Additive manufacturing (AM), or 3D printing, is a fast-growing field that offers many advantages over traditional techniques. It can create more complex parts than either machining or casting, can fuse different materials together, and is sometimes less expensive in low-volume or prototype applications. The consensus is that it is a niche technology for niche applications, not suitable for volume production, at least not yet.
On the other hand, some companies are finding it lucrative to build businesses, practice areas or departments around AM. A key to success is ensuring quality systems are properly adapted to the new technology.
One such company is Star Rapid Ltd. (Guangdong Province, China). “We make prototypes in plastics like silicon and urethane and produce a wide range of 3D printed parts in metal, ranging from stainless steel to maraging tool steel,” said Gordon Styles, president and CEO. The company has thrived with AM despite known disadvantages, such as slower production rates. Other disadvantages lie in the very newness of the technique, such as knowing the density and shear strength of a final part. Finally, there is always the need to check dimensional measurements against GD&T.
Making a part means knowing what you have. “If you can’t measure it, you can’t make it,” stated Styles.
A crucial measurement is the material itself. Metallurgical testing is essential. “The system we use for printing AM parts is more akin to welding than sintering or a binding process,” he said, referring to the company’s powder bed system, which uses lasers to melt and fuse metal powder in successive layers. High-powered laser melting changes the metallurgy of the part compared to the powder that it started with. “There is a myth that a machine arrives, you turn it on and run it with factory settings. That is not the case. The geometry of the part and the parameters used can change the metallurgy quite a bit. I have a full-time metallurgist on staff who trained for three years before joining me,” he stated. An optical emissions spectrometer features prominently in Star Rapid’s measurements, among other pieces of equipment.
Dimensional measurement is, of course, always important. Like a casting operation, many printed metal parts require post machining to achieve critical tolerances, especially mating surfaces. Most of the dimensional checks that the company makes today can be satisfied with existing metrology equipment, such as portable arms equipped with touch probes or laser scanners—even hand tools such as calipers and micrometers. “The exception is the emerging CT scanning technology. It is perfect for checking occlusions and the dimensional properties of internal cooling channels,” he said. However, Styles is also wary of the high cost of such systems, at least at present.
One thing that he is adamant about is the need for speed and automation, even with existing metrology techniques. He is candid about the fact that his operation is based in China, which requires he be ever vigilant about his supply chain. “We have 275 people in my company and 23 are [dedicated to] quality control, a ludicrous number based on my experience in running a shop in the UK,” he said. Quality control is a bottleneck, even for articles produced in-house with trustworthy processes and machines. “It can sometimes take a half a day for first-article inspection, which means a machine is down waiting for that ‘okay’ to begin production,” he explained. “That is lost money.” Styles needs speed. He needs automation. For example, he is looking forward to delivery of a white light scanning system capable of collecting hundreds of thousands of data points per second.
Does AM substantially change how metrology should be adapted to the products made with it? Not necessarily, according to Carl Dekker, president of Met-L-Flo Inc. (Sugar Grove, IL). It is less about the equipment and more about the data used to set expected tolerances, in his opinion. “It is a question of the CAD model used for the design and the STL file used to create the part,” he explained. While STL is useful as a neutral file format that allows these systems to create a part, there are a plethora of problems with it. It is a surface geometry approximation of a solid, it is not intelligent, and its resolution may at times not be adequate.
Under many conditions “you can really have a lot of discrepancy in the manufacturing process,” he said. “Managing deviations is critical from the standpoint of getting predictable and reliable parts.” The reason this is so important in the context of metrology is not every customer supplies the CAD model—sometimes they only provide the STL file. That means there is a built-in deviation from which it is difficult, sometimes, to determine tolerances and GD&T.
As for measuring critical dimensions, Dekker’s company uses standard inspection techniques, including laser scanning, touch probes, CMMs, and optical comparators. Because of the resolution of most AM processes, nine times out of 10 the metrology system will have an accuracy greater than the machine that produced the part, unless the part is further machined to reach tighter tolerances.
“A lot of your standard inspection techniques will still apply,” Dekker said. “But while you are not trying to reinvent inspection, you do need to find some additional ones to adapt and apply.” Why new measurement techniques? The problem, in part, is the relative newness of AM. He agreed that the process will change the properties from that of the input material to the finished product. “There are deviations depending on what energy sources are being used and how they are being applied that could impact the viability of the part. Some of this is still being learned,” he said.
This may lead to a requirement for in-process monitoring of the melt pool, as well as possibly monitoring the gas formed when lasers melt material. “We might need to track what is happening to the chemistry of the part. A mass spectrometer might be needed if it could be made affordable,” he said. It is also known that the crystal structures formed in AM differ from the material used in subtractive machining because [AM builds parts] layer by layer. The cooling process is different; therefore, the crystal structures are different. “That is just one example of how it differs from our past experience. We may need to measure those somehow,” he said.
Renishaw Inc. (West Dundee, IL) is pursuing a strategy of providing comprehensive AM solutions with its newest platforms, including metrology in the build chamber. It bought an AM company with the intent of first improving machine quality. “From research, we determined that many AM machines are simply assembled from third-party components,” explained Stephen Anderson, additive manufacturing business development manager for Renishaw. “To make a precision production AM machine requires more rigor.”
The company could bring to bear to the AM machines the same components it uses in its metrology equipment to precisely measure and position probes. Components such as encoders provide greater accuracy to the newest lines of AM powder bed devices, according to the company. On Renishaw’s latest machines, it also uses its own optical drivetrains, sometimes known as galvo blocks, according to Anderson. “This meant we could reuse all of our expertise in metrology technology to create a very accurate, repeatable and robust 3D printer,” he said.
Renishaw offers four powder bed systems, with lasers ranging from 250 to 500 W in power. “Our very latest machine can apply four lasers simultaneously with all lasers capable of addressing the whole build plate for faster production,” said Anderson. He also added, with satisfaction, that Renishaw uses its own AM machines to print the galvo block that is then used in its latest AM systems. “Our printers are accurate enough for us to print one of the most critical parts of our own system repeatably with confidence,” he said.
Another advantage of using its own optical drivetrains is it can put metrology inside the build chamber. “We quickly realized growing parts organically in metal gives you a downstream problem—how do you mechanically qualify the melt density of the internal geometry that you cannot see,” he stated. “While that can be done with CT or X-rays, these methods are not scalable for volume AM production.”
For this, Renishaw has developed its MeltView and LaserView options for use with the InfiniAM Spectral software suite. The MeltView system is an in-line opto-mechanical module, monitoring melt pool emissions in wavelengths from 300 to 2000 nm at 100 kHz. The LaserView system measures the input of the laser in RenAM systems using an infrared photodiode. Sensors monitor both plasma and the reflectance of the material. “This gives us near real-time feedback to see if there are any issues in the melt process and to correct or stop it during the process without waiting for the build to complete,” Anderson said. This is especially valuable in a production environment when build times can stretch to many hours, even days. Here, finding a problem after a build is costly, whereas spotting the issue quickly allows the user to be more proactive. It also opens the future possibility of closed-loop adaptive control in AM.
Once the resulting part is out of the machine, Anderson likes to describe it as a near-net-shape digital casting. It may need more machining to meet GD&T requirements for mating surfaces. Metrology systems like Renishaw’s Equator on machine probes, as well as traditional CMMs equipped with Renishaw’s touch trigger probes, are ideal for this, he said.
Carl Zeiss Industrial Metrology Technology LLC (Maple Grove, MN) takes a comprehensive view of the quality management required for additive manufacturing. One of the challenges of providing a comprehensive set of solutions is not only the relative newness of AM, but also its diverse techniques. “There are seven different manufacturing processes for building up a product using different materials, heating sources and lighting sources,” said David Wick manager, product management for Zeiss Industrial Metrology.
He also described six basic steps in any of those AM processes, from developing new alloy powders (or any incoming materials) to the final process flow digitization. (See figure on next page.) The Industrial Metrology and Microscopy divisions of Zeiss, as well as other Zeiss divisions, provide quality measurement tools from inspection of incoming powder to the dimensional measurement of the finished part. Scanning electron microscopes (SEMs) and X-ray imaging devices are important in determining basic material properties. If a CAD file is not readily available, laser and CT scanning devices are useful to reverse engineer parts by creating STL files. The STL files are then used to make the part. To determine the inherent component quality, a manufacturer can again use an SEM or an optical device on sectioned surfaces. CT scanning is useful for determining density and presence of defects. Once that is done, one can then apply the familiar suite of metrology tools to determine component tolerances as well as CT scanning for interior features.
Another key factor that Wick is hearing from many of his customers reflects the general need for more speed and automation. Not only do additive manufacturers want automation, they also want automation to be easy to get. “They do not want to spend a year paying a software developer to write an interface between their automation tools, such as a robot, and our instruments,” he explained. “To say plug-and-play is too simple a term; they want it as straightforward as possible.” At the same time, the challenge for companies like Zeiss is that every automation opportunity is unique.
The final part of the process flow that Wick believes will need further adaptation is in-process flow digitization. “This is where we have to figure out how to connect all the quality and measurement tools that might affect your network,” he said. This could include Industry 4.0 or Industrial Internet of Things protocol or some other type of backbone. “The quality tools in that case would be talking to the same database, feeding and getting information from the tool right in front of it and putting measurement data into a database for customers to look at later. Not all quality tools are integrated like that today,” he stated.
A look at some of the more challenging AM applications, as envisioned by the US Department of Defense, illustrates the direction some of the technical solutions may take. According to Dana Ellis, senior project manager for the National Center for Manufacturing Sciences (Ann Arbor, MI), the DOD is heavily researching AM as a way of solving some of its unique problems. The most important is that the DOD has a widely scattered and worldwide presence. The logistics of shipping parts when and where needed are difficult. “There’s a lot of need to print parts in real-time, whether [they are for] a drone or other critical elements for the Armed Forces,” he explained. “The question is how do we do that in an environment to ensure and improve the integrity of that data. [We need to know that] the data is not tampered with or disturbed and the data provenance is secure.” He reported NCMS is looking at novel ideas, like adapting the block chain concept to ensure integrity and security of the resulting part.
However, final part integrity will also require knowledge that the material and dimensional properties of the part also meet specifications. Inspection criteria, such as GD&T and in-process material specifications, will also need to be in the data set. Metrology devices optimized for automated use will also be crucial in remote operation. No doubt these will become equally important in any commercial setting.