The Casting Requirements for a Humanoid Robot

When we talk about future castings for humanoid robots, it is tempting to create one specification for the entire machine: lightweight, strong, stiff, and cheap. In reality, this would miss the most interesting aspect of the use case. The requirements are highly localized, and a casting in the arm requires almost the opposite properties of a part close to the base of the robot.

This means there will probably not be one ideal humanoid casting alloy or one casting process that wins the entire application. The opportunity lies in understanding each subsystem and producing the combination of geometry, material, and process that solves its individual problem.

Lightweighting matters most where the Robot moves

The further the arm extends from the robot’s center of mass, the more its weight affects the upstream actuators. A heavy forearm adds more than its own weight. The elbow and shoulder also have to accelerate, stop, and carry that mass while the robot is handling its payload.

Consequently, thin walls become extremely attractive in the arms. Every reduction in structural mass increases payload or reduces the actuator effort required to achieve the same effect. This creates a natural application for die-casting, which is commonly used for complex, thin-walled structures.

Near the bottom of the robot, however, the calculation changes. Additional weight in a wheelbase lowers the center of gravity and increases stability. There is much less reason to fight for every gram in that part of the structure, which opens the possibility of thicker sections and completely different manufacturing processes. However, Rheocasting fits perfectly for both regions, as it can cast thin and thick sections in one process.

Stiffness is more important than Maximum Strength

Another interesting difference from automotive is the relationship between strength and stiffness. According to Wilberth, once a minimum strength level of around 200 MPa has been reached, stiffness becomes the more important design target for the robotic structures.

This changes the discussion. Much of the solution therefore has to come from geometry. Ribs, intelligent load paths, generative design, and topology optimization increase stiffness while minimizing the weight.

For foundries, this is particularly interesting because casting gives the design engineers great geometric freedom. The challenge is to cast the optimized structure economically at the required volume. The best robot casting is therefore unlikely to come from taking today’s machined part and simply reproducing it in aluminium. It has to be designed around the properties of the casting process from the beginning.

A Robot has to Survive Falling Over

For bipedal robots, stiffness is only part of the problem. They can fall, and an 80- to 100-kilogram robot hitting the floor creates a completely different load case than normal operation.

This makes elongation and toughness important because a brittle casting can survive all normal operating loads and still fail catastrophically during an impact. Elongation values of 0.5% aren’t enough for these applications. The structure needs enough toughness to maintain its geometry when something goes wrong.

That also challenges the traditional approach of simply offering a standard AlSi9Cu3-type HPDC alloy for every new application. Humanoid robotics require a different balance of strength, stiffness, elongation, and manufacturability, and a foundry that provides this combination has a much more compelling offer than a supplier competing only on piece price.

Thermal Management is Part of the Structure

Humanoid robots contain many actuators, batteries, and AI chips, and all of them generate heat. Highly loaded joints are particularly demanding, and some robot concepts use heat pipes around the knee area to move heat away from heavily loaded actuators.

A casting must consequently become more than a structural housing. If the alloy and geometry also need help distributing and dissipating heat, thermal conductivity becomes another useful material property.

This creates interesting alloy-development opportunities. Instead of optimizing only strength and elongation, a future casting must combine sufficient mechanical properties with high thermal conductivity, and good castability.

Magnesium is Interesting, but not Automatically Better

Magnesium naturally comes up when lightweighting is important. Its lower density is obviously attractive, particularly in moving structures. However, magnesium’s lower stiffness means the casting needs to be thicker to achieve the same structural performance, reducing some of the theoretical weight advantage.

Magnesium creates challenges around machining and coatings, while its other advantage is vibration dampening. Magnesium therefore should not be viewed as an automatic replacement for aluminium but as another option where its particular property combination matches the subsystem.

It’s all about Appearance

Another requirement that foundries accustomed to hidden automotive structures easily underestimate is that many structural components of a humanoid robot are visible.

Blank surfaces, bead blasting, anodizing, and other finishing operations therefore are significantly more important than in body-in-white castings. At the same time, many robots will operate indoors in factories, warehouses, and homes, meaning that some of the extreme corrosion requirements known from automotive applications are less relevant. Joining also relies heavily on bolts and actuator connections rather than extensive adhesive bonding. 

The result is a surprisingly open design space. Humanoid robots require lightweighting, stiffness, toughness, thermal management, dimensional accuracy, and surface quality, but not every component needs all of these properties simultaneously. So, the foundry is not being asked to produce another standardized casting. It has the opportunity to develop the ideal casting for every function of the robot.