Earthquake Testing Structures with Glue
I have always been curious about structures. It wasn't too long ago that I was fortunate enough to be able to build my own shop. It was during that build that I began to think about – how well are all these houses built that we live in? When you are faced with building your own roof you tend to think about the beam and rafter placement, the deflection of certain woods and how to fasten together the whole thing. So with this in mind I set out to create a little test, and answer the question whether it is reasonable, or necessary to use any adhesive in home construction.
My hypothesis is simple: A structure that is reinforced with glue at the nailing joints, wall connections, foundation and roof will be much stronger in the event of an earthquake, hurricane or other more typical stresses such as snow loads. The basic tenant I am going into this with is that when two boards are attached, with any mechanical fastener such as nail, or bolts you multiply the number of fasteners to obtain the total friction and strength of the bond. When you use an adhesive, spread out evenly along those two boards the stress is distributed in a manner that creates a more unified and stronger bond than any mechanical fastener could, no matter how many fasteners are used. The only variable is the length of time the bond would last, which is a test for another time.
My method was to have two structures built with miniature framing lumber, on foundation plates of plywood that, while identical in every way were distinguished by the addition of yellow carpenters glue at all the joints, including the sills which connected them to the foundation plates. The foundation plates were divided in half to simulate the movement of the earth. The structure itself was designed to be 1/12 scale. For simplicity the wall studs, roof rafters and ridge beam were all based on 4x6 lumber, and the spacing was nearly 4 feet on center. While this is not a typical application it is difficult to cut, and nail framing lumber that is too small. In any event it would make no difference because both structures are designed the same. The finished framed structure is 16 feet x 20 feet in 1/12th scale.
There were two main tests applied to the structures. The first was a sheer test where, using a clamp I applied pressure to one of the foundation plates while the other plate was fixed to the bench top. The goal was to simulate an earthquake, or other ground disturbance that moved laterally.
The second test was the equivalent of a projectile test, like when in a hurricane a piece of loose lumber is sent flying into a nearby structure at high speed. I set up a ball to swing into the 16 ft side where the top wall plate meets the ridge beam of the roof.
Not surprisingly the glue helps enormously to make the structure very rigid. But there were a lots of surprising results on both the glued and non-glued structures.
Even though using 5/8” brads to simulate 16d common nails was not perfect, the structure remained very much intact after both tests. If there were any inhabitants within the building they would have survived without any roof rafters falling.
The sills, being fastened with 3/4” screws to simulate j bolts into a concrete foundation were strong enough to resist the movement of the plates, while the brads connecting the wall plates to the sill plates were weak enough to bend. The result was that the sills cracked and the walls were able to float along the cracked sills freely. One of the walls caved in slightly as the sill twisted the adjacent wall and the brad nail wall-to-wall connection became weak.
Overall the foundation plates were pinched at one end due to the fact that the far sill failed and forced the plates to move, levering against each other.
The thorough application of glue at the sills connecting them to the foundation, and the additional glue added to the wall plates formed a nearly solid beam running all along the base of the structure. This was one of the substantial differences in the two structures. While the nailed building relied on those single points of contact to connect to the foundation plates, the glue formed solid beams which were very firmly in place. When the clamping force was applied the building transferred all the energy into one single point on the sill and cracked it completely, separating the wall and sill from the foundation.
In terms of the rest of the structure, the walls and roof system were nearly intact. Only during the projectile test did anything break, and then is was completely demonstrating the rigidity of structure and the upper wall plate cracked through, while in the rest of the wall not a nail was disturbed.
In terms of survivability, both structures performed properly and did not fail completely.
In terms of repair, the non-glued building needed substantial repairs to the walls, roof and most of the foundation connection.
On the glued building much of it was left intact, except the substantial failure of the long-side sill and wall plate, which would be much cheaper and more manageable.
If you have any comments please leave them below, and remember this is not a perfect scientific test. It was for fun and meant to be informative.