Load-Bearing Cantilever Truss
105.5 g balsa cantilever that took first place, with rivets placed by analysis
What I did on this
I shared the force analysis and the comparison of the two candidate designs, and produced the CAD model and the member and node tables the team built from. We built and load-tested it together.
final assembly mass
members across 11 nodes
statically determinate — 19 = 2(11) − 3
in the class competition
The problem
Design a cantilever truss from popsicle sticks that maximises load-to-weight ratio, then load it to failure in front of the class. A cantilever is the unforgiving case: there is no support at the loaded end, so everything is carried back to the wall through the members.
What I did
I solved every joint by hand using the method of joints, expressing each member force as a multiple of the applied load P rather than a number. Working symbolically meant the analysis stayed valid no matter what load the structure eventually saw, and it made the comparison between candidate designs direct.
That produced a force for all 19 members, their sense — tension or compression — and, usefully, seven that carry nothing under the design load: C, K, O, P, Q, R and S.
I verified the structure was statically determinate before trusting any of it: 19 members = 2(11 nodes) − 3.
Then the design decision the project actually turned on. We had two candidate configurations and a limited number of rivets, which are heavier and stiffer than the plain fasteners used elsewhere.
Six of those seven zero-force members — K, O, P, Q, R and S — meet at just three nodes: N9, N10 and N11. That is where the rivets went.
That reads backwards at first. The reasoning is that a zero-force member carries nothing under the design load, but it is what stops the structure folding once geometry shifts or the load comes in off-axis, and it is the connection rather than the stick that fails first. Spending the reliable connections where the analysis says there is no load buys insurance against the load case the analysis does not cover.
How it turned out
The finished assembly weighed 105.5 g and took first place in the class competition.
It failed in tension, which is what the analysis predicted. The post-failure photograph shows members separated at their joints and several snapped through — not a buckling collapse of the compression members, which would have meant the slenderness was wrong rather than the material.
What I'd do next
The analysis was static and the failure was a single load-to-failure pull, so nothing here says anything about repeated loading. I would also like a load cell reading rather than a bucket of water, so the load-to-weight ratio is a measurement instead of an estimate.