( h o w . t o . m a k e ) a mold (and a cheap 3d scanner!)
responsive environments group :: MIT media lab
This week we learned about molding and casting. The assignment was to make a wax mold, specifically by milling it, and then to cast rubber in the wax, followed by plaster in the rubber. I wanted to replicate the 3d scanning and printing week using the far cheaper wax medium. First, I tried using stl2png.py to convert my face (in 3d scanned stl format) to cad.py to a bitmap, for milling with cad.py. After running for an hour, the process ran out of memory. I tried vastly reducing the number of points, but the script wouldn't budge. Next, I tried using the Modela Player software, which can natively read stl files. I was really impressed with the simple interface for setting up the job and moving the collet around, but unfortunately the machine wouldn't move at when I finally sent the job to print. John and I spent three hours on it before giving up (update: it appears that John and Hannah ultimately got the software working on another computer).
Returning to the world of cad.py, I created a simple 3d scanner using two flashlights and a webcam in a dark room. By pointing the flashlights directly at each other and placing an object (my face) halfway into the beams, I got an 3d scan in bitmapped form: the "tallest" point, at the tip of my nose, was brightest (being closest to the center of the beam). Of course, dark spots (eyes, etc) appear as holes, but the overall 3d contour of my face is approximated.
I then milled the image out in wax. Short on patience after my experience with Modela Player, I chose to cut only 20 steps in z, resulting in a pretty voxelated mold after about 45 minutes. The flashlight 3d scanner works!
The rest of the process was fairly straightforward and fun: mix, pour, and bake for an hour.
And the final step is much easier (and cheaper) than 3d printing: mix, pour, and wait 20 minutes; repeat indefinitely. I even stuck some LEDs in the mold as it hardened.