Q&A Jop-Site Glues
We tested nine glues for strength: construction adhesive
(MACCO's Liquid Nails heavy-duty construction adhesive),
contact cement (DAP Weldwood), cross-linking PVA glue (Franklin
International's Titebond II), cyanoacrylate (Pacer Technology's
Zap-A-Gap), epoxy (System Three Quick Cure), hot-melt glue
(generic general-purpose glue sticks), polyurethane glue
(Gorilla Glue), white glue (Elmer's Glue-All), and yellow glue
(Franklin International's Titebond).
The simple testing protocol used common tools, not
laboratory testing apparatus. We glued together nine pairs of
2x4 blocks, 6 inches long, applying the glue to only 7 square
inches of each test assembly. (Each 2x4 face received a
1-inch-wide strip of glue on either end.) Before the glue-up,
each 2x4 block was fitted with a 1/4-inch eyebolt with a
countersunk nut and washer (see Figure A). All of the glues
were applied according to the manufacturer's instructions.
After clamping for up to two hours, the blocks cured for five
days at about 70F.
A. Nine pairs of 2x4 blocks were fitted with
eyebolts and glued together with different
The test was set up as a series of elimination rounds,
somewhat like the TV show Survivor. In the first round of
testing, each of the nine test assemblies was suspended on a
chain and weighted with an increasing number of 42-pound
concrete blocks, up to a total of four blocks (Figure B). This
round eliminated two glues (Figure C).
B. The test blocks were suspended on a chain and
gradually loaded with an increasing number of concrete
C. The first test sample to fail was glued with
contact cement. The joint began to separate when loaded
with three concrete blocks.
For the second round of testing, the surviving seven test
assemblies were soaked underwater for 24 hours. Then they were
again weighted with an increasing number of concrete blocks (up
to five). This round eliminated one glue. In the final, brutal
round of testing, the six surviving samples were linked
together, chained to a maple tree, and pulled apart with a
come-along. Two glues were eliminated by this test. When the
hardware failed (the eye-bolts eventually opened up), the test
was declared over, leaving four winners (Figure D).
D. The last six glue samples were fastened to a
maple tree and tugged with a come-along (left). Two
failed, while four survived until the eyebolts deformed
Although this test was intended to measure glue strength, it
must be emphasized that the strongest possible glue is not
always the best glue. A glue only needs to be strong enough to
perform the job it's used for. Furthermore, if a glued joint
never gets wet, strength after soaking is irrelevant.
The Results: The Bonds That Held
Samples weighted with concrete blocks
Contact cement (Failed when weighted with 3
Hot-melt glue (Failed when weighted with 4
After soaking underwater, samples weighted with
Construction adhesive (Failed when weighted with 4
Samples linked together and pulled apart with a
Titebond, Titebond II, polyurethane glue, and
Nevertheless, some general conclusions can be made. Common
hot-melt glue forms a relatively weak bond. Household white
glue is surprisingly strong, even after 24 hours underwater.
And yellow glue, polyurethane, and super glue are so strong
that in most cases, the materials themselves, or the hardware
attached to them, will fail before the glue joint. -- Martin