A Feintool plant in Vaihingen, Germany, makes rotors and stators for the motors powering BMW electric vehicles. I filmed this video during a visit to the plant. At about 0:21 in the video, you can see a drone’s-eye glimpse of all the production this part producer has ready to ship to its automaker customer.

A rotor or stator is an unusual part. What looks like a solid is really an assembly. The basics of motor operation demand lamination layers with an air gap between each layer. So: Each rotor and stator seen here is a stack of identical stampings, each stamping about 0.3 mm thick, hundreds of stampings high.

And because even small imprecision leads to electromagnetic losses, there is a lot the manufacturer can do to deliver motor performance:

• See the tooling used to control stacking. Repeatability here directly translates to lower losses and a more capable motor.

• The air gap must be there but it can be small. Feintool does not mechanically join the laminations for these EV motors, but instead uses a precision gluing process the company developed itself. Closer bonding = smaller air gap = more steel per stack = more power output per motor volume.

• Even the precision of the assembly hardware does not represent all that can be done to limit variation. To even out any small thickness variation across each individual stamping, the robot rotates lamination substacks as it combines them.

Why this matters:

The advance of electrification means motors will be more and more central to our world.

And with motors, the manufacturing process directly determines how efficient, capable or small the motor can be.

EVs are a beneficiary, yet it’s not just EVs. The sophisitication of motor manufacturing will have a lot to say about how powerful, economical and compact the compressors, robots, heavy equipment, hand tools, and even recreational vehicles of the future can be.

To state the matter another way, the power output, efficiency, or package size of machinery of the future will be shaped by process precision in making motors.

Share

Subscribe now

This, by itself, makes the challenge of validating an additive part different from validating a part made another way. Every other part-making process had an opportunity early on to validate itself by doing. A robust measurement standard applied to castings, when this first began to happen, had to contend with the fact that cast parts were already in use and had already succeeded enough to earn trust. The measurement standard thus had to accept or concede to what had already been proven. By contrast, additive is the first such process to confront a modern measurement regime from its beginning, without a legacy of use cases behind it. In effect, this leaves additive held to a higher standard.

Ankit Saharan of EOS and I talk about this in the video below, about the challenge or realities of measurement and qualification for laser powder bed fusion (LPBF) additive manufacturing. This video is part of a series in which he and I explore myths and misconceptions around LPBF (also called direct metal laser solidification or DMLS). See other installments from this series in the form of our conversations on the repeatability and the speed of laser powder bed fusion.

From the video above, here are three ways part measurement is different if the part happens to have been made via LPBF:

1. Additive Makes Material and Geometry at the Same Time, and Both Have to Be Measured

To anyone accustomed with additive, particularly LPBF, the point above is obvious.

Still, how different this is. In machining, the metal properties are locked into the bar or billet before any cut is made. Machining at best changes surface properties, otherwise leaves properties untouched. Casting, by contrast, involves a phase change of the metal, but the geometry is defined by the mold. Nothing about the way the material solidifies will change the geometry (unless the mold is not filled).

Measurement in additive manufacturing—and by extension, process control in additive—therefore (among other things) seeks to account for the interplay between external geometry and internal properties, and how thermodynamics and other factors of the build may affect either or both. This can lead to measurement considerations for which other part-making processes do not offer precedent.

2. The Qualification Standard Derives from the Reason for Using Additive

In an intricate metal part made additively, there is potentially too much to measure. Yet not every measurement is meaningful to the application.

Laser powder bed fusion has a range of application niches that is diverse compared to other metalworking processes it competes or coexists with. LPBF might offer substitutes for parts made through other processes when this delivers the lead-time win. Or, it might fill an application by creating a part that could be made no other way.

If parts made by more established processes seem easier to qualify, this is partly because the application windows for those processes are better known and defined. Qualifying the additive part should focus on the use of that part, Saharan says, likely deriving from the reason why additive was chosen in the first place. What performance outcomes are necessary or valuable? The answer to this question should suggest the qualification requirement. There is not, and may never be, a universal way that all production LPBF parts are measured.

3. In Additive, We Are Measuring Too Much

…and we can’t help it. Since additive is the process invented after the maturation of modern industrial measurement systems, it gets subject to all the measurement we can bring to it.

And in most cases, there is not a history sufficient to pare the measurement down. Again, other manufacturing processes have the legacy of use cases before modern measurement systems began.

One specific way this plays out is in material property test coupons made to validate laser powder bed fusion builds. It is easy to plausibly say that industry is overproducing test coupons. However, by how much it is overproducing them is impossible to say. With much more confidence making LPBF production parts, it will become clearer when the test coupons are wasted metal.

All of this is ultimately to additive’s benefit. “It’s good that additive is held to such high standards because it helps ensure that we don’t leave any box unchecked,” Saharan says. The overmeasurement (if that is what it is) presents LPBF and additive in general with high bars or a multiplicity of bars to clear. This is a challenge today, but it will make additive that much stronger as it clears them.

Ankit and I discuss all of this in the video:

Subscribe now

Share

I filmed this video in Tokod, Hungary. The location is a Feintool production plant making rotors and stators for electric motors. Feintool is a global high-volume metal part producer with plants in various parts of Europe—Switzerland, Germany, Czechia, Hungary—as well as around the world.

Relative to the locations in Europe, Hungary is a lower-labor-cost region. If the purpose of automation is to replace labor, then we would expect the plants in higher-labor-cost areas to rely on automation while the low-labor-cost plants rely on labor instead.

That is not the case for Feintool. The company has robots and other automation systems in use here in Hungary, as well as throughout its operations, in all the regions in which it produces.

In this video, the robot cell performs brushing of rotors for AC induction motors to remove excess material after metal casting. Humans could do the flash removal, but there is a distinctly tight tolerance on the processing of this particular part.

And I see that crucial detail getting at the real role of manufacturing automation and what purpose the robot cell serves. Whether labor cost is low or high, the primary role of automation is to perform operations a human cannot do well.

Robots and other manufacturing automation are good at consistency in terms of accuracy, repeatability, and steadiness of production rate. Humans are poor at this.

Humans are good at flexibility, adaptability and responding correctly when signals are ambiguous or subject to change. Robots are not suited to this.

Thus, adding automation is not a means to do the same work a different way, with machines replacing people. Anyone analyzing automation this way is likely missing its point.

Adding automation is instead the way to expand the possibilities for what the plant can do. The right mix of team members plus automated systems opens the way for production to be responsive, productive, efficient and tightly repeatable all at once.

Share

Subscribe now

I’ve spent time with contract manufacturer and product development specialist Lifestyle Additive in Tucson, Arizona, as the team here experiments with black PA-12 nylon material developed by Advanced Laser Materials (ALM) for SLS machines from EOS.

The new material represents a new possibility. Conventional SLS nylon is white. Another color comes either from dying the part, or from an additive mixed in with the powder. The new ALM material features the black color as an inherent property of the powder.

Lifestyle is learning the processing parameters for building with this material. And more: The company is discovering that black nylon has special functional properties that go beyond the aesthetics.

Precision is part of the black magic: The black material is proving more capable for achieving fine-detail features than the conventional white nylon. It is not that one material is capable of tolerance the other is not. It is more that the new material allows the team to arrive at a precise process faster, with fewer of the iterative adjustments to design, orientation or parameters needed to get to the process that can deliver a needed precise feature.

In the video above, I discuss the material, Lifestyle’s experimentation, and how the team is seeing the precision difference.

Potentially, there is yet another unexpected functional advantage of black SLS nylon beyond this precision difference. Lifestyle is conducting tests to validate and measure another performance difference the team has observed. If those tests confirm the observed effect, I believe I will have a chance to report on this in a future post, so stay tuned.

Subscribe now

Share

Several thoughts about this:

1. More Wins in the W Column

In machined workpieces, the use of titanium alloys and nickel-based alloys has increased. That the application range for these metals has grown can be seen in their name change. We used to call them “aerospace alloys,” but now they are more typically “high-temperature alloys,” because the materials are not so strictly for aerospace anymore. Yet that expansion also expands a dependency. Working with these hard metals comes at the price of rapid wear of even high-performance tools. The cost of doing business with these capable metals is paid in tungsten carbide. Indeed, expanded use of these metals arguably leaves industry dependent on one element in particular, a win for chemical symbol W: tungsten.

2. Will Work for Carbide

The carbide-saving methods that become more attractive as prices rise include (A) regrinding solid-body end mills and, where possible, (B) using inserted tools in place of solid-body tools. The first choice adds the handling and management cost of getting tools reground and returning them to production. The second choice turns each end mill into a little assembly, with each cutting edge a separate component. Through both options, machining facilities to some extent to swap labor cost for carbide cost.

3. The Limits of Lean

Another option is inventory. In an environment of anticipated price increase, stockpiling to lock in today’s price may make sense. This imposes an inventory cost, which is potentially worth paying if the expected price change is steep enough. The manufacturers that are the most aware of inventory and its costs are the ones running lean. Stockpiling is contrary to lean, but this scenario reveals one of the limits of a lean model. The assurance of cost saving through inventory reduction works only if there is relatively stable pricing for the inventoried item.

4. Both Reducing and Demanding Carbide

Additive manufacturing is in an ambivalent spot with regard to increasing carbide tool costs. On the one hand, AM is the nearest of the near-net-shape metalworking processes. It reduces cutting tool expense by reducing how much finish machining is needed. On the other hand, AM enables new uses of high-temperature alloys. Every breakthrough new component in a hard-to-machine alloy realized through additive manufacturing is now a little more costly. Because that part still needs some machining, the cost of a tool has to be factored into the arrival of this new part.

5. The Exotic Becoming Ordinary, and China’s Interest

If the rising price for tungsten carbide rod continues, will exotic tool materials become less exotic relative to tungsten carbide’s widespread use? Cermet, for example, is a not a replacement for carbide in general, but can be a replacement in some cases. (BTW, read about an application of 3D printing cermet.) There are other non-carbide high-performance cutting tool materials as well that, while established, are significantly less commonly used and narrower in their accepted application range. Will it become cost-effective to develop them into more widespread use?

This question about cutting tool material replacement directly points to another question: Does China benefit from high tungsten carbide prices? Seemingly it does, as 80% or more of the world’s “W” supply comes from China. Again, tungsten pricing does not appear to be the sole factor in the price rise we are seeing. However, to the extent tungsten price could be used to elevate carbide price, there is a limit beyond which the tactic stops working—related to the question of replacement. Eventually, inventors find the alternative to a precious resource that becomes more costly in relative terms and stays that way. Rather than this happening, China’s interests arguably are best served by tungsten carbide remaining industry’s go-to cutting tool material.

Orthopedic implants are typically made to standard sizes because established production methods demand scaling and standardization. In most production methods, changing a design detail of the part imposes a cost for the tooling or setup change needed.

Additive manufacturing (AM) transcends this limitation because there is no cost for design change from one part to the next. In this digital production method, a batch of parts in which every one is different can be made (at least as far as the additive process goes) as efficiently as a batch in which every part is alike.

The scatter chart below suggests how valuable this freedom will ultimately prove to be for orthopedic manufacturing. This chart comes from ALM Ortho, a Maine-based company specializing in patient-matched implants designed from patient data and made through additive manufacturing (electron beam melting) according to the instructions of surgeons.

The chart shows the cross-sectional areas for 47 patient-matched implants. The implants are all of the same type: bone-anchored osseointegrated implants for limb-loss sufferers, allowing for a secure connection to the prosthetic limb. (See an example in this article.) Yet as these data show, there is no sameness to be found in the distribution of the needs of these individual patients.

Orthopedic implants obviously are not “one size fits all.” However, even the model of “many sizes fit all” does not conform to this data set. If “many sizes fit all” could serve this patient set well, then we would see clustering around certain sizes. The scatter is more diffuse than this, revealing how varied human bodies are—or how valuable additive manufacturing will prove to be in making medical devices that serve individual patients better than has even been economically feasible before.

It did not used to be this way. If an alternative method presented itself, one could clock it with a stopwatch. Or count parts at the end of the day. Whichever option produced the fastest or the mostest was the winner.

The major difference today is in context-redefining manufacturing alternatives that touch operations outside of part making, even outside of manufacturing.

For example, additive manufacturing looks slow. In any observable sense, it is slow! As I describe in the video below—and as the photo above illustrates—it builds parts a hair thickness at a time, so speed is determined by how many hairs high the part is.

By comparison, pressworking processes look fast. In every observable sense, they are fast! I was in a setting recently where I had the chance to study a stamping process running at 400 parts per minute. (Why I was here and what I was looking at relates to a story I will report later. Subscribe below!)

One of these processes will always lose the stopwatch test and one will practically always win it. The challenge with either of them is the speed or the slowness that is not seen.

A few points draw out what I mean:

1. Leadtime IS production time. We treat them as two different things, as though, once high-speed production begins, all the leadtime needed to get to that production is forgotten and can be overlooked. The problem is, the leadtime to deliver a tool such as a mold or die for a high-speed production process can be long indeed, and must be amortized. There are relatively few production opportunities that are allowed to run so long that the cost in time of waiting for the tooling becomes inconsequential as a factor in the timing of the work. Every molded part produced fast in this current moment carries the unseen time cost for its share of the past moments spent waiting for a mold.

2. Assembly is the time consequence of part-making limitations. In the perfect part-making process, the entire product would produced complete, and produced in one step. This practically never happens because individual parts can do only so much, meaning there comes a point at which separate parts must be combined into larger assemblies. But this is time-consuming, and in a sense, assembly is the unseen time cost resulting from the failure of part making or the limitations on what part making can do. If part making could do more—that is, if part making could deliver more in terms of realizing the form and functionality of the product during the part-making step—then less assembly time would have to be paid.

3. The process swooshes when it is nothing but net. When we speak of parts that are “near net shape,” a common phrase, the meaning of just one word in that phrase determines the scope and extent of another realm of unseen time cost. That word is “near.” How near is a consequential matter. The 3mm remaining stock envelope vs. the 0.5mm remaining stock envelope potentially dictates not just more time spent in finish machining passes, but perhaps an entirely different strategy and workflow for cleaning up critical surfaces.

And now here is the big, big problem with the points above: They occur in different places. Toolmaking and process development; assembly; machining as postprocessing following a near-net-shape operation—all these steps occur in different departments, involving different teams having different conversations with one another, perhaps separated into different facilities if not different companies, meaning the time expense is not occurring all in one place.

So: How fast is the manufacturing process? If the process wins its speed by reducing leadtime plus reducing assembly time plus reducing postprocessing time, then the total speed gain might be huge, and the total speed gain might also be hard to see.

There is no single place to click the stopwatch, and completed products arriving at a higher rate might do so along a different path or in a different location than where parts have usually been made.

This is both the promise and the challenge of additive manufacturing, particularly laser powder bed fusion for making near-net-shape metal parts. The idea is bigger than additive manufacturing, because AI and reconfigured global supply chains also present options that may produce and clock very differently than what was done before.

But watching the lasers trace their way through layer upon layer of metal powder metal powder is in a sense a picture of this idea.

It seems like a slow way to make a part. Is it really? Ankit Saharin of EOS and I talk about this:

Subscribe now

Share

However, a recent conversation with Fazal Mahmood, CEO and cofounder of Phaseshift, opened my eyes to a different and perhaps more resilient optimization path for AI and its use for AM.

Rather than tuning the process, why not use AI to tune the composition of the material? With AI optimizing material composition for desired properties, printability can be one of the properties optimized for. A suboptimized hard material might need just the right build design and parameter set during printing, or else it will crack. But a material optimized for both hardness and printability is more forgiving of variations in parameter or process choices.

AI Balances Multiple Goals

Based in Toronto, Phaseshift is a materials engineering startup focused on this work: using simulation plus AI to design new materials matching user-defined needs. That includes tailoring the material for success in its intended production method, which for most of the company’s work is AM, specifically laser powder bed fusion. Rather than discovering just the right combination of laser power, scan speed, hatch spacing, layer thickness and orientation choices for a given material, and fighting these parameters to make the material behave, the company can engineer materials for compliance within a generous working range.

Process optimization is therefore NBD.

Getting to this goal is the chief use of AI. Simulation, guided by the human expertise of Phaseshift’s team, establishes the physics landscape governing the new material.

Mahmood says, “AI doesn’t do simulation. AI is more like: How can I balance all the different aims together? The main thing we use AI for is optimizing chemistry for multiple objectives.”

The photo shows an example. The parts are made of a cermet—ceramic-metal composite—engineered with printability as one of the objectives. The parts were 3D printed successfully (no defects or cracking) via laser powder bed fusion without special process considerations, and without even follow-up operations.

“We reached 97 to 99 percent density with a simple 250-watt laser printer, with no base plate heating and no post processing,” Mahmood says.

This is a notable success for cermet even apart from additive. The conventional method of manufacturing with cermet generally involves powder metallurgy followed by sintering.

With additive, he says, “One of the biggest issues right now is that most materials are incredibly picky. If the printer settings are off by even a tiny bit, the part cracks or fails.”

That is where AI comes in. “AI scans through a combinatorially explosive space, a puzzle too big for human intuition, to find the rare recipes that are stable across a broader range of processing parameters.” That stability is what is meant by “forgiving.”

“The success we have had with cermets is a great example of this,” he says.

Simplicity Scales

Mahmood believes the real value of this printability optimization, and the simplicity it brings, may be in distributed manufacturing. That is, the promise of being able to quickly manufacture a part close to its point of use by sending the file to an AM machine in that location. Optimized processes do not help with distributed manufacturing, because the process then has to be perfect in every location.

“If you need a perfect, lab-controlled environment to print a part, you can’t use 3D printing in the real world, like at a remote mine or a military base,” he says. But materials optimization can deliver parts forgiving enough to be printed anywhere.

“You don’t want a solution that works in a very tightly controlled set of parameters,” Mahmood says. “You want something that a broad range of suppliers can print and supply. That’s what will really help us scale.”

Share

Subscribe now

That additive manufacturing and artificial intelligence go together is, in an intuitive sense, not hard to see. One is a digital manufacturing technology, while the other is the analysis capability arising from new ways to process lots of data. However, just how they fit together has been less clear. One obvious scenario is this: AI could use the measured outputs of AM to better control the process inputs for precise and nuanced effects related to material properties and thermal management during the build. But as it turns out, that scenario does not sufficiently describe the AI/AM interconnection. Or rather, that scenario does not capture the sole link between AM and AI, or even the primary link as yet.

Download the report here. I detail a variety of different applications in which AI is aiding the advance of AM, all of which connect to projects or technologies presented at the recent International Conference on Advanced Manufacturing, ASTM’s annual manufacturing technology event.

Here is a summary of different areas in which AI is not only helping AM succeed, but also helping it attain to successes only AI and AM could realize:

1. Going from design freedom to many design freedoms

Additive manufacturing allows for geometric complexity. How can that freedom be directed to specific end uses where the information driving the design complexity is itself complex? Two applications I describe are (A) the use of AI to rapidly design military drone airframes for needs identified during a conflict, and (B) tailoring the osseointegration geometry of orthopedic implants to match the bone structure characteristics of an individual patient.

2. Expanding the scope of AM process control

There are more “knobs to turn” in controlling the outcome of an AM process once AI can define them, and in essence chart the wiring for these new knobs. Example: AI is helping researchers understand the effect of dwell time within laser powder bed fusion additive builds, by discovering and mapping the repeatable and controllable effects on part geometry and material properties that dwell time might be used to introduce.

3. Making postprocessing as unattended as the additive build

3D printing can produce a batch of parts with varying geometries all in one unattended build, but this leaves a batch of disparate parts in need of postprocessing operations such as machining. How can this varied work be performed in a way that is also streamlined and largely unattended? As I report, AI-enabled robot cells provide an answer.

4. Normalizing and systemizing AM process development

Given the role of support structures and the interrelationship between part and build design choices and thermal effects, additive manufacturing presents a more complex design challenge than, say, designing for machining or molding. And yet, with AI, “more complex” is a surmountable challenge. New AI-related developments are revealing how the design rules of AM are as defined, predictable and “automatable” as the design rules for other processes.

In the report, I summarize it all this way:

[O]ne of the most significant distinctions separating AM from more conventional processes might be one we are only now coming to see. In design, process control and process optimization, there is not just an AI opportunity, but a role for AI at seemingly every vital stage of the AM value chain.

Download the report: AM and AI

Since writing the report, I have discovered another interesting shade to this idea. That “obvious” application I mentioned at the beginning—namely: using AI to optimize the process inputs for the desired outputs in AM—might actually prove to be one of the less needed or less meaningful uses of AI for AM. Reason: Using AI to tailor material properties instead might bypass the need for this process optimization.

That will be the subject of my next post.

Subscribe now

Answer: This additive manufacturing method widely used for production is as repeatable, if not more so, as any other operation used to create a near-net-shape metal part.

This fact flies in the face of what some believe about metal additive manufacturing. Here is why:

Repeatability of the operation itself is not sufficient for production. Repeatability for production requires every step in the production sequence to be performed repeatably.

While manufacturers take this for granted with other processes, AM calls for the same discipline. Laser powder bed fusion is production-ready—so long as the steps upstream and downstream from it are performed in a controlled and consistent way.

I discuss this in detail with Ankit Saharan of EOS in the video below.

I am excited for this to appear, because it is part of something more: an entire series that follows from this installment. Here, Ankit and I examine the myth that laser powder bed fusion is not repeatable. Other videos in which he and I appear expose and examine other widely held myths about metal additive manufacturing.

Here is our video on repeatability:

My favorite part of this conversation is the insight about shipping. In a process involving a relatively new part-making technology, it is this new technology that draws the blame in case of a non-repeatable outcome. However, what if a far more ordinary operation like packaging the part for shipment (in this case, between 3D printing and heat treating) is actually where the variation occurs?

Again, we take process discipline for granted when it applies to the metal part-making methods that we also take for granted. But when it comes to the newer part-making operations that bring new possibilities, even these operations still do not change the need for old-school systematic thinking across the entire process overall.

In future posts, I will share thoughts on other intriguing points I see in other videos in this series.

Subscribe now

The case has to be made to non-manufacturers as well, the ones who benefit from the new products and new functionality that a new method of manufacturing makes possible. Indeed, this is the more important group to reach!

I co-wrote an article with Brian McLaughlin, co-founder of medical device manufacturer ALM Ortho, with this aim in mind. The piece aims to reach orthopedic surgeons and the stakeholders near them to explain what today’s manufacturing technology is doing for osseointegration—or the capacity for the natural in-growth of bone to provide secure fixation of orthopedic implants.

ALM Ortho is an osseointegration specialist. The company works with surgeons to create patient-matched implants leveraging the trabecular lattice surface geometries encouraging osseointegration that the company can achieve through additive manufacturing—specifically through electron beam melting on machines from Colibrium Additive.

But a major obstacle is perception-related. Just how good additive manufacturing is for osseointegration is hard to appreciate from a perspective removed from the design and production of the part. Just how far additive has advanced osseointegration in recent times is hard to see. Indeed, many in the medical field do not recognize that osseointegration is a frontier. They see it as an attribute instead, something the implant either has or doesn’t have. In fact, osseointegration is a technology that has advanced considerably now that we have digital methods to produce precisely the intricate surface geometry that can encourage bone to grow around and through it.

One recent result is varieties of implants that could not work as effectively as they do without the advances osseointegration has made. An example is patient-matched bone-anchored implants that have brought new possibilities for secure prosthetic limb attachment for some limb-loss patients. The article describes additive manufacturing’s role for these implants in this way:

Additive manufacturing allows for complex geometry not only in the sense of the intricate lattice forms encouraging osseointegration, but in another sense as well. Additive manufacturing also enables the complexity of manufacturing implants that are different from one patient to the next.

Example: A bone-anchor implant manufactured to suit a particular limb-loss patient identified by a surgeon will perfectly fit that patient, following the length and curvature of that patient’s (remaining) bone. The implant is unique to the patient and fits that patient alone.

These two uses of additive manufacturing go together. Patient-matched design enables a fit that serves osseointegration, because the tailor-made form of the device helps maximize contact between bone surface and implant surface. Meanwhile, the patient-specific device generally needs osseointegration, relying on this for fixation. The departure from a standard design means departing from established connection points for standard fasteners such as screws through a titanium rod. For the orthopedic implant that has been suited to the patient’s bone, the patient’s bone itself is the best and potentially essential means of affixing the device.

The full article is worth a read even if you are not in the medical field. Additive is realizing major new possibilities in different ways in various specialized fields. As with the outreach to medical professionals here, we need to go into each of these specialized fields to explain what options the changes in manufacturing are now making possible.

Read: 7 Points on Osseointegration Today

ICAM (held in Las Vegas this year) drew about 1,100 attendees who came to see hundreds of generally engineering-focused manufacturing technology presentations. The conference is hosted by the ASTM Additive Manufacturing Center of Excellence, and in the past, the event has been strictly about additive manufacturing. The “AM” in the name used to stand for this. The show is still predominantly about additive, and in part this is because the welcoming of other technologies related to AI and automation serves to address additive manufacturing even more thoroughly. These digital manufacturing technologies interrelate and aid one another. For example, the link between AM and AI was apparent throughout the event, with many presenters sharing use cases or research leveraging both in one way or another.

Indeed, one of the important AM-related developments I identify below involves an advance that—seemingly—has nothing to do with 3D printing. Yet I see the connection to additive, and I talk about it in the video on this development below (point 5).

Taken together, these points I was able to report from the show offer a picture of the current state of AM’s advance, and where it is delivering new promise or feeling new pressure right now.

Here is some of what struck me during the week-long event:

1. The Cost Premium Separating Additive from Other Processes Is Coming Down

Cost economy often limits where additive can find its wins, but the cost gap separating additive manufacturing from other metalworking processes is getting smaller. I saw that theme come out in various ways.

Vladimir Navrotsky of Siemens Energy, in his keynote reporting on his company’s advances with additive, commented on the price gap between AM and investment casting. New productivity enhancements such as beam shaping in laser powder bed fusion will go a long way toward closing it.

And John Calhoun of CADmore Metal described how the company is bringing Cold Metal Fusion to North America. This is the process leveraging polymer selective laser sintering to produce metal parts, sometimes at lower cost compared to other metal AM options.

More:

2. AM Offers New Possibilities for Responsiveness Addressing Military Need

In a keynote talk, Jason Bridges of Lockheed Martin spoke of the current predicament of the military industrial supply base. It has become leaner since the end of the Cold War, but also brittle. Now, in the current world, we see we may need the supply base to scale up again quickly.

Toolingless manufacturing via additive offers a way for military suppliers to scale production with little lead time, he says, if we can get out of additive’s way and let it do so. Questioning and limiting data reporting for additive parts is one aspect of this he sees.

In another presentation, Benjamin Wolf of the German military spoke about additive for responsiveness in a more focused way. The military needs to own drone production, he says. The war in Ukraine taught this, how drone and drone component suppliers can be cut off. The German military is developing its own drone production capability using additive manufacturing.

More on all of this here:

3. Platform Change in AM Still Offers Significant Room for Capability Advance

Michael Tucker of ETH Zürich is part of a team that has worked to develop an idea in laser powder bed fusion (LPBF) additive manufacturing that is outside the box in a literal sense. The company’s Rapture system is a cylindrical LPBF machine. The recoater moves continuously in a rotary path. The laser scans in polar coordinates.

Tucker has spoken about this system at ICAM in the past, but what is new this year is the way the reconceived platform is revealing its promise for capability advance. The continuous recoater is ideal for multi-material builds, he says—an option more difficult to achieve using the back-and-forth recoater of a box-shaped LPBF machine. Cylindrical rocket engine parts could more easily be made from copper and Inconel together, the two metals combined in a solid part made from a single build using the cylindrical machine.

More in the video:

4. A Scaling Opportunity Waits for Increasing Systemization of AM Machines

To scale additive, centralize programming the machines.

In his keynote talk at ICAM, Thomas Pomorski of Ursa Major described the company’s success using a universal slicer to program LPBF machines from seven different OEMs. A build defined once has been successfully translated for seven different machines. They’ve proven this with various builds. (In the video, I mistakenly give a smaller number than seven.)

Programming for machining already works this way. That is, one environment creates programs for many different machine tools. Shouldn’t additive work this way as well?

Ursa Major worked with Dyndrite on the solution it has validated. Pomorski sees this kind of centralization as part of what it will take to scale up production by dramatically increasing the number of AM machines in use in one network or facility.

Bonus: The centralized programming opens the way for AI to be part of the solution as well. More in the video:

5. Systemized Subtractive Operations Offer Promise for AM Postprocessing Needs

Are robots ready for AI?

Mostly no. Many robot applications are silos. But that may change. What will happen once robot cells integrate not just hardware, but data as well?

Tyler Bouchard of Flexxbotics spoke about his company’s data centralization system for automation, in which read/write capability for every machine in the robot cell is integrated and consolidated. He described the coming AI opportunity for robot cells this will bring.

I see this directly aiding additive manufacturing. For AM to be cost-effective for more of production, the steps downstream from the additive build will have to be cost-effective. That is, postprocessing either needs to be minimized or eliminated, or else it needs to be as automated and as lights-out as the 3D printing. Material-removal cells that are robot-handled and AI-controlled offer the way to achieve this latter need.

More:

6. AM Offers an Economically Viable Solution for Reducing Plastic Waste

Can 3D printing reduce waste plastic heading to landfills? Yes. There is a path and it’s working.

Carolyn Seepersad of the Georgia Institute of Technology spoke about her work with ReCreateIt Labs, which is making home goods out of waste plastic that are sold through a store run by Habitat for Humanity International.

The ground plastic from bottles and other sources is 3D printed on a Gigabot machine from re:3D Inc.

There are challenges. They include:

• Engineering challenges. How can the plastic “flake” made from ground waste plastic be made more printable?

• Product challenges. What are the items to be made from this material that different customers with different interests most want to buy?

The team is rising to these challenges, and while the solution they are finding cannot account for all waste plastic, it clearly could be scaled beyond the scope of just this project. More in the video:

7. Geometric Freedom Can Deliver Mass Customization at Minute Levels

“Patient-matched” implants: The term usually refers to individualized orthopedic implants tailored to overall dimensions of a patient’s bone. But what about also matching the porous structure of an individual section of bone?

According to Jacob Peloquin of North Carolina State University, additive manufacturing can do this, too, and AI will get us there. He gave a talk on research into patient-matched osseointegration geometry.

Our bone structure is not identical from person to person, he notes. It is not even identical throughout one person’s body.

The system the NC State research is developing relies on high-resolution CT scans. AI agents work from this data to tailor an implant’s bone-ingrowth lattice to match the distinctive bone geometry, to better encourage natural bone growth into and around the geometry of the implant. The result is an implant that is patient-matched in more ways than one.

More in the video:

As part of my reporting from ICAM, I also filed a more detailed article on the growing and unexpected array of interconnections between AI and AM. ASTM will be posting this article. When it appears, I will share key points here, as well as a link to the full article. Subscribe to be alerted when this piece appears:

Subscribe now

In the video, you hear me use the words “print farm,” a phrase commonly used to describe certain additive manufacturing applications like this one. I thought more about the term after recording this. The phrase is both fitting and curious. It is fitting because of how well the agricultural metaphor describes the promise of production on a collection of small 3D printers. In a manufacturing system such as this, the parts grow, the process is gradual, it is quiet and safe, and a single person tending the work can gather the parts when they are done.

But the phrase is also curious (A) because it is not clear exactly when it applies (both lasers and large machines would seem to move the process in a direction that feels less like farming, yet some would still use the phrase for these machines), and (B) because of the way the “print farm” idea still draws on promise to a great extent. Even though the phrase has been with us a long time, what it describes it not common. I have seen few instances of, say, 15 or more small 3D printers all at once running simultaneous production as part of the same “farm.” With its six 3D printers, the application in the video is perhaps more like a print garden. However, part of what caught my attention with Angel Aerial Systems is how commited they are to the farm concept. They have created their product to be brought to market with a production system that will entail no factory as we imagine it, but instead space and tables for many more of these small machines. With a few more printer rows like the one described in the video, all producing parts for harvesting later, soon Angel’s operation will fit what the phrase has long described.

Share

Videography: Hannah Zelinski

This newsletter will be equal parts reporting and commentary. My aim is to explore how manufacturing is changing, how the technology of manufacturing is changing, how these changes interconnect and affect our world, and what each development might offer or mean for those who are building businesses or careers connected to manufacturing’s success.

If you have subscribed, thank you. If you have not yet subscribed, please consider doing so. I will post when I post—subscribing ensures you won’t miss it when I share something new.

Here now is the Zelinski Report’s first post, about a technology startup advancing two interrelated ideas: a new concept for long-endurance drones and an unusual plan for how to produce them.

Angel Aerial Systems is reimagining the way drones fly.

The challenge with drones, notably the quadcopter design that is common among drones today, is flight time. How can the drone be reconceived to extend its flight endurance without increasing cost? Can this be done by increasing battery size? Answer: No, not by enlarging the battery….

But instead by transforming the entire drone body into the propeller that keeps it aloft.

Angel CEO and co-founder Will Hasting (left in the photo above) is working to develop this design, leading the team now bringing this concept to market. The quadcopter and drone concepts like it, he says, “are based on conventional airframe designs conceived in the twentieth century.” That is, during the time of manned flight. The expectation of a pilot or passenger still haunts design assumptions, without this unnecessary consideration being seen. Take away this consideration, fully accept unmanned flight, and, he says, “It reopens the design space. The time is ripe to imagine what the drone airframe looks like for the twenty-first century.”

And the company realizing this idea is doing so using a production system that is almost entirely vertically integrated, while also imposing little in terms of footprint or capital cost.

A row of desktop 3D printers produces all the drone structural components, along with the shell for the controller.

So while this startup is rethinking how long a lightweight aircraft can remain in the air, it is also reevaluating how far low-cost equipment can go to support the company’s production.

Anything Can Be a Weapon—What Else Can You Do?

Hasting, a Navy veteran and engineer who in previous roles has worked in both aerospace and additive manufacturing, was inspired to launch this startup by the war in Ukraine. He says he was moved by the suffering of this war and wanted to be of help. At the same time, he learned of a Defense Advanced Research Projects Agency (DARPA) request for low-cost, long-endurance, loitering drones for Navy and Marine missions.

Hasting pursued but never won the project award. And during the pursuit, he elicited feedback from a mentor. University of Cincinnati aerospace engineering professor Kelly Cohen offered the perspective that changed his direction.

“Dr. Cohen said, ‘You can make a weapon out of anything,’” Hasting recalls. “He challenged whether we could ‘do something good in the world’ instead.”

Another use for long-endurance hovering drones presented itself. Hasting learned of a police search in Indiana for an elderly man lost in the woods, a search that was confounded by limited drone battery life and the resulting inability to fully cover the search area, even with three quadcopter drones in the air.

Angel redirected its efforts toward local-community first responders: police, rescue, firefighters. This is the market—safety and protection—that the company is now manufacturing long-endurance drones to serve.

On Standard Drones, Too Much Mass Is Non-Performing

The Angel drone design, named the Trio, is a departure from the quadcopter design. On the quadcopter, four propellers carry a roughly square-shaped drone into the air.

The Trio is shaped like a Y, with three propellers, one at the end of each arm. The propellers provide for the initial takeoff and vertical ascent.

Then, the drone enters “windmill mode.” The propeller arms all pivot, driving the drone to rotate, and each arm of the Y serves as an airfoil as the entire drone spins to move through the air. Play the video to see this:

On a quadcopter, much of the drone mass is non-performing, carried by the propellers. During the Angel drone’s windmill mode, as the drone spins, nearly the entire airframe is employed to generate lift.

The difference translates to a dramatic difference in flying time. Compared to the quadcopter drone model most widely used by first responders, Angel’s Trio can remain in the air 3 to 6 times as long before recharging.

The spinning drone can still carry a payload. This is attached via a turret at the center that is counter-rotated via a fourth internal motor. The turret is spun in a contrary direction that perfectly offsets the windmill. Thus, a mounted camera remains fixed and stationary while the drone hovers for hours in windmill mode above whatever scene the camera is watching.

How AI, AM and Amazon Transform Possibilities for Technology Startups

This new method of making a drone fly was developed and brought to market within a window of about two years. Hasting describes why engineering entrepreneurs now live in a new world of accelerated possibility. At the major aerospace company for which he worked until 2019, the challenges of a large organization might have extended a project of this ambition to over a decade or more, he says, and even the nimbler metal AM company for which he worked until 2022 would likely have had difficulty iterating at Angel’s speed.

He credits three advances for the development speed now possible, at least one of which is unsung. They are: artificial intelligence, 3D printing, and overnight shipping sources following the example of Amazon Prime.

What all these resources have in common is the way they collapse delays that would otherwise halt progress.

“AI is helping us diagnose and overcome engineering problems much faster than traditional analysis methods,” he says. Generative AI can point to answers that would otherwise require effort and delay to search and synthesize. Meanwhile, “3D printing and the new sophisticated filaments available for it let us speed through 30 design options in a couple weeks.” And specialized hardware that previously would have imposed a long shipping wait is now frequently available overnight, sometimes through Amazon, or often through industrial providers operating at the same speed.

The rapid access to needed hardware is the unsung enabler, just as valuable as AI and additive manufacturing. Angel’s vertical integration extends to producing its own electronics in-house, including developing circuit boards that conform to geometric spaces in the drone. Overnight delivery of electronics parts supports the speed of this development.

All these resources in combination—AI plus AM plus Amazon-like delivery—have proven to be combustion for accelerating Angel’s innovation.

If You Engineer Cost into a System, It Is Hard to Get It Out

The drone components largely must be 3D printed, not injection molded, because large hollow spaces for weight savings create geometries that require 3D printing to produce.

In the meantime, 3D printing creates opportunities for assembly consolidation. The Angel drone body as it currently exists consists of 12 major components, all 3D printed. Assembling a drone is a straightforward task easily performed by the company’s one 3D printer technician as the machines are printing.

Materials are ASA and nylon. The Prusa HT90 machines Angel uses feature heated chambers, though the company is operating them at their temperature limits to work with the material as productively as it can with the needed part-to-part repeatability.

These are, to an extent, high-end consumer or engineering 3D printers. An industrial-production polymer AM system could possibly print each part faster—though Hasting is not sure of this. He notes the longest Prusa print cycle for one of the drone components is 2 days, a cycle time not set by the printer itself. “For our carbon-fiber-filled materials, the filament limits the speed. So you’re speed-limited regardless of the printer, unless you use multiple extruders,” he says.

Meanwhile, each of these 3D printers costs about $10,000. The opportunity Hasting sees in this is that, for the price of an industrial polymer AM system, he says, the company could add somewhere in the range of 10 to 40 more of these machines. Hasting therefore wants to design the drone and engineer its components to accept the constraints of the Prusa machines and continue to rely on a printer like this for production.

He says, “Once you design cost into a product, it’s hard to get it out.”

His expectation is that Angel will scale up drone production through the addition of perhaps dozens more desktop 3D printers.

In the follow-up to this article, I plan to consider that choice. I will be posting a video filmed at Angel Aerial Systems that analyzes the print farm as a means of production, considering cost, labor and scaling among other factors. To make sure you are alerted when that video appears, be sure to subscribe:

Subscribe now

Share

Photography: Audrey Zelinski