Tag Archives: diy

Lego Robot

InstagramCapture_5d7b9f9a-5f85-4ac6-87aa-5cafbe142fb2Conceived as a step towards enabling interactive use of the architecture program’s new KUKA robot, the Lego robot was designed to mimic the proportions of the actual robot so that it could be used to test ideas and procedures at a small scale prior to implementing them at the scale of the actual industrial robot arm.  Additionally, if the Lego robot could be built such that it could record movements and positions imposed by hand, the actions recorded with the small robot could potentially be played back on the large robot.  This would be all the more interesting if it could be achieved in realtime, translating motion from small- to large-scale on-the-fly.

This project was not without precedent.  KUKA-specific software development firm OrangeApps, located in Germany, has developed a six-axis Lego robot that can be controlled using the KUKA KRC controller allowing KUKA KRL code to be executed against the Lego model.  Their design employs KUKA OfficeLite software, which simulates the KRC controller software on a PC, along with a custom plugin to control two Lego Mindstorms EV3 controllers, which in turn power six motors to drive the axes.  A seventh, smaller motor controls the gripper end-effector with which the robot is equipped.

The Lego robot pictured above was developed over the course of the initial month and a half of the Fall semester and represents the extent of its development to date.  Constructed from Dan’s personal collection, which consists primarily of a Lego Mindstorms NXT robotics set and three large Lego Technic sets, the robot implements all six axes but is not yet functional.  In order to keep the wrist construction compact, the design incorporates a custom, 3D-printed gear that facilitates the translation of rotation of the fifth and sixth axes across a single shaft that passes through the center of the fourth axis (see images below).

A few outstanding obstacles remain before the project can move from hardware to software development.

  • Since the Lego NXT controller can natively only control up to three motors, an alternative controller will have have to be used to drive the robot.  An Arduino-based solution will be implemented using the third-party Bricktronics Megashield.   The Arduino and add-on shield have been purchased and assembled, but have not yet been tested.
  • The current design is not well balanced and may require modification before the motors and gears will be able to handle the torque required to operate the second and third axes reliably.
  • Due to the need to amplify torque using gear differentials and due to the complex series of gears used to implement the wrist, backlash may become an issue resulting in reduced precision and stability.

 

Final 3D printed gear
Turntable through which gear would need to pass
3D model to print
Formlabs Form 1+ printer used to print gear
Robot wrist assembly

 

Mounting Assembly

The first step in utilizing the KUKA KR-60 robotic arm for digital fabrication required the development of a mounting assembly. Without it, the end-effectors we develop in the future would require custom attachment solutions independent to each end-effector. To resolve this, the mounting assembly was constructed of 80/20 Inc. aluminum extrusion and steel profiles. The mounting assembly mounts direct to the end of the KUKA and acts as a machine table for other custom end-effectors to be bolted and secured to the aluminum extrusions.

 

Router End-Effector

A router end-effector has been designed with the hope to soon be utilizing the KUKA’s six axes for milling. This low budget router end-effector is assembled out of one $30 hand router from Harbor Freight Tools and a custom holder to be 3D printed in two pieces with our Form 1 printer. When finished, the router will be secured to the mounting assembly. This end-effector, although relatively crude, is a mean to understanding the KUKA’s capabilities for 6-axis milling and a temporary solution until a more refined end-effector can be produced.
Watch for more updates on the router end-effector in the future.

 

Hot Wire Foam Cutter

While assembling precedents for our research with the KUKA we found many examples of automated foam cutting with a robotic arm. A hot wire foam cutter end-effector looked as if it would require little expense to develop so we began designing and fabricating the tool. With steel angles, 1″ pipe, nichrome wire, a guitar string tuning mechanism, and a power supply we were able to fabricate the tool with less than $90.

The end-effector is designed to cut up to 40″ wide and 16″ deep in one motion. The most challenging aspect of the end-effector has been producing enough current to get the nichrome wire up to temperature for rapid foam cutting. Initial tests of the wire’s cutting temperature were done by hand. We found that the wire was capable of cutting the foam, but only at slow speeds. Attempts to wire through the foam more rapidly resulted in the wire dragging through foam and large deflection in the wire. Once we have a larger power supply we expect to have no issues with this.

Watch for more updates on the hot wire foam cutter end-effector in the future.

 

First hot wire cutter test
Variable voltage power supply

 

Robot Work Surface

While fabricating the hot wire cutter end-effector we knew we would need a suitable work surface to shape our material. Research into industrial standards for work surfaces only revealed options that cost thousands. With little time and money to work with we set out to design and fabricate our own custom work surface.

Using Ball State’s Thermwood mill, the finished work surface was able to be constructed entirely of laminated and pressure fitted MDF pieces. A box raises the work surface 14″ above what the robot recognizes as the world plane. Here, material rests on a perforated surface that is itself above a vacuum chamber located inside the box. After a ShopVac is connected to the vacuum chamber material is held in place for the robot to begin work manipulating its shape.

Assembly and testing of the work surface has taught us a lot about where we could improve the design for the future. We hope to do so, but until then this inexpensive and efficient design should do just fine for some projects to come.