Previously on our foray into making our own chess set, we had modified our CNC positives with wooden dowels in order to create registration keys and a pour spout (See Caz’s post for images). With the positives all set up, it was time to create the rubbery negative mold that we would cast our chess piece from.
I cast the negatives using SmoothOn rebound 25. First I poured equal volumes of parts A and B into separate containers and stirred them before combining them into a single container and mixing the two parts thoroughly. I spent several minutes combing and mixing the two parts to get an even orange color because both components were quite viscous and poor mixing would result in a low quality or failed mold. I then slowly poured the mixture over the two positive CNC’d pieces, making sure to avoid air bubbles by pausing the pouring process and tilting the piece around to let the rebound 25 settle. The result was a nice positive mold on the first try, although in the future we will use mold release because it was a little difficult to to pull the mold out of the recessed wood block and we had to pry them up using popsicle sticks.
Right: non-parallel cut mold. Left: parallel cut mold
After repeating the same process to create the second half of the mold we post processed the pieces by trimming the overflow with a utility knife. Upon test fitting the two halves we realized we had made a mistake and the through holes we had made for registration keys didnt line up because we had cast the same design for both halves of the mold. To rectify this, I placed the one of the halves from each mold over the CNC’d positives with the indented side face up and marked where the registration keys lined up on the mold. I then used a drill to drill out holes in those places. The drill passed through the material easily but didnt make a very clean hole. Regardless, wooden dowels could now pass through both halves and line up the two halves.
Now it was time to put the molds together and cast the chess pieces. We decided to try both molds and see which one we liked better. After putting the two halves together with the registration keys, we clamped the two halves between two pieces of wood to hold the mold together. For the first cast we used Smooth-Cast 325, combining equal volumes A and B and pouring the mold in two pours, taking a break with the mold half filled to swirl and tap the Smooth-Cast to remove air bubbles before filling the rest of the mold. Due to clamping down the molds too tight in fear of spilling, our first casts came out oblong:
The halves were also misaligned, so for the future casts we clamped the pieces minimally and spent more time aligning the pieces because our registration keys were only marginally successful due to only having one for each mold. In order to improve the alignment, I would align the bases of the two halves together by pinching the bottoms of the mold together and bending the tops away from each other (similar to peeling a banana). Once the base was a aligned I would slowly bring the top pieces back together and make sure the edges of the of the pieces were lined up as the two halves of the mold came together by peering down into the mold cavity. Repeating the process from before along with the lining up technique listed above and using Smooth-Cast 300 and 300Q for the faster set time, I got a good cast of the non parallel cut piece in two more casts.
On the third cast of the parallel cut mold I noticed that the alignment was not improving between casts. So I removed the registration key pegs and lined the parts up as best as I could. To my surprise I got a great fit but I had under filled the mold:
Spurred on by my success I tried another 4 times to get it right. Although I poured enough material each time, i must have gotten lucky because I could never get the alignment again so Caz and I agreed to give up on the parallel cut mold. Our final cast was the black chess piece where we mixed the 300 and 300Q with some black die. For post processing, we used snips to trim the air vent and pour hole protrusions and then sanded the surfaces to make them smooth and flush with the rest of the model. The bases were sanded using the belt sander and the body of the piece was sanded with 220 grit sandpaper. We also cleaned up the point where the molds met to get a nice smooth surface:
At the end of the day I learned there were several main errors that occur when casting. First, poor mixing or improper proportions and the material casts quite globby and/or sticky:
Second, as discussed before, alignment. If we were to do this project over again, we would make a CAM file with corresponding registration keys and cast two molds that line up properly as well as an air vent and a pour hole. The time it would take to set up the file and mill it would make up for the time lost pouring and casting many times to get a good piece:
Two of many misaligned pieces.
Lastly, under pouring or pouring too quickly so there inst enough time for air bubbles to escape, which results in unfinished pieces or bubbles once the material sets:
All in all, this has been a great experience and I am confident that if I needed to create a mold and cast an object using this process again, I could do it much faster and create a higher fidelity end product.
