For our final project, we were put into pairs and had to create a chess using the CNC mill, silicon and easyflow liquid plastic.  Franck and I decided to craft the Queen, and we decide to represent it as the Empire State Building. We decided on this model we found in Thingiverse since it looked nice and seemed relatively ease to work with.

First, we had to take the .stl file from Thingiverse and turn it into readable g-code that could be understood by Easel, the program that controls Carvey the CNC Mill. For this, we used Autodesk Fusion 360.  Now, in retrospect, Fusion 360 is relatively neat program that I think we understand pretty well, especially for the great price of free (for students). But at the time, FUCK FUSION 360. When we tried to do it in class, so much of what we were supposed to do went nothing like it was supposed to in the instructions that we ended up losing a day and becoming very frustrated in the process. Thankfully, next class we finally figured it out and had a piece ready to be carved. Since it was our first attempt, we accidentally made the piece with no pegs or pouring hole.  We realized midway through the cut that it might be a problem, but thought that we could create a way to pour easyflo in the piece later.

The next step after carving is making the negative molds. Since our piece is symmetrical, we needed to make two silicone molds using the same wooden positive.

However, another problem arose. We did not leave enough space between the top of our carve and the object, meaning that unless we overpoured we would have no base on our negatives.

So, we improvised.

Once we had working silicone molds, we needed a pouring hole. We tried lots of different ways, including the ones shown below, but there were no clean ways to make a hole to pour the easyflow, or liquid plastic, into the mold before it solidified.

We had to go back to Fusion 360 and fix our file. Easier said than done. It took hours, but we finally were satisfied with our file.

To the Carvey we went again, hoping it felt merciful. It must have been in a great mood that day because our carve went off without a hitch. Time to make the molds. If you didn’t know, silicone takes around 4 hours to solidify. We had to make two silicone molds before we could do anything else. Spoiler alert: this wasn’t the last time we had to use silicone.

While we don’t have photos of our second molds, we do have photos of the results. After the molds were ready, we mixed the Easyflow chemicals, added a few drops of green and blue dyes to make aquamarine, and poured them into the molds as such:

We got a few prototypes in before we realized that the molds were simply not aligned with each other. Somehow, the code was wrong. Again. Which means we had to use Fusion 360. Again. Which meant we had to carve our files. AGAIN. We must have used all our good luck on the first days, because we were struck with delays after delays. First off, Fusion 360 cannot easily modify extrudes for some weird reason. We spent an afternoon just figuring out what we could do to our file. Then, Carvey messed up our carve. TWICE. And also we used the last piece of wood available. Really, it couldn’t be more perfect an example of Murphy’s Law.

Things started to look up after Tori cut us more wood (thanks, Tori!). The carve worked, the silicone molds came out nice enough, the days were nice. Then our luck ran out. Unveiling the first mold was actually painful. It was also mismatched. With our expectations low, we unveiled the next one. Gloriously, this one was actually somewhat presentable. Unfortunately, this was the highlight of its group. We went through more plastic molds and did not get any acceptable ones. Clearly something was wrong, and upon further inspection, it seemed that the silicone molds were slightly off center with each other.

Unfortunately, this late into the week we couldn’t do another carve. So we did what we do best: adapt. We cut the pegs and manually, slowly, carefully, meticulously, we joined the two halves until we were sure to get a useable plastic positive. We repeated this process until we had five total possibly-acceptable positives, which were swiftly moved to post-processing. The only thing we really did was sand the side joining marks and bumps, and belt sand the bottom, and we were done.

Overall, this final encapsulates this class.  We started early, found errors, adjusted to those errors, and then it was the day before the project was due (unless that was just me, in which case, oops).  But in all serious, this class was one of our favorites, and a provided a great outlet to be productive and do real life work whenever some of the more…conceptual classes (impractical nightmares (@CAAM 336)) where annoying.  Most of all, this class was one of best we’ve taken at Rice. Signing off for the last (and first time) time.

Carlos & Franck