Project 6 : Molding and Casting

Motivation

Over the Summer I had the great honor of visiting professional bellringers in Russia. One of the highlights of the trip was visiting a bell foundry, where we saw the whole process behind the casting of the sacred Russian Orthodox bells, and even managed to see them pour the molten metal into the the molds. So for this week's molding and casting project I decided to make a bell with the eutectic metal at our disposition.

The process by which the Russian Bell foundry makes the molds is very similar to the one we are about to apply in that they:

First make a positive of the bell; they have a basic metal shape of the bell to which they add decorations and letters that are handmade and then duplicated with a plastic mold and cast with wax.

Second cast the mold; they use higher temperatures and make a ceramic mold. The downside to this is that as the ceamic cools it gets a lot of cracks which will appear as ridges on the final bell that will have to be taken off by hand. The inside of the bell also requires a mold that can be less precise.

Third they cast the bell with molten bronze previously prepared in a furnace. They are then left to cool for at least a day, depending on the size of the bell.

The Project Design

I brought a small booklet from Russia with a description of the shape and structures of Russian orthodox bells. I photographed a drawing of a bell profile and then imported it into Inkscape. There I used the "trace bitmap" tool, which converts the .jpg photo into a .svg vector file (the threshold was set very close to the minimum so that the blurry spots were still included). The result still contatined too many vectors, so I drew a simple vector curve on top and edited it witht eh bezier curve editor until it looked good. I started again with a different picture to get the profile of the inside of the bell.

The result still contatined too many vectors, so I drew a simple vector curve on top and edited it witht eh bezier curve editor until it looked good. I started again with a different picture to get the profile of the inside of the bell.

I then exported the vector file to rhino in order to model the bell in 3D. At first this did not work, and I tried many different file formats (at the best I could get my file imported into blender). Finally I found that I could export as .svg, then import into a program called Scribus, where I would export the curve as a .pdf file which then can be successfully imported into rhino.

In rhino I first made a model of the bell itself. First I revolved the profile to make the bell, then I added a hook, decorations on the front, and finally made the models of the pieces to be cut out of the wax to make the molds.

Apart from some of the basic commands that I already knew (Polyline, Mirror, Rotate[3D], Show, Hide, Scale[nD], Explode, Group, Join) the following rhino commands came in handy, so I'll stash them here:

SetView: lets us set a perspective
CPlane: create a construction plane
PlanarSrf: make a flat closed curve into a surface
Extrude / ExtrudeSrf: axtrude objects
Revolve: make a solid from a closed curve revolved at a certain angle
Project: Project a curve onto a surface
BooleanUnion: a composite command that runs the following four commands:
1. Intersect: find where two solids intersect
2. Split: split the solids at the point of intersection
3. Delete: the pieces of one solid that are contained in the other
4. Join: the two solids left
BooleanDifference: Same but instead subtracts one object to another
Trim: select a trimming object, the delete the parts of a solid that stick out
SplitFace: divides a planar face with a line or an existing curve
Offset: copies a curve at a distance from the original curve (nice for making walls)
MoveEdge: Allows edge manipulation of a polysurface
ExtractSrf: seperate a surface from a solid
Cap: fill holes in an object
SelBadSrf/Object: find the buggy objects
ExtractBadSrf: Seperate the buggy parts from an object
Showedges + check "show naked edges": verify that a mesh is closed

Making the model was fraught with difficulties. First adding the Harvard logo to the bell model was a nightmare: unexplanable bugs (probably due to the recent release of rhino on mac) obliged me to try many different ways of doing the same thing. I first obtained it by the same method as the bell profile. Here is what inkscape gives with the different "trace bitmap options"

I then projected the logo curve onto the bell:

I then used the curves on the bell to split the curve. I then extruded the surfaces where needed and then joined with the bell (SplitFace + ExtrudeSrf + Join). Below the different stages from left to right: original curves - simplification - projection - surface making - extrusion:

Originally I used a different method: after the projection I extruded the curve in both directions and boolean unioned with the bell. This generated some strage bugs: in particular some pieces would not boolean union to the bell. I split this buggy parts into smaller and smaller pieces until they were too small to even be noticed. The following shows the beginning of this tedious splitting process (the yellow part is the one that still refuses to union):

I finally obtained a first version of the bell model (I engraved my name on one of the sides). Here is the surface:

And here is the visualization of the mesh obtained in MeshLab:

However I soon realized that this model contained too many details, small ridges and intricacies. These would not come out well in the molding process. Hence I started anew and simplified all the elements of the model. I traced a simpler profile of the bell shape, then simplified the Harvard logo. The new version in the construction process is shown below:

I then had to make the mold parts from this model of the bell. I planned out a three-part mold that would use lips to align with each other. I first had to cut the bell in two: of course BooleanDifference crashed so I went for trim + cap.