50 0.75 I00 I25 L50

DEFLECTION - Inch“

FIGURE 2| COMPARISON OF DEFLECTION 0F POINTS ON
TRUSS No.7 - GLUE'NAIL CONSTRUCTION.

 

 

 

 

 

I.50

 

 

 

 

 

 

 

 

 

 

 

 

 

I.75

I. 25

LOO
) DEFLECTION- inches

0.25

2.00 2.25

0.50 0.75

52

22 DEFLECTION 0F PANEL POINT (CI (TOP CENTER)‘ RING AND BOLT TRUSSES

FIGURE

53

mummam... Flam oz< 02:“. I Bub-50¢ sz<d “.0 20....0quwo MN wag;

0N.N 00.N n».-

on ._

«one... IZO_._.0m.._umo

mm... 00..

05.0 00.0

0N0

 

0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

\NNN
\\\\ \
\\\\
._\..\I\I.\.. \\
N\\\\
\

 

 

 

 

 

 

_-<

o t

0.

ON

0m

0w

0.»

00

I001 'PSI'SQI- OVO'I

SII

in the glue-nail trusses. The deflection curves established
from the results of two or three tests of a truss design do
not correspond well with the curves for other trusses of the
same design. In some cases, individual test curves of the

same truss did not compare well with other tests.

In plotting the deflection curves for ring-bolt trusses,
the points did not readily form a smooth curve. For a partic-
ular increment of load, the deflection frequently was less
than the proceeding increment. In some cases, a truss would
have come almost to equilibrium with the last increment of
load, then suddenly deflect an additional amount. This erratic

behavior occurred after the design load had been reached.

The points established by the glue-nail trusses always
produced easily established curves. Very seldom did one

point deviate greatly from the established deflection curve.

Method of Failure

The load at failure point varied with different trusses.
The location andmethod of failure, however, were more uniform
within types of fasteners than in truss designs. 'Table 3
shows the location and method of failure of each truss.

In general, the ring-bolt trusses failed at or very near
a joint. The glue-nail trusses failed at weak points in the
members.

There were two failures in the plywood gusset plates of

the glueanail trusses. The failure in truss Number 7 was due

55

to the surface grain of the plywood running the wrong way in

reapect to the stress on the gusset. Failure was due to

rolling shear within the gusset plate. The other failure in
a gusset occurred in truss Number 11. This truss had interior
plywood that did not have a grade mark. Failure in this case
was due to glue line shear between plys of the gusset.

The other failures in glue-nail trusses occurred in the

bottom.chords where deflects, such as knots or cross-grain,

were present. The lowermost member of the bottom chord in
the trusses was the most frequent point of failure.

In the ring-bolt trusses, failures occurred at the joints

along the lower chords of the trusses. Most of the "B" type,
ring-bolt trusses failed at one or both of the lower joints

along the underside of the truss. One of the "B" trusses
failed at the joint near the upper reaction. The ring-bolt,

type "A" trusses all failed at different locations, although,
the lower center joint of this truss had the most cracking--

even when failure occurred at some other location.
A great deal of the failure in ring-bolt trusses could

be due to the eccentricity of the joints. There was a severe
tendency of the lower joints to pull apart. The washers on
the end of the bolts would be pulled into the wood. The
result of this was that the total bearing area of the rings

was not-in use. The washers also were pulled in considerably

‘when.2-inch washers were used in some of the later tests.
No failures occurred in the top Chords of the trusses.
The center diagonals did not have any failures, except in the

ring-bolt trusses where the lower joints failed.

S6

Maximum.Loads and Load Factor of Safety

Load at failure was always well above the design load
for all the trusses. The test loads were, however, short-time
loads. Failure probably would have occurred sooner if the
duration of load had been increased. '

Since it would not be proper to compare the loads at
failure with the design loads, a reduction was made for longer
loading. The basis of this reduction came from Figure 2. The
test load was taken as being fromESto 10 minutes in duration.
Test loads were reduced to loads of 2 months' duration.
Considering the test load as. 165 percent of a permanent load
and the 2 months' load as 125 percent of a permanent load,
the tests were reduced by 125/165. This information is taken
froniallowable stresses and applies to individual members,
rather than to a complete truss. The effect should be nearly
the same except where strength of the rings or bolts are
effected. In any event, it more nearly signifies the load
factor of safety than the test results alone would do.

When the maximum.loads are averaged for both methods of
fastening in each truss design and the reduction. factor is
applied, the trusses still are stronger than the design load.
The average load factor of saféty is given for different

fasteners and designs in Table A.

57

TABLE h

LOAD FACTOR OF SAFETT BASED ON Two MONTHS
DURATION OF L0AD

j
T

 

Average Reduced Load
Truss Truss Maximum Maximum. Load For Factor
Type Number Load Load Longer of
#/rt.2 #/ft.2 Duration Safety
_ 125/165
RB-A 1 66.2
13 55.2
In 66.2
62.5 h7e3 1.57
RB-B 2 58.8
A no.6
5 37.9
6 h0.6
hu.5 33.7 1.12
GN-A 7 69 . 8
8 58.8
12 91. *
' 73.5 55.7 1.85
GN-B 9 95.6
10 58.8
11 8h.6
76.3 57.8 1.93

 

* Truss Number 11 had not failed at this load.

CONC LUS IONS

1. Both 30-foot Span, single slope, truss designs
analyzed are practical for farm construction. Either of the

two designs tested in.this research work will support 30 psf

and still have a factor of safety greater than one.

For the designs tested, the glue-nail trusses are superior
to the ring-bolt trusses. In each basic design, the glue-nail
trusses will support about half again as much load as will
the ring-bolt trusses. The deflection of the glue-nail

trusses is not as great as the ring-bolt trusses, even when
they are carrying greater loads. At an equivalent load of
50 psf, the glue-nail trusses are deflected about 3/11 of an
inch. With the same load, the ring-bolt trusses show deflec-
tions in the range of 2 inches and more. This same deflection

pattern holds true with smaller loads.
2. The factor of safety remaining after the test loads

were reduced to correSpond with a 2-months' loading ranged
from 1.12 to 1.93. A factor of safety of 1.12 may be a little
low for some areas of. northern Michigan; however, a roof load
of 30 psf is unrealistically high for most of Michigan. The
higher factors, 1.85 and 1.93, in the glue-nail designs may
be unjustifiably large for farm purposes. In this case, the
truss Spacing could be increased to take advantage of the
increased strength of the glue-nail trusses.

3. Both designs are easily fabricated and assembly

time is held to a minimum. A farm builder or a farmer, with

59

a small amount of building experience, can erect either the
ring-bolt or glue-nail trusses. Due to the Spacing of h feet
on center, the trusses also are light; therefore, they are
easily put in place without costly equipment.

The most critical part of- construction with rings and
bolts is locating the bolt holes so that the joints do not
have to be forced together. In the glue-nail trusses, the
shear strength of the glue line is several times greater than
the shear strength of the wood itself. This fact nearly
eliminates any possibility of weakness due to glue starvation
in the joints.

Full strength advantage of glue-nail fastening will
result if some additional time is spent in selection of the
lower truss members.

A. Larger span, single slope, glue-nail trusses can
use either of these basic designs. Both designs do a very
good job of distributing the stresses evenly. An extension
of the ring-bolt trusses should notbemade without T's-designing
the joints. The joints made with ring-bolt fasteners are the
weak? joints in the 30-foot trusses. Larger spans would require
larger rings and members. This would result in.much higher

building costs than can be justified for farm purposes.

SUMMARY

The major objective of this investigation was to analyze
and design 30-foot, single slope trusses that are suitable for
farm construction. Inthe preliminary planning, consideration

was_given to truss Spacing and the economy of using readily

available commercial grade lumber. In addition, designs that

allowed both glue-nail and ring-bolt fasteners were used.

Stress diagrams were drawn.f0r many different truss

designs. With the aid of the stress diagrams, two designs
‘were selected for further analysis. Both designs selected
:made construction with 2" x h" and 2" x 6" material possible.

The theoretical deflections of the panel points of both
trusses were calculated by the use of Williot-Mdhr diagrams.
The deflections were based on.the design load of 30 psf of
roof with the trusses spaced h feet on center.

In order to compare the actual trusses with the theoreti-
cal analysis, three trusses of each type fastening and design
were tested. The total numbercnftrusses tested were thirteen.

The test method consisted of loading the top chord of
the trusses by use of 2-inch hydraulic cylinders placed 2 feet
apart. Simulated wall reaction plates served to hold the
truss in place while a motor-driven.pump applied pressure to
the cylinders. Ames dial gauges were used to measure the
deflection of points on the trusses. Two or three loadings
were made on each truss before it was taken to the failure

point.

61

The test results indicate a greater difference between
type of fasteners then between the two designs tested.

The glue-nail trusses for both designs ' compared
favorably with the theoretical deflections. The ring-bolt
trusses deflected considerably more than either the glue-nail
trusses or the calculated values. When maximum loads were
reduced to simulate a two months' loading, all the trusses
produced load factors of safety larger than one. The averaged

results, reduced to correspond ' with a two-month roof load,

appear in Table II.

SUGGESTIONS FOR FURTHER STUDY

1. Investigate the possibilities of using different
types of materials for use as gusset plates.

2. Investigate the effect of larger washers or steel
plates on the ring-bolt trusses.

3. Investigate designs of single slope trusses that
have the lower wall rigidly connected to the truss.

h. Investigate the possibility of basing. design of
farm buildings on frequency of occurrence of snow storms for
different areas of Michigan. This would be similar to water
retaining walls that are designed for a number of years based
on past precipitation data.

5. In conjunction with Item.h, investigate the amount
of snow that accumulates on unheated metal roof farm buildings
in different areas of Michigan.

6. Investigate the effect of long-time loads on the

single slope trusses.

9.

10.

BIBLIOGRAPHY

Books

Barre, H. J., and Sammet, L. L. Farm Structures: The
Ferguson Foundation Agricultural Engineerigg Series.

New York: John Wiley & Sons, Inc., 1955.

Grinter, Linton E. The Theory of Modern Steel Structures.
Vol. II: Staticalgy Indeterminate Structures. New

York: The Macmillan Co. 1919.

Johnson, J. B. et a1, Modern Framed Structures (Part I)
StressesthSimple Structures. New York: John.Wiley
8: Sons, Inc.

U.S. Department of Agriculture. Wood Handbook. Madison,
Wisc.: Forest Prod. Lab., 1955.

Wang, Chu-Kia. Statically Indeterminate Structures.
New York: McGraw-Hill Book 00., Inc., 1953.

Articles, Periodicals, Reports
and Unpublished Material

Acre, Leon. Investigation of Failure in Plate Joint of
Ring and Bolt Trussed Rafters; Unpublished research

work, East Lansing, Mich.: Ag. Eng. Dept., Mich.
State University, 1957.

American Society of Testing Materials. ASTM Standards
on Wood, Wood Preservativesg and Related MateITaIs.

Philadelphia: Am. Soc. Testing Materials, Mar. 195A.

Bigalow, L. W. Pole Barn Construction. New York Agri-
cultural CoIIege (Cornell), BfilIetin 1101, 1950.

Boyd, James S. Secondary StressesinTrusses with Rigid
Joints, Special Application to Glued Wooden Trusses.
Unpublished Ph.D. Thesis, Iowa State College, 19511.

Cockrell, R. A., and Bruce, H. D. Effect of Thickness
of Glue Line on Strength and Durability of Glued
Wood JointsfMadison, Wisc.: Forest Prod. Lab.,
No. 1616, l9h6.

ll.

12.

13.

1h.

15.

16.

17.

18.

19.

20.

21.

22.

6h

Douglas Fir Plywood Association. Technical Data on
Plygood. Tacoma, Wash.: Douglas Fir Plywood Assoc.

Dryden, J. L., and Hill, G. C. Wind Pressures on
Structures. U.S. Bureau of Standards Scientific

Paper No. 253, 1926.

 

Dunsan, Catherine G. Effect of Moisture on Bacterial
Weakening of Casein-bonded Plywood. Madison, Wisc.:
Forest Prod. Lab., No. 2077, 1957.

Esmay, M. L., and Boyd, J. S. Gable and Single Slope

Wood Truss Design. Paper presented at Winter Meeting
of ASAE, Chicago, 1957.

, , and Cargill, B. F. Clear-Span Roof

 

Construction, East Lansing, Mich.: Mich. State

University and Cooperative Extension Service, Farm
Buildings Series, Circular 732.

Fenton, F. D., and Otis, C. K. The Design of Barns to
Withstand Windloads. Kansas Engineering Experiment

Station, Bulletin U2, l9hl.

Forest Products Laboratory. A roximate Methods of
Calculating the Strength of_W003. MEdIson, Wisc.:
Forest Prod. Lab., R-l630.

. List of Publications on Glue,gGlued Products

 

and Veneer. Madison, Wisc.: Forest Prod. Lab.,
Report No. 513, 1957.

. Casein Glues, Their Manufacture, Preparation
and Application. Madison, Wisc.: Forest Prod. Lab.,
Report No. D-280.

 

. Synthetic Resin Glues. Madison, Wiscr:Forest
Prod. Lab., Report No. 1336, l95h.

 

. Drying. and Conditioning of Glued Joints.
Madison, Wisc.: Forest Prod. Lab., Report No. L75.

 

Freas, Alan D. Factors Affecting, Strength and Design
Principles of Glued Laminated Construction;
Madison, Wisc.: Forest Prod. Lab., Report No. 2061.
1956.

23.

25.

26.

27.

28.

"29.

30.

31.

3h.

65

Giese, H. The Structural Application of Glue in Frmuhn;
Farm Buildings, (Part I). Ag. Eng. Journal, Vol. 21,
pp. A7-50. 19h0.

, and Henderson, S. M. The Structural Applica-

 

tion of Glue in FraminggFarm Buildings, (Part II).
Ag. Engr. Journal, Vol. 26, pp. 507-510, l9h5.

, and Palmer, E. D. The Structural Application

 

of—Glue in FraminggFarm.BuildingsL (Part III). Ag.
Engr. Journal, Vol. 31, pp. h55-h57, 1950.

, and Hamlin, C. E. Effects of Length and

 

Width on Strength of Glued Joints. Ag. Engr.
Journal, Vol. 33. pp. Ell-Elk, 1952.

Johnson, R. P. A. WorkingAStress for Struchuufl.Lumber.

Paper given at Wood Symposiump-One Hundred Years
or Engineering Progress with Wood. Chicago :
September 1952.

Kaufert, F. H. Increasingthe Durability of Casein Glue
Joints with Preservatives. Madison, Wisc.: Forest
Prod. Lab., Report No. 1332, 1953.

Markwardt, L. J. Wood gs an EngineeringgMaterial. Am.
Soc. Testing Materials Proceedings, Vol. R3,
pp. MOO-A35. 19A3.

McLauglin, C. F. CEsGin Glue. Gluing Short Course held
April 1950. Paper presented at No. Carolina State
College, 1950.

McLeod, A. M., et al. A Comparison of Shearng Strengths
of Glued Joints atharious Grain. Directions as
Determined by_Four Methods of Test. Madison, Wisc.:
Forest Prod. Lab., Report No. 1522, l9h6.

Molander, E. G. Wind Loads on Farm, Buildings. Ag.
Engr. Journal, Vol. 28, pp. 22-2h, l9h7.

National Bureaucu?Standards. American Standard Building,

Code Requirements for Minimum. Design Loads on
Buildings and Other Structures. Washington, D.C.:

Government Printing orrioS: Misc. Pub. 14-179, 1957.

National Lumber Manufacturers Assoc. National Design
Specifications of Stress-Grade Lumber and its
Fastenings. Washington, D.C., the Assoc., Rev. 1955.

35.

36.

37.

38.

39.

AD.

hl.

N2.

AB.

66

Radcliffe, B. M., and Granum, H. A New Low Pitched
Roof Truss with Nail Glued Connections. Ag. Expt.
Station, Purdue University, Bulletin 617, 1955.

Suddarth, Stanley K. A Hydraulic Test Facility for Wood
Structural Elements. Journal of Forest Prod. Res.

Soc., Vol. 6, No. 5, May l95h.

Traux, T. R. The Gluing of Wood Madison, _Wisc.:
Forest Prod.Lab., Report No. I500, l9h5.

, and Selbo, M. L. ResultscT’Accelerated Tests

 

and Long Term Exposures on Glue Joints in Laminated
Wood. Am. Soc. Mech. Engr., Paper No. h7-A-85,19h7.

Timber Engineering Company. Design Manual for TECO
Timbers Connector Construction. Washington, D.C.:

The Company, Rev. l95h.

. Typical Designs for Timber Structures. 3rd
Ed., Washington, 0.0.: The Company. 1955.

 

Wood, L. W. Basic Principles of Structural Grading.

Paper given at Wood Symposium-~0ne Hundred YGars
of Engineering Progress with Wood. Chicago:
September 1952.

. Evaluatibn of the Factor of Safety in Struc-
tural Timbers. Madison, Wisc.: Forest Prod. Lab.,
Report No. 2068, February 1957.

 

. Relation oertrength to Duration of Load.

Madison, Wisc.: Forest Prod. Lab., Report No.
R1916, 1951.

 

. Variation of Strength Properties in Woods
Used for Structuralijurposes. Madison, Wisc.:
Forest Prod. Lab., Report No. R1780, 1951.

 

.I. .I USE (:2 III