Embedded Files
During the semester and summer of 2025, I worked with a start-up developing a concrete 3D printer using robots that move and extrude wet concrete. I was a mechanical engineering intern, where I helped design, analyze, and manufacture crucial parts for the robot. I designed the full CAD assembly on SolidWorks and designed new parts such as the printhead mechanism and conveyor holder and manufactured them using 3D printers and welders with sheet metal and metal beams.
Reverse Engineered CAD Assembly
Reverse Engineered CAD Assembly
Joining an early-stage startup building construction 3D printers, the lack of mechanical engineers on the team meant that electrical engineers, civil engineers, and the business operations team handled the CAD design of the robots. Noticing a lack of cohesive assembly for these individual components, I wanted to create an overarching CAD assembly for the robotic 3D concrete printer to manage changes, complete motion assemblies, and design more accurate parts for the future. Measuring existing parts in the prototype, I created an assembly with 50+ unique parts, ensuring accuracy using calipers. When designing and implementing new parts, the overall CAD assembly was used to verify fit.
For instance, the z-axis gantry with the threaded rods needed to be adjusted to reduce bending and shear stress as the chassis moves up and down in the z-direction. The mounting surface for the bearing pillow was verified in CAD so that the threaded rod was perfectly concentric with it. As a result, friction and bending within the axis were reduced.
The tensioner idler position was set to create the ideal tension for the available belt. We initially encountered the problem of the driving wheel not effectively transferring rotation to the driven wheel using the threaded rods. I designed a tensioner idler mounting point and verified the belt length using the CAD assembly.
Extruder Printhead
Extruder Printhead
The helical screw within the printhead.
While testing the prototype of the robot, a problem we ran into was the slow print speed. The concrete would not extrude out of the printhead at the desired speed, causing problems forming and building the concrete layers properly. I identified the problems as being associated with the design of the printhead itself. There was not enough pressure being generated to rely solely on the gravitational force to push the concrete downwards and out. I tried to come up with a mechanical solution to push and extrude the concrete. Looking into several dispensing mechanism designs, I decided to use an Archimedes screw. A major reason for this is its ability to move concrete forward to extrude but also mix and distribute concrete in the back of the printhead. This prevented concrete from solidifying too early, which greatly reduced extrusion speed.
An initial sketch for the screw was sketched alongside my supervisor and a civil engineer. Several factors were considered, including the volume of concrete within the print head, the viscosity of the concrete that would be used, and the desired print speed. Using the sketch, I translated the sketch into CAD using SolidWorks. I utilized the sheet metal feature to provide an accurate flattened sheet metal DXF. In turn, this provided an accurate profile that could be waterjetted and bent to form an accurate screw shape. The CAD mockup is shown below.
A Profile of Flattened Sheet
All Flattened Sheets
Because I used the sheet metal feature, I was able to easily flatten the model and obtain an accurate profile to cut out, such as on the left. This ensured accuracy during the manufacturing process. I worked with a technician who water-jet cut the part and bent it by hand to match the pitch as designed in the CAD.
Results of Testing
After integrating the components with the rest of the robot, our team tested it to observe an increase in efficiency and print speed. Despite a near 30% increase in print speed, I observed an inconsistent output flow from the concrete. This was caused by concrete being unevenly moved from one side of the printhead to the other. As a result, concrete was more heavily concentrated on one side of the printhead over the other. Due to the viscous nature of the concrete itself, I decided that any traditional mixing methods for fluids would not be effective. Since concrete vibrators were already solutions used within the industry, I decided to implement this as the solution.
A new design simply involved having these concrete vibrators attached to the printhead to liquefy the concrete, forcing it to pour out through the output. The concrete vibrator is essentially a hose with a vibrating tip. I created mounts on the printhead to fit the hoses through. Although my internship term had ended before I could conduct vibrational analysis on the printhead, using the concrete vibrator handheld showed a more consistent concrete extrusion.
Conveyor Belt Holder
Conveyor Belt Holder
Alongside the robot that creates the concrete walls, there was also to be a mobile conveyor belt that followed the robot. I designed the holder that held up the conveyor belt, which transported the concrete. I first gathered available material, which included a 30x30 square steel beam and a quarter-inch sheet of metal. I first created a simple vertical beam and horizontal surface to understand how the design would fail. I ran a buckling simulation and a static stress simulation as shown below.
In order to prevent buckling, I created supports for the vertical beam at various connections. The factor of safety rose to much more than required. The design was changed iteratively to maintain the safety factor while decreasing the material use. Afterwards, I manufactured the conveyor belt holder by using a horizontal bandsaw to cut the square pipes and a sheet metal shear machine to cut pieces of the sheet metal to the desired shapes. I then welded some pipes with supervision from my manager.
Page updated
Google Sites
Report abuse