Throughout the design process, we used a software called Linkage. This program helped us manipulate the mechanism link lengths and joints to achieve our desired motion path.
From our initial design, we knew the sweeping mechanism had to move parallel to the table at the bottom of its motion in order to maintain contact with the surface and sweep it properly. After some research, we found the Chebyshev lambda mechanism to match the exact path we wanted, in which the longest link (with the broom attached to the end) is exactly twice as long as the long leg, and five times as long as the short leg. We tested these lengths in Linkage, and the motion path was perfect.
For the bucket, we knew it had to turn 180 degrees at the end of its path, then quickly reset to its starting position for the next sweep. We first found a complex 8-bar mechanism that would achieve a complex translation and rotation motion, but later simplified it to a simple 6-bar half-rotation motion with a quick return. This allowed us to cut fewer links, minimized our chance of error in joint motion, and minimized the space needed to offset each link.
Once we had our desired motion for each mechanism on Linkage, we noted the ratios of the lengths of each link in the mechanism with respect to the length of the sweeping motion so we could scale it to any size we wanted. Our limiting factor was the size of our bearings, which were 22mm outer diameter and 8mm inner diameter. With this in mind, we knew the smallest link we could make was 36 mm in length from edge to edge, including a 5mm buffer for the wood to surround the bearings and axle holes.
Once we knew our link length rations and our bottom limit, we began building our first prototype. For our first 3D prototype construction, it was easiest to operate at that smallest level for the sake of simplicity and testing the practicality of our design. Evaluating possible materials, we settled on plywood since the laser cutter let us “print” out the exact link shapes and lengths we wanted very quickly, and it was cheaper than acrylic.
Our first working model delivered the exact motion we were looking for, but was a little tedious in the assembly. To ease the construction of the final prototype, we scaled our links up slightly so the smallest link was 40 mm in length and much more comfortable to handle. We also discovered that wood worked very well with our bearings and axles, so we decided to stick with plywood for the entire project. This also added to the mechanical art aspect of the piece, giving it a da Vinci-esque aesthetic. With TA permission, this inspired our decision to stick with a hand crank as our power source rather than a motor, just as they would have in the 15th century.