CLUID ANID NAILED iCCIIE TIUSSIS
CKl IOUSE CM-STRJICTyION
UNITED STATES DEPARTMENT OF AGRICULTURE
FOREST PRODUCTS LABORATORY
In Cooperation with the University of Wisconsin
GLUED AND NAILED ROOF TRUSSES FOR HOUSE CONSTRUCTION
R. F. LUXFORD, Engineer
OTTO C. HEYER, Engineer
Forest Products Laboratory,! Forest Service
U. S. Department of Agriculture
This is a digest of a paper presented at the fifty-seventh annual meeting
of the American Society for Testing Materials held in Chicago, Ill., June
13 to 18, 1954. Publication of the complete report will be made by the
To simplify house construction and thereby to save on costs, the use of
lightweight roof trusses is becoming a standard practice. By using light-
weight trusses, the exterior walls and roof can be erected without the
placement of any interior partitions. Thus the interior becomes one big
workshop. The entire ceiling and walls can be finished as one unit with-
out interruption by partitions. If dry-wall construction is used, it
means that sheets as large as 4 by 12 feet can be placed without any
cutting. Similarly, the finish floor can be laid over the entire area
without the cutting and fitting necessary when partitions are in place.
By such construction interior partitions can be moved without affecting
the structural stability of the house; thus any floor plan can be changed
at any time to meet changed living requirements. Also, since no bearing
partitions are necessary, large unobstructed areas, in conformity with
the present trend toward open planning, are more easily obtained.
The lightweight roof trusses now in rather common use are of nailed con-
struction. When of adequate design and well manufactured, they are giving
good service. The question has, however, been repeatedly raised as to
whether glued trusses would not give better service; possibly with some
saving in material.
It is known that a well-made glue Joint will withstand a very high load
immediately after manufacture. It is not, however, known how such glued
lMaintained at Madison, Wis., in cooperation with the University of
Report No. 1992
Joints in truss will withstand the stresses caused by the shrinking and
swelling of' the component parts of a truss under service conditions. The
shrinking and swelling, caused by changes in the moisture content of the
wood, may be very considerable where large changes in relative humidity
occur. This is likely, for example, in attics that are not well ventilated
and where an inadequate or no vapor barrier is placed in the ceiling.
Therefore the Forest Products Laboratory has been making an investigation
to obtain information on the relative merits of nailed and glued light
roof trusses under simulated service conditions.
Construction of Trusses
The trusses tested were of the "W" type. The full length of each truss
was 17 feet 6 inches, and it was tested over a span of 17 feet. The dis-
tance between the panel points of the lower chord was one-third of the
span, and the Joint in the upper chord was at the midlength of the chord.
The slope of the upper chord was 5 in 12. The trusses were shorter than
usual in house construction, but they were the maximum length of a truss
that could be placed in the conditioning room available at the Laboratory.
The material was high-grade Douglas-fir at a moisture content of about 12
percent. High-grade material was used to insure failure at the Joints.
The material for the upper and lower chords, as well as for the diagonals,
was nominal 2- by 4-inch stock, the net sizes being 1-5/8 by 5-5/8 inches.
Plywood gusset plates were used for all of the panels. This plywood was
The nails used were ninepenny common of sufficient length (2-3/4 inches)
to extend through both gusset plates and the central number. This type
of nailing was used because it is more economical than nailing from both
sides of the truss.
The glued trusses employed considerably smaller gusset plates than the
nailed trusses. A resorcinol type of glue was used to eliminate the pos-
sibility of glue failures during subsequent exposure of certain panels to
high and low humidities. Pressure was obtained by clamps during the set-
ting of the glue.
In service some drying and shrinking normally take place, causing a
separation of the several parts of a nailed Joint. This lowers the
strength of a nailed Joint, since there is no friction to overcome. To
simulate service conditions, a 0.027-inch-thick spacer was placed between
gusset plates and the central member during nailing, thus creating a
separation of 0.027 inch between the parts.
Report No. 1992
Prior to test one nailed and two glued trusses were subjected to high and
low humidities. One cycle consisted of exposing the trusses to 80 percent
relative humidity for 30 days and then to a relative humidity of 20 per-
cent for 30 days. This complete cycle was then repeated, and the trusses
were tested immediately after removal from the kiln. During the condi-
tioning period the panels were loaded to a design load equal to a roof
load of approximately 55 pounds per square foot and to a ceiling load of
10 pounds per square foot.
All panels were tested in bending by dead load. The upper chord of the
truss was first loaded to a design load of 1,200 pounds, which is equiva-
lent to a roof load of 55 pounds per square foot. In addition, the lower
chord was loaded with six 50-pound weights placed at the quarter points
between panel joints, which approximated a ceiling load of 10 pounds per
square foot. After this loading the roof load was removed and residual
deflections were read. The ceiling load was left on the truss during the
entire testing. The roof load was then increased to approximately 2-1/4
times the design load, or to a load of 2,800 pounds. This load was then
removed, and the residual deflection measured. Roof loads were then
applied until failure occurred. Loading was in 200-pound increments up
to the design load, and then the load increment was increased to 400
1. Well-designed and constructed nailed trusses should give adequate
2. Glued trusses, because of the rigid joints, are much stiffer than
3. Glued trusses fall without warning, whereas nailed trusses will deflect
considerably when approaching failure with little increase in load.
4. Glued trusses show some loss in stiffness and considerable loss in
maximum load from exposure to low and high humidities. The same exposure
reduced the stiffness of nailed trusses up to and slightly beyond design
load, but with no loss in maximum load.
5. Even after exposure to low and high humidities, glued trusses deflect
less at design loads than unexposed nailed trusses.
6. Although the glued trusses showed considerable reduction in maximum
load from high- and low-humidity exposures, the failing load was still
several times design load.
Report No. 1992
UNIVERSITY OF FLORIDA
G general DiIrsc uissio nl 11111I1I1IJ l ll 11 11 JA1 l AUMIIII 111A! h11 h ll ijrllii
3 1262 08924 2621
It is well established that a glued joint is much stiffer than a nailed
Joint. Where low deflection is of importance -- for example, in the
lower chord of a truss that forms the framing for a ceiling -- a glued
truss may be of considerable advantage. It has a further advantage in
that there would be less risk of racking a glued than a nailed truss out
of shape during transportation and erection. In the manufacture of a
glued truss it is, of course, necessary to have pressure on the Joint
during the setting of the glue. This pressure can be obtained through a
press, by clamps, or with nails. In the manufacture of the glued
trusses Just discussed, clamps were used for obtaining pressure. It
would have been much easier to have used nails for obtaining the neces-
sary pressure, but then the question might have been raised whether the
efficiency of the glued joints was assisted by the nails.
In commercial production a nailed, glued truss might be the most econom-
ical to produce because nails are a simple and cheap method of obtaining
pressure on the glue joints during the setting of the glue. The glued
joints would certainly produce a truss with low deflection under load.
In addition, the nailing would have a psychological value in reassuring
those who doubt the adequacy of a glued jointed truss under large changes
Report No. 1992