Designed to be Printed, a 3D Printing Panel With Digital Engineering, Xometry, Velo3D, and Ultimaker

Kenneth Wong: Hello everybody. Good morning, good evening, good afternoon, depending on whatever time zone that you happen to be joining us in. Welcome. Glad to have you with us for another episode of DE Hot Seat Webcast. Today, we are going to be talking about 3D Printing. Specifically, this episode is dedicated to Designed to be Printed, so we will be talking a lot about Design for Additive Manufacturing or a skill that is known as DfAM. This episode is made possible by Velo3D, Ultimaker, and Xometry, so we'd like to thank all the three sponsors. So to get you the best possible experience before we start, let me give you some tips. Okay. I want to remind everybody that if you have questions for the panelists or questions about the topic, you can submit them into the Q&A box at any time. We'll address them in the Q&A session after our discussion.

 Kenneth Wong: We'll take as many questions as time permits, and for the ones that we cannot get to on the air so to speak, I'm pretty sure that we can work with the panelists to make sure that you get a written response. You can also download a copy of today's slides for reference, and also check out the resources list for resources related to today's topic on the left side of your slide window. If you are experiencing any technical problems, you can visit the webcast help guide by clicking on the help question mark icon at the bottom of the console. If slides aren't advancing, you might want to refresh your browser or try pressing F5 key to do that, because that usually solves a lot of the problems. All right, let's talk about today's topic. A lot of the people are exploring additive manufacturing, no longer for prototyping, but for mass production. 

Kenneth Wong: And that's where the term DfAM or Design for Additive Manufacturing is becoming very relevant, and it's gaining ground. It means to design part with topology that are specifically best suited for AM, for 3D Printing. Today, with the help of the panelists we'll explore the best practices in DfAM and the rules for programming, the rules that can be programmed and coded into the hardware, and the rules and the programs that cannot be coded, so you have to know as a designer. So let's start with that. For a start, I like to call our first presenter, Greg, from Xometry. Greg, welcome. Do you mind giving us your opening speech for three minutes please? 

Greg Paulsen: Yeah, absolutely. And Kenneth, this is just a fantastic panel and Design for Additive Manufacturing is something that I've been living and breathing for the past 14 years. Again, my name's Greg Paulsen. I am the director of application engineering at Xometry. And Xometry's pretty unique in this area because we're the largest digital manufacturing marketplace out there. And what that really means for people who are designers or engineers, or people exploring new technologies is we make sourcing parts easy. And we actually are kind of an Amazon for manufacturing and almost an Uber for manufacturers across different technologies, including eight Additive Manufacturing technologies. I'm going to take my three minutes and just show you what that looks like. I think that's the easiest way to show versus tell on here. And let me see if this should be pushed to the audience here. And I just have a quick web demo showing what Xometry is all about as far as like, I have a design, let's get it made. 

Greg Paulsen: In this case, if you go to Xometry site or you could actually do this with our CAD add-ins as well, it all starts with the 3D model. Once you upload, you're able to configure between different processes. So we do have additive, we also have subtractive options, molding, casting as well, but in this case I'm just going cross additive and very quickly look at things like pricing, look at things like design feedback, and even in different configurations. So I believe this one, and we started on MultiJet Fusion. And just as quickly, now I'm on vapor smoothed dyed blue selective laser sintering, kind of catching up here. But let's say I'm looking for a show piece model or something I could just as quickly go and look at pricing lead times for steel [inaudible 00:04:39]. 

Greg Paulsen: Even give it a quick clear, so it has more of like an ice cube clarity to it versus just a matte finish. See how prices changed dynamically over time. And it's a resource because usually what I'm doing is trying to source between this. Oh yeah, this is printing to metal. So let's print this in aluminum, just as easily can click and see pricing, and you'll also see that price go up a little bit, because one does not simply print in metal there. And ultimately, for you as a tool for designers, we're there to back it up on our side too. So everything that you're looking at, everything that we offer, we have design guides. We have feedback on our site. 

Greg Paulsen: We have application engineers like myself and our team to help customers choose their way. So that was a very quick run on, on the demo, but it's there to just show what you can do there. And the last thing to go before we go to the next panelists is we're starting to meet customers where they are, so in that design experience. So these questions that are answered today, as you're working on your designs, you can actually see how that prices, what does that radii do to my project? What does this lattice structure do? Get pricing feedback, and we run on Inventor Fusion 360 as well as SolidWorks platforms. 

Kenneth Wong: Really good. Thank you very much, Greg. And also the plugins are really good because a lot of the time when people are designing, they're not committed to buying or ordering yet. So they want to be just able to play around with it inside the CAD program, so the plugins are wonderful. All right. Let's call and meet the next presenter. The next presenter is Matt. Matt is from Velo3D. Matt, welcome. Would you also give us your three minutes opening remarks please? 

Matt Karesh: Hey Kenneth, thank you, and thanks for having us. We here at Velo are of upmost interest with this topic. It's something we actually work to get rid of. So the topic today, Designed to be Printed, we actually look at that internally as we want to print what your design is, and we don't want you to compromise your design for the printing technology. So everything we do is to achieve the goal of solving your engineering challenges and printing the actual Geometry that you're interested in creating to execute the function of your application. So what we do and where we came from, founded by Benny Buller back in 2014, was really a mission to democratize Additive Manufacturing. Seeing all of the challenges that existed specifically in metal, Additive Manufacturing, and the shortcomings that existed with all the other systems, and coming up with a way to make it easy for users, engineers, designers, to use the technology to deliver new high end, high value parts that couldn't do things or solve problems that couldn't be addressed with any other types of technology. 

Matt Karesh: We went public towards the middle end of last year and have been growing fairly rapidly ever since then. The printers that we offer, and the way to think about our solution, the printer is merely just the tool that delivers the hardware that our solution enables you to create. But we have a family of printers that we refer to as a Sapphire and a couple different flavors of those really just changing the fundamental size of which you can build. So a 315 mm diameter build plate by 400 mm tall is our base version. We have a full one meter high version of that. And then just as of last year, we shipped our first Sapphire XC. So it is a eight laser, 600 mm diameter by 550 mm tall build volume. These are all running one kilowatt lasers, which we use the full one kilowatt to do our bulk process, and have all kinds of metrology built into these systems to make sure that it is a true production ready solution. 

Matt Karesh: The types of parts we print, usually going after high value parts that are moving fluids, exchanging heat, that are performing complex functions that are generally very hard or impossible to manufacture any other way, including the conventional metal AM systems that struggle with angles below 45 degrees and certain features that are very delicate and fragile. Some of the customers that we tend to work with today, a lot in the space, aerospace, oil and gas energy sectors, but very rapidly expanding beyond these industries as well as we roll out our new products and improve our technology. And then the materials that we've developed today, so something to understand about how we do materials is we develop all of the processes internally. 

Matt Karesh: And when you purchase one of our systems or use one of our systems, everyone is running to the same recipe. And it is a very well refined and completed solution for all of these materials that has material property data behind it as well. So when you go and you use the Velo system, let's say for Inconel, you know that no matter what you're printing on, you're going to get the same product back out because they're all running on a common parameter set and using the same calibration, same process monitoring to ensure that these systems are indeed producing the same components, regardless of location or machine serial number. And that's a quick minute for Velo. Thank you. Kenneth Wong: Thank you. Thank you so much for the rundown, Matt. So let me welcome Dylan. Dylan is our third presenter, and Dylan represents Ultimaker. Dylan, welcome. Your opening speech, please? Dylan George: All right, thanks. So yeah, as Kenneth said, my name is Dylan George. I'm an application engineering manager for Ultimaker here in the America's region. And starting off with talking about how we got started and lets us find how we got to where we are today. We were founded in 2011, Ultimaker. And even by 2012, just a year later, we came out with our first software pack of Ultimaker Cura. It's when Ultimaker Cura was born. 2013, our Ultimaker 2 line was announced and with great success that we moved from our headquarters in the Netherlands. We opened up a North America headquarters in the US. By 2015, we had 1 million downloads for Ultimaker Cura, which was a huge milestone for the company. And just a short while after in 2016, brought out our Ultimaker 3 line of printers. 

Dylan George: This was a really big event, because this is the first time we offered dual material printing. So you could choose two different colors or two different types of materials, even water soluble material, as an example, and then your build material on top of that. And then recently we've had our S line of printers. This is really our position that we've moved into the professional user space and offered a professional level desktop 3D printer. And this has been a really successful product for us. It's opened up a lot of different application avenues for us, and been able offer a very wide variety of materials on top of that. In 2020, we actually hit 2 million slices per week with old Ultimaker Cura, which is pretty incredible if you think about how just not too long ago, where we came from. And then in 2021, we celebrated our 10 year anniversary, first of many more to come, so we're really excited about that. 

Dylan George: And then if you take a look at today, where we are today, just last week, we have four and a half million slices last week alone. And we've seen this hit consistent numbers from here on out. From that we have a half of a million active users. So not only do we have a large number of slices that use our software each and every week, but our user base is strong and growing every day. So we're really excited about that. Ultimaker Cura keeps getting better and better, and we have a lot of really exciting things to come this year. 

Dylan George: The digital factory is one of our newest products and something that would allow users to collaborate all across the globe, and work on projects together. But not just software and hardware, Ultimaker offers an entire platform to really work from. The goal is really to make everything easy and accessible for all of our users. So we have Ultimaker Cura as our software line, digital factory for a cloud solution as I said previously, along with our S line of 3D printers. We have our own materials that customers can choose from with preconfigured profiles and NFC chips on every school to make everything from even just simply loading the material, the printer will auto recognize that material. 

Dylan George: But not only that, we also have an entire material marketplace where the largest material manufacturers come together to provide these preconfigured print profiles for our machines. Again, making everything accessible and easy for the end customer. But on top of that, we also have an academy, so we really believed in sharing knowledge and sharing progress with our community. And so we have an entire academy course set where you can learn all that you need to know about FFF Design for Additive Manufacturing and even operating the printers themselves. And then last, but certainly not least, we have dedicated support team for any issues that might arise. We want to make sure to take care of you. 

Dylan George: And then my closing slide for the opening here is just, I wanted to showcase a fun slide that really, I think shows the full gamut of where we exist. We get to work with a lot of really great customers, especially on the applications team. And we service customers, anything from Coca-Cola to Tesla, so we really cover the full gamut. And that's

Kenneth Wong: Very

Dylan George: it for me [inaudible 00:15:39]. 

Kenneth Wong: Thank you so much all three of you for giving us the overall picture of the topic and why you are interested in this topic. So let's start our discussion here. First of course, we are talking about Design for Additive manufacturing or short DfAM. I'm going to start with you, Dylan, is it an important thing? Is it something that you need to explain to your customers? Or what would happen if somebody doesn't have DfAM and they start working on it? General thoughts first. 

Dylan George: Yeah. General thoughts. So for DfAM, it is incredibly important, it opens up a lot more doors, but it's not required. There's a lot of things that are capable of being manufactured with 3D Printing that are already currently produce with traditional manufacturing. As an example, something that is injection molded or something that has been a machine with CNC capabilities. If a customer is looking to replace tooling, jigs or fixtures with something that is traditionally manufactured, then that's generally speaking very easy to reproduce with 3D Printing. 

Kenneth Wong: Very good. So I'm going to move on to Matt. Now Matt, is a lack of DfAM skill or expertise a barrier to entry? Just because you don't have that in your in-house staff member, would it prevent you from designing something that is 3D printable? 

Matt Karesh: Yeah, I think from Velo's perspective, in general, in the metal AM industry DfAM is required and not having that scale does make it a challenge. But again, our mission is to eventually get to a place where DfAM doesn't exist. So having the ability to print very low angle surfaces without requiring support structures and create geometries that you desire, things like [inaudible 00:17:46] pillars, where most systems are putting supports inside the flow paths, where at the end of the day, that's just material that needs to be removed, and in some cases is physically impossible to remove. So having the ability to create these components without having to use a DfAM or compromise on your design is where we see the industry going. And basically getting to a point where DfAM doesn't exist and isn't required by engineers and designers. They can use the tool to create the Geometry they want. 

Kenneth Wong: Great. Now, if somebody submits something that they want printed, and it's not exactly designed to be optimal for 3D Printing, how does your automated submission system deal with that? 

Greg Paulsen: Yeah. And this is actually what's pretty cool, because we are a fairly agnostic manufacturing system. So we're looking at all different technologies, and we're putting AI and machine learning to work on these. So one of the things, that quick demo I was showing, I really didn't highlight this purple box that shows up sometimes when you upload your file, and it says recommended process. Now, it's not necessarily a DfAM thing, but what it's trying to do is saying, parts like this that are designed like this in this type of marketplace are typically purchased with this process technology. So if you have something that is big, bulky and has counter boards and stuff, it's probably a CNC machine part. But if you have something that has much more organic features, it'll probably actually tune towards a powder bed fusion process, for example, like selective laser sintering and MultiJet Fusion, and just help tune you there. Greg Paulsen: It's loaded question, because sometimes when a customer comes in, they may not know what they're choosing, but at least we try to help guide them saying, "A lot of times, CAD like this is purchased in this process." You can change it if you want. Just touching based on what Matt was saying though, about metal because I had a little example here. This is the classic GE engine bracket, and we were having a pre discussion where sometimes the things on my desk are things that failed. And the stuff with metal printing is you do need support structures and that support structure is the metal itself. So instead of me getting my file or sandy down, or having something soluble, although there's some cool stuff happening right now, you have to remove that. In this case, this part I'm pretty sure was actually just a file error, but it had a plane going across the bottom of it. 

Greg Paulsen: And it's hard to see on this camera, but it's just super rough. It is just incredibly rough, and it would not be desirable because when you're thinking about how you're growing parts, you're not just magically having a part appear here, unless you have something underneath that support structure or in some cases like a powder bed. And when there's nothing there, the metal just wants to do anything, but stay still. So yeah, metal is very, very sensitive, but some processes like laser sintering or Multijet Fusion, where you don't have support constraints, you could just go nuts. You could go nuts with design, and I call it very forgiving processes, because you could do some really unique things like that. 

Kenneth Wong: Wonderful. It's great to see some of the example objects. Let's move on to actually the second part, which is, let's see if we can share some practical tips about DfAM and help people understand this. I'm going to start with you, Dylan. Like you just pointed out, of course, if a part is originally designed for CNC cutting or molding, you can 3D print it if you want to. But if it is a simple geometry, maybe it really doesn't make sense to 3D print. So what are the kind of geometry that is ideally suited for 3D Printing would you say? 

Dylan George: Well, there's usually an overall theme, generally speaking, let's say with additive manufacturing in that there's a big focus on support structures overhangs. So as new users to additive manufacturing, I think it would behoove them to understand those capabilities and those limitations in the same way that fundamentally with any CNC or milling operations, you need to clamp down your feed stock in order to remove that material. You need to be able to support the structure that you are printing during that process. But there's examples like this one here that I'll show of a robotic gripper, where in this case, the traditional material was chosen as aluminum, very inexpensive, it's lightweight, it's stiff, it's strong, and it's very easy to machine. But in this particular case, they didn't necessarily need to use aluminum to be able to print the part. 

Dylan George: So this particular customer very cleverly used a glass filled nylon to make sure that it's stiff, and it's still a relatively strong piece of equipment. But then they used steel inserted rods or dows into the wear surface. So they have still that metal to metal contact in order to extend the life of the tool, but they're able to print still with all structure. So there's certain adaptations that you can make, especially further and further as you become a more advanced user, but your creativity really lends well to these types of things. Whereas with maybe some more advanced manufacturing techniques, there are things like what we have here, which is heat exchanger printed in BASF 17-4 on an Ultimaker, divided and sintered to create a stainless steel component. 

Dylan George: And this uses a lot of the methodologies that are akin to additive manufacturing, keeping an eye on overhangs, of course. But even with that being said, you can do a lot within those geometric constraints. So some of the things that you no longer have to pay attention to are internal cavities. You can actually create internal cavities of course, with additive manufacturing, whereas you couldn't with traditional means. And then you don't have the limitations like broad dye limitations, as well as sharp internal corners. These are all things that you can account for with additive. Kenneth Wong: Very good. Good point. Great. These of course, are what Dylan has shown the type of geometry that actually makes the most sense for 3D Printing. What advice do you have for some people who are designing, who traditionally design for the traditional kind of machining or things so that they might end up accidentally including some features that when it gets to you, you would have to tell them you can't really print them? 

Greg Paulsen: Yeah, absolutely. And it's really interesting because it does depend process to process. But for example, I use a lot of analogies when I talk to my clientele who may be used to a more traditional process like machining. DFM and DFAM for process like FDM or FFF, particularly if you're running a soluble material where support structure trapping is not as much of a challenge, is almost identical to design for CNC machining. It's like the thin walls access. You could go more with FDM. You can, but if I've taken a CNC part and I'm moving it to FDM because that's field and based process, I can go big. I can go chunkier. I can go broader. This is a CNC prototype in FDM, and I can have these thicker features. 

Greg Paulsen: However, that's because this thing is essentially planted down on a build plate, so all the stress is being held by the build plate and it stabilizes as it's heated up and cools in place. Other processes, those thermal constraints become much more of a challenge. And that's where the analogy from traditional to additive from a start. Like lattice and topology activation and generative. That's awesome, but just getting started, it starts to feel a little bit more like injection molding. I want those even wall thicknesses. I want something that in an expanded safe, when it shrinks, doesn't have this hot core in the middle that makes everything want to bend around and warp, for example. So usually if I'm moving something too additive, I like to look at the feature and be like, "Doesn't need to be there." 

Greg Paulsen: If I have mass on my part and I'm growing it, I'm 3D Printing it, I'm making my machine work for it. I'm adding price and overhead. If I don't have mass there, I don't need to deposit it, so it actually reduces cost. Whereas in CNC machining on the inverse, if you remove material, you're removing material with a tool where you leave it there, you're simplifying the part and simplifying the overhead. But some of those notes just on what Dylan was saying, orientation does matter if you're thinking about supports. That layer type, is this the filament? Is this a powder bed? Is this a liquid resin being cured? May affect how you do this. Think about making features that grow on top of each other, self-supporting features like think Corel, or Tree versus traditional things that may have much more of a square jagged cross line to it. 

Greg Paulsen: And I will also say as someone who has cleaned and shipped a lot of parts that clean ability and ship ability for prototypes, okay. Yeah. We could do that for this one part, make sure it's cleaned out, get a dental pick in there and work through. But if you're looking at designing for additive production, you do have to think about, is there a place where there may be a sacrificial material like support or powder that's trapped? Can I put a clearance hold there? You have to be mindful of that because that gives you that quality assurance, that consistent part time to time again. So if I'm moving from, at quantity one, I could take more risks, or I shouldn't say risk. I could be more relaxed on that. But at quantity 100 or 1000, I'm probably going to sharpen my pencil a little bit. And we'll give that feedback too and when we talk to the customers, if we move to production with them. 

Kenneth Wong: That's a good point because you don't want to be sending technicians to clean your part every three months or so, because you design it that way. 

Greg Paulsen: Or yeah. You make a part that has a confined hollow, and you don't have proper clearance, so the thing becomes a salt shaker. It looks great, and then one day it just decides to dump powder on you. And you're a designer too, so you're the designer. We're fabricating on the best process and a little extra clearance hole or wider hole in that area makes a remarkable difference on cleanability. 

Kenneth Wong: Very good point. So, Matt, you earlier talked about wanting to actually make DfAM irrelevant with the hardware's intelligence. How do you imagine a hardware handling the kind of geometry that we talked about that are, I guess, problematic? 

Matt Karesh: Yeah. it's more than just a hardware, it's having an entire solution. So making sure that the software that you're using on the front end is bringing in the most representative geometry. So ideally, your native CAD file, and then applying the appropriate parameters and working with the hardware to give the appropriate instructions. And then when you get to the hardware, making sure it's well qualified and running in spec, not only before the build, but during the entire process. 

Matt Karesh: So making sure you have the right monitoring in place and you're collecting the right data. And then on the back end, ensuring that everything went as inspected or as anticipated, and having the ability to go back and parse through that data, if something did go wrong, and identify where it went wrong. So more than again, just looking at the system, the hardware itself, it's having this full solution that integrates from CAD to printed part, making sure you're not having discontinuities during that process. 

Matt Karesh: And then once you have that stable platform, that's more comparable to a CNC machine that runs repeatably over and over and over, you can then go define and develop these very high end process parameters that are more or less operating on the edge of stability. And that's how you then unlock being able to create these very high end complex geometries and being able to create these surfaces that don't require supports, which also requires a very deep understanding of the melt pool behavior and how the laser interacts with the powder, how the powder interacts with the melt pool, all those types of things, which we try and take out of the equation for people. 

Matt Karesh: Let's say your business is making widgets, right? You don't want to go have to start up an organization that then becomes process parameter experts just to be able to make your widget. Your value is creating the widget itself. So if there's a solution out there that the hard part is already done, and you can just deliver the part that you want to make, that's the direction you would go. So Velo's approach is to do all that hard part, that deep knowledge of process development, and then provide that as part of our solution so our customers can then go make the parts that they need, want to make. 

Kenneth Wong: Okay, let's talk about hardware, software, processes, and services altogether, that ecosystem that is going to support you, if you want to go into mass production using AM. So let me start with you, Dylan. What's a good strategy to match a particular type of machine, a particular type of materials with what you're trying to accomplish with that design? 

Dylan George: Yeah, absolutely. So I'll use this visual as a use case here. So with matching the end use case, with Ultimaker for instance, we have hardware, software materials, but with the end hardware, or excuse me, the end material you choose, let's say that you're going to end up with a nylon print. You do have the flexibility to first print those first one to seven prototypes in an inexpensive PLA for instance, something that's more forgiving, something that you can learn from, maybe something that allows further overhangs without support structures, or maybe a water soluble or easily break away material as a support structure. And then you can graduate on either as you become more proficient with the technology and that material to something that will be your end product, such as a nylon or some sort of glass filled polypropylene, whatever suits the needs of the application. 

Dylan George: And then once you have your design, you can figure out just how many machines you need in order to keep up with the production that you require. So knowing the build volume that you require, as well as the overall quantity that you're looking to produce, you can decide whether or not let's say, keeping within our product line, as an example, you can choose to use an S5, which has a larger build volume, or if you're producing and only producing smaller components for your desktop, then that could be an S3, for instance. So really matching what it is that you're intending to produce with the type of machinery. 

Dylan George: Like Greg spoke about this many times about actually picking the print technology that matches with what your requirements are. And I think that's a really empowering thing for customers. So being able to preview what that is in your software, and then choosing the material, learning from that material, and then maybe picking your final end geometry and end material to produce from is I think, a lot of power for customers and something that I always recommend that our customers explore before deciding on their final choice. 

Kenneth Wong: Very good. I'm going to bring you into the discussion, Greg. For people who already have their own in house hardware and default AM expertise, how does on-demand printing fit into that process? 

Greg Paulsen: Yeah, absolutely. And Xometry, I mean, we are a digital manufacturing marketplace. And vast majority of our customers do have manufacturers. A lot of them are other manufacturers, but we give the digital version of if they open up their shop garage and there is a thousand other shops on the other side, right? Before Xometry existed, I was working in product development, and actually I was spoiled. I had a selective laser sintering machine, I was working on an industrial printer, but that was our in-house. And that was something that we used to iterate rapidly and work through it. I was running it overnight for internal projects. And Ultimaker wasn't around, and Dylan, so sorry about that. We're running this industrial machine, but at the same time for this product development firm, I was out sourcing easily six figures of work every year to other manufacturing technologies. 

Greg Paulsen: I needed to do a seal check, let's do stereolithography. I needed metal parts, let's do a DMLS or DMLM, or SLM, or whatever you want to call it today. And we are able to access that through services. Xometry, we've gone a step further because we're a service of services. And so we give you that speed. We're running every part starting the next day, because we always have parallel capacity of redundant manufacturers. When you talk about a world of supply chain resilience, literally the whole coast of the US can shut down and you could still click drag upload it from Xometry because we have the rest of the world available too. So we don't have the disruptions that you would have with a typical single shop or single supplier, your colleague has called in sick today, so you can't run your printer that day. 

Greg Paulsen: So there's a lot benefits to having a manufacturing marketplace as part of your digital strategy, as part of your production strategy. And let alone, I was joking. I was looking at this. This is a MultiJet Fusion part that is chemically vapor smooth dye black, and has fantastic results to it. And this purchase price is about $40 or so. So instead of jumping in with several hundreds of thousands of dollars of infrastructure to get the equipment, get the material, get this running up, you could just buy the part being built on the machine, being built by experts in the industry. And so it's really powerful to extend out and utilize manufacturing capacity and expertise as there. 

Kenneth Wong: Very good. Audience, just a reminder. I'm going to have one more question for Matt. And after that, we'll start moving into Q&A, so we'll start addressing your questions. So if you have questions for the panelists feel free to type them in. Matt, obviously, we have explored the idea that there is a huge difference between prototyping and using AM for mass productions. For AM for mass production, how's it different from prototyping? And what are some of the process elements that people need to pay attention to? 

Matt Karesh: Yeah, and I think Greg probably hit on this as well of the ability to have a little bit more flexibility on a one off part versus a part that you're going to run tens, hundreds, thousands of. So when you're looking at one-offs versus a production run, you want to make sure that on the higher volume side of things, that you've really sharpened your pencil as he said, and got all of your designs dialed in. But the other piece of it is, so your design styled in, how are you going to then scale that up? And I think part of the struggle with the general additive industry today is that a lot of the machines run differently, operate differently, and everyone has their own flavor of parameter sets. A lot of people view that as their intellectual property. 

Matt Karesh: So let's say you needed to go from 10 units to 1000 units and that requires a couple extra machines. And that's, especially in the metals world, a very large capital expenditure. So you would want to look through for other contract manufacturers or service providers who have that capacity and utilize that. But the fear then becomes, "Am I going to get the exact same quality out of all the different manufacturers?" So what we at Velo have tried to do, and I hinted at a little bit is a couple things. Make sure all of our printers run the exact same way, so we have one click calibrations where we've made it extremely easy to calibrate our machines such that we recommend they calibrate once every single week, if not before every single build. So you set the baseline for your machine very often, make sure it's running optimally. 

Matt Karesh: We then have that in process monitoring to make sure that during the process, everything is within spec, and if something does go wrong, it alerts you and tells you exactly what has gone out of spec. We also have taken all the parameter development in-house, and then everyone is running the same parameters. So now that you have this combination of hardware running well within spec because you've done calibrations, you're monitoring during the process, you're all running the same process parameters, you can expect to get the same parts out. The final piece there is the build file, right? So you've got your design sharpened up. You've got all your hardware lined up. 

Matt Karesh: Now we have the ability to create this build file, which we refer to as a golden build file that you set it up once, whether you have our Flow software or we at Velo help you do that, you can then create a print file that is completely self-contained, has all the instructions, send it to any printer in the world, obviously running the same material, and expect the exact same quality back out. So it makes a very easy, very quickly scalable solution to go from one-off wrap prototyping to mass production. 

Kenneth Wong: Okay. Very good. Thank you very much. This is a very good discussion, and it's now 40 minutes into the hour, so it's a good time for us to move into Q&A, so let's just go ahead and do that. The first question, this is actually one of the questions that came in through the registration process. So let's tackle this one. Cause comparison between 3D Printing and conventional manufacturing, current limitations in technology. Why don't I start with you first, Greg, because you are in the business of cost estimation. And then we'll ask the other panelists to chime in. 

Greg Paulsen: Yeah, absolutely. And as I noted before, there is an inverse in paradigms when you're thinking about what drives costs, when I'm thinking about traditional processes like direct manufacturing process, like [inaudible 00:42:22] cutting, CNC machining, et cetera. The cutting action is the work. So the more stuff I need to remove the higher the costs are driven, where in additive, the actual deposition of that material is the work. So the more stuff there is, the more work there is to it. So you always want to put your additive parts on the diet in order to help save costs and optimize a little bit. But there's also just general things about design, Design for Additive Manufacturing really, really shines when you highlight what additive can do, especially when you're thinking about an end use part that will be ultimately additively manufactured. 

Greg Paulsen: Can I consolidate part features? Can I creek off angle features? Can I go organic or optimize in a way that would just be very difficult to machine traditionally? One of the greatest challenges that I usually see is a super large organization that comes up and they're like, "I'm going to hire a full-time person to get into this additive manufacturing." And the very first thing the person does and says, "We've been CNC machining or turning this part for the past 30 years, and it looks like this. And they're like, "Can you make it cheaper and additive?" And you just look at them and you're like, "No. No, I can't. You have cost optimized this and bid against it. And there's probably a work cell that lives and breathes to make this fit widget, so no, I can't." 

Greg Paulsen: Now, if you're working on a new process or this is part of a greater assembly, where all of a sudden this could also have a complex gear function to it, or the weight savings that could gain from additive will save this airplane $10,000 a year, then that's the excitement, right? Apples are apples. If it looks like a machine part probably is a machine part. With additive, you could explore a lot more options on design. But ultimately, if your design intent starts at additive, that's where you really drive the value add there. And it could be more than just the manufacturer. It could be what else can it do for my assembly. Kenneth Wong: Very good. Very good point. Dylan, do you want to add something, cost comparison between traditional manufacturing and 3D Printing? 

Dylan George: Yeah. I think Greg highlighted a lot of really great points, but when you look at a part that you have in your hand, like the one that Greg just showed, a lot of times things look that way because that's the way they've been manufactured for a very long time. And it's difficult for us as human beings to look past that traditional form and really look to what are the design requirements, what are the requirements of this application? And that is really where you can pull a lot of the value. So you can choose what material you need to use. You can choose, of course, the geometry that really needs to be there, but the overall form does not always need to stay the same. 

Dylan George: In fact, you can find a lot of optimizations, a lot of cost and time savings just from changing that form a bit to match the new additive process that you're trying to adopt. So I just want to highlight that as we see more and more of this in industry, we will see more and more organic structures or structures that we're not traditionally used to seeing, because we're becoming more and more proficient in matching the requirements of the applications we're trying to make with these more advanced pieces of technology. Kenneth Wong: What are your thoughts, Matt? Do you really need that alien geometry to make the best use of AM? Or what are your thoughts? 

Matt Karesh: No, we have a probably inappropriate phrase at Velo, so I'll sanitize it. But it's why Velo? Or why AM? And you guys can probably pick the rest up, but there's always got to be a motivation for why you're making that part and the technology that you're choosing. And it's really no different than looking from a casting to a machine part, to an injection molded part. They each fit their own unique purpose and they do different things. So with additive, just like Greg mentioned, you're not going to replace a simple machine turn part with additive and make it cheaper. But if there's a reason, maybe it's a supply chain reason, or a quality reason, or a functional reason that you're doing something else to add value to these components, that's where it starts to get interesting in looking at additive and really using the advantages of the technology, not just trying to get costs out of an existing part. 

Kenneth Wong: All right. Greg, I think this question is the volume question. You might want to address that. How do annual volume affect design and printing process, and material use? So 10,000 units versus 100,000 units? 

Greg Paulsen: Yeah, absolutely. And I've looked into cost break evens with additive, particularly talking polymer additive manufacturer parts. So making some plastic process, whether it's a production photo polymer like carbon digital light synthesis, or like MultiJet Fusion or SLS, the summer plastics versus what if I just mold it? And I put the assumption that yes, there are design differences needed, but let's just say they're optimized for that. 

Greg Paulsen: You will be battling traditional process at that point. So when you are at a 10,000 or 100,000 life cycle, and especially if your design can be modified in a way for a tooled process, it is really difficult to battle that injection mold tool. And that's because some of those things like the greenness, we only use the materials you need when I'm additive. Well, at that point, the tool of monetization is so much on those volumes, and the injection for injection molding is exactly what you need. All of a sudden, you're getting one to one comparisons between molding and additive. Here's what I can do though. So let's say I'm working at that 10,000 skew, and my rev changes. So something else changes and my part fits too, and the rev changes. In traditional tooling, I have to change my tool. I have to remill or remake, or replace my tool because you can't just tweak. 

Greg Paulsen: For an additive manufacturing platform, you just put in the new file and you start printing more. And so when I look at additive production, it is not the 10,000. It is not the 100,000. That's not really how you think about it. You think in a modified just in time. So if you're looking at a project that's additive production, ultimately you may be making 10,000 or 100,000 of this, but you're usually optimizing the batch size based on the demand of your product, the lead time that'll take to produce those, and a lot of times are running several machines in parallel to produce a parallel for you. And you're going to get smaller batches more frequently versus just one giant load of parts at a time. And that gives you some benefits from inventory space. It gives you some benefits from, like I said, that design resilience. 

Greg Paulsen: So if my rep changes, I'm not screwed. I'm not eating large costs there, but just the mentality changes. I think to Dylan, this point too, that's something you have to do. You have to have this paradigm shift about what additive is, because all I've learned in school I've gone to is about traditional manufacturing, smooth surfaces, because everything's derivative of machine surfaces, all that stuff. So you have to start loving the layers. You have to start thinking about additive at a very different perspective. Yeah. 

Kenneth Wong: Very good. Good point. I'm learning to love those membranes and letter structures, definitely. Dylan, I think this is a question for you. Some have heard about integrating solvable plastics into design to create structures which could not otherwise be printed. After completing printing and setting the product is drench insolvent, for example, water, to remove the no longer necessary structure. Can you elaborate on this technology? So pros and cons, can you comment on that? 

Dylan George: Yeah, absolutely. So really the biggest con, we'll start with that, of using a secondary support material is simply the time used for tool changing. So if you need to move from one nozzle to another, you incur a longer production time to make that part, but you usually save that time on not having to perform any post process or manual post process. The benefits are, as an example, here's one where we have on the left, a traditionally manufactured CNC manifold. 

Dylan George: And with this manifold of single inlet and four outlets, all variant sizes, this particular engineer wanted to actually use advanced CAD tools to do a flow simulation and optimize for pressure drop across those four exhaust ports here. And so if you'd imagine this as an internal structure, you need to support these overhangs with, let's say, FFF printing. So this particular solution you can dip in water, and the PBA soluble support material will simply dissolve away, and you're left with the structure itself. So there's huge benefits that really unlock geometric limitations that we'd have with traditional means. 

Kenneth Wong: All right, very good. Matt, do you want to add something to it, for example, in the powder bear situation, what are the pros and cons and the advantages as opposed to the method, the water syllable method that Dylan has just talked about? 

Matt Karesh: Yeah, it's interesting because obviously the range of materials is much more limited in metals, but there's a couple different ways to look at it. One is using the advanced parameter sets to entirely avoid support structures, but also there's some interesting work that goes on out there of using postprocessing particularly like chemical etching and things like that. And if you have the right control over your support structures in your designs, you can get clever enough to use a chemical etching process to dissolve a way that connection points and effectively have the supports fall off assuming you needed the supports in the first place. So I think it's just a slightly different way of looking at how metal behaves, how to support it, and how to best utilize those different techniques to either not have to use supports or get them back out. 

Kenneth Wong: All right. We've got about six minutes to go. So let me see if we can take one more question. Maybe two, if we can squeeze in, but let's do one more question. Can you elaborate on the usefulness of AM in supply chain disruption? Sadly, we are now looking at a regional conflict in Europe. So this question becomes much more relevant perhaps starting today. Greg, do you want to start us off? Do you have any comment on that? 

Greg Paulsen: Yeah, absolutely. And we saw this also with COVID-19, especially really like March, April, May, June, at least in the United States, where we were hit hard with large supply chain disruptions, reliability, and non dispersed geographic locations. And all of a sudden, you're shut down. And additive really did pick up then. And it was a really exciting time to have access. And even people with home machines were actually able to work and help on PPE effort. One of the big challenges though, and I think hopefully our wake up call for this is the bare minimum requirement for additive manufacturing is a 3D CAD bottle. Period. Like the bare minimum requirement. 

Greg Paulsen: And one of the largest challenges with disruptions is usually not can we make it? But is it modeled? Do you have access to that technical data package? Because absolutely, we do have redundancy. We do have supply chains, but this is something we've seen in the past and unfortunately, I think we're going to see it in the near future until we create a digital inventory because it's just as important as a physical inventory of goods. So you can adapt very quickly and utilize top processes like additive for stop gap or final reliable solutions. Kenneth Wong: All right. Very good. Dylan, how about you? What are your thoughts on disruptions and how AI might fit into the supply chain to deal with these disruptions? 

Dylan George: Yeah, these past couple years have been really exciting for additive. That's been a positive, I'm sure. It's flexible in nature in that you can produce parts on demand. A nice part about having a professional desktop solution has been that even when businesses have had to maybe close doors temporarily for outbreaks and things, the engineering staff can still be sent home with these printers and produce parts, and innovate, and update designs as they need, and print, iterate, and make that tooling for when they get to go back into the production facility. 

Dylan George: Not only that, but as far as digital inventory, I couldn't agree more. That's something that is going to be huge coming up here in the next few years, as far as developing those digital inventory pieces, having hose parts made it a manufacturer in Wisconsin that needs to share that design and ultimately that physical part in Oregon and all across the globe, we're seeing companies collect these digital assets and make them available to be printed anywhere in the world at any other facilities. 

Dylan George: So I think from that standpoint, whether they need to share it globally or simply the production facility is in a very remote location like a lot of our customers are, it makes it difficult in these times to be able to order that material or those goods and bring them to their facility. Whereas, if they have the material available and printers to produce those components, then they can make it there in-house. And so I think depending on what the use case is for the customer, whether they order the printed part or produce it themselves, ultimately it has to work for their workflow. And I think that customers have more opportunities or choices than ever before. So it's a really exciting time for additive, especially specifically around supply chain. 

Kenneth Wong: All right. Very good. Matt, how about you, do you want to add your thoughts to this question? Or would you like to use the remaining maybe two minutes or so to address one of the remaining questions that you like? 

Matt Karesh: Either way? I think supply chain disruption is probably a pretty powerful tool that additive does. And I think a really easy way to look at how additive can disrupt the supply chain is look at all the pain points of the supply chain today. And so for very complex parts, it's things like tooling and lead times. So being able to eliminate all of that and deliver parts much faster, complex assemblies and things like that, where imagine you start with a forging and you do a rough machining operation, and then an inspection, and then maybe a heat treat process, and then a final machine operation, and then a coding. And then you assemble that. 

Matt Karesh: And then imagine taking that process and multiplying it by a bunch of parts that are all coming together, just the number of people involved to pull all that together, the number of drawings, the number of CAD files, the number of times things are shipped from point A to point B. You can effectively eliminate all of that. If you are able to combine all those components into one and the amount of time and money, and energy, and resources saved, it's an interesting challenge because it's very hard to quantify. It's a very complex problem. If people tend to look at just end part costs at the end of the day, but there's things like how many engineers did you have to have to maintain all those drawings, which doesn't necessarily directly go to the cost of that part, right? So being able to address a lot of those things is hugely powerful with AM technology, regardless of if it's metal or plastic, or some other technology. 

Kenneth Wong: Very good. Thank you very much gentlemen. With that, we actually hit our one hour mark right on the spot. So it's time for me to wrap up and say goodbye. To the audience, thank you very much for sharing your precious time with us. We'd also like to thank of course, our three sponsors, Velo3D, Ultimaker, and Xometry that makes this particular live discussion possible. I also want to thank our invisible hero, Steve Paul, who's our production manager who works behind the scene to make us all look good and sound good, and everything goes smoothly. So with that ladies and gentlemen, I'm Kenneth Wong for DE 247. Until next time, we are out. 

Matt Karesh: Thank you.