Reverse engineering is becoming a multifaceted and complex topic. Potential applications are growing, including creating models for parts built before CAD existed, matching human limbs for orthotics and prosthetics, updating CAD models to as-built after a production process, repair processes, moldmaking—and the list is getting longer.
What is driving this are new metrology sensors and more capable software, enabled by ever more powerful and cheaper computing. With the technology moving fast, there may be an application that’s right for you.
First, let’s look carefully at how we go from a physical part to a digital model. The data chain in reverse engineering starts with collecting data points using a metrology-type sensor, converting those points into a connected mesh, and then (often) processing that mesh into surfaces or a solid model. Faster, cheaper computing and better software has enabled improvements in each of these steps. Computers are computing faster, so data collectors collect more data, meshers mesh faster and CAD converters convert more accurately.
Not only have these improvements advanced the state of the art in high-end, high-accuracy applications, they have also opened new low-end uses operated by people less skilled in the nuances of reverse engineering but more attuned to their own profession. “You can think of all the possible uses today as a pyramid, where at the bottom of the pyramid are many, many applications but the technical challenges are relatively low,” said Scott Green, principal solutions leader at 3D Systems Inc., Rock Hill, S.C., maker of specialized software used in reverse engineering.
He noted that orthotics and prosthetics are examples of applications on the bottom end of the pyramid. There are many people who need a low-resolution scan and resulting model of a limb—and the tolerances required are not particular tight. “A couple of millimeters of error is not going to hurt anyone,” said Green. “The human body is squishy and can tolerate [that level of] inaccuracy.” Today, simple scanners attached to a tablet computer operated by most anyone can be used to create a model good enough for orthotics.
At the next level up are applications like consumer product design, where tolerances need to be tighter than a millimeter and the process may result in producing a million copies. This could include cell phones, shoes, bottles or other items with an artistic flair (hence the need to reverse engineer a physical model), and assembly is required. “The level of skill needed increases as well for reverse engineering, because it is a mix of scanning and original design work,” said Green. Accuracies are in the 50 μm range or so.
The next level of skill needed beyond that is for maintenance, repair and overhaul—a prime use of reverse engineering. Tolerances and resolution are tighter, down to single-digit microns, requiring more advanced equipment, with high-accuracy LiDAR or white-light scanners. Scanning an aircraft landing gear for repair needs single-digit micron accuracy; getting that accuracy requires a higher level of skill in operating the scanning equipment and using the software at the back end to create an accurate CAD model.
Finally, the highest level of equipment and skill is needed when reissuing legacy parts, parts made from paper blueprints “back in the day” that most likely don’t exist or don’t match what was made. “Often, people want to make those parts better, maybe lighter or stronger or both. So now I’m taking a part I’ve never designed before and redesigning completely, and now I may add topology, optimization or design for additive manufacturing on top of that,” said Green.
Bottom line? Vendors are providing more specialization, in scanners and software, for applications ranging from orthotics to building information management, and any number of specialized manufacturing areas. Name your application, then look for vendors or a service bureau to help you meet that specific purpose.
“Keep your mind open to what you can do with 3D scanning. Applications for it are almost limitless. There’s a scanner for every situation,” said Ronnie Hensley, manager of reverse engineering for Exact Metrology, Cincinnati.
In his 13 years in reverse engineering, he has seen vast improvements in scanner speed and accuracy. “Accuracy has improved tenfold,” he said. “Now, a scanning arm equipped with a laser can provide 0.002" (50 μm) accuracy, and the resolution is amazing. There used to be a stark difference between [structured] light and laser scanning, and [structured light] is still more accurate, with cleaner, more accurate data. But laser scanners are catching up quickly.”
Another application angle that engineers should at least think about is how artists are reverse engineering their sculptured pieces. “We have had artists come in and create one-off pieces that are fragile and require gentle handling. We scan it immediately, and then they have a digital archive of what they created. If anything happens to the original, it can be recreated via 3D printing,” he said.
The concept applies to engineering. He noted that customers will visit with mock-ups of a part that needs fitting into an assembly. “We scan it and give them a model, and then see if it fits in the assembly. If it does not, they shave off areas, we rescan it and repeat the process until it fits, then we give them a 3D model of the part,” he said. Common file formats for delivery include STEP, IGES, or native Parasolid models.
But can it help? Hensley recalled helping a customer in the hydro power industry by replacing fins in the turbines of dams used to generate power. Such fins get damaged over time and with use, and until now, that damage was repaired with technicians welding on and then grinding off metal until they could fit their original purpose. “During a proof of concept, we scanned one of their blades and I provided them with [a 3D digital file] which they plugged into a six-axis mill,” he said. “They milled out a piece that was an exact match. It fit like a glove and eliminated 10,000 man-hours over the life of the project,” which included many more turbine fins than the original. “You can use this data in so many ways. Because it’s not always just printing a part for somebody—it’s also helping them show them what their part is, how to repair it and use it in so many other ways.”
Another application of reverse engineering that proved useful is in the automotive industry, according to Johan Gout, chief operations officer of Capture 3D Inc., Santa Ana, Calif., the U.S. distributor of GOM structured light 3D scanners. By scanning manufactured parts and reverse engineering them (as opposed to their CAD nominals), users are documenting actual parts that fit into an assembly, using them as the baseline for new parts.
“They now document what the as-manufactured process is delivering. It’s improved their ability to have the fit and finish they want because they are now using the CAD definition of what their manufacturing processes are actually delivering. This has proven to be especially useful for companies that have internal combustion product lines and now want to make an EV version in that product line,” he said.
With the variable scan capabilities of today’s GOM ATOS scanners, millions of X-Y-Z coordinate points of an entire car body can be quickly captured to 30 μm of data resolution within a 1-m measuring volume. “But that may not be good enough for many of the important fine features, so by switching lenses to use a smaller measuring volume for higher resolution, important smaller features can be collected at seven to nine μm and merged into the larger data set,” said Gout. Think of it as a variable mesh model. What is the best sensor for reverse engineering? “I think a sensor that delivers a variety of resolution and accuracy to capture all part features is the most useful in reverse engineering,” he said. This process is also advantageous to EV startups to document their as-manufactured configurations.
Yet another application is to support Industry 4.0 initiatives, where as-built parts are scanned not to update a baseline but to become the baseline, the digital twin. “Companies can then track how, over time, those parts vary in production and use,” Gout explained. This is a more proactive use of capturing as-built data.
Industry 4.0 initiatives are also spurring improvements in integration. “That means the seamless integration of scanning software and the sensors, as well as the sensors having higher-quality cameras, better lens optics and more competitive pricing,” said Gout.
Lower-cost, more-powerful computing is increasing the performance of both the sensors and software that support advanced functionality. “The GOM Suite software that operates the ATOS scanner now automates its exposure setting to help with the data capture, along with an automated, intelligent STL processing algorithm,” he explained. This makes data capture more accurate. How? The software that creates the STL data recognizes the type of geometry it has captured and then adjusts its data collection plan to create the STL file with the highest fidelity and lightest weight. “That is a mode most easily supported in the reverse engineering process,” said Gout.
He was quick to point out that this advanced computing is accomplished with Windows-based, off-the-shelf, laptop and rack-mounted computers and not limited to special high-end models. “We scan a whole body-in-white. You can imagine how much data that is, but with newer software and [fast computers] we don’t try and store those millions and millions of points. We process them into an intelligent STL file to take up as little file space as possible.” He also noted that the specialty software companies devoted to reverse engineering seem to do better at handling large data sets and converting them to STL and CAD.
David Olson, director of sales and marketing for Verisurf Software Inc., Anaheim, Calif., one of those specialty software companies, agreed that the cost of scanners and specialized software has improved, opening new applications. “Reverse engineering has become more accessible to many, many people,” Olson said. “Now you’re starting to see scanning technology for niches. People are saying, ‘Boy, if we could only scan and reverse engineer this kind of crazy part’,” And they want to scan these crazy parts easily, with easily obtainable solutions.
His company is one of many suppliers responding with whole solutions, not just one piece of a system. For example, Verisurf software is now offered as a complete solution bundle with several 3D scanner brands, including ScanTech, peel 3d and SHINNING 3D. One-stop shopping has come to reverse engineering.
But it’s not just price and packaging that is improving. As the manufacturing world is beginning to embrace attached product manufacturing information (PMI) and geometric dimensioning and tolerancing (GD&T) to CAD models, a clean, high-quality 3D CAD solid model is important to attach PMI data to. “We, and others, have perfected processing of the scan data to a watertight STL, then to a NURBS surface,” Olson said. “One thing that makes us different is we [go a step further with] solid modeling, which is included with every license. Most reverse engineering tools just have surfacing or create a watertight mesh. We can take it all the way to a full-featured solid model.” Final file formats include STEP, IGES, Parasolid or other popular formats.
Advances in computing technology for basic consumer electronics have opened the field of reverse engineering for a broad range of new users—like almost anyone. The iPhone 12 Pro has a LiDAR sensor in it, in addition to its three camera lenses. According to the Apple Insider website, the iPhone 12 Max and iPad Pro also sport LiDAR. There are many apps available through the Apple app store; some use LiDAR, others use the cameras with photogrammetry software.
LiDAR is not something Apple explicitly touts; rather, it expects app developers to employ LiDAR without users having to know anything about it. For example, the Measure app that comes standard with the iPhone reportedly uses LiDAR now, as well as camera autofocus features.
For explicit scanning, the author’s own attempts at using LiDAR-based scans showed some potential on large objects and whole rooms. But only some. Scanning with a hand-held device led to many errors, gaps in coverage, and, well, some “wow” factors, but not much practical utility for engineering—yet.
The author also explored a photogrammetry-based app for reverse engineering called Qlone. This is a little closer to being useful for engineers looking at smaller parts. It uses a grid printed on paper to set the coordinates of the object in space. Again, it requires a steady hand to collect all the data. An inexpensive smartphone tripod improved the process immensely. It is also limited, like many of its more expensive structured light or laser scanners, in that shiny or transparent objects are not scanned well.
All processing for the scans is done on the phone, in near real-time, even for the highest density of points. For artistic objects that do not need to fit into an assembly, it seems adequate. The resulting scans can be used in AR or VR, or even used to feed a 3D printer.
The importance of this is not in what it can do now, in the year 2021. This kind of processing 10 years ago would have been beyond even a laptop computer, let alone a hand-held smartphone. What should we expect 10 years into the future? Expect quite a lot.
As technology improves, Olson sees more applications emerging. “There are many soft parts made of elastomers that cannot be touched when measured because they deflect,” he said. “Hot parts is another one. These are ideal for optical scanning, and I see future reverse engineering applications in these areas.” The other, broader application is following the trend in mass customization. “Mass customization is the future,” said Olson, where end customers want products unique for their personal fit or taste. Special seats for cars, shoes made to order and sporting equipment are tailor-made products best served by some form of reverse engineering.
Perhaps the biggest trend—artificial intelligence (AI)—is also enabled by faster, cheaper computing. It is easy to see AI being used for collection planning, model creation, even various forms of generative design and then coupled with reverse engineering. Look for AI to make reverse engineering even easier. “Software algorithms, ala AI and machine learning, will drive improvements in the technology,” said Green of 3D Systems.
Connect With Us
