Manufacturing guy-at-large.

Filtering by Tag: topper

Build process simulation

Added on by Spencer Wright.

Early last year, Andrés Bellés Meseguer reached out to me with a proposition. He had read my piece in Metal AM magazine, and wanted to use my printed parts from DRT Medical Morris to verify a build process simulation workflow that he created using Abaqus. I agreed, and with Dave Bartosik's help I got him the build files necessary to simulate the print.

Andrés' full results were published in a paper titled Prediction of Distortion of a Titanium Bike Part Built by DMLS, which he presented at a NAFEMS conference in November. The simulation used a fine hexahedral mesh at the part itself, and a coarser mesh for the surrounding powder bed and the build platform. At each timestep, heat (representing energy applied by the laser) is applied to nodes throughout the model; it then dissipates throughout the structure. Below, see a thermal map of the part about 70 minutes into the build:

Image courtesy Andrés Bellés Meseguer and Prime Aerostructures

You can also use this simulation to model distortion in the part - seen here at the end of the build:

Image courtesy Andrés Bellés Meseguer and Prime Aerostructures

The distorted areas in the simulation correspond well to the as printed part, but Andrés notes that the magnitude values don't match perfectly; it's likely that some of the discrepancy can be narrowed by adjusting thermal coefficients.

This field - simulating additive processes to predict and compensate for built in stress and distortion - is one that I've been excited about since I began working with AM. Thanks to Andrés for sharing his work - I'm looking forward to more progress on this soon.

My writing in Metal Additive Manufacturing Magazine

Added on by Spencer Wright.

I'm extremely proud to say that my writing is featured in the current issue of Metal Additive Manufacturing Magazine. The article summarizes much of my work over the past two years, and includes many of my thoughts about how the industry can be better in the future. You can read the full article here.

Thanks so much to Nick Williams and the rest of the Metal AM for asking me to contribute!

DMLS vs EBM titanium parts

Added on by Spencer Wright.

Note: Below, I use the term "DMLS" to refer to laser metal powder bed fusion. Some will have issues with this usage; I encourage readers to ignore the (possibly trademarked) terms themselves and focus on the information contained within them.

Over the past few months I've spent a lot of time trying to determine whether DMLS or EBM would be a more suitable technology for the titanium bike parts I'm working on. Early on (in my collaborations with DRT and Layerwise) I focused on DMLS, and more recently I've worked with Addaero on EBM. These two technologies are in many ways complementary, and in truth I expect to work with both of them - but their design constraints and total manufacturing process chain are a bit different, and I wanted to spend some time understanding how that will affect the cost and quality of the parts I'm working on.

In the video below, I show some of the basic differences in surface quality and support structures. Of course there are many other factors that are worth exploration, starting with minimum feature size and built-in stress, probably; I'll get to those in future updates.

Incidentally, I'm *really* interested to see the differences between DMLS and EBM in printing the lattice structures I've been working on. More soon :)

3D printing titanium: Learning to learn from success

Added on by Spencer Wright.

Dear reader — 

This report is an update to my experiences in metal 3D printing; it describes a good chunk of what I’ve been working on over the past four months. While I’d like to say that it stands on its own, I think there’s some context — especially if you’re just beginning to explore metal 3D printing — that can be gained from reading my earlier posts (starting with my long "Bin of broken dreams post") first.

For background: I believe that functional, engineered consumer products made by additive manufacturing are an inevitability. In order to prepare myself for that inevitability, I’ve been developing a line of bicycle parts made by metal powder bed fusion, a process that’s used extensively in aerospace, medical, and tooling applications. My last post described the difficulty and constraints I’ve experienced in part geometry and build orientation. Since then I’ve turned a corner, and today I can say that I’ve successfully designed, built, and tested a functional product — which at last puts me in the position of needing to learn from success.

As I’ve documented the process — and frustrations — of developing metal 3D printed parts, I’ve been pleased and surprised at the number of people who’ve reached out to me to commiserate (n.b., if you’re reading this and want to do so yourself, please drop me a line). Without exception, they have expressed solidarity. “We share all of your frustrations,” one person said. “I have been through the same pain as you,” said another. 

One of these people was Tom De Bruyne, General Manager at Layerwise. Layerwise is a Belgian company which was started out of the Catholic University of Leuven (one of the premier centers of additive manufacturing research); it was acquired by 3D Systems in late 2014. They’re famous for being one of the few service providers who built their own laser metal powder bed fusion machines, and have a ton of experience making 3D printed parts at both prototype and production scale. We struck up a conversation, and soon agreed to work together.

While popular opinion would have you think that quantity is a non-factor with 3D printing, the realities of running a service bureau are much to the contrary. To job shops, quantity is a critical factor; if a part will be produced at large volumes, every detail of its design and manufacturing life cycle must be examined. If, on the other hand, you’re printing a tool — or a prototype of a part that will be manufactured conventionally — then most shops will focus on getting the first print right without modifying its underlying geometry.

My project falls somewhere in the middle: while my design is certainly imperfect, there are many aspects of it which are very close. Moreover, it poses challenges (most notably its opposing cylinders, oriented 90° apart, and also its thin-wall construction and bolt boss) that will exist throughout any redesign, and solving them now will only improve my ability to deal with them in future iterations. 

At the current juncture, the key questions to test are:

  • Can we reliably build my current design with minimal post processing?
  • Does my current design meet the necessary performance standards (strength, security, etc.) for bicycle seatmast toppers?

In practical terms, the first question boils down to whether we can build a part that can be installed on a bicycle. This means two things: maintaining inner diameters which are round and dimensionally accurate to within +/-.006", and having a bolt boss on the long cylinder which, when tightened, is capable of securing the part to a bicycle’s seatmast.

My last prototype was built on its side, and the long cylinder’s aft wall distorted as it was sintered. We added a series of solid ribs in order to counteract the built-in stress in that wall, but it ended up like a game of whack-a-mole: each reinforcement just moved the stress somewhere else. A new approach was needed.

Unlike most of the US job shops I’ve spoken to, Layerwise bundles together many low-volume customer orders into each build; my part was printed alongside a handful of other titanium parts in one of the 15–20 machines that Layerwise designed and built themselves. This means that Layerwise is able to process a large value of parts at once, even if they’re only building prototype quantities of each design.

This poses a significant challenge: if one part fails, it can jeopardize many customers’ orders. To mitigate this risk, Layerwise constantly monitors a slew of process signatures, and can adjust machine parameters on-the-fly if they detect problems. They’re also working on a layer-based deposition control & verification system (the details of this are secretive, but it sounds similar to the optical tomography systems that other machine manufacturers are working on today). Still, each part is evaluated carefully beforehand in order to anticipate and avoid failures. Especially for short-run prototypes, it’s usually better to over engineer the support structures (and ensure a successful build) than skimp out and need to rebuild the part later.

Layerwise is secretive about their custom-built machines, but they did tell me a few details. They each contain a single laser, and are built around roughly the same form factor (275mm x 275mm x 420mm) as most of the industrial machines on the market today. I also understand that Layerwise uses a recoater blade (unlike 3D Systems’ ProX line, which uses a roller), though I couldn’t confirm whether it’s a hard material (like the high speed steel blades that most EOS machines use) or a soft one (like the polymer wipers often used on Concept Laser machines). They’re able to print in about 15 materials, with the majority of their work being done in titanium. They monitor temperature, pressure, and oxygen content on-the-fly, and are working on additional variables — including full melt pool analysis.

Layerwise has in-house wire EDM, and has a daughter company that does 5-axis CNC machining. But most prototype parts, including mine, are finished by hand. When validating a part for production, Layerwise tests the full manufacturing lifecycle, building multiple full batches of parts and sending them out to be post processed. Once they’ve validated the process, they will in some cases hand off production to other 3D Systems divisions.

Reorienting

I worked with Martijn Vanloffelt, a project engineer at Layerwise, to prepare my part to be built. He used a few key tricks:

In order to maintain roundness in the saddle clamp cylinder (the shorter of the two cylinders, which was was going to be oriented more or less parallel to the build plate), Layerwise reinforced the inner diameter with three serrated discs. 

They also oriented the part slightly off-axis in both the X and Y axes. This brings me to a point that’s worth highlighting: In metal powder bed fusion, a part’s orientation has a number of effects. First of all, any surface with an angle of less than about 45° (depending on material) must be supported. As a result, it’s generally better to orient a part so that all overhangs are as steep as possible.

From NIST Technical Note 1801, a good primer on laser metal powder bed fusion.

But in addition, one must consider the angle between the part and the recoater blade. If the part lifts up at any point during the build, the recoater blade will strike it. The longer the area of contact is, the worse the result will be. Some machine manufacturers offer alternative recoaters to lower this risk (3D Systems ProX uses a roller; EOS offers a carbon fiber brush; Arcam uses a metal comb; and both Concept Laser and SLM offer soft polymer blades), but most use a piece of high speed steel (or, in the case of older EOS machines, a ceramic blade) to spread powder across the build platform. Regardless, it is usually better to orient parts slightly off axis in the XY plane, so that the blade doesn’t contact the part’s walls all at once. 

Orienting parts off axis can also help improve surface finish. When a cylinder’s axis is aligned in the XY plane, the top face will exhibit an undesirable stepped appearance; my earlier prototypes all had this feature. When a part is oriented off axis, the surface finish is generally more consistent.

I should note that none of these techniques is guaranteed to work in all cases. Layerwise has a lot of experience building a wide variety of geometries, and has developed a sense of how to anticipate and deal with issues as they come up. I got the impression that the techniques they used on my part are things they’ve used in the past, but each design is different — and even a tried-and-true method of dealing with one design isn’t guaranteed to work well on another.

The Layerwise team also put a lot of care into generating mesh supports. Like most of the additive metal industry, Layerwise uses Materialise Magics for their final build prep, and they’ve developed expertise in exploiting the software in creative ways. I’m not able to share a detailed description of the supports Layerwise created for my part, but I can say (and anyone in the industry could confirm) that they were needed in four areas:

  1. Underneath the part to anchor it to the build plate. 
  2. Inside the saddle clamp cylinder. 
  3. Inside the window in the center funnel area. 
  4. Inside the seatmast clamp bolt boss. 

Layerwise took great care to orient the part such that it didn’t require support structures inside the hidden voids in its center section. This is something that designers and project engineers alike need to think about as a part heads into production. Not only can powder bed fusion not make fully sealed voids (if you printed a sealed sphere, the entire center would be full of trapped, unmelted powder at the end of the process), but many geometries will require support structures in areas where they’re difficult or impossible to remove. For instance, a Klein bottle could be printed in metal — but no matter how you oriented it, there would likely always be support structures stuck inside its fat end. Because of the angles in my part, it was possible to avoid this — but a different design might not fare as well.

The Parts

The first part Martijn printed was a big step forward: The build completed successfully without collapsing. However, a new problem arose. The windows on the seatmast clamp area caused the two “leaves” of that cylinder to twist as they were built. By the time the window closed back up, they had become misaligned, and a witness was clearly visible where they joined back together. The part had a clear flaw — and it wouldn’t be acceptable for production.

In the next build, Martijn added a curved, perforated disc to each of the seat mast cylinder’s windows, keeping them aligned as they grew. The part that resulted was a full success, printing with clean, smooth surfaces and good near net dimensions.

Layerwise’s second build.

Considering how much support material needed to be added back into the seatmast clamp area just to get it to build properly, I’m struck again with how inefficient my design is. The windows in the sides of the part are meant to reduce both weight and cost, but a bunch of energy is put into supporting them — and then cleaning the temporary supports out again. Instead of windows, I could just as well have replaced the walls with a lattice structure that would both decrease mass and be self-supporting during the build process, bringing the part’s cost down.

This hammers home a point that has plagued my design process: Without knowing — and, optimally, having input into — how a part is going to be oriented and supported during its build, designers are doomed to creating inefficient designs. Designing for manufacturing requires an intimate knowledge of the manufacturing process, including direct access to detailed information about how the part will be oriented and supported. But in most designer/service provider relationships today, that information comes well after many of the important design decisions have been made — if it comes at all. As a result, it often takes a large investment (both in time and money) just to prove whether additive can possibly be used to create the part at hand — and once that’s been proven, many additional iterations are sure to be needed.

This is a key problem in today’s additive manufacturing supply chain: while parts are usually designed in a solid modeling environment (often Autodesk Inventor or Solidworks, each of which cost between $5–10,000), builds are oriented and supported in Materialise Magics — which costs an additional ~$20,000. As a result, independent designers are stuck with a disjointed process, which requires costly iterations and lots of communication with the service bureau who’s preparing the part to be built.

Regardless: At this point in the process, it didn’t make sense to redesign the seatmast clamp area to reduce supports. Martijn’s build had a very high likelihood of completing successfully, and it was time to put it to the test.

It worked.

Post processing 

After printing it, Layerwise did a bunch of post processing before sending the part to me:

  1. First, the entire build plate was stress relieved. Layerwise’s stress relief process is proprietary, but a typical process would involve putting the build plate in a furnace and bringing it to 600°C over a period of an hour, then holding it there for three hours before turning it off. In theory the furnace is either argon purged or vacuumed, but in practice it may contain small amounts of oxygen too. Layerwise says that the vast majority of the stress relief that they do is performed in a vacuum, but argon is typically used on prototype parts.
  2. Then the parts were removed from the build plate. Like most shops I’ve spoken to, Layerwise uses wire EDM — though bandsaws are also common.
  3. At this point, each customer’s part is separated and processed on its own. Supports are removed by a totally unsexy manual process, often involving wrenches, picks, and mallets.
  4. Where support structures have been removed additional cleanup is usually necessary. On prototype parts, Layerwise makes extensive use of rotary grinding bits.
  5. The inner diameters of my part were both ground to their final size. Layerwise told me that this process was 100% manual, and I was blown away at the precision and consistency of the surface finish.
  6. Then, any remaining features were micro shot peened with a nonabrasive ceramic medium.

At this point, Layerwise sent me the part. Still to be done, however, was to tap the female threads in the seatmast bolt boss.

Herein lies an important point: metal 3d printing does not, in general, produce usable mechanical features like threading. In conventional manufacturing, threading is often just another step on the same machine: mills and lathes can both easily create female threads. But with additive, threading almost always requires secondary processing. As a result, the design files that are loaded into Magics only contain plain-bore thru holes; any threading specifications must be documented (and manufactured) separately.

So, the part that I received simply had a 4.2mm hole in it; it was up to me to cut the M5 female threads. “No problem,” I thought. I’ve got a tap handle right at my desk, and am more than comfortable using it. 

At this point, I became painfully aware of what’s called alpha case. Alpha case is a very hard, brittle layer of oxygen rich titanium in a part’s surface (for an interesting study on alpha case depth, see this); it’s the result of the titanium having been processed at high temperatures in environments where oxygen is present. And as I tapped the hole in the first part that Layerwise printed me, I realized that it’s very, very difficult to cut.

In order to make my job easier, I purchased a set of custom progressive taps from Widell Industries. Progressive taps cut threads in three steps, increasing the thread depth as they go. As a result, the cutting force required is generally much lower.

Even using progressive taps, I was shocked at how difficult tapping the second part was. It was incredibly slow going, and produced a lot of heat. I used cutting fluid liberally, and 45 minutes later was done.

I should note here that titanium is a hard metal regardless of how it’s processed. Moreover, alpha case is preventable; in this case, it’s simply the result of the stress relief process being done in a furnace that contains some trace oxygen. Annealed titanium 6/4 has a typical Vickers hardness of about 349, but when a part has been stress relieved in an oxygenated environment, that number might jump to more than 412. By comparison, 4130 steel and 6061-T6 aluminum (both of which are used extensively in the bicycle industry) have Vickers hardnesses of around 207 and 107, respectively. In future prototypes, I would probably specify that the stress relief should happen in a full vacuum; that would at least make the tapping a bit easier.

Regardless: Finally, the part was ready to assemble:

After a total of eight build iterations, I could finally have the part tested — and learn whether my underlying design worked.

Testing

To help understand if my design would handle real world performance requirements, I worked with EFBE Prüftechnik, a German bicycle & component testing facility. EFBE tested the part to ISO 4210–9:2014, 4.5. That test entails:

  1. Clamping the seatmast topper onto a 34.8mm pillar angled at 73°, and fitting a dummy saddle rail into the saddle clamp.
  2. Applying 100,000 cycles of a test force of 1230 N, at a distance of 70mm to the center of the rail clamp, with the saddle rail tilted down/backwards by 10°.
  3. Applying a vertical static load of 2050 N to the center of the saddle rail clamp.

Marcus Schröder, managing director of EFBE, put my part through the dynamic test first. It passed.

Before he went through the static load test, Marcus asked whether I wanted to make sure I got an intact part back — or if I would rather find the failure mode in the static test. In the latter scenario, he would apply the maximum force his rig could handle and see if he could get the part to break — allowing me to redesign it accordingly. Wanting to know as much about my design as possible, I chose the latter option.

Marcus’s test fixture was capable of applying 3750 Newtons to the part. My part withstood the whole thing.

So.

Marcus used penetrating dye to confirm that the part didn’t have any micro fractures, and it came back negative. The part had met and exceeded the requirements for parts like it.

My part, covered in penetrating dye after being tested.

It’s worth noting that this test is simply that: a test. It’s meant to simulate real world conditions and guarantee that the part meets generally accepted standards. But it simulates those conditions only generally; manufacturers of these kinds of parts will often have their own in-house spec that to tune the characteristics they optimize for. But in general, a designer needs to choose a test, and then optimize his design such that the part fails just beyond the test’s requirements. If I trust the ISO spec implicitly then it stands to reason that I should remove more material from the part; after all, it passed the test with a wide margin.

Regardless, my part could be further optimized. What I’ve done to date was prove a basic concept: That metal powder bed fusion can be used to make thin walled bicycle parts. Which, to be honest, isn’t a particularly surprising result; after all, metal powder bed fusion has been used by others to create all kinds of bike parts, including both road and mountain frames. 

The question is: Can I make it commercially viable?

Practical matters

With the current design and an order quantity of 10 pieces, the as-printed parts cost about $500 to make. Meanwhile, the most expensive commercially available seatmast topper I’m aware of (made by No22) costs $300, and the fanciest seatpost I’ve ever seen (made by AX Lightness) was under $600.

Now, there are a number of interesting things to note here. First, I’m able to buy in fairly low quantities. It’s not unreasonable for me to purchase parts in batches of 10, which is about as low as any non-stock commercial product in the world— and much lower than most products that involve forging, casting or CNC machining. If I can sell my part at a high end price point, then it wouldn’t take much cost reduction before I’ve got a reasonable margin — even with a strikingly low order volume. And there are a number of ways that I can reduce cost on this part:

  1. Even keeping the part’s design the same, I can reduce the cost by 25–40% by doubling the layer thickness. This will result in a rougher surface finish, but it’s possible that the difference will be acceptable.
  2. A significant amount of time and effort can be saved by redesigning the underlying model so that the inner diameters need very little — or even zero — post processing. It’s unclear exactly how much work this will take, but it could reduce the price significantly.
  3. Moreover, the entire part can be redesigned in order to reduce both the end mass *and* the amount of support structures necessary. Both of these have a big effect on price, though it will be time consuming to find an optimal design.

All of this assumes that I stick with a laser powered process. Electron beam melting (which I’ve been experimenting with) might reduce cost further, though the design constraints there are quite different. 

It’s also possible that while this part isn’t the absolute best application for additive manufacturing, there’s another place on high end bicycles where additive works better. This is key: The cost of this part is in the right order of magnitude as my goal. Any number of small changes — whether to the machine’s build parameters, the design, or the underlying cost of the technology — could easily put the numbers in my favor.

Industry observations

As I hope you would expect by now, a few things have jumped out to me in this past few months of work — some relating to longer term questions I have about the industry.

Manufacturing design

As a designer, having more — and earlier — access to support structure generation software is incredibly helpful. Today, countless design decisions are made on little more than a hunch; it’s not until much later that the ramifications of those decisions become evident. This leads to an inefficient design-for-manufacturing process, where it’s difficult for product designers and manufacturing engineers to communicate all of the nuance needed. Until professional grade design software (and here I’m looking at Autodesk, Dassault, Siemens, and PTC) allows these implications to be understood early on, this will be a big problem. In other words, I should be able to set build orientation and design support structures directly in my CAD modeling environment.

Documentation

The designs that are loaded onto a metal 3D printer are often very different from the finished part, whether due to the addition of stock material to be removed subsequently (like my female threads), or the support structures necessary to hold the part on the build platform. But today’s modeling software generally only shows one state for each 3D model; those intermediate, near net shapes aren’t linked to the end design. This makes the design for manufacturing process even more disjointed and awkward, as it means that I’m never working on the same design documents as my manufacturing partner is. This communication structure is bound (cf. Conway’s Law) to be replicated in the end product, and the result is bad. To fix it, we need a new class of software that blends CAD (computer aided design) and CAM (computer aided manufacturing), allowing designers to understand a part’s production cycle with perfect clarity.

True design optimization

In the work I’ve described so far, I’ve relied exclusively on conventional volumetric and NURBS modeling techniques. But a new wave of design tools is out there — topology optimization and lattice generation, for instance — and they promise to create designs directly from functional requirements. Presumably these techniques would be just as applicable to my part as they are to the applications they’re already used on (mainly aerospace, biomedical, and other high tech applications). I’ve begun to explore this space, but the fact of the matter is that I’m not aware of a single consumer product today that was designed with these tools. My hope is that they both remove weight and make the part more visually appealing, but it could take a lot of work and some expensive (and experimental) software to find whether I’m correct.

Blatant editorializing: Gongkai for industrial additive manufacturing

Today, the viability of additive processes is totally opaque, producing a chilling effect on the creativity of both designers, service bureaus, and machine manufacturers. It’s my strong belief that that will only change by better understanding the efficiencies (and inefficiencies) of the additive manufacturing toolchain, and through my own work I hope to do just that.

In a series (one, two) of recent blog posts, Bunnie Huang describes the way that Chinese electronics manufacturers have been able to drastically decrease the cost of consumer electronics. To me, they provide an example for how additive manufacturing could advance much more quickly:

My most striking impression was that Chinese entrepreneurs had relatively unfettered access to cutting-edge technology, enabling start-ups to innovate while bootstrapping. Meanwhile, Western entrepreneurs often find themselves trapped in a spiderweb of IP frameworks, spending more money on lawyers than on tooling. Further investigation taught me that the Chinese have a parallel system of traditions and ethics around sharing IP, which lead me to coin the term “gongkai”… this copying isn’t a one-way flow of value, as it would be in the case of copied movies or music. Rather, these documents are the knowledge base needed to build a phone using the copyright owner’s chips, and as such, this sharing of documents helps to promote the sales of their chips. There is ultimately, if you will, a quid-pro-quo between the copyright holders and the copiers.

It would be an understatement to say that industrial additive manufacturing hasn’t adopted gongkai. Today, trade secrets & patents are the name of the game; the access I’ve been permitted (by Layerwise, DRT Medical-Morris, and countless other friends across the industry) is rarely afforded to others. It’s my feeling that this is bad, both for the technology as a whole and for the long term interest of individual players within it.

For instance: Anyone with experience could, given a part geometry and its build orientation, surmise more or less what the support structures will look like. If you have a physical part in hand, it’s even easier to reverse engineer; a part’s layer boundaries reveal its build orientation, and even with careful cleanup it’s generally possible to tell which surfaces have had support structures removed from them. In short, manufacturing forensics is, with enough experience and care, fairly reliable. And yet orientation and support structure setups are almost always treated as closely guarded secrets.

With so much uncertainty in industrial additive manufacturing today, firms are caught in something of a prisoner’s dilemma; their protective IP strategies prevent the industry from moving forward in the way that everyone ultimately wants it to. 

All around the world, intelligent, hardworking people trying to solve the most basic problems in additive manufacturing. Everyone in the industry knows what these are, and all of the major players are fighting tooth and nail to solve them first. And though it seems a long way off, I think most of them genuinely believe that they’ll see fully automated orientation & support structure generation within the next decade or two.

But today, the process of printing a part is decidedly hands-on; expertise is critical. “The people who are good at this stuff are good at it because they’ve been doing it for eight years,” one industry veteran told me recently. For sure, this industry is full of valuable intellectual property. But in most cases, *craftsmanship* is central to most firms’ bottom lines — and it is protected at all expense.

I believe that industrial additive manufacturing needs far, far more knowledge sharing. We need an environment closer to the one that Bunnie describes: 

We need “a ‘network’ view of IP and ownership: the far-sight necessary to create good ideas and innovations is attained by standing on the shoulders of others, and as such there [should be] a network of people who trade these ideas as favors among each other.”

In the coming moths, I’m looking forward to working on just that.

This post is just one in a series, and I remain convinced of what I’m working towards. With any luck — and more open collaboration with intelligent, committed people in the industry — I’ll have more to report soon :)

Thanks to

First, thanks to Martijn Vanloffelt and Tom De Bruyne, of Layerwise, for both their hard work and their willingness to help me understand how they work.

Second, thanks to Marcus Schröder, of EFBE, for both his hard work and the enthusiasm he had for testing my part — and discussing at length the testing & engineering cycle he sees in the industry today.

Thanks also to Dave Bartosik, of DRT Medical-Morris, and Dustin Lindley, of UCRI. Their continued technical help has been an incredible asset, without which I may never have begun this project.

Thanks also to Clay Parker Jones and Bradley Rothenberg for reading early versions of this piece.

First EBM prints

Added on by Spencer Wright.

A few weeks ago I visited Addaero Manufacturing, one of the very few EBM (electron beam melting) service providers in the US. After my recent trials (and successes) with laser powder bed fusion, I wanted to try building parts with EBM. EBM is used extensively by aerospace and medical OEMs, but its penetration into the job shop world is way behind laser. Addaero, whose founders (Rich Merlino and Dave Hill) both worked at Pratt & Whitney before striking out on their own, is located just a few hours from New York City, and they were gracious enough to build two parts for me to evaluate the process.

I'll be writing up a longer post on the unique design considerations that EBM poses, but for now I wanted to share the pictures I took while there:

At this point, the parts Addaero made for me still need post processing before they can be assembled and tested. I'll be working on that over the coming weeks, and will update soon on my progress.

CT Scanning of 3D printed parts

Added on by Spencer Wright.

A few weeks ago I visited CIMP-3D by invitation of its co-director, Dr. Tim Simpson. I was there partly just to visit (I love these kinds of places), but also to see first-hand the role that CT scanning can play in non destructive testing of additively manufactured parts.

CIMP-3D is located at and operated by Penn State University, and serves as part of Penn State's Applied Research Lab - and as an Additive Manufacturing Demonstration Facility for DARPA. In aggregate, they help both government agencies and commercial partners qualify and improve parts made by powder bed fusion and directed energy deposition. In their well-equipped shop, they have two powder bed fusion machines: an EOS M280 (EOS calls their process "DMLS", a term that I continue to get flack for using generically :) and a 3DSystems ProX 200 (3DSystems calls their process, which was developed out of their 2013 acquisition of Phenix Systems, "DMP" - for "direct metal printing). For their work on directed energy deposition, they also have an Optomec LENS MR-7 (a laser based powder deposition machine), and until recently had a Sciaky EBAM (a large scale wire fed electron beam welding machine, which had been sold just prior to my visit).

While I was excited in see their directed energy deposition machines, the real attraction was their GE phoenix v|tome|x m300 CT scanner. This machine is made by GE Measurement & Control division, which is part of GE's Oil & Gas business unit (it should be noted that I've done consulting for both M&C and O&G, though not for the people who make CT scanning equipment). CT scanners are *expensive* (close to $1M, depending on options), and are basically unheard of in private service providers. They can be used to analyze both the as-built form of a part (which will often deviate from the as-designed form significantly), and also any flaws (cracks and voids) which would make it unusable.

Before I visited CIMP-3D, Corey Dickman (an R&D Engineer there) was kind enough to print one of my seatmast toppers, in aluminum, on their EOS M280. It came out well, with only a small defect in the seatmast clamp area. Corey used some pretty clever support structures, tapering them in order to provide a balance between a solid grip on the plate on the one hand, and relatively low material usage on the other:

CT scanning uses a series of 2D X-ray images to reconstruct a 3D part. In CIMP-3D's scanner, the part is placed on a turntable in the middle of the machine. The X-ray projector, on the right side of the machine, shines X-rays through the part onto a sensor on the left side. Solid parts block X-rays, creating shadows on the sensor, and the result is a greyscale image where dark areas correspond with solid mass and light areas correspond with empty space. 

The scan moves pretty slowly. My part was scanned in 3500 slices, or one scan every ~.1 degrees. At this rate (and at a voxel size of 58µm), the total scan time was about an hour. Each scan takes about a second, and between scans you can see the turntable rotating slightly.

Fixturing the part in the machine presents an interesting challenge. You want it to be held securely, but you don't want any other solid things touching it - as they will cast their own shadow in the X-ray images. As a result, parts are often held in place by simply sticking them into a piece of styrofoam - as mine was.

Once the data is captured, it's loaded into a separate workstation to be reconstructed. The amount of data that needs to be processed here is staggering - my scan generated about 25 gigabytes of image data, which reconstructed into a 5.7 gigabyte model.

Once the reconstruction was complete, Griffin Jones (the R&D Engineer responsible for CIMP-3D's CT scanning) was able to do a visual analysis of the as-built part, checking it for voids and flaws. The as-built model can also be overlaid on top of the as-designed model, allowing for deviations to be easily quantified. The model can also be explored layer by layer in any orientation, allowing for a really complete understanding of what the solid part looks like:

A word on resolution: this scan was performed at a voxel size of 58µm, and each voxel is assigned a greyscale value that corresponds with the material's radiographic absorption coefficient at that location. However, any given voxel is subject to some amount of randomness as well; if a voxel has a vastly different value than its neighbors, then the operator needs to determine whether that's a result of a microscopic void, or a result of random variations. 

As a rule of thumb, Griffin assumes a void when he sees three voxels in a row with low grey values. Interestingly, the scan did reveal a few tiny voids in my part. They're mostly near the edges - specifically, the zone right at the boundary of the profile scan (the perimeter of the part's cross-section) and the infill hatching. Since the scan was performed at a voxel size of 58µm, and Griffin was looking for three voxels in a row with low grey values, the voids we detected were about 180µm - just larger than the diameter of a human hair. 

My suspicion - which would need to be verified by destructive testing - is that voids of this size are probably well within the functional requirements of my design. Of course, this particular model is aluminum, and the design is meant for titanium - but I'm looking forward to having a ti model scanned and destructively tested in the future.

For most product development teams, non destructive testing (NDT) is just one part of the process of qualifying a new part. My part, for instance, is being put through physical load testing this week - and I'll use the data I get from that test to improve my design. But for early on in the build planning process, having a tool that allows you to dive inside otherwise obscured areas of your part is incredibly helpful. Especially in the case of complex, topology-optimized parts with organic forms, it can be difficult to impossible to measure a part's deviation from the underlying design. Moreover, there may be regions that it's impossible to inspect without destroying an expensive prototype. My part has just this: the front of the neck section contains a completely hidden hollow zone. And as I move into redesigning for EBM, knowing the areas where powder tends to cake up will be even more helpful.

Thanks so much to CIMP-3D for hosting me!

A successful print

Added on by Spencer Wright.

The other day I got a package from Layerwise. In it was the second titanium seatmast topper of mine that they printed, and this one is ready to ride.

...but actually, this part might not actually be ridden - it's off to Germany to be tested. I'm in the process of writing up a longer report about how the project has gone over the past month or so - expect that soon!

Tested.

Added on by Spencer Wright.

After more than a year and a half of research, modeling, procurement, site visits, redesigns, and batches of failed parts, I've finally got a functional, 3D printed, titanium seatmast topper.

Yesterday morning Clay and I took it for a 20 mile ride, and aside from some cosmetic issues (he *really* needs a ti stem now...) it worked well. We'll road-test it a bunch more over the next few weeks.

I'll go into detail in a long post soon, but the short story is this. This part was built by Layerwise, a Belgian startup that was acquired by 3D Systems last year. While Layerwise has a bunch of IP (software + hardware) that allows them to tune the process parameters, the main difference between this part and my earlier prototypes is the build orientation - and some clever use of temporary structures and supports. This part was also shot peened, which (along with the orientation change) improves the surface finish noticeably.

I'm expecting another copy of this part in the next week or two; it will go to be destructively tested in Germany. It's only slightly different than this one: Layerwise is adding some additional supports in the the seatmast clamp window, which will help it from distorting slightly during the build process.

Once the part is destructively tested, I'll get a better idea of the areas where I can remove material in order to make the part lighter. I've been wanting to redesign the part for a while - partly to reduce the need for support structures, and partly for aesthetics. The most likely path for both of these is to introduce a number of lattices, which will likely be lighter, be easier to build, require less post processing, and be more visually compelling and distinctive. I've got a few thoughts on what this should look like, but will also be working with some software & design companies who have more experience with lattices.

Having a working (albeit imperfect) part in hand is a really validating step. I'm *really* looking forward to more of these - and to critically evaluating their performance.

3D printing titanium and the bin of broken dreams

Added on by Spencer Wright.

Note: What follows here describes my early experiences designing and building metal 3D printed parts. Since this post was published, I've had a number of parts printed successfully; you can read more about the current state of this work in a recent update, here.


Last December, in an industrial park in Cincinnati, I watched as Dave Bartosik set up a build platform on an EOS M280. The part he was printing is one that I began designing a full fourteen months earlier, before I had any idea of the intricacies of metals 3D printing, nor the complexity in bringing an additively manufactured part from design to prototype.

My part, being sintered from titanium 6/4 powder.

My part, being sintered from titanium 6/4 powder.

While consumer 3D printing community has coalesced around openness, the industrial market tends to be opaque. Sure, anyone can design parts with free software. But the process of bringing those designs to life requires hundreds of engineering hours, tens of thousands of dollars worth of software, and millions of dollars worth of tooling — and even then, developing a single working prototype is a long and expensive process.

To many people, 3D printing offers a new manufacturing paradigm. It stands for rapid iteration, customization, and distributed fabrication. But today it’s arduous and costly, and the knowledge base needed to enter the market is concentrated in a select few people.

Here, I’ll share what I’ve learned about 3D printing titanium parts so far — and what other designers should expect out of the process as well.

A primer on 3D printing metal

Within the maker and startup communities, a 3D printer is essentially an extremely accurate, robotic, hot glue gun, capable of making complex plastic parts within a few hours. For some people that means making functional models more quickly, and with less interaction with outside suppliers, than if they were machined or injection molded. For others it means a couple of desk trinkets, and for the really ambitious it provides a peek into a new way of bringing products to market, replete with on-demand, custom-to-order parts, produced just walking distance from where they’re used.

My background is in short-run manufacturing. I spent a few years building custom bicycle frames, and later ran a small prototyping shop, where I designed and tested robotic door assemblies on fast development cycles. When people talk about the need to iterate rapidly, I get it.

For five years, I worked in a shop that could make machined and welded parts on demand. There were always hiccups — prototype development is often more art than science — but in general we could, with a day’s notice and a bit of help from McMaster-Carr, take a 3D model and create a useable part from it. So when I think of 3D printing metal, I keep those kinds of capabilities (which are shared by easily thousands of shops around the US) in mind.

Metals parts are 3D printed in one of three ways:

  1. Binder jetting, in which powdered metal is sprayed with glue to get it to stick together — and later infused with a second metal to make the bond permanent. The most common binder jet printers are made by ExOne. In general, binder jetting is used to create prototypes or parts that require low strength.
  2. Directed energy deposition, in which metal is sprayed or fed at a part — and melted to that part by electric arcs or lasers upon contact. DED machines vary significantly; see DMG Mori and Sciaky for context. Directed energy deposition is even more niche than binder jetting; its application is mostly limited to repairs and very large aerospace parts.
  3. Powder bed fusion, in which a bed of powdered metal is selectively fused (through sintering or melting) by a laser or electric arc. The most common powder bed fusion machine is probably the EOS M280, but Concept Laser, Arcam, and Renishaw (among others) all have their own offerings. Powder bed fusion has a variety of uses in both production and prototyping. It supports a wide range of materials (the most common being titanium, stainless steel, and cobalt-chrome) and with care can be used to create lightweight, strong, and highly customizable parts — just like mine. But the process is far from easy and is definitely expensive.

Though it’s used extensively within the aerospace, medical, and tool & die industries, there are no metal powder bed fusion products on the consumer market today. Which made me think: why not?


The powder bed fusion process

When it comes to machine design, humans aren’t nearly as creative as you might think. The basic model for milling machines has been around for more than a century, and despite the fact that the process has changed significantly during that time, the overall layout of industrial tooling today is much the same as has been since the Civil War. Today’s powder bed fusion machines are no exception: If you squint just right, they look a lot like CNC vertical mills.

Powder bed fusion machines consist of three core parts: The build platform, a material source & recoater, and a source of thermal energy.

During operation, a thin layer of powder is spread by the recoater blade across the build platform. Then the heat source — either a laser or an electron beam — scans across the build platform, melting powder selectively as it goes. Where the metal is melted, it fuses to the metal around it, creating a solid part. After one layer is scanned, the recoater blade spreads another layer of powder and the process repeats.

Printing things this way is slow; a part the size of a juice glass might take ten hours. The recoater takes a few seconds to spread each layer, and although the laser is moving incredibly fast, it takes some time to scan the cross section of a part line by line. Add to that the fact that each layer is about eight ten-thousands of an inch thick, and you see how anything larger than a thimble would take a long time to complete.

Although powder bed fusion can be done by both lasers and electron beams, lasers are far more common. Electron beam melting (EBM) is notoriously difficult to control, and although it has some advantages, EBM parts tend to have very coarse surfaces and require more post processing as a result. EBM also suffers from relatively low market penetration; by my count, there are fewer than five service providers for EBM in the US.

By comparison, laser sintering (which I’ll refer to as DMLS, for direct metal laser sintering, though that is technically an EOS trade name) is almost ubiquitous. I’m aware of about seventy US shops offering metal laser sintering in-house, and even consumer-facing providers like iMaterialse offer DMLS. And although EOS sells far more metal laser sintering machines than any of their competitors, the market is still competitive — and that competition is beneficial to the industry as a whole.

Within the aerospace industry, DMLS has a high adoption rate — at least in an R&D context. In fact, my collaborators and primary tour guides to the industry (Dave Bartosik and Dustin Lindley) both began their careers in additive at Morris Technologies, the aerospace DMLS giant that was acquired by GE Aviation in 2012. Between the two of them there’s about as much experience printing titanium parts as anyone else in the world.

But for all the work that has gone into understanding the properties of additively manufactured parts, the process is still very much in its infancy. It is not in any sense a mature technology, and the result is that each new part you design for DMLS — and, indeed, each new copy of the part that you print — is very much an experiment. Small variations in geometry and orientation can have huge effects on the way that a part prints. The laser’s scanning path is a closely studied subject, but much is not yet understood about it. Even keeping those variables constant, it’s often the case that building the same part on a different machine will produce very different results.

All of which is to say that DMLS is anything but plug-and-play. Even when a design has been optimized specifically for the process, it often takes dozens of tries before a functional part comes out of the printer. And the process of troubleshooting a failed build — even at the most advanced DMLS shops in the world — still involves a lot of trial and error.


Part constraints

In general, parts that benefit from 3D printing tend to have the following traits:

  • Benefit from weight reduction
  • Benefit from customization
  • Complex geometries
  • High inventory costs and/or long lead times

My seatmast topper, in full and half-section views. Note internal cavities in the center neck area.

The part I’m building is a seatmast topper for high end road bicycles. Cyclists want lightweight, custom parts. Custom bicycles are increasingly popular with consumers, and they carry high price tags and long lead times. Broadly, it’s my suspicion that more and more bicycle components will be produced on-demand through 3D printing — if only for the simple fact that within high end cycling, sexy sells.

At about 60 grams, my part is fairly lightweight. It’s also relatively small, and fits easily within nearly every DMLS machine’s build platform. And because of its function (seatmast toppers are used to hold a bicycle saddle onto the frame) its structural requirements are fairly predictable. These factors, plus the fact that seatmast toppers are easy for almost any cyclist to install on their own bike, make it a good candidate for 3D printing.

But that doesn’t mean it’s easy to print. My part consists of two cylinders, oriented 90° apart and joined together by a funneled neck. The part’s wall thicknesses fall between 1mm and 1.75mm — roughly .039"-.068". And it’s critical that these walls not vary much in thickness; if they end up just .010" thinner (for comparison, a sheet of paper is about .004" thick), the part could be unusable.

Harder yet, the inner diameters of both of the cylinders must be accurate and consistent. Again, variations of just .005" can have a big effect here — and if the cylinders end up with oval cross-sections, the part won’t work at all. And the titanium 6/4 that my part will be made of (which is named for its 6% aluminum, 4% vanadium content) is notoriously prone to built-in stresses, meaning that we’ll have to be very careful setting up the build parameters and support structures to prevent the part from turning into a pretzel during the process.

As with almost all 3D printed metal parts, mine will require some degree of post processing; at an absolute minimum, the clamp bolt threading will need to be tapped. But because of the physical tolerances listed above — and the mechanical and aesthetic properties of DMLS parts, which tend to be rough and unpredictable — it’s likely that extensive finishing will be required on both the inner and outer surfaces of the part.

In short, my part’s manufacturing process chain was always going to include some subtractive steps. There are many feature types that 3D printing simply isn’t intended for, and I knew going in that this part would require more than one process as a result. But until we picked a build orientation — and built parts that passed fit & finish tests — we wouldn’t know for sure what our total process chain would look like.


Build orientation

My prototyping partner, DRT Medical— Morris, prints titanium parts on an EOS M280 — the workhorse metal 3D printing machine for American job shops. Based on my research, EOS’s market penetration outmatches each of their competitors by five to one. As of November 2014, EOS only lists eighteen service-ready M280s in the US, but their data is clearly incomplete; I wouldn’t be surprised if the real number was triple that. Moreover, the vast majority of DMLS machines are purchased by OEMs, who use them for internal capacity only, and my suspicion is that the proportion of EOS machines behind closed doors is similarly large.

The M280's build platform measures 250mm (x) by 250mm (y), and is has a maximum build height of 325mm (z), including the build plate. My part is about 70mm long (x), 37mm deep (y), and 90mm tall (z).

bk1033 drawing.jpeg

A common misconception about 3D printing is that the unit cost doesn’t vary much with quantity — that printing one part is just as efficient as printing a thousand. Granted, 3D printing doesn’t require tooling per se, but non-recurring engineering costs are absolutely to be expected. In addition to the time spent determining an optimal build configuration, there are significant changeover costs when a DMLS machine operator goes from printing in, say, titanium to stainless steel — which some low-volume service providers might do on a weekly basis. Moreover, metal 3D printing providers don’t normally print multiple orders in one build. In other words, if I buy a titanium part at the same time as another customers does, they’ll usually run those orders in two separate builds — even if both parts might fit on the build platform at the same time.

The exact math is hard to reverse-engineer, but there are generally four variables that determine the cost of a DMLS part:

  • Finished part mass. There are two subcomponents here: Raw material cost, plus the time it takes the laser to sinter the part. Raw powder costs between $300 and $600 per kilogram. My part weighs about 60 grams, which puts the material cost in the neighborhood of $30. But that 60 grams will take about eight hours to sinter, and the cost of sintering time adds up quickly. As a rough guide, expect to spend on the order of $100–200 per hour for part build time.
  • Support structure mass. DMLS parts require solid support structures to tie them to the build platform, and those structures are made of the same metal powder that the part is. If your part has a lot of overhanging geometry or requires additional support structures for other reasons, you’ll pay for those (and the time it takes to sinter them) as well.
  • Part height. Across all types of manufacturing, capital expenses (tooling, etc.) are paid off over a period of years. In DMLS, part height correlates directly with time spent recoating the platform with new powder, and that in turn equates to a higher tooling cost that the supplier needs to pay off during your build. As a result, designers are incentivized to orient their parts as close to the build platform as possible, to reduce build height and hence reduce recoating time.
  • Number of parts per build. Because the setup and powder recoat time can be shared across multiple parts, buying in batches (when possible) will generally be less expensive than buying one-offs.

Build configuration 1: Part on it side. Ten parts, ~40 hours. Image courtesy DRT Medical — Morris

In my case, those last two factors act directly against each other. If I orient the part on its side, I can fit about 10 parts per platform, with a ~40 hour build time — about 4 hours apiece. But if I orient them vertically (upside-down ends up being more favorable), I can print 24 parts in a ~85 hour build — about 3.5 hours per part.

Build configuration 2: Part upside-down. 24 parts, ~85 hours. Image courtesy DRT Medical — Morris.

At this point, Dustin and I spent some time thinking through the manufacturing process chains for each of these configurations. It’s very likely that we’ll end up needing to machine the inner diameters of both of the part’s cylinders, and we wanted that process to be straightforward and involve as little custom tooling as possible (post processing DMLS parts often requires extensive custom tooling). As far as we could tell, build configuration 2 was going to be slightly easier — mostly because the long ID could be machined while the part was still on the build platform. But the difference was very difficult to quantify, and in the end our build orientation was determined for a much simpler reason: powder availability.

The DMLS powder market is, like most things in this industry, changing rapidly. Powdered metal is expensive to produce, and the particle size, shape, and consistency are critical to finished part characteristics. And while prices are going down rapidly (double digit percentages year-over-year, I’m told), service providers are still stuck with having a large chunk of money tied up in raw powder powder at any given moment. Add in lead times, and the fact that titanium powder isn’t particularly fun to handle, and you can see why job shops would only want to keep as much powder on hand as they absolutely need.

In the end, we ended up building my part on its side simply because DRT was getting towards the end of a batch of powder, and the taller orientation was going to require more than they had on hand. Which is possibly, given the small quantities these parts will probably be produced in, the right orientation anyway — and was a less expensive build to boot.

At this point, our goals were explicit: Determine the minimum amount of post processing necessary to produce a working part.


Stress & build failure

It’s critical to remember that 3D printed parts move as you print them. The same is the case with other manufacturing methods as well (injection molded parts shrink as they cool, machined parts warp as they’re cut, etc.), but DMLS creates extreme thermal gradients, and the net effect is that stresses are built into the part layer by layer.

Build orientation affects stresses in *huge* and unpredictable ways.  See this paper , by Amanda Wu, a researcher at LLNL, for more.

Build orientation affects stresses in *huge* and unpredictable ways. See this paper, by Amanda Wu, a researcher at LLNL, for more.

Predicting built-in stresses is an incredibly difficult task, and is the topic of a lot of basic research. At the moment, the best we can hope is to analyze and understand the stresses that are built into parts once they’re complete; predicting them before they happen is still a long way off.

Despite the fact that stress prediction is very much a dark art, the correlation between laser sintering and welding is not a trivial one; many of the same principles apply to both. But when approaching a new part, it’s almost always the case that the best way to deal with potential problems is through trial and error — and then adding and removing support structures as necessary.

This is a key point: powder bed fusion involves welding your part to the machine while you build it. The build will definitely fail if the part lifts off the build platform (when this happens, the recoater blade strikes the part. In general it doesn’t damage the machine, but I’m told it can be… exciting), so a lot of effort goes into designing clever — and hopefully not too massive — solid and lattice support structures to keep the part where it’s supposed to be.

Even if the build itself doesn’t fail, internal stresses can still render it unusable. This is why most parts are stress relieved (a heat treatment process) after they’re built and before they’re removed from the build plate: doing so allows the crystalline structure to relax, preventing failure later.

Going into the build process, I was warned many times that cylinders oriented parallel to the build plate are notoriously difficult to build. The stress profile around the circumference of the cylinder will tend to vary widely, and the result is that you generally wind up with a big oval. But other than orient the part at a 45° angle to the platform (and risk ovalizing both cylinders), our options on this part were limited. So, we started as simply as possible, and iterated as needed.


Support structures & Iteration

Once the build is complete and the part is wire EDM cut off the plate, support structures are removed manually.

Once the build is complete and the part is wire EDM cut off the plate, support structures are removed manually.

While I was in Cincinnati, I visited MicroTek Finishing — a major player in the metal 3D printing world. While there I spoke with Tim Bell, who related an anecdote about his time at Morris Technologies, the aerospace 3D printing giant that was acquired by GE in 2012. Tim was a product development leader at Morris, and he talked of a large bin that they had in their shop. It was called the Bin of Broken Dreams, and into it went an endless stream of failed parts.

My part has now been printed in six different build configurations. We (and by we I mean Dave Bartosik, whose creativity and enthusiasm for getting the build to work was inspiring) added solid supports in a number of places, chasing built-in stresses around the part with each iteration. The latest prototype, although nonfunctional, is nevertheless a big improvement on the earlier builds — and the process has taught us a lot about the idiosyncrasies of my design.

To begin, Dave let Materialise Magics (the industry standard for support structure generation software) do its thing with no manual intervention. Magics generates mesh support structures, which are scanned every other layer of powder (solid regions of the part are scanned every single layer). As a result, they’re very easy to chip off the part — but don’t have the same strength that solid supports do. As internal stresses proved to be an issue, Dave added solid supports to keep the part undistorted and tied to the build platform.

Build 1

In this build, the part is laid on its side and supported only by mesh supports. The build failed at only 15.6mm in the z-direction, when the recoater jammed on the saddle clamp end of the part, which had lifted from the build platform.

Build 2

Here, the seatpost clamp cylinder is firmly fastened to the build plate. But the stresses just concentrated on the other end of the part, pulling the bolt boss and some of the front edge off of the platform at a height of 22.7mm.

Build 3

Both ends of the part — the saddle clamp and the bolt boss — are firmly anchored to the build platform. But this created a complex bending moment, pulling the center of the part upwards; the build failed at 22mm.

Build 4

Here we’ve got solid supports on both the saddle clamp cylinder and the bolt boss, and added an additional solid rib to the middle of the part, tying it down there. This is the first build that completed; all of the others had failed midway through. We’re clearly getting closer, but the bottom of the part has distorted, pulling in and looking like a big “D”.

Build 5

To prevent the bottom of the part from distorting like in Build 4, we added a second solid rib. It helped, but only below the centerline of the cylinder; above that, the wall still pulled in.

Build 6

Build 6 finally produced a part that’s generally round and complete. This was achieved by extending the lower rib up the side of the part, giving external support to the entire bottom edge of the seatmast clamp cylinder. But although the top and bottom of the seatmast clamp are both basically round, the internal stresses still needed to go somewhere — and ended up bulging out the middle of the tube instead.

Throughout each of these builds, three things have remained consistent. First, the surface finish on the exterior of the part leaves much to be desired; it will definitely need to be finished in a separate step. Second, the surfaces that needed to be EDM cut from their solid supports (the saddle clamp and the bolt boss) are irregular, and will need to be smoothed into the rest of the part. Third, the internal diameters will almost definitely need to be post-processed by machining or EDM — even the saddle clamp, which overall had passable surface finish, was undersized by .020" — about four times the desired variance.

The net effect is that after six build iterations — each of which took almost two full days to set up, build, stress-relieve, and cut off of the build plate — we still don’t have a functional prototype to test.


Takeaways

What to take away from this? Well, prototyping is hard — but everyone knows that. My primary observations have more to do with the state of the industrial marketplace, and the maturity of metal 3D printing processes, than with the fact that we’ve now put six parts into our own bin of broken dreams.

File processing

As with consumer 3D printing, industrial 3D printers work exclusively from STL files. This produces a total break in the design-to-manufacture process. When I export an STL to send to a manufacturer, all of the underlying feature data is lost; all that’s left is a shape. This is drastically different from the conventional manufacturing world, where parts are regularly built directly from underlying design files.

Tolerances

For the vast majority of machined parts, any single dimension is expected to be accurate to within .005", regardless of size; in other words, a quarter-inch hole should be between .245" and .255", and a one-inch hole should be between .995" and 1.005". For a relatively small cost, designers can specify even tighter tolerances, and the means of achieving them are predictable and not overly complicated. But with additive, tolerances accumulate across the part at a rate .005" for every inch of distance. That’s fine if you’re building a one-inch part (whose dimensions will be between .995" and 1.005"), but larger parts can be problematic; a ten-inch part will be between 9.950" and 10.050" — a decidedly generous tolerance. Moreover, these tolerances don’t always stick; many of our early prototypes didn’t come close to meeting them. And when a part prints out of tolerance, the way to fix the problem is essentially to fiddle with the underlying design and then build it again.

Intellectual Property

Across the metal 3D printing industry, a stream of contract manufacturers told me the same thing. DMLS build processing is hard, they say. And the only way to maintain a competitive edge is to invest countless time and money into R&D — and then guard institutional knowledge vigilantly. On many occasions this is referred to as intellectual property, but the truth is that it’s closer to expertise; what’s being developed is craftsmanship, not patentable tools or methods. But whatever the name, the effect to designers is stifling. Regardless of manufacturing method, the design-to-manufacture process benefits from transparency; if a build fails, then I as a designer want to know the reason — and adjust my underlying design accordingly. Until the additive supply chain opens up to sharing its experience in the design-to-manufacture process, new DMLS products will be few and far between.

Undistributed Manufacturing

Today, 3D printing metal parts via a distributed supply chain is a myth, full stop. And while I’m as excited about that vision as the next guy, distributed manufacturing will continue to be a pipe dream for the foreseeable future. A distributed manufacturing ecosystem can only exist once there’s a robust network of suppliers capable of making parts repeatably. And while it’s my sincere feeling that the most hardworking, intelligent, and visionary people in manufacturing today are working in 3D printing, there simply isn’t currently a rich network of DMLS suppliers. For instance, the closest DMLS-equipped shop to New York City is a 200+ mile drive away. Meanwhile, MFG.com lists 68 machine shops within a 150-mile radius. If distributed 3D printing is to become a reality, the install base must increase by orders of magnitude — and the reliability and repeatability of the processes must improve dramatically as well.

In-Process Monitoring

In conventional manufacturing, parts are checked between operations to ensure that critical dimensions will be met. But the current generation of industrial 3D printers have little in the way of in-process monitoring, with the result that distortion isn’t detected until the build fails altogether. Although there are hints that this may be changing (B6 Sigma has announced some ambitious plans recently, and a lot of primary research is being done on the subject), the fact remains that until we’re measuring and analyzing the factors (thermal gradient, sound, vibration, etc.) that indicate build failure before it happens, trial-and-error will be the only way prototypes are developed.

The Process Chain

3D printing is very, very good for certain things. But it is not a one-stop process. For now and the foreseeable future, additive manufacturing will be a poor method of creating a number of important mechanical features, including many aspects of fastening and articulation. In addition, the surface quality of 3D printed parts will be unacceptable for anything requiring tailored aerodynamic features, and will be similarly poor for products whose fit and finish are of high value for aesthetic reasons. This is not to say that those aspects won’t improve; they will. But while I expect additive manufacturing to be an important part of the way parts are produced in the future, it’ll be a long time before it’s used to produce a wide range of products. And for those products which are well suited for 3D printing, their total manufacturing process chain will include subtractive tools (machining, honing, polishing, etc.) for the foreseeable future.


Next steps

My part has come a long way. Just having a physical prototype in hand makes a huge difference in understanding its benefits and drawbacks, and I continue to believe that with continued research and prototype development, I will find a way to make it commercially viable and attractive to high end cyclists.

But there’s much work to do. Moving forward, I see three primary directions to explore:

Keep the current build orientation, and continue to iterate on support structures as necessary.

At this point, it’s clear that we need to rethink the way we’ve been mitigating internal stresses. The external ribs are working somewhat, but even if we can add enough of them to make the build work, they leave ugly marks on the outside surface which require additional post-processing. Instead, I plan to experiment with reinforcing the inner diameter of the seatmast clamp cylinder. One thought is to create an internal lattice (like those that Frustum’s software creates), which would provide rigidity during the build and then be removed via machining afterwards.

Change the build orientation

Turning the part so that it’s upside-down on the build platform — with the seatmast clamp on the top — will offer significant advantages. The saddle clamp already has a thicker wall than the seatmast clamp, and is likely to resist distortion more easily. And with the seatmast clamp oriented in the z-axis, it’ll be in much less danger of distortion.

Try EBM

The electron beam melting process preheats the entire build platform to just under the melting point of titanium, and so generates much lower thermal gradients — and as a result less internal stress — than DMLS. EBM also generally requires fewer support structures, which is helpful for part cleanup. However, the surface quality and minimum feature size of EBM is significantly worse than DMLS, so EBM would probably require a longer overall process chain, with more material removal than DMLS would.

Regardless, I’ll be continuing this work over the coming months. These technologies are changing rapidly, and any ambitious product designer would be wise to pay close attention to their development. And only by experimenting with actual parts can anyone hope to keep up.

I believe that functional, engineered consumer products made by additive manufacturing are an inevitability. But as a product manager today the viability of metal 3D printing is totally opaque, and that will only change by careful study of the efficiencies (and inefficiencies) of the additive manufacturing toolchain.

Join me in working to make that a reality.


Thanks

First, thanks to Dustin Lindley (of UCRI) and Dave Bartosik (of DRT—Morris), without whom all the cool stuff described above would have never happened. Thanks also to Greg Morris (who originally connected me with Dustin, Dave, and Chuck Hansford at DRT), to Clay Jones and Jordan Husney for their creative inspiration and infectious enthusiasm throughout the process, and to Clay Jones and Mike DiGiulio for reading early drafts.

Lastly, thanks to Undercurrent, which is providing critical funding for this project — and which I am proud to call home.

This article originally appeared in three parts on 3D Printing Industry.

Topper drawings

Added on by Spencer Wright.

While I've been writing a much longer post about DMLS, I made a couple of additional drawings of my topper. The longer post is coming soon, but I'll drop these here for now.

DMLS in process

Added on by Spencer Wright.

This is my seatmast topper being printed from titanium 6/4 powder on an EOS M280 at DRT Medical - Morris.

This video clip is about 1-2mm into the build, so this is the very close to the bottom edge of the part (which, to be clear, is lying on its side). A lot of what's being printed during this clip is support structures & ribs that will help hold the part to the build plate, but you can clearly see the general shape of the part already.

This is 6th iteration on this build. In other words, we printed 5 parts before this one, and each of them failed for one reason or another. We've (and by we I mean mostly Dave Bartosik, the head Additive Technologist at DRT, with me trying to look over his shoulder) made a bunch of modifications to the build to help the part come out within spec, and I'm hoping that today we end up with something that has consistent inner diameters and is more or less useable.

Anyway, what you're seeing here is a 400 watt ytterbium fiber laser in the process of melting 30 micron layers of titanium powder. The recoating arm spreads a thin layer of powder, and then the laser scans a cross sectional slice of the part, and then the process repeats. 

When the video goes into slow motion, notice the smoke that's coming off of the weld pool. The machine has a laminar flow of argon gas that's blown across the build platform (from top to bottom in this video) that pulls the soot away and filters it outside the machine.

More soon.

Initial topper builds

Added on by Spencer Wright.

This.

Today I stopped by DRT Medical - Morris to check out the first builds of my titanium DMLS seatmast topper. I'll be writing up a longer report in the next few days, but the short story is that it's moving forward - but that the process of printing a part is *not* at all straightforward, even when you have some of the most experienced people in the business working with you.

More soon. And - fucking cool, right? This thing started out as titanium powder, and was literally made by laser sintering. Nuts.

Topper progress

Added on by Spencer Wright.

This is still happening.

It's been slow, but my metal powder bed fusion (aka DMLS, LaserCUSING, selective laser sintering, etc.) seatmast topper is moving forward. With any luck, I'll have a part in production in a week's time.

I've made some small changes to the design. The biggest thing is the hole in the back of the part, which is meant to reduce mass. I also modeled the threads in the clamp, which will help my manufacturer print the threads.


Choosing a job shop for this has been interesting. Since my post on DMLS pricing, I've had a bit of interest on my project. My hope has been that I'd be seen as more of a partner than a customer,  but the extent of that remains to be seen. Selling a partnership is something I'm green at, and companies that deal mostly with corporate and institutional buyers don't necessarily think of investing time into a project that has an indirect upside.

Nonetheless, I think there's something to it. This project is partly product, partly experiment in advanced logistics. The information I'm learning on the subject is free for anyone to see, but the partners that I'll end up working with will develop unique experience working on a thin-wall, consumer facing part. There aren't a ton of people working on that kind of thing, but I expect that'll change in the near future. My hope would be that my partners would agree with me, and would see this project as an opportunity to develop additional capabilities at a relatively low expense. 

Nevertheless, I'm determined. And I'm looking forward to having a piece of laser sintered titanium in my hands, too :)

What It Would Be

Added on by Spencer Wright.

I've spent much of the past few weeks thinking about how to proceed on my DMLS seatmast topper project. Since I posted my findings on DMLS pricing in mid January, it has became the most viewed thing I've written here. It's also become (aside from my home and "feed" pages) the most popular landing page on my site, due to both high search term ranking and the fact that it's been shared around the 3D printing community extensively. 

There are a few distinct goals moving forward, which represent partially overlapping ideas of what I'm working on.

  1. Develop a body of knowledge and understanding which encompasses the advanced manufacturing industry and the supply chain logistics required to sustain businesses selling engineered 3D printed consumer products. 
  2. Develop deep experience operating the toolchain required to design said parts for manufacturing. Includes parametric solid/NURBS modeling, organic T-spline modeling, topology optimization modeling, and DFM/CAM software.
  3. Design said parts, and have them manufactured & tested. Use the resulting data to develop an understanding of the mechanical properties and DFM guidelines for parts made via DMLS and other advanced processes.
  4. Develop & launch distribution & sales operations (likely e-tail), and sell physical product there. 
  5. Do some/all of the above with bicycle parts as a specific focus.

Many of the points here are not mutually exclusive; point 5 certainly entails many/all of the previous items. But my path forward depends crucially on whether I approach this as primarily a research project - or as a business.

One possible build orientation, courtesy C&A Tool

On one hand, I'd like to move forward in a rapid, directed manner. Doing so will require resources, however, and may inevitably constitute a full-time job. In order to fund such an effort, I'd need to show a near-term market fit - which will require me to approach this specifically as a business.

On the other hand, I'm aware that there are benefits to approaching this primarily as research. I ultimately want to learn; building a business is just one of many ways of doing so. Moreover, there are any number of businesses which address parts like mine, and selling bicycle products doesn't apply to all of them. It's possible that advanced supply chain logistics is a better fit for my knowledge and skillset. Focusing on the bike market might not be the best way to approach such a goal.

Regardless, it's likely that I refine and then purchase a seatmast topper in the coming weeks. This will require a small investment on my part (a few thousand dollars), plus about a week's work. I'll learn a few things about the process, and going through a build will give me an opportunity to cement relationships with suppliers and processing & testing partners.

Assuming the test part is functional, the next step would probably be to feel around the market a bit. This topper will be on the expensive side, and though I'm confident it'll sell, branding it will be delicate. Finding an audience, and defining the product in a way that they can relate to, will be an interesting exercise. 

With a bit of luck, I'll have a working assembly + a landing page with preorders by Memorial Day. Stay tuned.

DMLS Pricing

Added on by Spencer Wright.

Note: Updates on this project can be found here; or sign up for updates by email.

Over the past few weeks, I've collected a handful of quotes for the seatmast topper. All are for DMLS laser sintered titanium. I've had direct contact with 12 potential suppliers and received 9 quotes back. They range from $987 to $2377 for one finished part.

First: Although I posted the part on MFG - and they distributed it to 110 suppliers - I received *no* responses there. In fact, only 8 of those suppliers ever even saw the RFQ. I also posted the file to Elihuu, and later received a personal email from the founder. That RFQ hasn't been live for quite as long; I'm hopeful it develops into something. I also spoke with Jonathan Placa and ProtoExchange, who was able to source me a few competitive quotes.

Separately, I've tracked down DMLS leads around the internet & through personal connections. Both of these channels have been as fruitful as they usually are. My crash course in the economics and availability of DMLS parts has been quite fun, and want to share a few of my findings here. Some of these are obvious, but worth noting anyway.

The number of DMLS suppliers is small; even fewer are printing in titanium.

This became evident when the second shop I sent my files to replied that they had already seen them. Presumably this is because the first shop I contacted had some relationship with them - shared production, engineering, or some supplier/buyer relationship. The deeper I got, the more evident the small size of the industry became - as I asked more experts for the names of qualified job shops, I was often referred back to people I'd already spoken to. This kind of pattern is common in small industries - I had similar experiences when I was looking for inflatable pneumatic seals - but when it comes to larger/more mature industries (e.g. CNC, stamping, etc) there are generally more shops than one person could keep track of.

There are only a handful of machines that print DMLS titanium. 

The firms who were able to produce my part had one of three machines: EOS M280; Concept Laser M2; Renishaw AM250. A few other firms had an EOS M270, which prints only mar-aging steel (though it can be reconfigured to print a special blend of ti with different mechanical properties). All of these has a build volume roughly the size of a 10" cube.

The next generation of machines will be bigger and faster.

In particular, the EOS M400 is closer to a 12" cube and might reduce build times by as much as half (though one supplier seemed suspicious of that claim). It will be released Summer 2014, though, and likely won't be in wide use for a while after that.

Note: As multi-laser machines become available (the M400-4, a 4-laser machine, is scheduled for 2015), one would expect that cost would come down even more - especially as, I hope, more and more parts are being designed for this process.

Speed & build volume aren't everything.

The one supplier that suggested they'd be getting a M400 noted that even if it *does* reduce build times by half, that would still only reduce the cost of the build by something like 20%. Factors like handling time (this is *not* a hands-off process), material mass (titanium powder isn't cheap), and machine cost & maintenance will keep prices high for a while. 

Nobody seems to be making consumer products this way.

Most of these firms' clients are aerospace, manufacturing & biomedical companies. They're often buying up the whole build platform on a machine - running something like $30K - to print a huge part that would otherwise be made up of numerous (and expensive) smaller parts. Mold & die work is a decent part of the industry, where 3D printed internal cooling chambers on complex molds can decrease cycle time for a molded consumer product significantly. 

Excess Capacity isn't built into the current marketplace.

In traditional manufacturing, a significant portion of the cost of the part is fixturing and setup/teardown between different parts. Some advanced manufacturers (think Shapeways & i.materialise) are able to capitalize on 3D printing's lack of tooling, but most of the job shops I talked to aren't operating that way. Instead, they're performing builds per-order. In other words, they don't bundle multiple orders on one build plate - if I order one of my topper, they run a titanium build just for my part. That's hugely inefficient for small parts like mine, but because of the low volume of orders there's not much these shops can do about it. In a few cases, a supplier mentioned that they could set up standing orders for a part, and piggyback those orders into other clients' builds... But that sounded like a bit of a longshot - this particular supplier told me that it's often months between titanium builds.

The main way to decrease price is still to increase quantity.

Because most DMLS shops aren't bundling orders from multiple clients, each client is essentially paying for a bunch of unused build volume. The unused powder is recycled, of course, and the laser isn't running over unsintered parts of the build plate, but the setup/teardown time and material deposition runtime are still allocated to  your job. As a result, a part's price will decrease significantly - perhaps as much as half - by ordering in larger quantities (one supplier quoted me $569 apiece if I ordered in quantities of three, and $491 if I filled his plate with 6 parts).

This is a bit dismaying for single-piece-flow geeks, though I'm sure at some point in the near future it'll start to change. Shapeways, for instance, doesn't operate this way at all, and I expect that as job shops see more orders for DMLS parts, they'll largely follow suit.

The net effect is that the promise of advanced manufacturing - where you don't worry about tooling/setup/teardown costs, and small batch, JIT parts delivery becomes a reality - is still a bit off. 

However.

Consumer-ready DMLS parts are a good bet for a few reasons. First, the quotes I received aren't - if you just squint at them a little bit - all that bad. Sure, three parts at $600 apiece doesn't give me much margin if I'm selling them to consumers for $400, but that was never my intent. Plenty of high end seatposts retail for $300, and this part offers a few distinct features (lightweight; customizable geometry; "cool as shit;" etc.) that those can't.  And if the improved build times of this year's machines result in a price drop of 15%, and if the next two generations of machines - ones built with multiple high-power lasers - cut build times in half one or two times over... Well, all of the sudden I'm looking at a pretty affordable part - at least for high end customers, who tend to exhibit price-elastic spending habits. 

It's also the case that my topper's design doesn't utilize the technology as efficiently as possible. It's still largely a tube-to-tube structure, which just isn't the best use of additive manufacturing. In order to improve its strength:weight ratio and decrease cost, I'll be exploring lattice structures in the next week or two. My hope is that by working with Within Labs, I should be able to reduce cost by an additional 30%.

I've also been working on a few smaller parts. Because of the way most DMLS suppliers are operating, having a variety of differently sized items to fit on a build plate could increase the output of a build significantly without much effect on cost. 

A note on i.materialise.

Honestly, I was shocked at the price they quoted me ($617 per part). It was less than 2/3 the cost of the next lowest quote, which led me to suspect that their process was different in some way. I contacted their support team, who didn't go into specifics but did reply that their "machines are the same as for industrial projects but the approach and handling of the orders in different, which results in a different quality."

Moving Forward.

At the moment, I've got a basic proof of concept (SLA model) and pricing that puts me about 150% over budget. I'll be visiting at least one DMLS shop later this month, and will also be making big changes to the current design. I'd also like to explore the possibility of integrating additional components - saddle clamp parts or (my dream) a lightweight saddle frame - into the topper itself. The more parts I'm able to reduce with my design, the more I expect to close the price gap between DMLS and traditional manufacturing methods. 

Expect updates.


Thanks to the following people for helping me get this far (in no particular order): Kane Hsieh, Jordan Husney, Clay Jones, Dorian Ferlauto, Scott Miller, Jen McCabe, Shane Collins, Robert Hassold, Duann Scott, Greg Irwin, Jonathan Placa, Siavash Mahdavi, Kaveh Mahdavi.