Which robot arm is suitable for a SCUTTLE chassis? Almost any, but many collaborators asked for our recommendations on robot arms to add to their SCUTTLE. The following options are recommended for low cost, DIY-friendliness, and true open-source designs. Soon we hope to establish a true favorite and offer detailed resources for a SCUTTLE standard Mobile Manipulator.
I made a short list of open projects that contain hardware (not just circuits) on a gist here, but it is only updated through 2021. This post is being written in May 2023.
SCUTTLE’s ideal robot arm (not yet found) has:
- Truly Open hardware
- Kinematics lessons taylored to the arm
- Realistic payload above 100 grams
- Modular design (eliminate some joints if needed)
- popular COTS hardware,
- Truly 3D Printable designs (easy tolerances, no supports required)
- Supported by an organization with true open source values
- a community of makers who contribute
by BCN3D & Generalitat de Catalunya: PRL 8
Probably the most refined robot arm design that also provides CAD models for hacking as of 2023. We hope to get one of these models added to a demo SCUTTLE soon. It’s fully made from maker-style products even down to the circuit boards. They are taken from common 3D printer components.
The only reason it does not get a project readiness level 9 is because it’s not (yet) designed for battery-operation. It just takes a little bit of electrical engineering for this to be SCUTTLE-ready for makers.
by Nyrio: PRL 8
This was our favorite design in 2022, but they have slowly become more of a typical product and less accessible for makers, and less 3D Printable. They now claim it is “based on open source technology.” Still a great balance of cost and performance for classrooms & makers.
Compared with Moveo, it’s more acessible to younger audiences & new programmers, but less documented for makers.
Small Robot Arm
by Skyentific PRL 8
Scyentific’s Adam Giunta is my all-time favorite youtuber for the mechanical side of makership. His arm project has not been updated since around 2019 but it’s good quality and very makeable. His main content for building the project is hosted on his Github.
by Dobot PRL 7
This unit is not open source but they were when they raised $615,000 on Kickstarter! It’s the only option today if you want something you can shop for at common distributors like Digikey. This arm was our selection for low cost educational machine at TAMU, since it sells for under $2,000. The accuracy is nice with repeatability of end effector in sub-milimeter ranges. Also, this option gets you the support of a large company for instructions, warranties, etc.
Compared with the above arms, Magician has finer precision, and comes with a suite of compatible modules such as a dobot-branded conveyor, etc. Still, not really open-source.
SCARA DIY Arm
by Levi Janssen, PRL 5
This unit is in early stages but has a very special advantage: If you add a gripper and hold an object in static configuration, it takes NO ENERGY. By loading the bearings instead of the motors, SCARA configuration gives a wonderful advantage when it’s running on a battery.
Printability: Levi uses metal components but they can easily be made with 3D printing. It’s also open and modular enough that you can add more linkages, and copy/paste the kinematics to new joints.
Education: In the youtube videos, Levi breaks out the white board and talks through the calculations for movement. That means even though this bot isn’t sold in stores, you can make your own variation quite easily.
by ON Development, PRL 6
The most affordable design in our list, with open model and some software distributed by the offering company. This model is very simple and made from hobby-grade servos plus lightweight NEMA 17 stepper motors. The special part of this design (why isn’t this more common??) is that linkages for actuate the joints and the motors are at the base. This takes the weight away from the extended end. So, if you were going to make a lightweight arm with actuators of 3D printer parts (limited torque, super available, super cheap), then this is the recommended option.
Documentation is not great, and it shouldn’t be in the same class as the stronger robots since the payload is probably under 100 grams.