Abstract
Direct digital manufacturing (DDM) typically leverages additive manufacturing (AM) to create complex, "nonassembly" mechanisms directly from CAD models. While AM offers geometric freedom, subtractive manufacturing (SM) provides superior material strength and isotropy. However, SM process planning for internal geometries is traditionally difficult. This paper presents a voxel-based computer-aided manufacturing (CAM) system that enables Direct Digital Subtractive Manufacturing (DDSM). By treating material removal as a voxel-by-voxel process, the system automates toolpath generation and collision avoidance. The approach is validated through the successful fabrication of a functional ball-in-socket mechanism from a single aluminum block.
Introduction
The rise of 3D printing has popularized the concept of "nonassembly" mechanisms—complex devices that function immediately after fabrication without manual assembly. While Additive Manufacturing (AM) excels here, the resulting parts often suffer from material anisotropy and lower strength. Subtractive Manufacturing (SM), such as 5-axis CNC milling, produces high-performance isotropic parts but faces significant hurdles in process planning for intricate internal features. Traditional surface-based (B-rep) CAM software struggles with the computational cost of collision checking in tight, complex spaces. This research introduces a voxel-based approach to bridge this gap, aiming for "push-button" CNC machining of functional assemblies.
Voxel-Based CAM and Data Structures
The core of this DDSM system is the shift from surface models to voxel-based representations. Using a custom CAM system called SculptPrint, the part is represented as a grid of 3D pixels (voxels). To manage memory and computational efficiency, these voxels are organized in a Hybrid Dynamic Tree (HDT).
This structure allows for rapid spatial queries, which are essential for the "reverse additive" logic of subtractive manufacturing. Instead of calculating complex surface intersections, the system simply identifies which voxels to remove.
Automated Process Planning
To automate toolpath generation, the researchers employed morphological offsetting—an algorithm inspired by image processing. By "dilating" the part volume with a spherical structuring element representing the tool tip, the system generates a safe "contact volume."
For every point along the toolpath, the system generates binary accessibility maps to determine safe tool orientations. This allows the 5-axis machine to reach deep into crevices while automatically avoiding collisions between the tool, the part, and the machine itself. GPU acceleration makes these high-density voxel operations feasible in real-time.
Results and Validation
The system was tested by machining a functional ball-in-socket joint from 6061-T6 aluminum. Unlike AM, where supports counteract gravity, DDSM uses tiny 1mm support tabs to resist the mechanical thrust of the machine and prevent internal components from vibrating or falling out during the final cuts.
Key findings include:
- Support Efficiency: The machined part required only 0.06% support material by mass, compared to 5-20% for FDM 3D-printed versions.
- Predictive Accuracy: The voxel approach provides a "What-You-See-Is-What-You-Get" simulation, accurately predicting toolmarks and scallops.
- Functional Success: The automated orientation selection successfully navigated low-accessibility regions to produce a moving assembly from a solid block.
Conclusion
This research demonstrates that voxel-based modeling can bring the automation of additive manufacturing to the high-performance domain of subtractive manufacturing. While challenges remain—such as CNC controller bottlenecks when processing massive numbers of tiny voxel-based movements—the DDSM approach opens new doors for creating high-strength, complex functional assemblies without the need for manual assembly.
Summary updated