Manufacturing and Assembly - SF

As mentioned previously, most of the custom parts on the mechanism (i.e. everything but the wooden dowels) were laser cut from 3mm or 6.35mm acrylic and plywood. The process of designing and assembling the mechanism was somewhat iterative, as the design goals were refined and each prototype was modeled and tested. Many parts of the mechanism were also built on top of existing parts from previous prototypes. As a result, not all the pieces in the final prototype were created at the same time, with some parts of the final mechanism being the second or third iterations of their respective parts. The basic steps I followed as I went through the design iterations for new parts were as follows:

• Create an initial model of the part in CAD
• Manufacture the part, usually by laser cutting
• Attempt to mate the part with the rest of the mechanism
• Realize that some improvement could be made
• Modify the part in CAD, or create a new version of the part
• Make the modifications on the actual part, or manufacture the new version
• Repeat previous three steps

Motors and Driveshafts

Since they were relatively permanent parts, the mounts for the motors and the bearing blocks were more some of the more straightforward parts of the mechanism. I first laser-cut the faceplates for each motor, which had holes corresponding to the preexisting mounting holes in each motor. Next, to mount them to the acrylic backplate, I cut some 8mm wooden dowels (the same ones used in the driveshafts) to length on the bandsaw, and drilled holes on either end for screws on the drillpress. At first, I fixed the dowels to the faceplates, but not to the acrylic backplate, as I had not yet finalized the location of each motor.

Figure 15. View of the motor mounts.

The driveshafts were manufactured in much the same way as the dowels for the bearing mounts. I only drilled holes into one end of each driveshaft, however, as I originally planned to mount the gears to the shafts using screws. The driveshafts were fixed to the mechanism using bearing blocks, which were also laser-cut using 6.35mm thick acrylic. I had originally planned for the ball bearings to be press-fit into the holes in each bearing block, but since the laser power had to be set higher to cut all the way through the thicker acrylic, the holes ended up a little too big. Thus, Loctite was used to secure the bearings to the blocks, which was a nonideal solution since the bearings were prone to falling out when too much force was exerted on them. Eventually, both the bearing blocks and the motor mounts were fixed to the backplate using only bolts, which threaded into the dowels in the bearing blocks and the acrylic backplate, respectively.

Since in the beginning I had planned on driving the mechanism using belts rather than gears, I laser-cut some acrylic pulleys which were essentially three acrylic circles with holes cut into their centers sandwiched together. These pulleys were to be mounted to the shafts of each motor (the pulleys for the motor had D-shaped cutouts in their centers to lock their rotations) and each driveshaft directly. However, this idea turned out to be a terrible one, due to a multitude of factors. First, since I did not order a proper drive belt (ideally a timing belt), the belts had to be be made from rubber bands, which turned out to be a terrible way to transmit torque. Since the belts were elastic, the driving pulley attached to each motor was only able to transmit torque to the driveshafts intermittently, when the tension in the rubber band was enough to overcome the friction in the mechanism. Additionally, because the pulleys were flat on their sides rather than crowned, and due to errors in the pulleys' alignment, the belts tended to ride up on the pulleys, thus further reducing their ability to transmit torque. To compound matters further, the rubber bands were fragile, meaning that whenever they got caught on something (usually the crack between the acrylic circles used to assemble the pulleys), they would wind themselves on the pulleys and eventually snap.

Due to all of these factors, I scrapped the idea of using pulleys altogether and laser-cut some new gears from 6.35mm plywood (I hoped the thicker gears would help them mesh better, even when they weren't fully aligned). I was able to re-use the motor pulleys to mount the gears on top of with Loctite, but the gears on the driveshafts were mounted only using copious amounts of hot glue.

Figure 16. View of the underside of the mechanism.

Gears and Linkage

The acrylic base of the mechanism that the gears rotate on was used from the first prototype all the way to the final mechanism, and new parts were mounted in screw holes drilled alongside the existing ones. Several iterations of the gears were manufactured as the design evolved, as well as the arms of the linkage. As I touched on in the Design Process section, the arms of the linkage and the non-driven gears were mounted to the acrylic backplate and each other using M3 bolts and nuts.

A significant challenge I faced during assembly of the mechanism was fixing the gears to the driveshafts so that they would turn together. Initially, I planned to screw the gears to the ends of the driveshafts using a hole drilled into the end of the shafts, but this turned out to be unfeasible because the screw torque required to fix the gears to the shafts was great enough to strip out the shallow-cut threading in the wooden dowels, thus rendering the connection unusable. After much trial and error, the solution was to widen the holes in the gears and put the driveshaft through them, and to file corresponding keyways in the driveshafts and gears to fit a key into. This turned out to be a challenging process, however, due to the fact that I had already drilled a hole into the end of the driveshaft, leaving little room for error when filing the keyway. In the resulting assembly, the end of the driveshaft was flush with the top surface of the gear, and the input link, formerly mounted to the gear itself, was mounted to the hole in the driveshaft instead.

Figure 17. Top view of driveshaft and gear assembly.

A second challenge I faced was the links of the mechanism interfering with the heads of the bolts that were used to assemble much of the mechanism. Having dealt with this issue in the first prototype, I resolved it in the final mechanism by manufacturing extra-thick spacers between to put between the links at the joints where the input arms connected to the rest of the mechanism. These spacers were created by making cuts across a 16mm wooden dowel (the same type used to secure the two acrylic backplates to each other) and drilling 3.5mm holes through the resulting discs. The spacers were each at least 8mm thick, giving the bolt heads plenty of room to move underneath the linkage arms.

Figure 18. View of the input links, with wooden spacer on the right.