About Virk I

Virk I is a 4 DOF mini scara desktop robot targeted to education and hobbyists. It’s a quality, robust, neat and cute robot, designed, machined and assembled in our workshop. It was fully functional by the end of 2019.

Main features are:

  • Clean design thanks to its hollow axis rotary joints
  • Wooden, slim structure
  • Precision cnc machined parts, hand finished
  • Harmless and will not break if stopped or moved by hand
  • Compact and easy to carry
  • Open design (understandable drawings pending, will send messy drafts by request)


There are just one working unit at this moment, and eight more being finished. You can buy Virk I HERE.

If you wish more information, just send an email to igarrido@automaticartisan.com. Check lastest blog entry about Virk I HERE.


  • Australian blackwood body (mostly).
  • Motors: Nema 17  stepper on rotay joints, Nema 11 stepper on z axis and wrist
  • Full extended range: 293mm (from shoulder axis to end center)
  • Z-travel: 70mm
  • Weight: (pending)
  • Z axis slide: 8mm linear rails
  • HTD3m timing  belt on rotary joints
  • 8mm pitch leadscrew on Z
  • Homing switches
  • Servo gripper
  • RAMPS 1.4 electronics – Marlin driven (modified vesion)
  • CV (computer visión) enabled through Python/OpenCV app.


Hollow Axis Rotary Joint


One of the most interesting part of this robot is our precision hollow axis rotary joint. Being hollow axis means that the wires can go hidden and not be a mess as in other designs.

This joint has three bearings inside: two trust bearings and one deep groove, thin bearing. It has three acetal parts and one PVC mount base. Also, there’s a custom made timing pulley attached to the rotating part of the joint.



  • ¿Could be used to engrave? I don’t think so. This arm has enough power to move itself and lift small things, but lower enough so it can’t damage itself or the user. Engraving requires more stiffness and power.
  • ¿Could be adapted for 3D printing? Maybe, but I guess this will not be practical. Belt transmission and arm mass add some “spring effect” on every movement, so the faster, the less precise. That is, of course, just theory.
  • ¿How much precision can you get from this? Repeatability tests are pending. Rotary joints are very precise, but final precision depends on a lot of factors: calibration, proper belt tension, load, belt fit, total stiffness, etc. Anyway I guess it will be precise enough to be a useful as learning tool and to give fun to robot lovers.
  • ¿Why wood? Because it’s beautiful, cheap, lightweight and easy to work. And because I love wood.
  • ¿What kind of machines and skills I need to make something like this? You can probably do all the parts using a cnc mill or a good cnc router. A manual or cnc lathe will be handy, too. This is not a simple project, and has a lot of parts, so you need to have experience using cnc  machine and be able to grind your own cutting tools. Also. you will need a lot of tools; drills, taps, etc.
  • ¿Could this be 3D printed? Maybe some parts, but not the rotary joints and other precision parts.
  • Will the design be available as open hardware? Yes!. However, make good drawings from my messy drafts will take a lot of effort, so meanwhile I will send drafts by request.
  • Will this be mass produced? Not in the short term, but we are considering to start a kickstarter campaign.
  • Where is your Marlin modified vesión available? Here.



3 Responses

  1. Hi,
    I have on and off followed your project.
    I have written some software that converts Cartesian gCode to SCARA gCode and would work with your machine if using GRBL with no problems.
    The code started out life as a gCode filter for a laser engraver, so it has options to reduce the number of segments.
    Basically, the controller (GRBL or other) does not know your machine is a SCARA.
    I built a simple SCARA prototype to test it so more or less it works.
    I have also modified the GRBL firmware to be more SCARA friendly for the homing process. So I can provide tips for setting up GRBL for your machine.
    Finally I wrote a viewer that models the SCARAs gearing and limits. Always nice to check your design before running it. Especially if the design is out of limits.
    The software is command line driven at the moment (i.e. use a batch file to run it), when I am 100% happy with it I will convert it to GUI and with a built in gCodeSender.
    If you interested, let me know (I am not asking for anything).
    You can have a look at one of my projects if you want to check (https://hackaday.io/project/164066-prototype-scara).

    Regards AlanX

    1. Hello! I’ve checked your project… nice!. As far as I understand Marlin scara implementation works in a similar way: every linear segment is divided into smaller segments, converted to angular through inverse kinematics and then sent to a standard cartesian engine (fooled). In your proyect this convertion is made outside; every way has his pro and cons. Not the best way to deal with scara dynamics but works. And we can always trust on linuxcnc if we really need near perfect movements, speeds and accelerations.

Leave a Reply

Your email address will not be published.