RR1: Real Robot One - DIY 3D-Printable Desktop Robotic Arm (C) Pavel Surynek, 2021 - 2024 ======================================================================= 0. Contents: -------------------------------- 1. RR1 introduces itself 2. CAD source files for rev.2 3. Conclusion ====================================================================== 1. RR1 introduces itself -------------------------------- RR1: Real Robot One is a dektop 6-axis robotic arm. The robot is based on 3D printed parts, uses stepper motors Nema 23 and Nema 17 and custom planetary gearboxes. Robot can be controlled using Arduino (I used Adruino Due), but you can use another single-board computer. The RR1 robot uses a so-called closed-loop, i.e. the control unit thanks the encoder mounted directly to the joints of the robot knows at every moment, what configuration the robot is in (i.e. what are the angles of its joints). In this repository, I try to post materials for building the RR1 robot. The published materials refer to the second copy of RR1, the so-called revision 2 (rev. 2, the orange robot). The first copy, i.e. revision 1 (rev. 1 - the black robot), was used to test and verify the possibility of a robotic arm using 3D printing to build. Because the materials for rev. 1 never were very well organized, I decided to post from rev. 2 including CAD source files, i.e. everything is completely open source. Design experience is such that the revision 1 of the RR1 robot (rev. 1 - black robot), which was printed entirely from PETG, was not ideal. It turned out that overall the robot is not rigid enough, especially it showed some flexibility in the gearboxes, although it didn't seem like it there was some noticeable backlash, which is typical for planetary gears. In the gearboxes of the RR1, the backlash is greatly reduced by the use of Herringbone gears. So I recommend printing mechanical parts such as gears from a stronger material such as polycarbonate with carbon fiber. This is how revision 2 is printed (rev. 2 - orange robot) and the overall rigidity is much better. There is practically no backlash in the gearboxes. In addition, the whole design includes encoders directly on the joints, so that any non-rigidity can be compensated by software, as we have very precise information about the configuration of the joint from the encoder. With the gripper, it is exactly the opposite, where we want a certain flexibility. For for the correct function of the gripper, it is necessary to print the rods from the flexible material such as TPU 98A. There is also a need from this flexible material to print the inner serrated surface of the gripper. Of course, 3D printed parts are not everything, stepper motors are needed, screws, bearings, cables, aluminum extrusions, etc. Bearing dimensions are marked in the drawings so that the correct bearing can be determined accordingly. The philosophy in the choice of bearings is that I want as many of them as possible from local Czech companies. In addition, an enormous number of screws of various sizes are needed, most of the screws have a diameter of 3 mm, several screws are 5 mm in size (installation of aluminum extrusions) and several 2.5mm and 2mm screws for mounting encoders and small parts. I recommend using Philips screws with a low head (most holes in 3D-printed parts are dimensioned so that the screw head is aligned with the surface of the part). The vast majority of nuts are 3mm square, which slide into the printed parts from the side (you can see a lot of small rectangular holes in almost all parts that are intended for these nuts). The links of the arm use aluminum profiles of size 20mm x 20mm (upper arm link) and 30mm x 30mm and 15mm x 15mm (elbow link), their length can adjust overall size of the robot, but I do not recommend making the robot too big, because every extension of the arm increases levers and torques. A total of 7 stepper motors are needed, four Nema 23 motors of different sizes – 112mm for the shoulder joint (this motor is probably a bit oversized, but gives the robot a strong look), 76mm for the base, 54mm and 80mm for the elbow joints, further motors It does not have 17, of which one is a large 59mm for rotating the wrist joint and two smaller 39mm for wrist rotation and opening and closing of the gripper. The gripper are moved using a threaded rod, so it's good to use a motor with a threaded rod preinstalled instead of a motor with standard shaft. The encoders are mounted directly to each single joint, 3D printed holders are ready for encoders AMT 103V and one AMT 102V for gripper. ====================================================================== 2. CAD source files for rev.2 -------------------------------- I decided to post the complete source files with CAD drawings for RR1 revision 2. I make the source CAD files available completely free of charge freely available to the maker community under the attached GPL license. Although I spent a lot of time creating the robot, I think as an academic I should to provide the results of my work to the public. The current state of the source files is somewhat messy and far from everything nicely named. Some parts even occur in several variants in the source files. However I think a bit of a clever maker can familiarize with the drawings and can build his/her own robot. It must be said that I am still developing the RR1 robot, at the time of writing I am working on the revision 3 (I repeat that the published version of the sources is for the revision 2). So more perfect documentation, naming and organization of files will come in revision 3 a later ones. For better orientation, I also add STL files in a separate directory. You can also be used directly for printing if no part modification is required. FreeCAD has been used to design the 3D printed parts in the spirit of my philosophy of maximum use of open source software that gives a guarantee of long free duration without ugly marketing tricks. For printing, I recommend Prusa MK3S+ or Prusa MK4 printers. As of materials, I recommend using Prusament PC Blend with carbon fiber for mechanical parts (gears), PETG or PLA is good enough for the supporting parts. I printed my copy RR1 rev. 2 using these printers and materials (see more details on https://surynek.net/_main/index.php?select=research#RR1rev2). As of mechanical parts, it is sometimes due to printing inaccuracies and various properties of the materials to adjust the size of the parts by approximately 0.1mm in order make everything fit together smoothly. So it is perfectly fine that a particular part, typically a mechanical gear, has accurate dimensions only after several printouts (the more printers you have, the better :-). Of course, most inaccuracies need to be tuned for planetary gearboxes. ====================================================================== 3. Conclusion -------------------------------- This is just a short introduction to the RR1 robotic arm. I would like to create detailed construction instructions, but I will probably leave it until the next revisions of the robot, because I am still improving the robot and making changes. I mean that revision 3 is already in development (you can try to guess what color it will be). As more files and documents are ready, I will post them as well. I welcome any observations and comments, or suggestions as what could improve the robot in the next revisions etc. ---------------------------------------------------------------------- Happy and successful robot construction! Pavel Surnek, January 2024