•
THE SIGNIFICANCE OF MECHANICAL WOOD JOINT TESTS
FOR THE SELECTION OF WOODWORKING GLUES
(Printed in Industrial and Engine- ,:fe ing
emis-Gry,
Vol. 21, No. 1, Janualy. 192)
By
T. R. Truax
Wood Technologist
F. L. Browne
Chemist in Forest Products
Don Brouse
Assistant Engineer in Forest Products
The testing of glued wood joints is well adapted
to the laboratory study of such problems as the usefulness of
new adhesives or new treatments of old ones, the technic of
gluing operations, the gluing characteristics of different
woods, and the causes of weak joints in specif 4_0 -vi)od_working
operations, Wood-joint tests are not suitabL?, fcr the grading of animal glues for woodworking, because in practice more
direct evidence of the physical properties of the glue is
afforded by the viscosity and jelly-strength tests. Strong
joints may be made with any animal glue suitable for woodworking, but different gluing conditions must be used according to the grade of the glue. If wood-joint tests are made
•
with glues of different grades under a uniform set of gluing
R825
conditions, the grade of glue that will give the best results
will be the glue best adapted to the particular gluing conditions and fair deductions regarding the intri ee'n "strength"
of the glues cannot be drawn. Wood-joint tests ale used to
some extent for selecting casein and vegetable glues, because
no better grading system has been worked out and because such
glues yield strong joints under a comparatively wide range of
gluing conditions. An unfavorable joint test indicates that
the glue is not reliable under the particular gluing conditions tried; it does not prove the glue to be lacking intrinsically in "strength" or "adhesiveness."
When a well-designed and well-glued wood joint is
torn apart, except where one or both of the joined surfaces
are end-grain wood, the wood splinters over all or a large
part of the joint area. Accordingly, a sample wood joint
that breaks by splintering of the wood has been considered
by many woodworkers as a reasonable proof of satisfactory
gluing. T e chnologists, however, have revealed that wood failure alone is not a satisfactory criterion of good gluing and
have therefore designed numerous special types of wood joints
suitable for testing in machines that indicate the load under
which-failure takes place.
It has been generally assumed that woodworkinA . glues
possess an intrinsic property known as "strength" or "adhesiveness," which is significant of their value for making strong
wood joints, and that a suitable technic for measuring this
property should provide a better standard of grading glues
for woodworking purposes than tests of viscosity and jelly
strength.1
Wood-joint tests, therefore, suggest themselves
as a direct method for measuring this supposedly characteristic property of glues; Thus, Sauerl would grade glues in
accordance with a wood-joint test and the viscosity of a
standard solution. The British Engineering Standards
Numbers in text refer to bibliography at end of article.
•
8825
-2-
410
Association-3_ specifies that animal glues for airplane propellers pass a certain wood-joint test with a breaking stress of
1100 pounds per square ; I nch or more. The specifications of
the U. S. Army and Navy= for animal glue formerly included
the requirement that test joints withstand an average load of
2400 pounds per square inch before breaking with no specimen
breaking below 2200 pounds per square inch. However, this 5
joint test has been abandoned and the present specificationsare based chiefly upon a comparison with a standard reference
glue with respect to viscosity and jelly strength. Proposed
specifications of the Federal Specifications Board do not
include joint tests for animal glue.
The attempt to grade glues on the basis of woodjoint tests presents a.dilemma. If, on the one hand, the test
specimens are well designed and depend upon side-grain gluing,
well-made joints fail very largely by splintering of the wood
even though a very strong wood and a comparatively low-grade
animal glue are chosen. Under such circumstances the strength
of the wood rather
than that of the glue seems to be measured.
If, on the other hand, the test specimens are poorly designed,
and especially if they depend upon end-grain gluing, the
failure is almost certain to take place entirely in the glue .
line, but the stress producing failure may be materially less
than that required to break good joints in side-grain gluing.
Furthermore, the wood-worker objects to an end-grain joint
as it is unreliable.
g,
Rudeloff,. using a test specimen of red beech with
end-grain gluing and carefully applying a load in tension
uniformly over the joint area, obtained 1630 pounds per square
inch for his strongest joint and considered values exceeding
about 1000 pounds per square inch representative of strong
joints; variations of individual tests from the average ranged
from
16 toby123 per cent. Hoppl and Gills avoided the dllemma
presented
w ood-joint tests by breaking s p
ecimens of the
dried glue jelly in tension, but the preparation of the specimens was a long and tedious task. Bateman and Towne- showed
that the strength of such glue specimens varies enorm?yly
with the relative humidity of the air. Other workers--- have
found it more convenient to impregnate such materials as
filter paper with glue and measure the increase in strength,
a'procedure open to much criticism.
.
411
McBain and his co-workers 11 overcome both experimental difficulty and mechanical objection by preparing thin
films of the adhesives for measurement of their strengths in
tension. However, the results of such direct observations
of the mechanical properties of glues present the following
facts, which are contrary to the theory that glues possess
an inherent "strength" indicative of the strength of the
wood joints that can be made with them;
R825-3-
(1)
Animal glues of the wide range in grades
employed for woodworking do not differ observably in the
strength so measured, though they may vary in a property
called by McBain and Lee "deformability."
(2) A high-grade animal glue remains practically
unchanged in strength after a solution of it has been kept
at 140° F. (60° 0.) for a month (a treatment that degrades
it enormously).
(3) Although animal glue of the woodworking grades
is much stronger than wood joints made from it, there are
glues which when tested alone by the method of thin films
in tension give lower strength values than wood joints made
from them.
Influence of Gluing Technic on Strength of Wood Joints
Previous workers have not fully recognized that
the strength_of a wood joint depends upon the care in making
it and have assumed that an arbitrary set of gluing conditions
can be adopted in accordance with which the test joints may be
prepared, regardless of the nature of the glue to be tested,
or the use for which it is intended, The fallacies of such
a procedure are that some glues give strong joints only when
gluing conditions prevail that differ materially from those
necessary to make strong joints with other glues, and that
the wood-worker cannot arrange the same gluing conditions in
all his operations. For example, in gluing plywood panels a
relatively long assembly tire elapses between spreading the
glue on the wood of the first panel and applying pressure to
the bundle of panels in the press. In edge-gluing lumber
cores, however, pressure is usually applied to the joint
within less than a minute after spreading the glue,, The
wood-worker using,animal glue must buy a very diffc,.rent grade
for the two operations if he is to obtain joints stIong enough
to withstand the maximum stresses that the wood itself can
sustain,
The several factors -- kind of glue, grade of glue,
temperature of room, temperature of wood, temperature of
glue, quantity of glue spread, time of assembly, and pressure
applied -- may be termed
collectively the "gluing conditions,"
These factors govern the rength of wood joints and have
been described elsewhere. It will therefore suffice to
illustrate with typical results the joint-strength values
and types of failure obtained with common woodworking glues
under different gluing conditions. Two animal glues, one of
•
R825
-4-
•
high grade and one of comparatively low grade, were chosen
because most of the published discussions of glue-joint tests
have been upon such glues. Animalglues, moreover, are more
sensitive than casein or vegetable (starch) glues to changes
in the gluing conditions, and there is a recognized system of
grading them independently of the making of wood joints.
Glue A was purchased to match the "standard animal
glue" used during the World War as the basis for comparison
by the U. S. Government inspectors in certifying animal glues
for airplane propeller manufacture. The viscosity of glue
A as determined by the standard c procedure of the National
Association of Glue Manufacturers- was 108 millipoises and .
its jelly strength 303 grams. Glue B was made by the same
manufacturer that made glue A and had a viscosity of 65 millipoises and a jelly strength of 190 grams. It represents a
moderately low grade of woodworking glue. To prepare them
for use, the dry glues were soaked in cold water until
thoroughly softened, melted at 140° F. (60° C.) and spread
at that temperature on the wood with a mechanical glue
spreader. The proportion by weight of water to dry glue was
for glue A, 2.25 to 1, and for glue B, 1.95 to 1. These
ratios gave glue mixtures of approximately equal viscosity
at 140° F. (60° C.).
The wood used in side-grain gluing was sugar maple
(Ater saccharum). Details of the gluing conditions are given
in Table 1, together with the results of the tests.
Photographs Of representative fractured test specimens are shown in Figure 1. Good, chilled, dried, and starved
types of glued joints are illustrated, all four of which have
been observed in commercial woodworking operations.
•
Good joints are characterized by high average
strength and by the occurrence of the fracture, at least in
part, in the wood rather than in the glue line Each test
specimen characterized in Table 1 as "good" failad in the wobd
at least.in part, indicating that there is little or no advantage in strength to be gained by using a "stronger" glue.
In good joints the difference between the average and theminimum breaking loads observed in the specimens tested is
usually less than it is in the chilled, dried, or starved
joints.
Chilled joints are obtained when the glue congeals,
but does not dry, during the time of assembly to such an
extent that the pressure applied proves insufficient to establish adequate contact between glue and wood and to cause
•
R825
P-4
O
drj
CL)
-P •0 IP 0 ' 1
CO +3
0
• 1-1 Z .
F-.4 ti-.1 0 Tj
cr3 0 -r-a (3 I
.0
Pi
. ..4
. ..om--;
..
H
,.
a)
0
dr-I
H
-0 0
0 -1--D cd
cd
F-I
+"
--:4
•.
s4
cd
0
0
g
;-1
•0
F-1
-I-)
•
C/)
0
..
0*
o.
0.
4-) T-1 ci-i
*0
0.
H
•
..
0.
N\ CV 150 r.--
0m.
va
0.
a.
0
0.
rd
A*
0.
rd
0 0 a)
Ow
..
F4 •
•
CA
CV H
rr•
0
0
.
00
.0
m.
0.
O.
CV CU
0.
wa
Om
0
H
1 -i-'
: -r4
0 P-1
• •
rd
...-.....-.....-...--.. a)
0000 a)
a) o o a) 0
CI P Pi Pt. r-I
0 0 0 0 0
-..........-----....- n.--
g00000
•
0.0
O•
Am
CU LC'‘ re\ CT CU
L.C1 CU I•rl reN CV
ri
'---,
f•i
H
0.
• ri.
0*
00
6*
0.
*a
EEE
0 0 0 -I-' •
4 >
rd rd ,c1 bf. cd
•ri . 1-4 •r-+
0) 0
0
a.* F-I CI)
00
0.
0w
PO
Q.)CO
F-1
..o,
.
1
n3 4-t
F-1 c) cdl
(1)
0
P, a) 0
9 9 9
.
..wm 00 00 00 0 0 0* m• 0. 0. 0.
ix,
0
l'i
00 00 Op
I 0
b0 0
-H -1-.4
CI) +'
.. ..
0 cd
A 0
• • • • o. .. ..1.
0
N 0 0 N-LC\
H
0
0 -I-) a)
0
W
I X
0
9-1
CI)
CO
0 N 0 pcn t•fl
11.0 00 Om 00 00 00 00 00 Ow 00 I-4 FA I-1
I-I I-I I-I
1-I I-1 I-1 I-1 FA
I-1
1-1
I-I
I-I
I--I
0* 1.0 00 00 00 0000 0. .. 1" •• .. " ••..
CO
cd
0.)
0
a)
4-3
4-3
1.)
-6-
4
4)
0 Cd
F-i
0
rd Ti
cd 0
0 0
fe)
P,
co
g
CO
•
it,
0
cd
0
04-i
0 -
-4-
4
1.-1
Fw
F-I
rd
cp
0 -P
al 0
r-I ti)
0
H
-r-1
4
CO
P, 0
cd (13 =
H a) rd
0 E (I)
ca
a)
0
0 r--I r-I
.0 0
+3 0 =
0
a) ca .,
r0 CO a)
-P cd 0
•H
•1-1
bl) CD = -P
-1-1
1:11 (X) Ili MI P74
4-4 -P
(i)
CIS
of)
-.4 •.4 •:4 -.4 --.4
-P
..
rd
a)
F4
H
4-4
CO -P
'ti
H
N
F-I
001),(.--!, - P
4. 0 -4-D cl.1
.,-1
of 00 00
E
0
4'
-I-)
CD Cd
04 4
0 4-'
a)
H
P.
F.C3;
P,4
co
0 0.)
.1-1
4 ›.. a)
H
4-3 +' 0 -I-3H P,
0
a3
0
E
a)
F-i
o
ca
M
0
0
0
0 ›.• r-I
¢,r-1 ,O
0
r0
1-I
CD -I-I Q)
4-1
=
›••
CD
H • co
,C) -P a3
E 0
CD 0 ,C CO . r-o -I-)
cd
(ll cri
,0 0
0
-P 0.)
o
co
F-I
-H
M0
a) 4-3
co a)
a3 F-I
0
H
0 •H
0 cd
0 c4-4
•I-4
0 4D
a)
P.I
-P
CO
-I-D
0 0
o •r-I
14
0 -1-3
F-1 g
0-, r0 '1'1
0
F-I 4 -,--)
0 0
H0
E-4 rz c
I r•I ,of oi
a)
p144
a3 +' 0
0 a) -I-I
a) -0
0 4-3cd
Ti
0 0 0
Cd
CD „CD
,-1
0
rc:I cd
rci 0 H
a) 0 P,
0 --1
H 0 ,M
tio as 0
0
0
k -P ,•-L'
'60 150 CU '60 '60
C)
R825
r.).., w 4:1
ai o a)
F.-1
-,--1
CD
r-1
- H
to
150 150 CU '60 150
a)
-P 0
9=3
g
0 H
0
4-3co 4-3 a)
P, 0 CO
-P cd 4
.H (1) +3
0 -P
0
co -H
0 a
cd 0
cd tiD
4 0 4.-3 4-4
c) qd-t 0 0 CA 0
Cb • 1-1 F-1 0 F a)
r-4 0 1-t d H
to
0 .0
mu 00 a* 0.
E-4 -1-D
00 00 00 00
,0
r.4
F-I
Cd
a)
+'
;A 0
4 a)
C.) 4 f-I
cd 4-3 a)
-4-D4-4
0 0
M 0 0 bD a)
0
rd •ri
a) H
0 P.
,-1
sa,
a) cd
I
'60 150 CU 150 CU
F-D
+3
H
•1-I
0 N 0 N\ 0
r-I
0
0) •
-P a) 4
4:1
cd
.
F-1 0
a) 0 F-I
0
-1-3
0
4
bD -H cd
r-I
Fl
a)
4
o
0 a)
0 -0
•ri 4-D
•--4
a
0
r..4(Ni 0 0 N-0
tO tO N tO N
a
I-I
r--1
-4'
-P
g
00
•H -,-1 • .1-1 > ›-
00
w
•
F-I•
+3
0 4
eg)
›, 4-3 O a) et-4
cd
F-1 ,c) 0
H 0 0
0 0
P-d 4
TS rd rci cd cd
dri 0)
0 o 0 o
1-4 M.4
01 01
.
•
C5 4 a) F-1
04-4 P, 0
F-1 0 $ -P
a)
.0
+
W
ma
Cu -H
0
0 0
0 • •1-1
0
CD
=-i ti
1rd0
1
.. .. .. .. . .. .. 00 00 00 or 00 04. 00 o. *0 6. mil. 00 60
, 0
•
' 0 0 0 0 0
0 0 0 0 0
I 0
fa0
0
0
CU CU CU CU CU
CU CQ CU CU CU
•.-.1
tip0
•-4
I-I H
H
r-.1 1-4
r-I
0
-r-1,zi E-4
,-_,
-ri
•-3 o
....................................
0
a) 0 I
H
-P
0
..-I
,C1
rd
a) cd
0 o
H F-4
H ta0 cd
0
00 00
O. O.
cd
O r-I 0 : 0
0 •r1 'CI
• H
CU r-I.-
0 0 Q) a.) o
0
VP- P- P, P-r-I
0 0 0 0 0
,:s
-.....--..----------_, ..-
.. .
4-D
cd
FA
0
rcp
. H
0 0 0 0 0
0 0 0 Lc\ 0
CU CU CU ---
0 0 0 0 0
•
E-I Ca
, CD
0
00 00 SO 00 0* **
..-.....-....--....-,. ti)
,-C)
..P
0 0 0
0 in 0 0 Lc)
H Cr\ N- CV CU
Isr\ CU H Ise"\ CU
,e3
o F
0 a)
-r4 M
4-1 EH
O 0 ,g
-4-D
-r-1 0 rd -P
H
4 0 a)
40 -P H
o--/
•
OV O.
0 H
Cd
0 0
0 0 0 0 0
of
ti--1 ?,
bD CO 0
CD
a)
'0 F-1 bD
0 -0 0
ta0
me 00
•
•.
Ln 0 Lf1 O
LCIIn
crn LC" \ L.C1N- 1"r" \
te's CU r-I CV CU
CO 0
CD
Om
CU
r-i 4-1
.,-4
Ca, o
al H
P. -P U)
CU
.0
-P -1- 1P 0)
rd 4
0
4-4 H
0 0.1_, LH
cd 0 0
0 bD
Om dbw 00
6-1
F-4 CO
114
o •rdl 0
0 cd H
Fd-I b.0
bD bD
0 I id-4
dr-I 0 Cti
0 rd 0
,..-1 9-I .1.-1
0 0
P •,-1
PI
F
ow
r`-- LC1150 N h-
\-0 LO H re \ H
VO 0 01 tr.\ rcl
00
cd •
0 4-1
F.-1 r-I
Cd a)
r--I
4-.) 0
cd 4
0 -P
• r-I a)
r-I
CV CV
0.
9
0
(IE5
0
0
4-4
0
• H F-i r-0
150 h-
r-I 150 CU 1--1..-
M. H
H H
•
co 0-
.0 CD
CO 0.
dri b.() 0
a)
t-I 0
ci-i •t-I •r4
_- LC'_-A- LC \ LC\
Cr) N r-1 rr-) H
•
r-I
•
P,
•
a) g
P- -1-I
•0 •• akif
0 -r1
0
cd
E-1
m•
In N-hr'.4- LC1
LC1 CU 1---1 Fdr \
CD CO
-0
0 rd
Oro"
a)
aa
qi.r.-4
•
• CS
/-1
0*
0
H
mo
a g
Ow 0.
U)
r0
4-1
0
.0
F4 4-)
0. 41
I`r \
.. ... .. .. .. ... .. .. .... .. .. .. ..
'0 I
c)-' I a)
a)
ho
Hco
FA
o
rd
a3
0
.4
drI 0
E-4
0.
a
F-I 0
H cd cd
4-1 4-1 0
O
.,-4 cf-i
o
00
a)
P-,
,-4
0 'd -H
0 F-4
0 0
CD CO
fl
LC111\0 0 0 '
r-1
F-1
0
CO P, a3
n3
a) 0 -1-4
m0
-r4
(-I
.0 0 0
4 rd
-P F-i it
Cd F-4
0 h0 elf)
0 0
CO ,.
CD
LID
0.
0
F•r
0 rd
0 C)
we 00 wa 0*
0 0 0 0 0
g
a)
0
bit) 0 0
rd
0 -P H H o
0
0*
CD
U)
0
-I-D
•r-I
R
+3
-P W
a)
P-t
9
.0
rd rd 0 1-1 F-4
0 0 •r-I •r-I CZ
0 0 H ..0 4-D
0 -I-D -P
P-1
rH
bD O 0
00 00
H
rd 1-4
0
a
•..i
`4::) i
a)
bu
cd
;-1
a)
O
0 0
-4-)
0
P-1 CO
›-.-P
-p g •
•r-1 0
00
a)
rd rd
0 0
C. CD n C) CD
.. .. ..
0
0 00
rd '0
0
H -ci f--1
rd H 0 r-I F-1
0 -r-I -rt •r1 it,
0 4 F-4 .0 -P
F-I
CD
rd
•
cd
r.r.4 rxi
-di 01
410
the glue to flow into a uniform film. Chilled joints usually
have a thick layer of glue, which may either shatter or pull
away cleanly from one surface when the joint fails. They are
more likely to be obtained in practice with a high-grade than
with a low-grade glue. A cold glue room end cold wood, a long
time of assembly, and a low pressure favor chilled joints.
Of the faulty gluing conditions, that causing chilled joints
is most easily overcome because the pressure may be increased,
the wood p reheated, the time of assembly shortened, or glue
may be placed on both surfaces to be joined instead of one
only. Only hot glues are subject to chilled joints.
Dried joints are obtained when the glue loses enough
moisture during the time of assembly through evaporation or
through absorption by the w ood . to,harden appreciably. The
fractured test specimens exhibit characteristic glossy areas
where contact between glue film and wood was not established.
A warm glue room, warm wood, and a long time of assembly
cause dried joints. All common woodworking glues are subject
to dried joints.
Starved joints are caused by the glue penetrating
excessively into the wood or by excessive 'squeeze out li of
the glue when pressure is applied and, contrary to a common
belief, they are favored by a too generous rather than a too
stingy spread of glue. The broken pieces of a starved joint
show ltle
ilt
or no glue between the joined surfaces and microscopic--1 observation of a cross section through the joint
reveals deep penetration of the glue into the wood. Warm
wood, a heavy spread of glue, a short time of assembly, and
a high pressure produce starved joints. They are more likely
to be obtained with a low- than with a high-grade glue. Hot
g lues are more subject to starved joints than glues that are
used cold because the latter usually are less mobile.
'
It may be noticed in Table 1 that some ,different
gluing conditions were chosen for glue A than for glue B.
This was necessary because the range of gluing conitions in
which joints of any of the four types are obtained varies
with the grade of glue. The wood-worker is concerned with
knowing which grade of animal glue can be trusted,to give
good joints under the gluing conditions attainable in his
particular operations. With the woodworking glues in common
use, faulty joints are far more likely to result from maladjustment of glue grade to gluing conditions than from inadequate adhesive power inherent in the glue itself.
R825
-7-
S
Since several of the factors entering into the gluing conditions are interdependent and compensatory, the woodworker has a certain latitude in adjusting them to meet the
requirements of various operations without the necessity of
changing to a glue of different grade. Changing the water
ratio for a given glue has somewhat the same effect as changing the grade. Table 2 shows a few of the many possible gluing conditions that give good joints with sugar maple using
glue A.
Glue Grade and Joint Strength
Proposals to determine the intrinsic strength or
adhesiveness of animal glues by means of wood-joint tests
usually set forth a standardized routine of gluing conditions
to be followed with all of the glues tested, except that some
of them permit adjusting the proportion of water so that the
glue mixture will have about the same viscosity at the temperature of application. Table 1 shows that such procedure
is fundamentally unsound unless the chosen routine of gluing
is definitely related to the p articular manufacturing conditions for which a glue is being selected, for no one set
of gluing conditions will p roduce joints that are representative of best practice with glues of every grade. Thus we
find;
Glue A
•.
Glue B
Under gluing condi : tions designated in : Average : Wood : Average • Wood
Table 1
: joint :failure; .joint : failure
: strength,:
: strength
:
:
•
.
: Lbs.per :Per cen t
: Lbsope_r
i-er cent
. sq.in..1;
sq.in,
:
..
.
•
I
:
3950 :
55 :
2950
75
II
2500 :
27 :
3100
84
9= 14.22 lbs. per sq. in. = 1 kg. per sq. cm.
Under one set of conditions glue A seems listronger,11
under the other, glue B. Neither glue appears to best advantage when applied in accordance with the best practice for
the other. With glue A chilled joints resulted under
•
R825
-8-
rd 1 m/
o r-I
•
o
m
a)
F4
i
C‘.1 O ri f r1 N---
rrm.---isok.o N—'50 +5
H
I's—
0 0
• Oe 00 0. 00 ee 0•
qi
f-r
0
•rI
Cr
a)
O a)
r-I
wo •* • 0
R:1
ID
0
•:4
•
P4
CO
4-1
*a
be
w
se
PA
a) G
ID- -H
.
u) 0
ee
we
.00
0
ee
OO
....
F-1
..-1
0 0 0 0 0 0
0 0 0 0 0 0
al
CU
cr3
P.
CU CV CU C1.I
a)
,P ta
0
o
Fl
00
-P
•
ti
el. 0• e•
se
rd
11.
...
N...
El
0
F-i
4-1
171 H
4-3
0.
•e
..... ►
0•
ee
go
00
e e g g e e
0
0
0
u)
0
OOPPOO
0 a)
H F-I
.4-1 H ,4 ,4 ,-1 H
a)
a) a) a) a) a) a) %
,-1
275 Ti 2:1 2:1 Ti zi
0 P_
cu
es
Ow
64
r-I r-I
a) cu
0
EV
----,--,.....
-,
....,
0=4
e•
WO
CU CU I'41 LC1 C1.1 1S0
-I—I
+5
*1 *.'
S.
0.41, .0
Oe
I
F4 Q.)
21 0.) F-I
0.
ee •es
so
Oe
0.
ea.
•ea
0
a)
P•
CO
-1-)
op
.
1
o0 0-P
.
0
4.0
•• ••
/-1
•.se
6.1.)
-I-D
ea
0 0 0 0 0 0
- H 0 0 irl 1.11 0
• • \-0 0 Crr r--- O r--
CO
NTh-..1-
Pr
rr's
ro w
o
bD
asa) bi)
• 1-4 CO
CD
0
n0
CO
•
bu
C
f-•
F -+
bp
111 7:1- N._ Lc.,
Prl N NM re1 CU N1
es
N.
CD
cd
• •
A-)
rrn Lfl ti50 Lrn 0
rcyco \x)r--
• 0'
01 CO
e.m.
4-3
•H
se .0 04,
F- 4 g
a) .1-1
P.
•
1
0
A-)
0
Os 41 .
so
•
a)
F-1
u)
a.)
ta0
P-1 •
Os
a)
+D
11100000
M
CO
a)
r*--- I'--150 150 01 01 0
-I-)
CD
1.
•
1114.es
CU
ea
0.1.
6.
0.
so
.4
.0
0
a)
000
E-4
0
0
150 01 01
0
a)
CO
s•
.•
Orr
ar
04
cd
e.
as
r-1 H rl ri
1111
CU CU CU CU CU CU
46 64 64 op 04 4P4o
•
r-1 H r-1 r4
R8 2 5
CO
a)
cdl
410
conditions II, which gave the best results with glue B. With
glue B the joints made under coAditions I, though classed as
good, bordered on the starved condition.
If the two glues are compared when each one is used
under its most favorable conditions, the chemist, noting the
higher average load at failure, is likely to conclude that
glue A is "stronger." The craftsman, looking at the fractured
test specimens and observing the greater amount of wood failure, would favor glue B. The differences appearing in the
table for these particular tests are due mostly to variation
in the strength of the maple blocks, because the wood failure
in both cases is high enough to indicate that the strength of
the wood rather than that of the glue joint governed the load
at which failure took place. In support of this interpretation the results of other tests may be cited. Eight pairs
of maple blocks were glued with glue A under eight different
good gluing conditions; ten test specimens from each joint -eighty specimens in all -- gave an average load at failure of
3300 pounds per square inch with 62 per cent wood failure and
a minimum load of 2035 pounds per square inch. Another series
of eight joints was made with a glue of still lower grade than
glue B, the viscosity of which was 42 millipoises and the jelly
strength 150 grams; the eighty test specimens gave an average
load at failure of 3500 pounds per square inch with 47 per
cent wood failure and a minimum load of 1927 pounds per
square inch.
What Wood-Joint Tests Tell
•
From these co nsiderations it is clear that woodjoint tests in side-grain gluing are not suitable for the
routine grading and evaluation of animal glues for woodworking. When used under suitable gluing conditions, all the
grades of glue used in woodworking give joints whose load at
failure depends chiefly upon the strength of the wood joined.
By testing each glue under a range of different gluing conditions wide enough to establish the zone of conditions within
which it will give good joints, it would be possible to group
animal glues in classes c
orresponding roughly to the grading
on the basis of jelly strength and viscosity as now employed.
But such an arduous task is one for the research, not for the
control,•laboratory. Viscosity and jelly strength, which
are quickly and fairly easily measured with precision, give
essentially the same information. Viscosity tells the woodworker how much water to add to get the right amount of the
glue on the wood at the least cost. Jelly strength, in so
R825
-10-
•
far as it is proportional to the chilling point, indicates how
long it will take to reach the right consistency for pressing
under the conditions prevailing and how long it will remain
in that favorable condition. The two properties together
serve admirably for checking successive shipments of glue for
uniformity and for comparing competing brands to insure
economy in purchasing.
On the other hand, the wood-joint test is essential
in the glue-research laboratory for studying such problems as
the practical usefulness of newly p roposed adhesives, the influence of new treatments upon or additions to established
glues, the gluing c haracteristics of different woods, the
technic of gluing operations, and for examining glued wood
products for their reliability or for the location of the
causes of manufacturing difficulties. Direct Measurements
of the strength of solid specimens of glues may not, in the
present state of knowledge, be substituted for the wood-jo:.nt
test because it has not yet been shown that the "strength 4 of
a glue so measured bears any essential relation to the strength
of wood joints made with it, nor do they appear to have a close
relation to, he viscosity or jelly strength of the glues
themselves.
Wood-joint tests were adopted by the U. S. Army and
as the foundation of specifications for caseinglues.
Most attempts to test vegetable glues also rely upon woodjoint tests. Some r epresentative results with these glues
are given in Table 3 and Figure 2. At present wood-joint
tests seem necessary for these types of glues, because knowledge of their significant physical and chemical characteristics is not yet sufficient to permit the adoption of more
refined testing methods. F ortunately, casein and vegetable
glues are much less sensitive to changes in gluing conditions
than animal glues and good joints are obtained with them
through a fairly wide range in conditions. There is no danger
of obtaining chilled joints with cold glues and starved joints
are c
omparatively rare. Dried joints, however, are often encountered.
Navy15
Navy
When making test wood joints with any glue it is
well to bear in mind that, while a favorable outcome proves the
glue capable of making good joints, an unfavorable test does
not prove the contrary. Before the glue can be rejected as
inherently weak, it must be tried out under conditions representative of the range under which it might be Practicably
employed.
•
R825
H
- P 0
o
r0
C.)
P.
q
,
0
E
oi
0 - r-1
P
F-1
0
- i-1
H
sA
+4
ct.
4-1
0
Oo0
rd CD
-P
W
•r-1
O
4-)
•
01)
0)
cf) H
T-4 a) be
-P
U) 0
(3)
P
CI)
rD
cd cd
0 C.)
o $.-1
cd
co bD
P
O
P
1
0
4-3
r-4
4-
a) bD
rd
Pr--.1
0 bD r-1
0 ,Q
0 03
P cd
-r-1 F-1 bD
b0
bO O)
rd 4
4-i -P
0 CD .H
Q) 4-I
0
a) a)
P
F-1
r-1
4-1 r--1
-H
•H C
a)
E-1
•
1'41
1)
Cd
E-1
i
..
-P
0
a)
0
F-1
0
P-1
..
-1-)
4.o 4...
rd
ce
rd 0 F-1
0 -H CG
0 F-1 +)
0A CO
o. .... a. D. ...
in 0 0 0 0
1.11r-I
0\
op
n
e. .. .... .. m. .. .... an ..
C3\ 0..d- 150 r-1 0
0\ N 15,0 Nk.0
F-1
U)
PA
OS • 00 0. 00 00 me Ow 00 Ow 00 0.0 .0 00 4.0
0
M
cd
0
F-1
4-i
0
rd O
0
4-)
0 • H
0
CD
a)
f:
0
(1)
0 tH
tH -P
o 0
bD
0 0 0
0 -P 4
+D
CL•4
+) -1-)
. ..
cd , g tH
0 be o
rd
-1-1 0
0
H o be
OS
4-1
P-- F-1 g
p ra •n-i
0
cd
P
co
-P P
H
cd 0
(I) CD
F- 1 - P
CC, •4
H 0
cd 0
-P •1-1
o r-i
-p
a)
a) 0
4 .--1
+' bD
rei 0 •1-1
g g tai)
cc,
a)
a)
4-4 a)
0 4
-P
GI) r0
0
rd F-1
P
g
0 0 a.)
0
4• ©1 • ri
+D
• ,.,,
+D
F-4
rd
. r8
0 0 0) -P
4 g co
-P P-,4
P o +D
cc, .4-D -,-1
•1-1 k
0 C.-i
r-I
0 P., F-1
o - r-3 a)
a)
..V -1-'
r0
(1) ,C4
0
0 -P
,g 4 -P
•1-1 0 r0 Q)
r-1
bn
k
rd
.4
4
-P 19 o OD
bD +D -P
c)
cd 4-' Hmi
0 4-♦
r0
H 0 CD
0
r
0 co
..
rd l
4-4 •
-1
P
CD rd
0)
F-1
r--1
r0
0
.0
-P
-H
-1-)
cu
P
•4
4
4-D
=--
H .
0 g 0 0 0 000
0 ill u-1
fa, •.-1
0 0 0
0110 I'+1 '60 150 l'el
• •
r-1
CU r-1 r-t
N
CO 0'
9
•4
,,.,..
F-,
:-..
-H
bit}
P
cd
a)
H0
a)
-4
'CI
rd
rd a) F-4
0 - H CC:
0 F-1 -P
0 CI CD
-1) o
cu -P P 0
F-1
-1-1 rci
CD 4-i 0 0
.4 o .,--- F-1
0
(1)
r0
rd
a)
rd
0
0. 00 00 00 ma
F-1 •
(1.) g
0
00
.
'0
0
-I-1
.0 or 00 0.
000
4-D
(I) ,. 0 4
+D
-r-i
-4-D 0 bD
•,-i
cd
-P 4-i cH g
ct 0 F-i F-,-1
p.t
0 r0
cd 0 CD
P. 0
a)
E -H o F-1
0 -P 1.--7 CD
0 cd + 0 'CI
0
0 1.1) 0
F-1 'd
Cd
4-I 0 a)
F-1
P
.0
0 0 -P
a)
co d-'
rd
-P
cd
..0 CD
M -P
cd
o
-4
•-A
cd al 4-1
-P E 0
i-
}
.. •
.... • • • • •• -- ••• So •* • • :;
•1-1
a) rd
CD -P td)
F-1 •
tH
F-1 (1)
a) g
0
0
g CD 0
0 (I) -H
cd H a)
0 -1-1
Pi-r1 0 0 0 0 0 0
r-1 ,0 0
,--1
CD r-1
0 0 0 in0 0
0
g
co p.,
• •
N N...-
N Cvl--
0 • P P,-P ,-0
•H g a)
-H (I) -H P.,
0 RI,
-P
0 cd -P
Cf) 0'
cd -P
0
g
-P 0
H cd
.0 0
ct, 4-,P. zi •,-4
$--i 0 rd
P4
I-1
.. .. .. .. .. 0. db. .0 Ow .0
lo
o g 4 CU
CD 0 H
cd P.
P. = 0 0 -P r0
01
'd
o E
a)
-P
-1--'
rdrdrd
CD
a)..---, Q)
0- (1)
C/3
k 0
41
.
o...,-c o -1-) W
CO0C0
CD0C0
0 r-♦
0 =
0 H
ca
ed 4 0 co rd
0 0 o
-4-D r--i ra
o 0 o
,0
0
• -1
0 0) -P ,.0 (1) co
r--1 P - r-1
r-1 0, r-1
ad
P
0
0 o
irN, 0 0
O
0 a)
0 0 c) 0 o 0
_•
4 0
--- ----- -,-- -------_,----
-H
•H W
CD
0
-P tz
0
0 bD
-P
E-4 CO
-P g -.1.--i
rd -,--I cd
-H
-r4 H 0
Cd
LC LCl cv CU LC\ r-4
P Cd
H
P
rd CD r0 =
H re ----,
o
cd
rd
H r--1
P.
rd
H
P- g a) P., 0
P
w.
..
41..
v.
a..
..
..
d..
40.
Om
es
Oh
•
•
•
•
.•
rci
CL)
M
^
CU
CD
0
CU
-r
I
00
(1) g o
.P/I
4 1--4 H En (1)
cCD-+?0. 0 ca v
0 -,-1 o
rd
-H - ›.
-H -H rti
ta0 , a) ccd
1H H -P
P
cd -H
'CI
P
0 I-I Co a) -P
bp 0
rd rd ti rd rd cd
0 a)
x a)
.r4
r.-1 F-1
0 (1) 0
0 0) 0
0 a)
=
0., -P CO +4 g >-
0 CI) F-1
cl,
M
::,4 M
CO
F - 1
0 (0 tt -1-) 0 r-1
H
0
00 04 00 . h. 00
(1.) CO r--1
0
rd
P „0
0 o 0
I
cd "H
-P 4-1 ,
0 0 0 4 = 0
0 0
cd
CD o di
M
4-4
0
F-1
0
-P
CO
0 0
-1-1
(1)
•.-1 •1-1 la
•1-1 -H P
H 0 H
CD
M 0
Co
0 0 0)
be CO • cd
rd rd q-4
FAO
0
rd rd -H
0 0 4 0 CD .Ct
P 0 r-1
0 0 4
P 4 - '
00
a)
cd
O g biC,
,•A
`- E.
''A 1.1
H .r-I-P
CD CL)
..
......................4D -r1 -P V o 4- Da)
4
a)
;-I
CD
cd
F--1
a.)
g-1-4 0 0 0 0 in in
N r N rn N r.-----
1.4" \ C \ .1 C l I
• .
rc \ CV CV
9 H ›-. 9 9
9 9 -
cd
I
Cd 4-1
H 0 0
0
o
r..•
P. a) o
0 $.4 H0
a) 0
El +D
.. . • 00 00 00
0
0
H
0
0
000 000
w 15o 'to 'Go 150 150
0 00 00 00 ow
4. 0. 0. 0.0 00 v. 00
0 0 0
•-1 r-1 r-1
r-C) ,-0 rC)
C0 cd cd
A-1 4-1 +)
a)
bD bD bp
a) a) a)
0 0 g
• -1 +4
o ce 0
0) ca cr)
•1-1
cd cd cd
0
0 CD
-12-
•f---3
4-i
a) 0
a)
p, 0
I-1 CO -r1
0 0 CC
P. 4
• cd -i'
o .4 r--1 r0
0 +' r.0 -P rd (1) ra CO
o 0
0
0 g 0
0 ''b M 0 0 (1) -r-Ii 4-i -1-) • H
4-1 -P -P =
0 0 P M -P -P 0
a)
-P P cd ca
0 -P
Co a) o cd a) 4
o 0) g
-H -P
H+.11-1 F-1 M -H
Ch CD
EI-4 0
,O •-I 0 0
P •4 == 0 CO H M cd H.
0 CD
,
CO -H 0 01 ..0
0)
O 0 y (D 0 -P CO
0 cd P, F-1 0 o k 4 4
a) 0 a) ..-.1 0, 0 cd 0 0+D
4
E-I
01
It E-i
Ir•--
--!:
rol 01
o cd
ix4 W
rdl (DI
•
Procedure of Forest Products Laborator in
Making Wpod-Joint Tests
Two types of wood-joint tests are made at the
Forest Products Laboratory. The test of plywood specimens,
designed especially to determine the i p,ter resistance of
glues, has been described elsewhere- The test of lumber
joints in side-grain gluing is made as follows:
Commercial 1-inch sugar maple (Ater saccharurn)
lumber is used and wood of average density or higher,
straight grained, and free from defects is selected from it,
This wood is seasoned thoroughly and stored in a room kart
at 30 per cent relative humidity and 80° F. (27° C.) un-c21
needed. The moisture content is then about 7 per cent of
the oven-dry weight of the wood. Just before gluing, -1.)4,%cs
3/4 by 5 by 12 inches are cut and planed to uniform 1,-(],if_7
and smooth, true joint surfaces. The temperature of -61-,C
gluing laboratory is subject to control within the ran g e of
conditions likely to be encountered in practice, and a steamheated chamber is provided for preheating the wood when so
desired. The two pieces to be joined are weighed together
on a balance, glue is applied to one of them with a singleroll mechanical glue spreader, and the two pieces with the
applied glue reweighed to obtain the weight of glue spread.
As a rule about 1-1/4 ounces of wet glue are appIjed per
square foot of joint area. The second piece of wood is put
in contact with the glue line at'once or at the close of the
time of assembly according as open or closed assembly is
decided upon. At the end of the allotted time the joint is
put in position in a screw press under caul boards, a
hydraulic pressure gage inserted, and the proper pressure
applied. Usually the pressure is 100 or 200 pounds per square
inch. If it falls off immediately after application by reason
of "squeeze out" of glue, the pressure is readjusted at once.
•
As a matter of convenience joints are left , under
pressure for 16 to 24 hours.' A few hours Would be sufficient.
They are then conditioned before test for 7 days in a room
at 30 per cent humidity at 80° F. (27° C.) and provided with
good circulation of air. From each joint ten test specimens
of the dimensions shown in Figure 3 are cut and torn apart
in the machine (fig. 4) and shearing tool (fig. 3) illustrated.
The speed of the machine is such that the shearing blade advances at a rate of 0.0157 inch per minute. Attention is
called to the self-adjusting bearing in the upper jaw of the
shearing tool, the purpose of which is to distribute the load
R825
-13-
•
uniformly over the upper and lower shoulders of the test specimen, even though they may not be truly parallel to each other.
The part P is a semicylindrical-shaped rocker having a flange
that fits in a groove of corresponding shape in the upper jaw
of the shearing blade. It is free to rotate around the axis
of the semicylinder within limits set by means of a pin inserted in the flange of the rocker and sliding in a groove
in the shearing blade. There is also an oil hole for lubricating the bearing.
For each specimen tested, notation is made of the
breaking load and the estimated percentage of the glue-line
area in which the wood splinters. Two or more duplicate
joints are usually prepared, each one giving ten specimens
for test. The average and the minimum breaking load anc_ ;he
average percentage Of wood failure are generall y taken
t]ie
final record of the test.
Conclusions
1.
The strength and character of failure of test
wood joints made with animal glue are greatly influenced by
•the gluing conditions under which the joints are prepari7ed.
2.
The conditions that produce good with the grade of the animal -glue and no one set
conditions gives good joints with all grades of gla;,
vary
3. Animal glues of all grades commonly ud for
woodworking, when applied .under conditions suitable for each
glue, produce wood joints whose load at failure depends chiefly
upon the strength of the wood rather than upon the grade of
the glue.
4. It has not yet been shown that animal glues
differ among themselves in an intrinsic property of "strength"
or "adhesiveness," which is as significant of their value for
woodworking purposes as the viscosity and jelly-strength tests
now constituting the principal bases of grade.
5. Wood-joint tests are used in the inspection of
casein and vegetable glues becnuse a more refined technic for
evaluating them has not yet been developed. Although a favorable wood-joint test proves that the glue is capable of making strong joints, an unfavorable test may not prove the
contrary.
•
R825
-14-
S
6. Wood-joint tests Pre well Pdapted to the research
laboratory dealing with problems involving the technic of
gluing and to the engineer in checking glue-room procedure.
The tests should be made in side-grain gluing p nd the nature
of the failure as well as the load at which it t a kes place
should be considered.
Bibliography
1.
National Association of Glue Manufacturers, Ind. Eng.Chem.
16, 310 (1924); Alexander, "Glue and Gelatin," Chap.12,
New York, 1923; Bogue, "The . Chemistry and Technology
of Gelatin and Glue,`" Chap. 10, Nevi York, 1922.
2.
"Kolloidchemische Technologie," edited by Liesegang.p.882,
Dresden and Leip z ig, 1927; Kolloid-Z., 33, 265 (1923).
3.
First Report of (British) Adhesives Research Committee,
1922, p.19.
4.
U. S. Army Specification 98-14,000 C (1923); U..S. Navy,
Bureau of Construction and Repair, Specification 11D
(1923).
5.
U. S. Army Specification 98-14,000 C (1924); U. S. Navy
Sp ecification 52 G 4 C.
6.
Rudeloff, Mitt. Materi paprIfungSamt Berlin-Lichterfelde
West, Nos. 1, and 2 (1918).
7.
Hopp, J. Ind. Eng. Chem., 12, 356 (1920).
7,
8.
- Gill, Ibid.,
9.
Bateman and Towne, Ibid., 15, 371 (1922).
10.
102 (1915); 17, 297 (1925).
Hauseman, Ibid., 9, 359 (1917); Heinemann, Chem.-Ztg.,
24, 871 (1900T; Neumann, Kolloid-Z., 33, 356 (1923);
Bechhold and Neumann, Z. angew. Chem., 37, 534 (1924);
Horst, Ibid., 37, 225 (1924).
11.
•
R825
McBain and Hopkins, Second Report of (British) Adhesives
Research Committee, 1926.p. 62; McBain and Hopkins,
J. Phys. Chem., 30, 114 (1926); McBain and Lee,
J. Soc. Chem. Ind., 46, 321 (1927); McBain and Lee,
Ind. Eng. Chem., 19,1005 (1927).
-15-
•
•
El, 80 (1929).
12.
Browne -nd BT01180, Ibid.,
13.
Truax, Timberman, December, 1922; Jones, Proc. Am. Soc.
Testing M p terials, 23, Pt. II, 583`(1923); Truax,
Furniture Mnnuf. Artisan, May, 1924.
14.
Browne rnd Tru n x, Colloid Symposium Monograph, Vol. IV,
p. 258 (1926); Weinstein, Ibid., p. 270.
15.
U. S. Army, Air Service, Specificrtion 98-14,020 C (1925);
U. S. N'vy, Bureau of Aeronautics, Specificrtion 52 G 8
(1924); see also Report No. 66 of the Nrtional Advisory
Committee for Aeronautics, 1920.
16.
Browne y nd Hrubesky, Ind. Eng. Chem., 9, 215 (1927)
R825
-16-
S
4.
•
r
1. 5hrAking •
appearanoe cf th,
epectimens of four Lype p
wcce
with lov grade animal gJ 7 2'
hili ves of five apeoiment
'raot-ary ir.
.de with h.
is mai3e
H E wood :oint.
N13630r
MIMIONIL MIME
S
VEGETABLE GLUE
CASEIN GLUE
GOOD JOINTS
DRIED JOINTS
PA el 2. -\
STARVED JOINTS
Fig. 2.--Showing the appearance of the fracture in broken test
specimens of three types of wood joints made with vegetable
and casein glues.
(Each group is made up of the two halves
of five specimens
cut from a single wood joint.)
ql,
Fig. 3.--Shearing tool, and test specimen
used for Forest Products
Laboratory's glued woodjoint test.
1
•
r15302F
Fig. 4.--Glue-joint testing machine.