As 3D printing continues to develop and evolve, we’ve begun to see it branch off quite a bit from the typical image of a machine depositing layers onto a flat build platform. 3D printing has been adapted to print onto preexisting objects, onto fabric, and in multiple directions. Machines capable of doing these fancy tricks are still rare and expensive, though, and not quite accessible to the average maker, but that may not be the case forever. A group of UK researchers recently published a paper about their efforts to develop an affordable system that can 3D print onto uneven, previously unknown objects.
In “3D Printing onto Unknown Uneven Surfaces,” the research team of Nils Bausch, David P. Dawkins, and Regina Frei from the University of Portsmouth and Susanne Klein from Hewlett Packard Labs explain how they developed a prototype system using 3D scanning, multiple-axis 3D printing and conformal printing, a method of printing often used to print conductive inks or sensors onto flexible or curved surfaces. As its name suggests, conformal printing enables the deposition of printed material in a way that conforms to the surface of an object.
“It appears that there is currently no solution for conformal printing onto unknown uneven surfaces or 3D objects, at any scale,” the researchers explain. “The only commercially available solutions for printing onto known uneven surfaces – aerosol jet printers (Paulsen et al., 2012) – are made for printing electronic circuits and requires inks to be transformed into aerosols. Starting at $250k, the printer is not accessible for those with a smaller budget. Moreover, for most industrial applications, not only accuracy but also speed is critical. Hence the printing technology needs to be scalable for applications on bigger surfaces than those in electrical circuits while maintaining accuracy and repeatability.”
The tops of these cones illustrate a newly designed object to be printed on top of a 3D scanned object, represented by the black mesh.
To begin developing an FDM machine that could be used to print onto an unknown surface, the team laid out their methodology in six steps:
Place and fasten an unknown object into a 3D printer.
Scan the unknown object.
Extract point cloud data.
Superimpose the point cloud data with a to-be-printed new object.
Generate the toolpaths for the 3D printer.
Print the new object on top of the previously unknown object.
The modified Prusa i3, with modifications illustrated in color
The researchers settled on a Prusa i3 for its low cost and its open source design, which allowed them to modify it into what they describe as a 7 DOF machine, DOF standing for Degrees of Freedom.
“The DOF include linear movements along the 𝑋𝑋, 𝑌𝑌, and 𝑍𝑍 axes as well as angular movements around these in 𝛼𝛼, 𝛽𝛽, and 𝛾𝛾 direction and the extruder output,” the researchers state.
One of the many modifications involved the incorporation of a DAVID 3D scanner consisting of a blue line laser and a monochrome camera. An object is placed into a gripper device and secured with bolts to ensure that it will remain stable while the machine scans it. Once the user is satisfied with the scan, the DAVID4 software processes it into a point cloud, which is then exported as an STL into Rhino/Grasshopper. That STL file is then combined with an STL of a new object to be printed on top of the scanned one.
By superimposing the two files, the researchers were able to line up the print job so that the first layer of the newly designed object could be printed accurately on top of the scanned object, held in the grippers. It sounds so easy, doesn’t it? Unfortunately, it’s not quite as simple as I’ve made it sound – I recommend reading the entire research paper to get a better idea of just how much work went into the project, and how much work is yet to be done.
A simulated toolpath output shows the scanned object as a smooth surface with the layers of the new object printed on top.
“At this stage in the project, the scanned objects are limited to simple shapes to allow fast prototyping and precision testing,” the researchers point out. “Moreover, the printer and scanner currently has limitations to the angles that can be scanned, brought about by the calibration panel remaining in place throughout the scan, limiting the movement of the object. Another limitation is the print head size, which hinders approaches to concave corners with sharp angles.”
Ultimately, the team wants to create an affordable desktop printer capable of printing directly onto any object, no matter how odd the geometry. Now that they’ve succeeded in printing the first layer of a new object onto a pre-existing one, they’re working on the subsequent layers. Future applications of the technology could include repairs and customization, the printing of electronics or security features onto existing items, or modifying prosthetic devices to fit individual patients. You can read the full study here.