9.3 - Design Process

When we started the process of trying to create an AFK device, we had a very different idea for how it would function. Our initial plan, shown in Figure 1, was to have a 5-bar mechanism with dual servo motor inputs on both ground linkages and a stylus attached to the joint between links 3 and 4. We would supply random inputs to both the servo motors and put caps on the motion to keep the stylus confined to the area of the trackpad. After conversing with the teaching team, we determined that this design did not fit entirely within the scope of the class, as it focuses much more on reverse kinematics and the actual linkage system is relatively simple. We also wanted to shift our focus to systems that consist of a single input manipulator as opposed to the dual inputs described here. Due to these reasons, we decided to scrap the design and pursue different ideas.

We began looking into the concept of stacking crank and rocker mechanisms so the output of one would be the grounded position of the subsequent one. We could also drive all the input links with a gear train consisting of varying tooth counts, all originating from a single input motor. This would make it so the input link of each crank and rocker would rotate at a different velocity. We initially believed that we could get away with only having two crank and rocker mechanisms stacked on top of each other; however, after doing the math to calculate the cycle time, we found that we would need at least three in order to achieve the cycle duration we were hoping for. Each crank and rocker will be driven by a gear with a different number of teeth such that their greatest common factor is 1. This will make it so the gears only realign themselves after T2 × T3 full cycles of the first gear. Our tooth counts for the three gears are 47, 43, and 40 teeth, respectively. This would mean that the first gear would need to undergo 1,720 full cycles before the output path repeats itself. If we run the motor at a slow pace so one cycle of the input gear is 10 seconds, this whole process would take approximately 4.8 hours to repeat, which we believe is more than enough time for our purposes. Despite all this, we constructed only a double crank and rocker for our prototype, shown in Figure 2, to simply prove that the idea was viable. Another aspect of the design we were still debating was whether to use gears or belts for our drivetrain. We decided to implement both in our prototype in order to determine which was the more reliable option. When it came to designing the gears, we originally tried to create a design ourselves but quickly found this to not be a viable option. Eventually, we used the SolidWorks gear presets and modified them to fit our needs in terms of tooth count. These two gear iterations are shown in Figure 3.

For our final design, shown in Figure 4, we took everything we learned from the first two iterations to create a mechanism that achieves everything we set out to do. The final design consists of three stacked crank and rocker mechanisms, each with an input link driven at a speed determined by the gear train ratio and a variable grounded link dependent upon the output of the previous crank and rocker. After constructing the prototype, we concluded that gears were a much more viable form of power transmission compared to belts, so we used them for the entirety of the drivetrain. Another change we made was to add a second layer to the base plate so the axles would have support at the base under the gears and also at their tops above the gears. This change aimed to prevent the wiggle of the axles that we observed in our prototype. We were also able to create a mount for the mouse, shown in Figure 5, that seamlessly attached to the output manipulator, allowing us to translate our random motion to an input on a computer. Finally, we added a mounting plate for our drive motor that connects directly to the 47-tooth gear on the bottom left that drives the motion of all three input links.