As the midterm project began, I decided that my illustrator project wasn’t quite exciting enough. I decided to check in the older book by Levy and I saw the steam engine model on the cover. It has many of different types of mechanism integrated into one. The crankshaft uses an older design called an Eccentric drive to transfer linear motion from the piston to rotational motion of the crankshaft as well as linear motion in the valve train without the need for a separate camshaft. For single-piston designs like many steam engines, this reduces complexity at a slight cost of additional friction. It is this Eccentric mechanism I decided to tackle first as it seemed like the hardest to get working smoothly. Because the tolerances must be so tight for it to be smooth, my rough prototypes were laser cut from the beginning. I cut the eccentric and strap out of 2 plies 3/16 inch birch and sanded them each with 400 grit sandpaper to get them smooth. I then cut a mock-up backing plate out of 3/16″ oak to test the layout and rotational smoothness. In this first prototype, I also made a first attempt at the piston connecting rod and guide. Using some clamps to temporarily hold everything together, this was the result:

I was happy with the smoothness of the eccentric and the strap, but the oak backboard, connecting rod, and guide were all causing a fair amount of friction. To remedy this, I re-cut them out of the same birch. For the next prototype, I began the process of making the valvetrain, a significantly more complex mechanism. The small connections required between the components of the valvetrain meant I had to make dowels that stepped down from 3/16″ to about 3/32″. This allowed them to drive the mechanism without requiring holes so large that they would threaten to crack the small components. In this iteration, I also began troubleshooting the box simulating the cylinder head. I found that the locations of the holes allowing the cylinder and valve dowels to pass through needed to be very precise, otherwise they would force those components off of parallel and caused them to bind up. After 3 or 4 iterations of the cylinder head, I ended up with this medium-fidelity prototype. Even though parts of it were still held together with clamps, and some of the valve train pins were substituted for Dremel sanding drums, the action was fairly smooth.

For my final product, I made quite a few refinements over the medium fidelity. Most obvious was replacing the wooden cylinder head box with an acrylic version, the sides of which I sandblasted to focus attention on the front. I also added a free-rotating sleeve to the handle to lessen the friction on the fingers. Finally, I replaced the Dremel sanding drums with stepped dowels and retaining pins to finalize the arrangement. To mount it all, I used 1/2″ poplar, which I had to cut with 6 passes on the laser cutter. The outcome wasn’t entirely square, but some filing fixed that.

I also used a linseed oil finish on all of the visible surfaces, the results of which are highlighted here: The linseed oil finish is on the left. It doesn’t darken the wood like most stains. Rather, it just brings out the depth of the grain and helps smooth the surface further. Overall, this project has been very rewarding. 20 hours in the OEDK later, I think made a final product I can really be proud of.