I led the whole mechanical design for this robot project while working with 3 other students. One of my most fulfilling projects.

Goal: Design and fabricate a robotic arm capable of playing checkers for you.


LINKS

I started by designing links in Solidworks then used 3D printer to print the designs. Design is an iterative process and I made around 4 design iterations before settling on an optimal link design.

I did static load case FEA analysis on the links based on the stall torque of the Dyanmixel servos.

GRIPPER

I made three iteration for the gripper design. Each of them were based on a completely new mechanism concept.

For gripper 1, I did the following design. The idea was to have the 3 fingers of the gripper get pulled in and out by sliding them on a slot. This was a bad design as the force vector is clearly not in the axial direction of the slot. Plus, the gripper would need to be fabricated using 3D printer which increased the friction immensely. I 3D printed and assembled the part just to try it out(unfortunately I did not take pictures of it). The fingers did not get pulled inward, rather the links connecting the fingers and the circle in the middle got bent and broke.

For my second iteration, I made my decisions based on designs that I knew existed. This gripper used slider crank mechanism which is one of the classic mechanism used to convert rotating motion to linear. I did the calculations for the links lengths of the slider-crank mechanism to get the proper motion distance. This design was looking pretty promising but turned out to be a failure as well. It was not necessarily the idea that was bad but the parts we had did not allow this mechanism to work as intended. 3D printing parts meant bad tolerances and the rod on which the fingers were supposed to linearly move on was off from its expected location. The rod also got bent after the fingers moved just a little bit because it could not handle the load from motors. Instead of a 3D printed rod, if there was a Al rod, this mechanism, in my opinion, could have worked. 

The third and final iteration of the gripper design was a simple geared mechanism. I used parametric spur gear CAD model that I had to make the gears in SolidWorks and added the links to it to get the mechanism. This simple mechanism worked very well. The only drawback of this gripper was the fact that it could not pick up the checker pieces from the top but rather would have to clamp the pieces from the side. Even though this limited out robot arm's capabilities, I was more than happy to have got something to work on the deadline day of the project.

PRISMATIC JOINT

The team wanted a prismatic joint to make the project more fun. I decided to use a pulley mechanism to make that happen. I sourced necessary parts from McMaster and built the pulley mechanism to move the whole top part of the robot.

Along with doing the mechanical design for the robot, I also helped in calculating the Forward and Inverse kinematics for the robot, and doing simulation of the robot in Matlab.