//* {Milling Sand & Differentiated Cast Components} ::
*//{2010.11.04 ~ Research}

Background ::
Sand casting has been around for over 5000 years, but has nearly always relied upon a positive pattern, against which sand is compressed, to create the mold cavity. There have been many advances in both patternmaking and mold-making, however the fundamental process has remained relatively unchanged since the bronze age, relying upon some sort of pattern to make a mold. A single wood or aluminum pattern can produce thousands of molds with care and maintenance, and it is not economically viable to produce and engineer a pattern for the production of one or two components. A component to be sand cast must be designed for both the molding, casting, and chasing processes - drafted for the release of the sand, filleted to aid the distribution of the liquid metal throughout the cavity, thickness transitions thoughtfully coordinated and minimized, re-entrant corners designed to facilitate access to finishing tools, etc. Moreover, the design of the gating (plumbing) used to distribute the hot metal throughout the mold cavity is specific to the volumetric distribution and geometric configuration of each individual component, and must be engineered through software, heuristics, or trial and error.

Abstract ::
If the patternmaking could be eliminated from the process of sand-casting entirely, and castable molds are produced directly from CAM tools, then the opportunity to produce castings in limited production runs becomes an economically viable proposition in addition to the production of families of differentiated components. Volumetric distrubution changes throughout differentiated components can serve as drivers to determine the necessary changes in gating design based upon a proven initial model. Waste in the production of patterns could be eliminated, as would be the time necessary to fabricate them.

Through the construction of a CNC router setup to deal with the invasive and abrasive nature of loose sand, and the testing of different sand-bonding resin chemistry and different types of sand, I was able to deploy a process for the CNC milling of sand molds directly from digital parametric models for the production of differentiated cast metal components.

Quick Exercise ::
I milled a few sand molds in an afternoon, that would have taken days or weeks to produce using traditional patternmaking processes, and brought them to an iron pour to cast the next day.

Using this process abnegates the need for draft angles because there is nothing that must be delicately removed from rammed sand, as well as eliminates any intermediary steps between machining and casting, freeing up the design process to express the tooling itself directly inside the mold cavity. I took this to the max (why not) in milling these ingot molds, which turned out a sweet, ugly, & medieval ingot mold... which also makes sweet, ugly, and medieval-looking ingots.