For this week, I found a cute whale design on Etsy, which I decided to develop with Blender. I took the original image and used it in an Empty object to guide me during modeling. The first stage was to draw the base polygonal faces making up a planar section of the whale. I did that by creating a basic quad, which I then cut into using the Knife tool.
I then extruded the main face (after removing the initial quad vertices and associated faces) to give a basic volume. At this stage, I used the sculpting tools to shape it into an approximation of the original whale.
My main sculpting tweaks were:
I then used solid geometry to create a single-piece mold (as a first attempt to molding). Initially, I didn't realize that we'd create two molds: the first one in wax to cast Oomoo, forming the second mold in which we would cast the real final material (Dry Stone or Hydro Stone). These latter materials are much harder to remove, so it's great if the mold itself is elastic, but then it's hard to use milling to generate anything in a rubbery material, so we have to do that indirectly through another mold (the initial wax one).
Solid geometry in Blender is easy but requires a few considerations for working correctly. First, the normals must be oriented correctly. When using a plane, extruded into a cube (for the molds), it can easily happen that the orientation is wrong, and the extrusion uses the orientation implied by the normal direction. This means that if you extrude in the back direction, the resulting cube has normals pointing inside of the solid. When this is the case, any CSG operation assumes that the real solid is the complement (everything but this), except that it won't work for union (since the union with everything cannot be represented into something useful). In this case, it suffices to invert the normals. For this specific case, the base plane is below and extruded upward so no issue arises.
One trick I learned about CSG and topology is that you want to simplify the solid before instead of after applying the operations.
In practice, I never applied the operations and had operations using objects that were generated through CSG.
For some reason, the topology got completely messed up when simplifying (and triangulating) after CSG,
whereas simplifying my mesh before resulted in a nice watertight valid mesh.
This took me a while to realize.
However, it was not necessary as I ended up using the Roland SRM-20, with Mods FabModules, which does not seemingly require even watertight meshes since it converts it into a heightmap.
I initially loaded the full resolution model, which had several too many thousands vertices, and had an uneven base surface because of my topology fixes that were somewhat destructive (Meshmixer? FreeCAD? ... or maybe it's Blender's fault). I had to fix the topology because I was trying to use on the Othermill machine which needed step files I eventually loaded the simplified mold (around 3000 vertices) on the Roland SRM-20 as it got free (topology does not matter now).
The cut did not have any major trouble. However I feel that several of the endmill motions were not doing anything (maybe bad settings? maybe too cautious path planning?). Multiple paths seemed to redo previous paths (without cutting much if anything). The fine cut was also quite slow and used half of its time going up and down at the borders of the mold, which could be sped up substancially since these don't matter unless there's something to be cut within the mold. We had two options for the endmill that depended on the cut depth. Since I had a shallow cut, I used the smaller endmill without fear of collision.
As a next step, we want to cast the oomoo mold that will be used to cast the actual whale. The mixing was somewhat easy although we were told to go as slowly as possible and even slowlier than we think so as to reduce the amount of bubbles. I always have the impression that I'm mostly introducing bubbles into the mix, probably because I like making tiramisu, where you have to mix egg white into the mascarpone mixture, and this requires introducing bubbles to make the mix fluffier. To my good surprise, I didn't end up having any bubble problems with my mixes. And then, as soon as it's out (it says 75mins before unmolding, and then at least 4 hours of post-curing, but you really need 2h before you can unmold, otherwise it's still sticky and not fully solid), I did it a few times again so as to enable casting many whales at once.
I also realized I could use the remaining of Oomoo for creating automated molds around objects put in a cup. Since this was not planned, I picked the closest thing I could find, and that ended up being a broken USBTiny ISP board with an ATTiny on it only. I didn't expect much out of it, and the traces came up very well! There were lots of bubbles due to the bad control of that environment. I should have poored into another cup, instead of pushing the board into the cup.
Finally, we have our flexible molds and we cast plaster drystone and hydrostone.
This step is simpler because the mixing is simpler: put water part, mix powder, done.
For safety reason, we all wore masks, even though the actual mixing was very short.
The interaction between the oomoo mold and the mixture is quite interesting. In practice, you can poor the mix into the mold, or you can also paint it / paste it, fill it. Eventually, the mix seems to move away and just does not stick to the mold. This is really cool!
I also printed a chocobo, and poored oomoo to create a mold for it, but it's still waiting to be born.
Finally, I made quite a few of these whales. I preferred drystone so I only made three in hydrostone (one batch at center of whale spiral below). The details are high quality. The bubble issue arose for me because I used the remainings of other people. They had probably done a good job mixing, but I was collecting their remainings, which had already started solidifying. It worked, but removing bubbles late in the process is hard if not impossible. Note: you can hide the bubble holes by using remainings of mixture to patch them!
The chocobo mold was quite complicated and I tried to cut it into two equal parts using its symmetry, but with few markers, the result was expectedly imprecise. Casting drystone was the most complicated part because of the leakage at the bottom (after creating a hole at both top and bottom sections to let the air evacuate). Unfortunately, the result was not as good as I expected, and this show how complicated geometry requires complicated molds.
I also tried to color my whales using uv-curable photopolymers from my lab's 3d printers. I simply dipped three whales into a container filled with a solution containing magenta pigments with our base rigid photopolymer. Then, after a few hours of rest, I extracted these and cured them using our Fusion UV machine. The solution did not stick well to the drystone finish, so it may require more complex processing or a layer-by-layer deposition and curing for real applications. Because of the dipping and direct exposure, the coating is highly non-uniform.
