• Background

    shawn.deol12/24/2020 at 01:08 0 comments

    Wow this is my first time documenting a project in a very long time....

    So basically this project takes inspiration from ZenXY, Sandscript and the Sisyphus table. I've done a lot of the groundwork and this post will be a summary of the background.

    Step 1: Core Components

    After doing a lot of research into existing sand plotter I've come to the following list of components as "crucial"

    • Gantry motion system 
      • Examples include cartesian (CoreXY, Hbot), polar (lazy susan+slip ring or scare robot) & many others

    Are CoreXY 3D Printers Taking Over? « FabbalooBeckhoff Information System - EnglishSisyphus: the kinetic art table - Raspberry Pi

    • Magnetically coupled ball bearing 
      • The gantry will carry a magnet under the sand bed that's coupled to the ball bearing
    • Cool LED strip 
      • To cast shadows & better see the patterns created!
    • Stepper motors
      •  Easiest motor to work with for this application
    • Motor control system
      • This can be a CNC SBC (single board computer) for easy GCODE motion or an Arduino+stepper controllers such as a motor shield if you want to make simple geometric patterns
    • Sand bed
      •  Type of sand & bedding material vary

    Step 2: Motor form-factor limitations

    So I want this thing to be TINY but just how tiny is tiny? Briefly looking at available components on the internet I quickly realized that the limiting factor will be the size of the motor. I essentially see 3 classes of motors suitable for the task of a sand plotter (open loop position control, easy location based movements)

    1. DVD drive micro steppers

    These guys are difficult to control, expensive to source & hard to mount (very low torque as well). For these reasons I've largely ignored them 

    Hack old CD-ROM's into a CNC Machine | Tinkernut Labs | The Curious Mind Of  A tinkerer

    2. Hobby/toy stepper motors (geared, eccentric shaft)

    So I tried these guys out for my first prototype, unfortunately the vendor I used has very poor quality control which resulted in multiple motor failures. They also don't have the best mounting features & the shaft can quickly overheat causing spur gears to slowly melt and deform (if printed from PLA)

    Hobby Unipolar Stepper Motor | Creatron Inc

    3. NEMA standard size stepper motors (no gearing, pretty beefy)

    NEMA motors come in a variety of sizes, all the way down to Nema 8. Unfortunately Nema 8 is pretty expensive and I wasn't able to locate any pancake (low profile) variants for that specific size. 

    What do NEMA sizes mean? | Applied Motion

    Fortunately Nema 17 pancake stepper motors are commonly used in 3D printing and I was able to snag a few for $10 a piece from Amazon.

    Step 3: Enclosure form-factor limitations

    With the motor issue being dealt with, the remaining form factor limitations are electronics & other general enclosure related factors. I plan on using an Arduino Uno to keep this project fairly simple. As for the enclosure, I plan on using an Ender 5 to print the round enclosure (yes I want round) which gives me a maximum bounding box for the sand plotter.

    Other enclosure related limitations include 

    • Size of viewing window/lid (I was able to find round discs of acrylic on amazon in a few different sizes)
    • Size of your office desk?
    • Honestly at this point there are no other limitations...

    Step 4: Motion system form-factor limitations

    OK so I lied, the type of motion system you use will heavily effect your form-factor and final footprint. While CoreXY is very efficiency for 3D printing & other typical cartesian applications, it leaves a lot of deadspace for a sand plotter which would require a larger bezel (not a big fan). Because my project is intended to be very tiny this issue is amplified so CoreXY/Hbot/Cartesian systems are out of the running.

    Next there is the variety of polar solutions, many of which require typically requires a slip-ring or less than 360 degrees of rotation (another non-negotiable for me), other polar solutions I've researched result in quite a high z stackup (total height) which also put me off.

    The final option (and the one I settled on) was a modified Scara robot. This modification is very similar Mark Roland's design which in turn in similar to Rob Dobson's design (funny how that happens)