Robotic Arm
By Harbin Singh
Built in the Flowers Invention Studio
I've been interested in robotics design for over five years, but I haven't had the chance to fully own all aspects of a robot (mechanical design, electronics, and programming) yet. For this project I wanted challenge myself by learning a new skill in each category of robotics.
My vision for this project was to make a 6-axis robotic arm that can easily grab objects of varying shapes and sizes. Therefore, a large reach radius and adaptable gripper was needed to achieve this at the cost of precision. I opted for low cost and lightweight servo motors to move all the axis of rotation as their torque to weight ratio was needed to achieve a total reach radius of 500 mm. To fabricate and assemble the arm, I used SLA and FDM 3D printing, soldering tools, fasteners, and power tools at the Invention Studio.
Some new mechanical design skills I used in this project are generative design and finite element analysis. I originally designed the key structural beams with an "H" cross section to maximize the stiffness to weight ratio. Later, I redesigned the beam with Creo's generative design software which allowed me to cut the weight of the beam in half and not sacrifice much stiffness. I first 3D printed the beam in draft resin to test the part tolerances and print quality. I then followed this process for the other beam and the gripper base.
With these complex geometries, I was able to take full advantage of the highly precise Formlabs SLA printers at the Invention Studio. The black support structure left a lot of scratches and bumps along the parts, so I did my best to sand them down for a smooth surface finish. The scratches were much more noticeable on black resin versus white resin.
The assembly of the gripper was the most challenging to part to get right. I needed a very tight clearance fit to get the gears to mesh well, but the 3D printers did not have a good enough tolerance to achieve this. I ended up printing, drilling, and sanding each component 3-4 times to perfect the fits.
I redesigned the smaller components a few times to ensure it would be strong enough and print well. This part especially failed a lot at certain print orientations. Some of the gear teeth on the left part did not print well. The flanges on the right part were curved downwards and the supports were difficult to sand down in the small cavities.
After fabricating, post-processing, and assembling the arm, I was able to start wiring and soldering. I gave every degree of freedom just enough slack to not limit its motion with cable sleeves and zip ties. I then soldered pins to the ESP32 board and used jumper wires to connect it to the servo wires.
Once I got the code running properly, the arm was swinging around well and was able to grab a lot of objects.
Budget and Parts List
Final Notes
Overall I'm very happy with how this project turned out. Most of the issues I encountered were with the new software and fabrication techniques I used. Especially with robotics projects, it can be pretty difficult to figure out if a bug is from the electronics or code. In that debugging process, I ended up learning a lot more about microcontrollers and programming fundamental than I would have if it worked on the first try.
One thing that I would change is designing the arm to be easier to disassemble. When there was a hardware issue, it was a lengthy process to reach the part that I would need to replace or fix. Placing bolts in more convenient locations would have definitely saved me a ton of time.
Some next steps for this project would be programming a more sophisticated control GUI for the arm, so it can be easily maneuvered. I also plan on adding a current sensor for the gripper servo. Then it can automatically stop actuating when it senses that it grabbed an object.
Funded and supported by the Invention Studio at Georgia Tech, a student organization.
