One of the major obstacles to the adoption of 3D printing in many markets has been the variability of its output — in terms of dimensional accuracy and material properties such as porosity, strength, temperature, and chemical resistance. But the additive manufacturing industry is launching a full-scale assault on the problem. It consists of a three-pronged effort using hardware, software, and management systems.
With 3D-printing technologies increasingly becoming a mainstay in modern manufacturing operations, original equipment manufacturers (OEMs), software houses, 3D-printing factories, and contract manufacturers are striving to fine-tune the efficiency and repeatability of these production methods. Variability in the 3D printing of products has been a major concern of management for decades. Production engineers and managers pay special attention to product consistency with respect to dimensional accuracy and material properties such as porosity, strength, temperature, and chemical resistance.
Current levels of consistency in 3D printing — also known as “additive manufacturing” — are sufficient for many products. They include molds, toys, dental devices, optical lenses, eyewear, printed circuit boards (PCBs), some antennae and sensors, and non-weight-bearing metal and plastic spare parts for locomotives, heavy industrial equipment, airplanes, and military equipment.
However, that is still a relatively small portion of the potential market where this manufacturing technology could be applied if the consistency of its output could be raised. Understanding this, the additive manufacturing industry is launching a full-scale assault on the problem. The assault is a three-pronged effort using hardware, software, and management systems to reduce the variability of the objects printed.
Hardware. It’s hard to improve the output of 3D printing without considering the hardware of the printers themselves (e.g., motors, print heads, lasers), as well as hardware devices such as temperature sensors, humidity sensors, and X-ray cameras to monitor quality and catch errors layer by layer during the printing process. Velo3D, a Californian printer manufacturer, is one example of a company whose machines can monitor metal parts during the printing process. Through the use of sensors, its printers can be augmented with a system that monitors things such as oxygen levels, humidity, and unused powder levels. This level of visibility and control allows them to achieve higher yields and greater repeatability for many types of products without the need for post-processing (refining the product after it has come out of the 3D printer).
Efforts to improve the printing process can also involve conventional (subtractive) manufacturing tools, which are being used to improve consistency layer by layer. The machines of 3DEO, another Californian printer manufacturer, have sensors to leverage real-time data on dimensional accuracy and process parameters to optimize prints. Based on this data, 3DEO machines utilize a cutting tool (a micro end-mill) that trims edges and internal features, such as lattices and holes, to achieve the required tolerances and desired geometries. The trimming is done layer by layer.
Software and data. Artificial intelligence and machine learning are also playing an important role in the drive to make 3D printed products more consistent. They are being used to optimize the configuration of materials, design features, printer settings, printing processes, and environmental conditions for making a product. These technologies can create production feedback loops that automatically eliminate defects as printing occurs and can significantly reduce inconsistencies of outputs across printers and over time.
For example, PrintSyst, a new Israeli software company, has developed artificial intelligence that synthesizes the results of thousands of print jobs. It looks for factors that can help achieve higher consistency, yield rates, cost savings, and any dimension of quality customers might prioritize. The software then suggests printing technologies, material choices, machine parameters, and even design modifications to achieve the goals selected for optimization.
Older software for orientating and stacking parts in a build chamber has advanced and now contains algorithms that do hundreds or thousands of calculations in a matter of seconds to zero in on the ideal way to make a particular geometry in a specified printer. Such solutions are offered by global 3D-printing-software giants such as Materialise (Belgium), Siemens (Germany), and Autodesk (United States). Pennsylvania-based software company ANSYS provides design-for-additive-manufacturing tools and process simulations for metal parts that allow customers to achieve “first time right” output.
New additive manufacturing platforms — e.g., 3DPrinterOS, which is made by California-based 3D Control Systems — are now capable of managing consistency for thousands of printers distributed over the globe. Such platforms can control multiple 3D-printers remotely and assign jobs based on the availability and capability of machines. In many situations, 3D Control Systems installs its own software on the printers themselves, which allows its platform to take control of the printers. This control is used to improve consistency by preventing or correcting common errors or mistakes, including the use of inadequate print files, unlevel build plates, and the inappropriate selection of nozzles for the desired output.
Management systems. This approach uses tried-and-true management techniques for improving consistency and reliability. For instance, transparent vendor-rating systems and pre-certifications encourage competition akin to a kind of Darwinian natural selection where achieving and maintaining consistency in quality is incentivized. The best firms win more business, and the others are driven to catch up fast or perish. Companies such as Xometry and Fictiv, on-demand providers of industrial parts, run tightly controlled supplier-qualification and certification programs to encourage selection of suppliers that deliver high quality and consistent parts. The best suppliers are given access to larger and higher priority jobs, which pressures other suppliers to improve their consistency.
One of the newer management methods includes “deliberate constraining,” which purposely limits the usage of printers to their optimal yield ranges. Not all technologies are equal, so knowing the intricacies of their performance within certain boundaries such as size restrictions, material choices, and batch sizes allows the producer to take full advantage of a given 3D printer’s strengths. For example, 3DEO limits its focus to small, complex metal parts that have a maximum volume of one cubic inch. This allows 3DEO to not only make the most of its printers by focusing on where they truly excel but also fill up their build chamber to achieve optimum economics.
Not all sources of variability have been fully addressed. These include how machines are brought onto the network as well as the form and structure of data and workflow in and out of nodes in the system. However, as firms develop protocols and platforms for managing their distributed manufacturing systems, these issues will eventually be solved. For example, additive platforms will eventually gather a standard set of machine-level data in systematic ways that will make it possible to reduce variability in production significantly.
In today’s world of uncertain demand, flexible digital manufacturing, pandemics, and other disruptions of global trade, there will be more need for distributed and localized manufacturing. Taking too long to shift output can cause firms to miss significant opportunities, especially if the new products are not consistent from the get-go. Consequently, it is reasonable to expect growth in the use of 3D printing, which will further increase the pressure to improve its consistency.