A couple months ago I stumbled across Fab@Home and knew right away I needed to build one of those machines - basically a 3D printer that lays down layers of silicone to build an object. Well, plans change, and I ended up building a CNC router instead, but it was close enough and my craving was nonetheless satisfied. Here’s a dramatic picture of the nearly finished product:
And a less dramatic one:
Now the longer story. I started buying parts with absolutely no idea what I was doing. Because my original intention was to build a fabber, I obediently ordered the main electronic items from the Fab@Home bill of materials, starting with a Xylotex 4-axis stepper controller, some stepper motors, an LPC2148 ARM processor and a JTAG adapter. While waiting for that stuff to arrive, I did some research and realized that although it would be super cool to make the Fab@Home machine, the blobs of rasterized silicone it produced looked like crap and I really couldn’t think of anything I wanted to make with that particular look. In the meantime, I saw all these videos of CNC routers - a much better fit with my penchant for woodworking.
After several false design starts, I found http://buildyourcnc.com/ with videos of a guy building a CNC from start to finish with just a vague idea of the dimensions he wanted and a measure-as-you-go attitude. This was perfect. I followed his general approach but with a much smaller footprint, ending up with about a 13″ x 11″ x 3″ toolspace, as well as a few other modifications worth mentioning.
I didn’t bother making notches for the bearings, opting to bolt them on instead, as others have done. I also used bigger bearings, which were the smallest I could source locally.
I embedded the transmission nut in a piece of plastic rather than bore a long hole through the MDF. I didn’t like the idea of the threaded rod going through the board - partially because the insertion of the transmission nut seemed to cause the MDF to split in the videos, and partially because I didn’t trust myself to drill those long straight holes. It turned out to be a non-issue since I was using a larger diameter threaded rod that would have required too large a hole.
I went with a solid block of wood for the motor mounts. Dangling the motors off the end of four bolts didn’t seem very solid. I didn’t even try it, so I could be wrong. For the motor couplers, I cheaped out and used clear tubing and small clamps.
For the router mount I went with two brackets on the front instead of one, making for easy plumb adjustment. Actually, I didn’t have enough room for cross dowels if the front bracket were on the same plane as the inside due to the bolts holding the inside bracket on, so this design deviation was a necessity.
Once it was mostly together, I took it apart to paint it, but not before some test runs with a Sharpie on paper, noting areas that needed tweaking during reassembly. It’s painted and back together now and running nicely. The only thing left to do is mount some limit switches.. and spend some quality time with the software.
Here’s a hasty video of it drawing in sand.
In the end I barely scraped the surface of CNC design and construction. I should really have spent a few days reading the sites and learning from the mistakes and wisdom of others, but as usual I just plowed ahead blindly. Seemed to work out this time.
Oh, and I didn’t end up using the ARM microcontroller, but I have other ideas for that. Stay tuned.
August 23rd, 2008 at 9:27 pm
Super Cool! You are Bill Nye and Norm Abram’s love child, I know it.