Sarah and Alice’s Chess Pieces

Hey everyone! For our final project we decided to use Thingiverse.com to make a bishop ghost chess piece that we would mold and cast using both the 3D printers and CNC machine at the OEDK. This project was fun for the both of us and definitely much less work than the midterm project, but overall we’re really happy with how our pieces and molds came out! Follow along below for all the details on our process and how we got to our final product!

GATE 1: Choosing our chess piece

When looking at Thingiverse.com, there were many different objects we could have chosen from, but we both agreed that we wanted something that would actually look like a chess piece rather than something turned into a chess piece. Due to this, we started our search by looking at existing chess sets on the website and letting our creativity go from there. We ended up choosing the “Cute Ghost Chess Set” by @ethanjurman because of the smooth/rounded shape of the pieces as well as the variety we could choose from since it was a full set. From here, we 3D printed both the king and bishop in order to have two options to choose from. We kept the scaling from Thingiverse.com for this iteration then picked them up later that night.

King and bishop ghost chess pieces ready to be 3D printed on the Bambu printers.

When the 3D prints came out, we found that the two pieces were fairly similar in size with the bishop being a bit taller and the kind being a bit thicker. We decided on the bishop because the king actually did not have a flat bottom. Due to this, there was an overhang that we couldn’t correct in our 3d prints. We also liked that the bishop piece had an interesting shape on the front with the two inlaid eyes as well as the curvature throughout so we went forward with that piece, but scaled it down to a height of 2 inches.

King and bishop ghost chess pieces 3D printed for comparison.

 GATE 2: 3D printing 1/2 of our mold 

Moving into Gate 2, our main goal was to use MeshMixer to attain the back half of the mold split down the middle, combine it with the provided mold base, and 3D print the combined shape on the Bambu 3D printers. To start, we imported our .STL file from Thingiverse.com into MeshMixer in order to configure it in a way that could be added to the mold base. We chose the backside of the bishop piece as we thought it wouldn’t need provide as much detail and we expected our CNC’ed part to provide more detail than the 3D printed one.

Bishop piece cut in half in the MeshMixer software

The MeshMixer software was helpful for us to visualize our part as well us understand the size and number of triangles it would be made up of. This was useful to make sure it could be exported and would send to the 3D printer. Our piece was fairly small and we were sure to import it as a solid shape so the number of triangles was insignificant and uploading it to the printing software was easy.

The total print time for this part was 37m 12s (model printing time = 30m 19s, prep time = 6m 53s) with the cost being $0.23 and the amount of filament used 151.50 in.

Back of bishop piece ready to be printed on Bambu printer with mold back attached

The design of the bishop backside plus mold base was based off of what was already on the mold base as well as additional features we added in order to make the molding process more efficient. The extruded circles were already on the mold base and are used to confirm that the front and back pieces are able to fit into one another in the correct position. We would hate to go through all the work of casting our pieces just to realize our molds weren’t held firmly in place and therefore aren’t able to make the object we want. The other two features we added were the air tube and pour hole. The air tube is used to allow any bubbles to escape if trapped in the bottom of the casted piece while the pour hole is to allow an easier pouring experience by making the space to pour into larger. We were not extremely precise with the pour hole, but could have centered it to make it match the other side if we felt necessary. In the end, it worked well so it wasn’t a big problem for it to be off-centered.

Back of bishop piece with mold base 3D printed

To make the mold of our piece, we then made a box of indeterminant size with hot glue and tape in order to make sure there was no leakage at the base. One way we tested this was by placing our piece over the ceiling light to see if any light shines through. No light shined through on ours so we went ahead and poured the mold liquids together. For both of our mold bases we used the white Magikmold P-525 Silicone Rubber Base along with the blue Magikmold P-525 Silicone Catalyst; these were measured 1:1 by weight, mixed together, then poured into the box. We let it sit for an entire day, although it probably would have been good after around 5 hours, and came back the next day to take them out of the mold.

Box of indeterminant size created to enclose 3D printed piece in order to make first mold

First mold (backside of bishop) setting

GATE 3: CNC’ing the other 1/2 of our mold 

For Gate 3, we worked together to take the front side of the bishop (which we had already cut and put into a file), attach it to the same mold base, then export that file as a .STL to the VCarve software. This step took a bit more math than we had expected in order to center the piece in a way that would match the other back half, but a bit more organization on our parts definitely could have helped with that when making the back half.

Additional math needed to match our front half piece to the position of our back half piece

When getting to the VCarve software, we made sure to measure our piece of wood with calipers in order to get the most exact measurement. We measured our block of wood to be 2.4995″x5.429″x1.352″, but had to play around with the coordinate system in order to make sure our piece was both oriented correctly on the block of wood and on the CNC’s coordinate system. We went through many iterations of our piece being squished, stretched, and cut off in certain places so definitely had to use some intuitive thinking to figure that out.

Setting block measurements on CNC machine

Playing around with the VCarve coordinate system in order for it to see the cuts we wanted to make

Due to the Shapeoko CNC mill being out of order, we also 3D printed the front side of our mold just to have a back-up plan if we couldn’t find another time slot to CNC. Our part ultimately was able to be CNC’ed on the Nomad CNC mill, but it was helpful to have the 3D printed piece just as a back-up. This only costed us $0.32 and around 30 minutes of print time to wait for so we thought it was worth it in the end.

3D printed front side of bishop

When working on the Nomad CNC mill, we first created the Roughing pass where we were able to CNC most of the wood that would need to be taken out without going into too fine of detail. We used a 1/8″ end mill for the roughing pass as it has a sharper edge on the end for cutting whereas we used a 1/8″ ball end for the Finishing pass. The entire cut took around 1 hour with the roughing pass taking about 42 minutes and the finishing pass taking around 20 minutes. We really liked how the CNC negative came out and did seem to have a bit more definition than the 3D printed back side.

Lots of wood shavings from initial roughing pass

Roughing pass

Finishing pass

The next day, we came in to create the mold for the back side of our piece just as we had done for the front although the mold we had created for our front side was actually a bit squishy so we decided to redo that one as well. We believed that the initial mold would have allowed for too much deformation when the mold halves were put together and flattened the back side of our piece instead of keeping its original shape.

We ended up creating molds for the CNC’ed front side, 3D printed front side, and 3D printed back side. We did the CNC and 3D printed molds of the front just to see which would be better and actually ended up liking the 3D printed one better so planned to use that one instead.

First pass at creating our front and back molds

Sadly, when we came back a few days later we were only saw our front side mold from the CNC’ed negative (which we weren’t planning on using). This was weird as we knew we’d left it there so we looked around and asked the TAs and Dr. W, but nobody seemed to know where they would be. Instead of wasting time, we decided to use the time to mold the front and back again so that we could come in the next day to do our casting. This was a bit inconvenient, but fortunately we were ahead of schedule so it wasn’t the end of the world.

Creating our chess pieces

When casting our pieces, we used rubber bands to secure the two mold halves to one another, poured 8 grams of the Castable Resin Part A and 7 grams of the Castable Resign Part B into a beaker, mixed it together, then poured the mixture into the pour hole.

Below is a summary of how our different casts came out:

Cast No.  Keep/Trash Notes
1 Trash Used deformed mold back side just to test it out, but found the back fairly squished in due to the rubber bands
2 Keep Air bubble near pour hole. Used 3D printed front so we can see the lines a bit more.
3 Keep Small air bubble near pour hole. Used 3D printed front so we can see the lines a bit more.
4 Trash Used the CNC mold and it leaked quite a bit without us holding it so the volume was less. Ended up having a hole in the bottom due to the leakage.
5 Keep Used 3D printed front but must have not filled enough. Filled the hole with excess and let it dry.
6 Keep Small artifact on bottom.
7 Keep Added blue dye. Overall no artifacts or bubbles!

Post-processing our chess pieces 

After we found the 5* casted pieces we wanted to use, we moved on to post-processing with 150 grit sandpaper. The only place it was difficult to sand was inside the eyes so we did the best we could there. We sanded all the edges and grooves of our pieces and are really happy with how they turned out!

The main artifacts we had were air bubbles at the base or not having enough casting materials which left a small pocket. Most of this was able to be post-processed away, but we did figure it out by the end on our blue ghost!

Overall, we really enjoyed this project and ended up with 5* pieces we were proud of! Below is an image of our last clean workspace of the semester!

* Due to our molds going missing, Dr. Wettergreen told us to just complete 5 molds instead of the original 8. This was very kind of him and we appreciated it a lot on the time end of things!

Cost Breakdown

Cost Type Cost Price Source Quantity Total
Materials Cardboard $2/box Packaging Hero 1 $2.00
Magikmold P-525 Silicone Catalyst $265.00/16 lb. kit Brick in the Yard Mold Supply 300 grams $10.95
Magikmold P-525 Silicone Rubber Base
TC-802 Castable Resin Part A  $129.00/gallon kit Brick in the Yard Mold Supply 105 grams $3.57
TC-802 Castable Resin Part B
PLA Cost from printer for both mold bases and the two test pieces.  $1.23
2x5x1 in Pine Wood $2.78/6-ft Home Depot 5 inches $0.19
Labor Prototyping Engineer (You!) $20/hour ZipRecruiter.com 17 hours*2 engineers $680.00
Overhead Facility Cost (Machine Time) $90/31 days TXRX Labs 1 month $90.00
Quality Control $40/hour Glassdoor.com 1 hour $40.00
TOTAL $827.94
Print Friendly, PDF & Email