|Table of Contents|
./ '1 L o
STUDIES Or Tlt STlINCTIT OFr 6LUU
LAMINATIEU WCOU CONSTRUCTION
UNITED STATES DEPARTMENT OF AGRICULTURE
FOREST PRODUCTS LABORATORY
Madison 5, Wisconsin
In Cooperation with the University of Wisconsin
Digitized by the Internet Archive
ST-uTl77 OF Tn STPHC-IH OF GLTED TATT. IT2.
W::?C C ,:i::S .UCTI '."-
By AlIi D.. F-.A3, Engineer
Forest Products Laboratory,2 Forest Service,
U. S. Department of Agriculture
Glued laminated construction has had a long history in E-rope, particu-
larly in Germany, Sweden, and Switzerland. A wide variety of applications
were observed by a member of the Forest Products Laboratory staff on a -uropeon
trip during which he inspected some 50 structures varying in age up to 25 years,
all laminated with casein glue. He reported that members used in buildings in
which normal atmospheric conditions prevail had not seriously deteriorated even
after considerable periods of service. Casein-glued members used under more
severe conditions, such as locomotive-repair shops, chemical plants, footbridges,
and railway-platform structures, also were reported to have given good per-
The use of glued laminated construction in the United States is much
more recent, so that examples of extended service are not available. While
definite information on the first uses of glue in structural members in this
country is lacking, it appears that no extensive development in this field
occurred prior to the installation of glued laminated arches in a building on
the grounds of the Forest Products Laboratory in 1935" Since that time, how-
ever, the number and variety of applications have grown so remarkably that a
new industry has developed,
Arches were among the first forms of glued laminated construction to
find wide use in this country. The inherent structural efficiency of this
form of support, together with the fact that it made available unobstructed
floor area usable to a considerable height, resulted in the use of arch struc-
tures in industrial buildings, grnnasiums, and aircraft hangars. The possi-
bility of varying the form to give the necessary pleasing and traditional
effects, resulted in the application of considerable numbers of laminated
arches in church architecture. The farm market in the past few years has
absorbed an increasing amount of glued laminated construction in the form of
curved rafters, particularly for barns, but for machine sheds, brooder houses,
and other farm buildings, as well.
LA summary of a paper of the same title prepared for presentation at the Pacific
Area -ational Meeting, A.S.T.M., Oct. 10-14, 1949; San Francisco, California.
2vaintaincd at Madison, Wis., in cooperation with the University of IJisconsin.
militaryy requirements during rorld War II gave considerable impetus to
the laminating industry. Not only did these requirements cover buildings of
various types, but many other uses as well, including parts for aircraft and
for wood vessels.
3xtorior applications in the United States are less common. There are,
however, a few instances of such applications. A footbridge in Madison, Wis.,
has given satisfactory service over a period of about 9 years. Laminated dredge
spuds in use on the Columbia River have apparently performed satisfactorily over
a period of 3 or 4 years. The railroads have installed a number of experimental
bridges with a view to gathering service experience as to the suitability of
laminated members because of the increasing difficulty of obtaining suitable
Advantages of Glued Laminated Construction
Advantages of glued laminated wood construction are many and significant.
They include the following:
(a) Ease of fabricating large structural elements from standard commer-
cial sizes of lumber. Laminntcd arches have already been erected that provide
buildings with clear spans up to 170 feet, and, also, laminated beams of 80-
foot span. Arches with sections as deep as 7 feet have been projected,
(b) Achievement of excellent architectural effects, and the possibility
of individualistic interior decorative styling.
(c) Freedom from checks or other seasoning defects associated with large
one-piece wood members, in that the laminations are thin enough to be readily
seasoned before fabrication.
(d) The possibility of designing on the basis of the strength of seasoned
wood, for dry service conditions, inasmuch as the individual laminations can be
dried to provide members thoroughly seasoned throughout.
(e) The opportunity to design structural elements that vary in cross
section- along the length in accordance with strength requirements.
(f) The possible use of lower-grade material for less highly stressed
laminations without adversely affecting the structural integrity of the member.
(g) The fabrication of large laminated structural members from smaller
pieces, which is increasingly adaptable to future timber economy, when more
lumber will come in smaller sizes and lowor grades from smaller trees.
On the other hand, there appear to be no disadvantages of lominatod
construction as such. Modern glues and gluing techniques provide both adequate
and effective means of bonding laminations into an assembly cqunl or superior
in strength to a single-piece member of equivalent section. They may be
selected to provide a laminated assembly that is water-resistant or waterproof,
Rept. To. R17L49
as conditions of use may dictate. When properly glued, laminated members
may be given preservative treatment by pressure methods much as solid mc-r-bers
are treated, which will improve their resistance to decay when used under ad-
verse e:posuro conditions.
There are, however, certain factors involved in the production of
laminated timbers not encountered in producing solid timbers. A number of
(a) The preparation of lumber for gluing and laminating usually raises
the cost of the final product above that of solid green timbers.
(b) For constructions in which green timbers are satisfactory, more
tine is required to cut and season lumber and to laminate the timber than is
required to cut solid green timbers.
(c) Since the value of a laminated product depends upon the strength of
the glue joints, the laminating process requires special additional equipment,
plant facilities, and fabricating skill not required for producing solid green
(d) Since considerably more operations are involved in manufacturing
laminated members than in manufacturing solid members, there are moro possibil-
ity for error, and special care must be exercised in each operation to insure
a product of high quality.
(c) Large curved members are difficult to ship by common convcyancos.
Background of Research
The first research at the Forest Products Laboratory on the strength
properties of glued laminated construction was begun in 1934 on curved members.
This research, which covered a period of several years, studied a number of
factors affecting strength, including the effect of curvature, end joints, and
defects. It culminated, in 1939, with the publication of U. S. Dept. of Agri-
culture Technical Bulletin No. 691, "The Glued Laminated Wooden Arch." This
bulletin not only presented the results of an extensive series of tests, but
presented also recommendations on working stresses and design procedures. It
has had extensive use, since its publication, as a basis for design ,.and speci-
fication of this type of construction.
During World War II, it bccame increasingly apparent that additional
research was necessary to answer the questions that arose with the development
of this method of construction. Accordingly, with the cooperation of the War
Production Board and industry, further work was undertaken at the Forest Prod-
ucts Laboratory, including the testing of a large number of full-size beams
and columns to provide additional design data and information for technical
phases of specifications. Factors investigated included the relative strength
of members containing end joints of different types, the effect on strength of
Ropt. 1o. R1749
defects in different lar.ini.tions, the effect of varying the thickness of
laminations, and like factors.
T.%ile this paper has for its major purpose the presentation of informa-
tion on the strength of glued laminated construction, the author would be
premise if he did not mention the developments in glues and gluing techniques
that have made this material no promising for the future. Prior to World *;Tar II,
casein Clue was the only adhesive reasonably acceptable for use in structural
nr!-bers, and its limitations under severe exposure were recognized. During
Torld W r II a number of synthetic resin adhesives suitable for use under
severe exposure and capable of being set at room or intermediate temperatures
(between room temperature and about 200 F.) were developed by the adhesive
industry. Coincident with those developments, the Forest Products Laboratory
pursued an active and continuous program of research on the properties of
these adhesives and on the techniques of applying them for optimum results.
Results of Research
Available lengths of lumber are frequently not adequate to provide
full-length laminations for the larger structures. It is necessary, therefore,
to join two or more lengths of lumber end to end to provide the necessary
Butt joints are simple to make, since they require no special prepara-
tion of the ends of the pieces to be joined. Obviously, however, they have
no strength in tension when not glued, and tests have shown that even when.
glued they arc very low and erratic in strength, even the best gluing tech-
niques affording no more than a fraction of the tensile strength of the wood.
Further, oven with the best techniques in fitting, butt joints can be expected
to be only partially effective in transmitting compressive stress.
The longitudinal stress in a butt-jointed lamination is, of course,
zero immediately adjacent to the joint and increases by transfer through shear
from the adjacent laminations as the distance from the joint increases. Hence
the excess of longitudinal stress in the continuous laminations over that in
the jointed ones is a maximum at the joint and, consequently, the shear stress
is a ma::inm at the joint, whore it is very concentrated. The result is that
shear failure starts adjacent to the joint and progresses os more of the
length of the jointed lamination is relieved of its stress. Progressive
failure of this character was observed in tests of beams containing butt joints.
Strain measurcmonts made in the vicinity of butt joints in laminated
beams indicate that the joints tend to cause considerable concentration of
stress in the unjointed laminations. For example, strain in the region of a
joint in the top laminr-tion of a beam was found to reach a measured value as
high as 2-1/4 times as great as the strain at the same vertical position in
a cross section that was at some distance from the joint and subjected to the
Ropt. iTo. R1749
Butt joints are undesirable, therefore, not only because they fail to
transmit longitudinal stress, and therefore represent an ineffective area in
a beam or column, but also because they concontrato both longitudinal and
shear stress. Butt joints arc additionally undesirable in curved laminations
because of their effect on intorlamination contacts in their vicinity. It is
impossible to produce curvature right to the end of a square-ended piece.
Hence, in the vicinity of butt joints, contact between adjacent laminn.tions
can result only from pressure sufficient to crush the wood and expel the glue,
an occurrence that makes the joints between laminations locally deficient in
resistance to shear.
When butt joints occur in adjacent laminations, tests have shomn that
the joints must be well separated if their effects are not to be additive.
Even with a spacing as great as 30 times the lamination thickness, it appears
necessary to consider both laminations ineffective in columns or in the com-
pression portions of beamso Vhon the joints arc in the tension portion of a
beam, at the spacing mentioned above, the strength is lower than would be
estimated by considering both laminations ineffective, a fact that indicates
the undesirability of using butt joints in members subject to tension stresses.
Tests have indicated that a scarfed joint, if carefully prepared, well
glued, and of sufficiently flat slope, can have a considerable proportion of
the strength of an uncut piece. Beams, arches, and columns containing scarf
joints gave test results nearly as high as did similar members having continu-
ous laminations. Tests of individual joints indicate considerable variability,
so that some conservatism is desirable inassigning allowable stresses to
jointed members. Tests of members containing scarf joints in adjacent lamina-
tions indicate the desirability of separating such joints, but have not pro-
vided a firm basis on which to establish spacing rules.
Many other types of end joints have been proposed and some are in use.
The data available are generally inadequate to establish the stresses that
should be assigned to them. These joints generally have a form that is
designed to facilitate alinement of the two pieces and to provide additional
gluing area. Difficulties in accurate machining of the mating parts, and
small changes in shape due to moisture changes, may actually result in loss
effective gluing area than in a plain joint.
Effect of Knots
In solid timbers, the position of defects within the cross section of
a member is essentially fixed by the location of the defects in the tree. In
laminated members, however, the location of defects can be more or loss con-
trolled. This fact opens up the possibility of producing members combining
high-strength and low-strength material in a single member. It would appear
that a bean of relatively high strength could be produced if the defective
material were placed near the neutral axis and the high-strength material in
the outer portions of the depth.
Ropt. No. R1749
This possibility was explored in tests of curved members by combining
defective and clear material in various proportions. The results indicated
that up to about 60 percent of defective material in the central portion of
a laminated arch would not seriously reduce the strength below that of members
consisting entirely of clear laminations.
The same problem was studied in somewhat greater detail by means of an
extensive series of tests on beams and columns containing knots in various
combinations of size and placement. The data from these tests afforded a
soeh'.,ht more complete relation between strength and knot size and placement.
Inasmuch as the strength of a beam depends upon the moment of inertia
of its cross section, it seemed valid to assume that the reduction in strength
caused by l ots could be related to the moment of inertia of the parts of the
cross section occupied by the knots. Obviously, knots that are reasonably
near to each other longitudinally would have nearly the sane effect as if they
were at the same cross section. For each beam, there was computed a value of
IK, the sum of the moments of inertia of all knots within 12 inches of the
central cross section.
A study of the data revealed a relation bet -con bending strength and
the ratio IK/I., where IC is the gross moment of inertia of the beam. Doth
modulus of rupture and fiber stress at proportional limit decrease with
increasing values of I,/IG, with the rate of decrease becoming larger at the
greater values of IK/IG. Contrary to experience from tests of solid timbers,
the modulus of elasticity was found also to decrease iith increasing values of
Recommendations for Design
Time limitations hI"vo permitted covering only two of the more irlortant
phases of research on this subject. A considerable number of additional
phases, however, have been studied in more or loss detail. Among those are:
effect of lamination curvature on strength, strength of curved members in
radial tension, the shear strength of members in which gluing pressure was
obtained by nailing, shear strengths of both knotty and clear beams whore
gluing pressure was applied with clamps, and the effect of both knots and
end joints on the strength of columns. The data from these studios form the
basis for a Trarnuzcript currently in preparation and presenting recommendations
of the Forest Products Laboratory with respect to the design of glued laminated
construction. It is expected that these recomntmendations, together with those
of the Laboratory on methods of fabrication, will be published as a technical
bulletin of the U. S. Department of Agriculture.
Rept. ITo. R1749
Vile the research discussed herein represents a considerable body of
knowledge on the subject, which permits the designing of laminated structures
with reasonable confidence, there still remain problems to be solved.,
During the past year, the Forest Products Laboratory has made additional
studies on the spacing of end joints in laminations. The results, while afford-
ing useful information, were not conclusive. Further, the possible combinations
of joints are myriad, presenting an endless array of possible tests if all were
to be investigated. Some attempt should be made to develop a theory of joint
action to simplify the problem.
It has long been known that deep beoms develop, in test, lower stresses
than do shallow be2ams, and that box or I-beams develop lower stresses than do
beams of solid section. Two independent investigations of the effect of depth
have developed two different expressions for depth effect, which vary consider-
ably from each other. The maximum beam depth tested was 14 inches. Yet beam
or arch depths of 3 feet are not uncommon, and an arch with a depth of 7 foot
has been projected. In designing at such depths the Laboratory is on uncertain
ground. It needs, therefore, to investigate the effect of depth over a consid-
erable range and to study the fundamentals of this phenomenon.
The weakness of some Douglas-fir in shear has booeen noted in test. If
it were found that Douglas-fir wore markedly stronger in shear in the radial
than in the tangential plane, it might be advantageous, for cases where shear-
strength requirements are high, to use laminations so cut that the piano of
shear would be radial. The Laboratory proposes to study this problem and to
investigate the possibilities.
The items mentioned above represent only a few of the more urgent
problems still to be solved. There are many others.
Ropt. HTo. R17-49
UNIVERSITY OF FLORIDA
3 1262 08924 4619