Making bits of the past

While researching historical topics, I love it when I come across plans for how things work, or explanations of how to make things. Old patents, instruction sheets, blue prints, and articles in magazines like Popular Mechanics. Google has made some of these sources more readily available through their services.

When the back catalogue of Popular Science became available via Google, I immediately began to look for old issues that explained magic tricks. An issue in 1944 had a small section on magic tricks that one could make, and the plans included diagrams with measurements. In particular, one puzzle toy stood out -- a cone that seemed to roll uphill.

The object was simple enough, and the details explicit enough, that it seemed to be something that would make a good test with the Makerbot. How easily would it be to take the full plans for something, and materialize them into a physical thing?

I turned to Google SketchUp to build virtual designs of the object. In order to print one, I had to take the limits of the Makerbot into mind. The build platform is only roughly 4x4x6 inches, and the ramp was larger than that. The geometry of the whole thing also seemed like too much to print, so I set about breaking the piece up into the most basic components. Due to the symmetry of the piece, only three parts were necessary: a cone, and two sides of the ramp.

For me, one of the strengths of Google SketchUp is being able to input actual measurements, and have the program adjust a component being made to that size. I had to convert the imperial measurements to metric, but was then able to draw each of the parts, and form them into objects in SketchUp. I exported the files as STL format, and set began to print them.

The first problem I ran into was that the length of the ramp portions were at the diagonal limits of the build platform. Building a raft for it in Skeinforge covered the entire build platform, and the extruder head ran into the bolts holding it together. It made a pretty bad sound -- not what you want to hear while printing. Adjustments to the Skeinforge settings fit it all onto the platform, but the next issue was warping.

ABS shrinks about 2% from when it is hot for extrusion to a cooled state. As the layers build, each shrinking 2%, this causes the piece to bend upwards from the edges. The thin bases popped off a few times due to warping. The company behind the Makerbot released a heated build platform that appears to eliminate the warping.

Eventually, I was able to get all six pieces printed, and was surprised at the consistency of the warping. I stuck them all together -- the two cones with double-sided tape, and some clay was nearby for the ramp. I set the cone on, and it rolled up the hill. Success!! The files are posted at Thingiverse, along with a photo.

While neat to look at, I was mostly wondering about what other sorts of plans could be used as the raw data for objects. Plans for things exist in a variety of sources, and it is getting easier to both digitize that data, and to materialize it in some form.