We’re back for another round of trebuchet testing, and this time the prototype is beginning to look like a functioning weapon of (small scale) warfare.
Above you can see the prototype assembly. The supports hold the arm assembly high above the floor and the arm is held in place pre-launch using the holes on the right hand side. Parts were printed to attached the counterweight to the arm to generate consistency in testing.
First, the assembly process. Thanks to the difficulty in achieving tight geometric tolerances using PLA the fit of several components was either too loose or two tight. We attempted to make the parts fit together via a interference fit (so we would not need to use any adhesive). This could potentially be achieved by sanding the parts to their correct sizes in post-processing. However due to shrinkage and inconsistency in the printing process we were not able to achieve this. The proper application of force and a little high strength adhesive took care of these issues.
Small locking brackets were added to a critical location of the new “wireframe” trebuchet to stiffen the design. Even with these new components the system was still a bit unsteady, so the team glued the frame of the trebuchet to a piece of cardboard for stability. One potential fix for this is to design a base with a large, flat, surface that would provide better stability but would not take long to print.
The weight brackets were then attached to the brass block and the assembly added to the arm. Finally, the arm was attached to the frame with a nail through the pivot point and printed spacers were used to center the arm in the frame.
Proto V2 includes the ability to hold the arm of the trebuchet at various starting angles before release for the purpose of optimizing the various parameters of the device for longest throw. Rails made of PVC pipe were also added to the frame to guide the basket as it is pulled out from the frame.
Testing began by implementing the same scoop shaped plastic basket (constructed from the bottom corner of a 2L soda bottle) as previous tests but with very thin thread replacing the paracord material previously used. This thread had less elasticity which was ideal for measurement and consistency, however, it was very difficult to adjust the lengths of the line (single member) and loop of the attachment system. This thread was replaced with a much greater diameter cord and larger holes were drilled in the scoop to accommodate. This larger cord allowed for easy and precise adjustment of the length of the loop and line of the basket.
Testing then commenced with trials at greater and greater initial angle of the trebuchet arm. We discovered early on that if tension is not maintained in the line attaching the basket to the arm (this is the string that is rigidly attached to the arm rather than the loop that slips off) the sudden impulse of the line being pulled taught will jerk the basket so quickly that the M&M flies out the back of the basket. To remedy this, the relative lengths of the line and loop were adjusted so that the line would always be just slightly shorter than the loop. As a result when the basket is pulled to the forward position the line is pulled taught before the loop. However, this created an issue with the loop falling off the peg of the arm due to lack of tension. For these tests we kept the angle of the peg constant, parallel to the length of the arm. Using a peg with a slight up-sweep would help the loop stay on the peg at more extreme angles. However, the team has hypothesized that adding an angle to the peg will detriment the release of the loop. This theory will be tested in the future but for the tests discussed here a straight (0 degree) peg was employed.
A table of the distances achieved at each angle position and the lengths of line and loop used to achieve these throws is presented below:
The team was able to achieve very consistent results with gradually increasing distances eventually plateauing at about 9.9m using launching position 11, which corresponds to an initial arm angle of 21degrees.
Moving forward the team will explore alternative peg designs, hoping to increase the initial launch angle further.
Final designs for the trebuchet supports were also discussed with several concepts on the table. The final selection will be made soon and the geometry modeled and printed.