Eight Charmander Bishop Chess Pieces

These past few weeks, we (Raeedah and Paula) have been working on the final project: making some chess pieces!

Steps: 

Two half positive volume pieces were made (left side with CNC and right side with 3D printing) from which the respective molds were taken. These were then used to cast the eight chess pieces. These are the steps followed:

1. Select a chess piece from Thingiverse
   1. We first looked into some options
   2. Then discussed with the professor and narrowed down to three options within a Pokemon chess set [1]. 
2. Downloaded files and 3D printed 
   1. Using the 3DPrinterOS program, we prepared and sent the files for 3D printing
3. Showed the three pieces to Dr.Wettergreen for feedback 
   1. We ended up selecting the pokemon Charmander, but we were told that we would have to work on the space later on
   2. We also chose (after printing two different sizes) the original size of the file as our 1:1 representation of our final piece
4. Modify the file using Meshmixer
   1. Uploaded the chess piece .stl file from Thingiverse to Meshmixer
   2. The first goal was to check that the file contained 10,000 triangles or less in the wireframe mesh. Since our file was under this threshold, we kept it at the maximum and did not have to reduce it. 
   3. The second goal was to cut the piece in half 
   4. We obtained two binary .stl files from this step (left and right)
5. Solidworks
   1. Imported previous .stl file as a solid body 
   2. Measured the width and height of the piece. Since we were happy with the size of the original Thingiverse file, we did not scale the figure.
   3. Downloaded a mold base from Canvas based on the dimensions of our piece 
      1. Our file was 2.5 in wide and 3.25 high,  so we selected a 2.5×3.5 base
   4. Assembled the base and half chess piece together following the given steps (we did not add a pour hole or an air tunnel)
   5. Save both files as an .stl (left and right)
   6. After this process was done, we had to modify the file further to add a block to the legs of the Charmander (described in Troubleshooting section) 
6. More 3D printing
   1. After we had both files, we 3D printed them. While we only needed one for later on, we printed both just in case.
7. VCarve
   1. Imported our file 
   2. The goal of this step was to add the toolpaths that the CNC machine would follow to cut the wood part, using both roughing and finishing steps
   3. We faced some challenges during this step (detailed in the Troubleshooting section) but obtained the .gcode files for this CNC process in the end
8. Used Shapeoko CNC machine to cut a one-half positive (left) of the chess piece 
   1. Uploaded the .gcode file into Carbide Motion
   2. Prepared for cutting: initialized, set the zero, changed tools between roughing and finishing, etc. (see Troubleshooting section) Video of CNC cutting

      

9. Molding process to make two half-negative molds 
   1. Made a cardboard ‘wall’ around both the CNC and 3D printed pieces 
   2. Calculated how much total volume we would need from the silicon mix
      1. First estimated cuboid volume 
      2. Then estimated (using water displacement) and subtracted the Charmander volume from the value above
   3. Added 209 g (based on 2x calculated volume) from both Part A and Part B of the silicon mixture
   4. Mixed both parts thoroughly and poured the fluid on both CNC and 3D printer pieces
   5. Left to dry and later on removed from the respective pieces for the casting step below
10. Casting the chess pieces from the molds
    1. Again we first estimated again the volume of the Charmander using water displacement (about 0.7oz).
    2. We first put together both of the silicon molds
       1. We aligned the parts and used rubber bands to hold them together (we had some troubles described in the Troubleshooting section so we later modified this method)
    3. Mixed 0.5 oz of A with 0.5 oz of B (we started with 0.4 of each to make a total of 0.8oz, but later decided to increase it) from the liquid plastic mixture.
    4. Added color or tiny glitter-like balls to Part B in some pieces 
    5. Then started a timer as we poured Part A into Part B and mixed both
    6. Poured the mixture into the silicon mold within 2 min 
       1. First poured a bit of the mixture and moved around
       2. Then poured the remaining of the mixture into the mold and waited for 10-12 min
    7. Removed the pieces from the silicon molds 
11. Post-processing
    

    Some pieces before any post-processing 

    1. Diagonal Cutters 
       1. Cut all extra material on the bottom base of the piece before we sanded it
       2. Cut some parts of the CNC left hand to look more like the right hand (3D)
       3. Used these to make small improvements overall (eg. improve the joint of the two halves within the piece)
    2. Sanding 
       1. Used three different numbers of sanding paper to make the bottom of the piece look smooth
          

          Before

          

          After

       2. Used sanding paper to sand some other small parts for some of the pieces
    3. Clear coat 

During post-processing we made our final choice on which 8 pieces to submit; we first had done 10 but then redid a few based on the TAs suggestion. Three of the final pieces were from the first batch and the other five from the second batch.

 

Troubleshooting (challenges and achievements):

Throughout this process, we faced some challenges that we had to overcome in different parts of the process:

• Charmander piece
  • The first challenge we faced was that the piece selected (Charmander) had some gaps in the legs which, as discussed by the professor, would not be able to be CNC’d (the tool can only move up, down and sideways as it removes material, but cannot go underneath any part of the wood). Thus, we added a block in this area for the two half pieces using SolidWorks using the TA’s help. 
  • However, we had not realized that this same issue would occur with the hand. We noticed when we had already gone through the SolidWorks steps and so, to avoid wasting any more time, we decided to keep going, hoping that the CNC machine would not give us errors. It did not in the end and instead cut the hand leaving the material underneath intact. This is why our Charmander has material under one hand. Based on these results, we could have also applied the same principle for the leg part, however, we decided to keep the block there anyways, not only to avoid modifying the files again, but also because it would look uneven (this way it looks like he is sitting on a block). 
• VCarve
  • Before detailing specific issues, our main challenge during this part was that we had to redo the file every time we made a mistake since VCarve does not allow you to go back, which meant that we repeated it over and over again many times (estimated 20, felt like forever; this was tiring but also made us kind of experts) 
  • Toolpath view cutting all the wood stock and taking a very long time: After following the given steps for the VCarve program, we also viewed the toolpaths that we had selected and realized that the preview was cutting away all of the wood stock as it cut the figure, which would not allow us to later on pour the silicon to make the mold. It was also very time consuming and unnecessary. Fixing this problem involved trying to shorten the stock dimensions used for the program. However, after we did this and added the roughing and finishing toolpaths, we saw that each step would still take around 17 hours. With further troubleshooting with the TAs, we modified the width and length to be the same size as the actual Solidworks piece and decided to do one rough and two finishing passes as this offered the best combination we could see to decrease the time. The first roughing was done using bit #201. Then, #102 and #111 were used for finishing. These changes saved a lot of time, and we ended up with a little over 2 hours only to cut the piece.
  • Gcode:  After we finished what we thought was the last VCarve file, we used the computer that connects to the Nomad3 CNC machine to try to upload our .crv file in order to convert it to .gcode. However, we could not open the file and so we decided to restart the process once again on this computer. We then managed to obtain three different .gcode files for each of the roughing (x1) and finishing (x2) steps. 
• CNC piece
  • The first issue we had was related to the material for the CNC machine. It was kind of messy downstairs, so we had trouble finding bits as well as long clamps to hold the wood on top of the sacrificial wood that was already there. After consulting this, we ended up using a thinner piece of plywood that we taped to the bed as well as to our piece. 
  • Wrong orientation: we did not realize that we first put the stock wood in the wrong orientation. We started the cut and had to stop it due to this since the file did not fit in this orientation (and it would not be able to hold the silicon). We thus flipped the stock wood to the other side and fixed the orientation and then restarted the cutting process.
  • Zeroing troubles: Because the final file that we made in VCarve was set to the piece dimensions, rather than the stock ones, the zero could not be started from the corner. Thus, we decided to do it manually using the probing pin by setting it in a spot in the middle. This also meant that we could not re-zero after starting the file. 
  • Short bit troubles: The roughing cut went smoothly, however, the first finishing cut started making a very high-pitch, loud noise when it was already done with the figure and was getting close to the wall. We did not know what was going on, but then we realized that the collet was cutting through the sides of the piece and burning a bit its way through. The moment that this was noticed, the machine was stopped to avoid any accident. Since the first finishing was almost done, we decided to keep moving and change the bit to do the second finishing and final step. However, we then stopped it once it got to the point where the previous step had finished and managed to obtain a finished Charmander part even though the rest was not done.
• Casting – The issues were tried to be solved as we went but were divided them into:
  • Figure alignment and leakages: Since the final cut of the CNC did not make the two pegs properly, it was hard for us to align the silicon molds together which gave. Also, at the beginning, we only used rubber bands and we had some issues with leakages and squeezing the piece (we had to add tighter rubber bands after having already poured the mixture into the mold!). We decided to use two pieces of wood to support the molds before adding the rubber bands. We later also tried with clamps and even holding the molds. However, our pieces still looked misaligned. As a result, we decided to use two curvy needles on the bottom of the silicon molds to hold the two molds as well as zip tie plastic straps, and then added the wood and rubber bands (and sometimes clamps). We also checked the alignment after adding all of this and held the piece with our hands while doing the steps below.

• • Right hand and tail: The first few pieces casted did not have the top of the tail as well as part of the hand from the 3D printed mold. One solution was using a thin syringe to add liquid directly into the tail (it was also tried with the right hand). Then, we poured the liquid onto the rest of the piece and hit against the table to ensure the liquid reached all parts of the chess piece and to try to remove bubbles. We also inserted a thin wire into the tail to try to get the material all the way to the bottom. This method filled out the tail and hand.

• Post-processing: 
  • While post processing the pieces, we had some small issues. For example, while removing material from the hand of the colorful (brown-orange mostly) piece, the hand broke. One tail also broke during post processing, so these parts were pasted back with superglue. 

• • After everything was finished and even after the blog post was published, we realized that one of the figures had not been sanded at the base (we had sanded instead one of the discard figures). Thus, after the clear coat had already been added, the sanding steps were performed in the base and then a new coat was added to this part. This last clear coat might also be a bit thicker than expected since we could not find the same clear spray at first and we then found one in the box for throwing away empty cans, so the expectation was that almost nothing would come out but quite the opposite!

Cost calculation: 

• Labor time: 
  • The total labor time for this project was about 27 hours 45 min (1665 min). However, we thought that this was a lot since it included a lot of learning curve and led to a very high price. Thus, we decided to make an estimate that would better represent the price of the figures using the more hands-on labor time (including from the moment we worked with the CNC to the time that we finished the post-processing of the figures, but excluding the software times as well as any time used to send the 3D printing files). This was about 17 hours and 5 minutes (1025 min) 
  • We mostly worked together for this project. There were a few short instances that we did not, but we still decided to add all of the times together. We believe that due to this, and for the labor of this project, we should earn $24/hour for both of us
    • Thus: $24/60min*1025 min = $ 410
• Raw materials:
  • EASYFLO 60 CASTING RESIN- $128 for 15.2 lb [2]
    • We made 16 pieces with about 0.5 oz from each Part for each. Thus, we used about 1 oz in total – somewhat less for the first pieces, but estimating with this value. 16 oz = 1 lb.
    • Thus: $128/15.2lb = $8.42105263
  • Platsil 73-25 Silicone Rubber – $ 299 for 16lb (assuming it means total weight for both parts) [3]
    • We used about 209g for each part, so 418g in total
    • This is 0.921532 lb, so $17,2211292 
  • SandPaper – $6.98 [4] (home depot)
  • Clear coat – $6.48 [5] (home depot)
  • Rubber Bands – $1.08 [6] (home depot)
  • Wood piece – $1.78 [7] (home depot)
    • Similar dimensions to our piece
  • Syringes [8] – $14.13 for 10 pieces
    • We believe that we used less than 10, but it is a good estimate of the cost
  • 3D printed pieces
    • The3D printed half pieces with the block were $0.61 (right side – which we used) and $0.44 (left side) 
    • Also for the last three pieces of the pokemons we printed before choosing, the cost was of $0.88 total 
    • We also printed some pieces previously, when deciding and making the 1:1 3D printed piece. Thus, we are adding a total cost for 3D printing of $5 since it should be lower than this amount 
• Raw materials total: $8.42105263 + $17,2211292 +  $6.98 + $6.48 +$1.08 + $1.78 + $5 + $14.13 = $61.09218183

Note: we are not including any machine time cost due to electricity cost, since this was shown to be low in the past. We also did not include the cost of some smaller materials. 

• TOTAL COST = labor cost + raw materials cost = $410 + $61.09218183  = $471.0921818 which is about $470 

 

References: 

[1] https://www.thingiverse.com/thing:3394719 

[2] https://www.brickintheyard.com/products/easyflo-60-casting-resin?_pos=1&_sid=5f68d1876&_ss=r&variant=42704770388 

[3] https://www.brickintheyard.com/products/7325u16?_pos=1&_sid=83bf067a7&_ss=r&variant=32052651622471

[4] https://www.homedepot.com/b/Paint-Paint-Supplies-Sanding-Supplies-Sandpaper/N-5yc1vZbobk?NCNI-5&searchRedirect=sand%20paper&semanticToken=d10r20r00f2200000000_202304302050262906132407196_us-east1-c36s%20d10r20r00f2200000000%20%3E%20rid%3A%7Bd2c05183a87554a6c8005a5f1def5663%7D%3Arid%20st%3A%7Bsand%20paper%7D%3Ast%20sc%3A%7Bsandpaper%7D%3Asc%20ml%3A%7B24%7D%3Aml%20ct%3A%7Bsandpaper%7D%3Act%20nr%3A%7Bsandpaper%7D%3Anr%20nf%3A%7Bn%2Fa%7D%3Anf%20qu%3A%7Bsandpaper%7D%3Aqu%20ie%3A%7B0%7D%3Aie%20qr%3A%7Bsandpaper%7D%3Aqr

[5] https://www.homedepot.com/b/Paint-Spray-Paint-General-Purpose-Spray-Paint/N-5yc1vZcgeo/Ntk-elasticplus/Ntt-clear+coat?Ntx=mode+matchpartialmax&NCNI-5&visNavSearch=clear%20coat

[6] https://www.homedepot.com/s/Rubber%20bands?NCNI-5

[7] https://www.homedepot.com/p/8-in-Wood-Shims-12-Piece-per-Bundle-WSSHW08/300723328#overlay

[8] https://www.walmart.com/ip/Slim-Nutrient-Syringe-1ML-10pcs-Kitchen-Cooking-Syringe-Disposable-Plastic-Nutrient-Syringe-Injector/1913372500?from=searchResults