MECHANICAL ENGINEER
Aluminum Cast Robotic Links (Neocis)
About the Project
The goal of this project was to design six custom links for a 7-DOF surgical robotic arm. These links make up the patient-facing portion of the robot, meaning they needed to satisfy both functional and aesthetic requirements. Visually, they had to match the design language of medical devices, with smooth, clean surfaces and no visible screws or hardware. Structurally, they needed to be extremely rigid, since any deflection in the links directly affects the arm’s accuracy, which needed to remain within a sub-millimeter error budget.
When I joined the project, the overall kinematic chain and approximate joint sizes had already been defined, but the link designs themselves had not yet been developed. This project involved taking these initial definitions and creating production-ready links that met all stiffness, aesthetic, and manufacturability requirements while integrating seamlessly with the rest of the system.
My Role
I was the sole mechanical designer responsible for the links. My responsibilities included taking concept sketches from the industrial design team and translating them into manufacturable parts, selecting fabrication methods, and performing FEA to ensure stiffness requirements were met. I developed solutions to hide all fasteners using screw caps that still allowed secure removal for maintenance, led design reviews to gain buy-in from cross-functional teams, and worked closely with our suppliers to address manufacturing challenges and successfully transfer the parts into production.
Functional Requirements
-
Aesthetics: All visible surfaces needed to look clean and match the medical device aesthetic, with no exposed screws.
-
Structural Rigidity: Each link needed to maintain high stiffness because any deflection directly impacted overall arm accuracy, which was required to remain below 1 mm. The stiffness specification for each link was initially estimated using the preceding cross-roller bearing as a reference, and then refined based on an overall system error budget.
-
Kinematic Chain Compatibility: The link geometries had to align with the kinematic sketch developed by the systems team, ensuring proper reachability and workspace coverage.
-
Electronic Housing: The links needed to fully contain all joints and internal cabling, routing wires through the arm to remain hidden from the surgeon.
-
Compact Distal Form: Particularly at the patient-facing end, the links had to be as small as possible to minimize visual obstruction of the surgical site.
Manufacturing Method
A major early decision was the choice of manufacturing method, which had to satisfy these requirements while remaining cost-effective. Three methods were evaluated: carbon fiber, aluminum sand casting, and split machined components. Machined links were ruled out because they could not be made as a single solid piece, which would have negatively impacted aesthetics and violated the requirement for no visible fasteners. Carbon fiber offered excellent aesthetics, stiffness, and low weight but was prohibitively expensive, with limited opportunities for cost reduction at scale. Ultimately, V-Process aluminum casting was chosen, as it allowed for high-stiffness, visually appealing links at a lower cost while remaining suitable for mass production.
Design Process
The design process for each link began with CAD models informed by industrial design mockups. These provided the desired exterior shapes but did not address mounting points or structural performance. I incorporated features for mounting the actuators and internal electronics, then performed SolidWorks FEA to simulate the expected moments and loads derived from the actuator specifications (as detailed in the custom joint project). While the links almost always passed yield stress checks, the initial geometries rarely met the stiffness requirements.
To address this, I iteratively modified the geometry to achieve the required stiffness. Minor stiffness adjustments were made by increasing wall thickness, which had minimal impact on aesthetics, while larger increases required enlarging the cross-sectional area of critical features, leveraging the fact that stiffness scales with the diameter to the fourth power. This process continued until all links met stiffness requirements while maintaining the original aesthetic intent from the industrial design team.
Fastener aesthetics were addressed with a custom system of magnetic screw caps, which concealed all mounting holes while allowing removal only with an external magnet tool, ensuring both visual cleanliness and safety.
Once the design was finalized, manufacturability was addressed. V-Process aluminum casting allowed for zero-draft-angle patterns, so the links could be cast as designed without modification. Critical surfaces, such as mounting interfaces, bolt holes, thin walls, and locations for screw caps, were designated for post-casting machining to ensure precision. I created both “as-cast” and “as-machined” CAD models, incorporating extra stock where needed for machining. Part drawings with GD&T specifications were generated to define critical features and ensure proper fit and alignment.
Throughout this process, I worked closely with our casting supplier, holding weekly meetings to address casting and manufacturing issues and making design modifications as needed. When challenges arose in controlling the quality of the outer surface to meet aesthetic requirements, I collaborated with the supplier to implement a 3D scanning method to distinguish acceptable parts from those that did not meet visual standards. This supplier collaboration ensured that the final parts were both manufacturable and visually consistent with the intended medical-grade appearance.
Testing & Validation
Independent testing of the links was limited due to their very high stiffness and the lack of in-house equipment to measure individual link deflection accurately. Instead, validation was performed by assembling the full robotic arm and confirming that the system could be calibrated to achieve sub-millimeter accuracy. Successful calibration demonstrated that the links were sufficiently stiff to meet the overall error budget, confirming that our initial stiffness estimations and FEA analyses were reasonable approximations for system-level performance.
Outcome
I successfully designed six production-ready links that met all of the project’s functional, aesthetic, and manufacturability requirements. The final designs were released to production and integrated into the robotic system, contributing to the arm achieving sub-millimeter accuracy during calibration. These links balanced structural rigidity, visual design, and manufacturability, helping establish a reliable foundation for the robot’s patient-facing arm.
Lessons Learned
-
Consider the pattern and core when designing cast parts:
Some early designs featured long, thin hollow channels. While these could be manufactured using 3D-printed sand cores, the transitions from large to small cross-sections made the cores prone to breaking during transport. I learned to visualize the core geometry and ensure it was robust enough to withstand handling without compromising manufacturability.
-
Account for variation and surface finish in as-cast components:
The interior surfaces and dimensions of as-cast parts are not tightly controlled, and surface finish is rougher than machined prototypes. Initial cable channels that worked on machined test links caused cables to catch or experience high friction when cast. Expanding the channels and machining critical surfaces resolved these issues in later designs.
-
Include sufficient stock for machining:
Casting variation can result in surfaces being undersized relative to their intended machined dimensions. I learned to add extra material in regions designated for machining to ensure there was always sufficient stock to achieve precise final dimensions and maintain alignment for mounting surfaces, bolt holes, and covers.
-
Industrial designs must be checked for engineering feasibility:
Industrial designers provide excellent aesthetic guidance, but every concept must be evaluated for mechanical performance, stiffness, and manufacturability. Many initial mockups required significant refinement to meet structural and functional requirements without compromising visual intent.
