It turns out that a lot can happen over the span of only a couple of months. A much accelerated and abbreviated version follows:
a) I designed my pieces (dinner plate, cappuccino cup, and a soup bowl) in Rhinoceros based on typical dimensions (diameter, height, etc.) of these types, as the invention of this set involves the patterning and surface effects rather than the form. I put these surfaces through Grasshopper to achieve the desired surface faceting . (Grasshopper script to follow)
Top view, showing the faceting
Perspective View
b) I prepared files to make the positive molds of these pieces, so I could simply pour plaster into the cavity and out would pop a perfectly slip-cast-able mold for each of these pieces. To make matters more difficult, I decided that achieving a crisp, faceted interior was as important as the exterior, so I would have to make multi-part molds and solid cast the pieces.
Pro - perfect edges and rims, sturdier final objects
Cons - complexity of molds, increased milling time, possibility of release issues with inside part of mold
Everything but the cappuccino cup was a relatively straight-forward 2 part mold. That required 4 pieces...
c) I CNC milled the pieces out of MDF, thinking it would deliver a higher level of resolution than foam and without the fuzziness, which proved to be mostly correct. As usual, milling these pieces lead to some unforeseen challenges that I could have designed for, had I thought more about the process. Some were too deep for a bit small enough to secure higher-resolution milling results, my registration marks were squares (impossible with a round bit...), and some of the patterning was simply too small. All in all, these parts came out of the milling process quite nicely.
The inside and outside parts of the plate mold during the milling process
d) Since MDF swells upon contact with water, it's extremely important to seal it exceptionally well before pouring the plaster in. The internet said that a 50/50 mix of wood glue and water painted on to the surface would work well, and I ran with it.
e) After sealing the positives with the wood glue mixture, it was time to pour the plaster and wait for it to dry, then simply pop the plaster out of the MDF and repeat the process to my heart's content, creating multiple molds from each positive and dramatically increasing my production capabilities.
f) DISASTER.
g) None of the plaster negatives released from the MDF. Well, none but one. I'm not sure if the wood glue sealer didn't work, or if the issue was in the design of the pieces (no draft angle, too much vertical face, small details to grab on) though it's likely a combination of the two. Needless to say, this was less than ideal, as I had the plaster that I needed stuck to the MDF that I did still want to be able to make additional molds from, especially after so many hours on the CNC mill...
If I had listened to my professor this would have been avoided. Simply run the pieces through the band saw to remove the vertical sides from the bottom and I'd essentially end up with my own mold coddles, which would have ensured that everything released nicely. In my infinite wisdom, I thought that everything would be fine and the plaster would release perfectly just by sealing and applying Vaseline as a mold release. Wrong.
h) A very hands on process to remove the very stuck MDF from the plaster ensued, using chisels and saws and taking the better part of a day. It made a bit of a mess out of the cappuccino cup mold parts, which will become slightly problematic later on...
i) It turns out that solid casting is another challenge... Ensuring that the mold has enough slip to cast solid proves to be another challenge and it took several tries to begin to get this process correct. The cappuccino cup mold also has some imperfections because of the struggle of getting it out. As such, it has to be packed full of clay around the exterior to not leak everywhere, making the casting process more involved.
The bottom of a non-solid piece removed too soon.
The aftermath of a bad cast.
j) The casts are turning out better with more practice but there's a good bit further to go and many many more to make! More photos and updates to follow.
Two of each before bisque firing.