I Designed The Worst 6 Axis Robotic Arm

A story about how I think I’m a better engineer than I really am.

This is the Makr Shakr. Its a set of robotic arms that make and shake drinks that people can order on their phones. When my brother came across a youtube video in August of 2019 displaying the many features of these bartenders, he reeled me into the idea of making our own version.

Here’s a little background from around the time this project started: I had been making remote-controlled airplanes from foamboard, racing drones from kits, and robots from LEGO kits. I had then joined my high school’s FIRST robotics club and participated in the FIRST robotics challenge. It was there I got a taste of what mechanical engineering and design. I learned SOLIDWORKS computer-aided design (CAD) and manufacturing techniques. But honestly, I had little to no experience with a personal engineering project. I had never done any physics, calculation, or design before.

The problem was that at the time I thought I was hot shit.

I told my brother I’d make it happen. And by September, I had my first design made in SOLIDWORKS. Here’s all the problems with it:

Motors

Motors are the core of what makes a robotic arm move — and I selected terrible ones.

Part selection is one of the most important parts of a project. One of the reasons a student mentor on our robotics team is so helpful is because he knows almost every legal part that we are allowed to use, and therefore can help select a part that gets the job done.

The only motors I knew at the time were motors from robotics and drone racing motors. Drone racing motors are based on high rotational speed and low torque, which is not at all ideal for a robotic arm. So I decided on the cheapest, most reliable option from the robotics catalog of motors: the 775 pro.

the 775 pro from vexpro

Using a brushed motor like this is not at all ideal. Let’s look at the specs and we can see why

775 pro details

We can see that the free speed is 18750 rpms. That’s insanely high. For a robotic arm like ours, I calculated the max rpm we would need is around 20. But that’s not that important, because we can run the motor slower and achieve the speed we want. What we can’t achieve is payload.

We wanted this robotic arm to be 34. inches in length total — 18 inches on the first arm and 16 on the second. We also wanted it to be able to carry a 1.5 lbs payload on top of its own weight.

Let's say that the arm weighed 5 pounds. At the stretched 34 inches, a 5-pound payload means a torque of 19.2 N-m (Newton-meters). That’s nowhere close to the 0.71 N-m ratings of the motor. dividing 19.2/0.71 gives us a ratio of 27:1. And this is if we run at MAX power.

Motor data from the 775 pro

According to this chart, if we were to run at max power, we would consistently pull 347 Watts — that's simply unsustainable and would heat things up a good amount. It would also require a decently large power supply.

At this point, I decided that I would attach another motor to the first arm and use a 14:1 reduction using Versa Planetary gearboxes — another part from the FIRST robotics catalog.

There’s not much wrong with these gearboxes other than the fact that they are extremely heavy and bulky. they added a good amount of weight and had I chose better-suited motors, I could have easily avoided the need for a high reduction gearbox.

PLA 3D printing

In order to print the parts from the project, I purchased a 3D printer. The easiest and most affordable material that is used by many for 3D printing is PLA. PLA plastic is very easy to print, as it does not require any treatment before use, nor does it need perfect conditions to come out nicely.

A section of the first arm printing

PLA also has its issues.

First, it needs to be printed in high density to be strong. PLA is a fairly weak material that can be bent and cracked, so it needs to be printed with wide layers and high density, which, as seen in the picture above, I didn’t consider. Adding more material would just increase the weight and print time, so I tried to stray away from having to add material to the arm.

Second, it doesn’t bond very well. 3D printing works by adding layers on top of each other, and for PLA those different layers don’t seem to bond too well to each other. Therefore, they often split apart easily on high loads.

Third, it is very brittle. PLA will not absorb any impact force. It is very hard, so it breaks upon almost any force. If you were to drop a printed part from this project, it would almost definitely crack.

The last problem with PLA and 3D printing as a whole is that 3D printers can only print parts that are flat on one side. In order to print vertical layers, there needs to be a flat base. That raised an issue for me as I had parts that had strange contours. So I split the parts into two different sections and used Acrylic Cement (pretty much glue) to stick the pieces together. This was a terrible idea. I have no idea what I was thinking and why I believed that it would be a good idea to glue together crucial parts of the arm that are under high-stress loads.

Using aluminum, carbon fiber, or other materials for at least the internal structure would have been a far better idea.

Wait what’s physics?

In August of 2019, when this project started, I was going into 10th grade. I had never seen a physics textbook, which became evident in my design.

The entire body of the robotic arm was 3d printed, and although all the torque force does go to the motor, I had no familiarity with the idea of inertia.

I witnessed the effect of inertia on the robotic arm as I accelerated a section of the first arm by hand and quickly decelerated it. I then watched it split amongst its 3D printed layers and crash into the floor, where it split into hundreds of pieces.

Making a body of a large robotic arm out of weak plastic was not a good idea at all. I had not considered the effects of any physics on the arm, and therefore many things obvious to any engineer just flew over my head.

Takeaways

So what did I learn from this project?

  1. I’m a terrible engineer with a lot to learn
  2. I need to research more on the parts available to use
  3. I should consider the reality of my design before I manufacture it
  4. It’s okay to fail

The failure of this robotic arm has already fed into my next design. I am using different motors, more aluminum instead of plastic, and have been doing the calculations, and accounting for inertia to make sure my next version will work.

Heres images of the old vs the new

images of the old version
images of the new version

A robotics and economics enthusiast || Junior at Saratoga high school || Writer @ Towards Data Science

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store