Molding and Casting

Design a 3D mold, machine it, and cast parts from it.

Learning outcomes:

• Design appropriate objects within the limitations of 3 axis machining
• Demonstrate workflows used in mould design, construction and casting

Concrete evidence:

• Explained how you made your files for machining
• Shown how you made your mould and cast the parts
• Described problems and how you fixed them
• Included your design files and ‘hero shot’ photos of the mold and the final object

A New Process

Molding and casting was an all new experience. Designing parts was straightforward, but I had some difficulty getting toolpaths for a Roland MDX-40A mill from FabModules.org or Mods. Because of the other time constraints and time spent working on trying to get the Roland mill and machineable wax to work, many constrained choices had to be made for the week.

I ended up milling out a design on the Shopbot, which I could get working quickly with vCarve. The design was milled out from 1.5inch pick foam insulation. Allegedly, the shopbot just can make as detailed features as the tabletop Roland mills are able to do, although my outcomes suggest otherwise. [During the composites week, I did see the ShopBot made a nicely detailed 3D form from an STL.]

Design Choices

I chose a rather simple design to make: the logo for the school where I teach. The choice for this simple model was for three reasons: 1) I was skeptical that the ShopBot would resolve good detail; 2) The foam I had was barely thick enough to introduce much vertical geometry; 3) I needed to make sure I made something that paid homage to the school district that was subsidizing my study in Fab Academy. This design is not inherently fully 3D, rather 2.5D, which concerned me for not meeting my weekly objectives. So I did take the time to ensure that there was an attempt to run the toolpaths in 3D over the design and that I would choose a 3D design during the composites week, to get experience milling a 3D shape.

Models to Toolpaths

I made my designs in SOLIDWORKS, and then exported an stl file to vCarve for the ShopBot. Again, I had difficulty getting the right outputs when I was working with the Roland MDX-40 mill. Neither Fabmodules or Mods seem to support this mill (as far as I can tell). I spent hours trying to figure out these processes, but time was tightly constrainted this week because of school duties. I am more familiar with vCarve, so I moved on to that workflow. Once in vCarve, it is easy to design toolpaths above/ around an stl file, with only a few settings. I generated three toolpaths that would:

1. Roughly remove material above the stl
2. Apply a 3D finish cut over the surface of the stl
3. Remove my section of foam from the machine

Milling the Forms

The rough cut toolpath (1 above) left the large, blocky design looking good. There was some roughness around the edges, primarily along any edges that were angled off the dimensions of the machine. My hope was that these would smooth out with the 3D finish cut (2 above), but the image of the toopaths on vCarve was too dense with toolpaths to tell for sure. When I ran it, the finishing toolpath started in the center with an interesting spiral-outwards technique to go over the entire top surface for the foam. This action started on top of the center island and when it spilled over to the lower levels of the foam features, it was not a seamless transition. The machine had taken what was suppoesed to be an agular step-down sort of profile and instead calculated the toolpath to be an angular descent. At this point I double-checked to make sure I had included the correct settings and mill bit for toolpath generation - I had. The machine just generated goofy toolpaths for the finish cut on this design. Instead, I decided to run another cut, but omit the finish toolpath. One of my classmates had done this and seen good success, so I would try it as well.

Also, it is important to note that since the second toolpath made such a mess of the form I was making, I did not run the third (cutout) toolpath. This was simply for the sake of time but, while pulling the form off the ShopBot, I realized that it would have been a problem if I had cut it: I needed walls around my form so that it could hold the Smooth-On two-part urethane. This realization had me add a footer to my model, so the urethane would have walls to hold the casting material. I also needed to sink my entire design deeper into the foam so that it could be properly enveloped in urethane when I got to that stage of the process.

Milling Solutions and Casting

The final toolpath combination that was run included only a rough-cut and a path to remove the section of foam from the larger stock, but with an offset to leave walls for the urethane. This gave a nice look to the design, even if it was still a little rough around the edges. It would certainly be good enough for a garden stone. [In a later week, it became clear that increasing the path overlap would render more detail.] From here, I applied two coats of release material and then made the urethane mold with Smooth-On two-part urethane.

After a 24 Hour curing period, I cast two materials, with my primary concern the whole time being speed. I started casting on Tuesday evening. First, I cast Plaster of Paris. I did this for two reasons: it is cheap and I knew it should be reasonably fast. I also embedded some metal screening into the plaster, in order to make it a little stronger. [later I would learn this made it actually a crude composite.] This cured for 2 hours in the mold and then was demolded to dry further. Only 2 bubbles.

The second material I cast was certainly non-traditional. I used Quickrete - yes, quick-drying concrete from Home Depot. I had used this material before to set posts in my yard, so I knew it was both fast to set and cheap. Only, I had forgotten that concrete and cement are not the same. A silly mistake, but all of this week's work is getting done under strict time constraints and with my head other places. Life is busy. So, before the concrete went in the mold, I ended up using some metal screening to pan out all of the pebbles. Throwback to a geology class... This left me with a powery quickdrying cement-like material. I mixed it in a 2:1 ratio with water and poured it into the mold. Two hours later, I came back and was able to demold the material. Once out, it also quickly dried on the front. I could see the color lighten as the moisture exited the material before my eyes. I even used a small soldering heatgun to expidite this. I was amazed by the process fo the moistue wicking out so quickly.

All of these last-minute decisions aside, the final cement cast came out nearly as well as the urethane would allow. There were hardly any bubbles, and all were very small. The fine texture on the cement seemed to be no more significant than that on the urethane. The shape ended up being quite accurate to what I had intended and the finished product seemed nice enough that it could be used for a reasonable gift or decoration in an office or garden.

======================= UPDATE ===============================

I later baked the cement in a toaster oven for three reasons: 1) we have a not-for-food tabletop toaster oven 2) I vaguely remembered a friend in college telling me he had to bake cement for a class, to make it stronger 3) I have no real imminent need for the cement part, and I can make more from the mold. Overall, it was an interesting experiment. I put the cement in the toaster the day after demolding, for 20 minutes at 350F. When I opened the door, I could see that the part had split into two large pieces. From that point on, the cement also became very brittle. Breaking when only small force was applied. The Plaster of Paris piece is still in one piece, and looks the same as when it came out of the mold. If/ when I do this with students in the future I believe I will use Plaster of Paris because it is both cheap and readily available.

You can see all images (used and unused) from this week's work HERE.