AM_F2019_07

Tom: So we have with us Alex Steph [Stephens], who is the vice president of business development at Digital Alloys, which is an additive manufacturing firm. And I passed around the little Boeing titanium piece that they're making. We've talked about them and his blog sites, so he's basically here to answer your questions. He can present some things. Alex has got a bachelor's degree in mechanical engineering from the University of Southern California. I understand they have a football team that defeated like MIT's football team. Maybe not recently. MIT gave up football in 1908 because they hadn't won a game for three years. Anyway, they've done better in recent years, so they come back. He has a master's degree of materials development. He's mostly worked on the business side of things, and you've got a few what we call NGO leaders for global operations with Sloan, people back there. Okay, and we've got some mechanical engineers — raise your hand mechanical engineers — and we got some materials engineers, so if you want to raise your hand. Okay, so you got a little bit of a mech, just mix. You have our good ChemE engineer, anyone else that I didn't include? Okay, so Alex, I'll let them to start out.

Alex: All right, thank you Professor Eagar. So I want to make this as interactive as I can. I don't want to stand up here and sort of try to give a lecture. Have you guys had a chance to read through some of the blog material? Okay, so I think a large reason I'm here today is to kind of help you guys sift through and understand some of that research we've been doing. But I wanted to start out by giving a little kind of background on myself, a little bit about Digital Alloys, the company I work for, and our technology, and a little bit about how I think about just kind of the whole additive manufacturing space as it applies to metals.

So, myself, I've been in the 3D printing space for about six years now. I've been on the kind of business development sales side of things. But I was at Stratasys, if you guys are familiar with that company, one of the first major polymer printing companies, developed FDM technology. And then they acquired a company that developed PolyJet technology, which is sort of like an inkjet polymer process. And then I was at Carbon, which is a very well-funded startup in the Bay Area that has developed a new liquid light based process they called DLP, it's kind of like SLA. And for Carbon I led their automotive vertical, basically all the business development sales into the automotive industry with their technology. And then about a year and a half ago I joined Digital Alloys, which is a venture-backed startup based just in Burlington here. Professor Eagar is one of our advisors. A lot of our research team came out of MIT, did their PhDs or postdocs in material science. We're about two-and-a-half years old. Have received investments from Boeing — everyone knows Boeing — Lincoln Electric, if you haven't heard of them, one of the largest welding technology companies, and they also manufacture wire, which is our feedstock, so that's a strategic investment. And we're about twenty-five people now. We're still sort of in this stealth mode, if you will. We haven't fully commercialized the process yet, but we're pretty open to talking about how it works and the markets that we see as immediately kind of attractive for it.

So that's sort of a little background. Let me kind of share with you guys sort of a little bit about an overview of the metal additive space from my mind. And you guys at any point feel free to raise a hand or interrupt me if you have any questions. Again, I want to make this more of a discussion than me just talking. But a lot of people, when they think about additive manufacturing, it's just like one manufacturing process, right? It's just one sort of category to a lot of people. Really, it is a lot of different manufacturing processes, and just like in conventional manufacturing, each one has its own set of pros and cons and has applications that make sense for it.

When we think about metal additive, there's actually about — I have to count all of the different processes here — something like fifteen different ways to additively manufacture, 3D print metal. And this is a group of consulting firm out of Germany that put this together, and I like the way that they kind of broke down the different technologies. So you can see, starting in the center here, the dark blue represents the metal material, the green represents processes that use metal in combination with a binder, and you can see it's broken down by feedstock. So wire based processes, a lot of these are under the category of direct energy deposition, DED, you guys have heard of that. Our process, our logo is here, under sort of this resistance welding top, and I'll get into more of how our process works.

There's powder based processes. A lot of these were sort of some of the first metal 3D printing processes which were officially called kind of 3D printing. Professor Eagar and I were talking about how in theory additive manufacturing has been around for eighty years or something if you consider just kind of welding on one layer on top of another. But from the perspective of this new 3D printing industry, it really started with laser powder bed about twenty years ago out of a lot of research in Germany with lasers and powdered metals. And then there's companies like Desktop Metal, also based in Burlington, they're neighbors, and I know a lot of ties to MIT, where you're doing things with binder and metal powder. And then there's some sort of kind of niche interesting processes that use in some cases large rods of metal that are doing like friction stir welding, basically spinning rods to generate heat to melt the metal in contact with the substrate. And there's things like ultrasonic sheet welding, basically where you're binding whole layers on top of another.

So my point is that there's a lot of different ways to process metal under this category of additive manufacturing. And to pick the right solution, or to figure out if additive manufacturing even makes sense, it really requires some level of understanding of all these different processes. Things like the economics, what materials do they use, what sort of volumes and throughputs are they capable of — the same way you think about any manufacturing process. Any questions on this chart or what I've said so far? Yeah. So now let me go into this deck.

Cool. So Professor Eagar basically grabbed a lot of sort of snapshots from this blog series that I've been writing on behalf of our company for the last year or so. I'll kind of talk to some of these kind of highlights here. So first of all, starting with — I think a lot of times when you look at a manufacturing process, from my perspective at least, you really have to start with kind of the business value. What does it do that's unique, what does it do that's valuable? I know that's not necessarily sort of an academic perspective, normally you want to start with sort of how it works and why and what are the physics. But from my perspective, working with companies in the industry, everyone is excited by this concept of 3D printing, but at the end of the day it doesn't really matter if they're not able to use it and it's not able to make some sort of real impact to their business. So coming from that perspective, they really think about 3D printing as another manufacturing process, albeit sort of a newer, more hype, more sexy sort of process, but they think of it as a manufacturing process. So what are the costs, what sort of materials can it process, what sort of designs can it enable that we couldn't produce before, and how valuable could those designs be to our customers, to our products? And what does the whole process look like? Is this something that we could really integrate into a factory that we're running, or is it something that we would purchase from a supplier and just build expertise in the process to understand how it works?

So on the topic of kind of value drivers from a business perspective, everyone's using 3D printing for prototyping. You look at anyone that's making a physical product today, any sort of established company, they have a handful of 3D printers, and they're all using it early on in the design stage, right, to make concept models. At least a good amount are also using it to start doing some real functional prototyping, assuming they're making a product or components of a product where there's a technology that can get them close enough to how the final product would be made and function. A lot of companies are using it for what I would just call sort of manufacturing tools, so things like fixtures, jigs, low-volume custom things that are used in the process of making another part to hold it or to measure it, things like that. And then there's people that are looking for sort of a lot of advanced design that they get from 3D printing, so things like lightweighting are especially important to industries like aerospace, industries like automotive, because the weight of a part is not just the cost of the material that went into it, it's really how that product's gonna perform. And there can be a lot of cost savings and performance benefits from removing weight from that part through designs that you can achieve through additive.

Conformally cool tooling — I guess quick show of hands, does that term mean anything to anyone? A few of you guys. So in a lot of tooling processes, as in injection molding, die casting — people familiar with those sort of general manufacturing processes — the rate limiting step is really how fast you can cool and solidify the material in a mold. And that sort of largely constrains the economics of the process. So 3D printing has been used — it's still sort of niche because it's been so expensive — but it's been used and proven quite a few times now in these processes where you can put cooling channels into a part that conform to the — it's called conformally cool tooling, to conform to the surface of the tool. It allows you to really optimize the cooling to remove heat from the plastic part in the mold or the metal part in a die as quickly as possible and in a uniform controlled way in order to solidify, finalize that part, pop it out, and start producing the next one. So that's kind of a hot area for 3D printing, I guess no pun intended there. But that's an interesting area for 3D printing because you make one part, you make one conformally cooled insert, and it could potentially improve the productivity and efficiency of making thousands, hundreds of thousands of parts.

People are looking at 3D printing as a way to improve the efficiency of machined parts. So a lot of parts in low-volume industries like aerospace and medical, they start out as a big block of metal, and then they take that workpiece or billet or sheet and they machine away, right, from the outside using CNC machining. They mill away, they cut away all that extra material to form the final shape. You can imagine how wasteful that is as a process. A lot of materials, you end up with a bunch of metal chips on the ground, and with a lot of materials are very hard and costly to recycle, so a lot of times that's just waste. It takes a long time to do that, it's a very slow process, it takes expensive CNC machines a long time to do that. So a lot of people looking at 3D printing as a way to make that process a lot more efficient, right, put material where you want it.

And in my experience working with companies like Boeing, oftentimes — do we have a slot that slide in here? Probably not. So oftentimes you will start with a big block, and you may be creating just like a T-bracket, alright, something that looks like that, and you'll have to mill away all of that material. Sometimes you'll start with a ten kilogram block, end up with a one kilogram part. And you think about a material like titanium, which aerospace is using more, and ten kilograms of titanium, $500, $600 worth of titanium, ninety percent of it ends up on the floor. Yeah, I don't know if you got that one, but so one reason Boeing invested in us is to think about how they can improve their — yeah, here's a good example of one. You know, to start out with a big block and then machine this part, you can imagine how much waste there is. And so companies like Boeing look at 3D printing as a way to save money, save time potentially. But if you can print this faster than you can machine this from billet, potentially there's time savings too. But also just sort of as the environmental footprint, they think about the energy use. There's a ton of energy that goes into producing the primary material, especially with a material like titanium. So if you can reduce a lot of that scrap, you can reduce your energy footprint too. And then of course if you can make a lighter part that goes into the plane, the energy savings there can be massive.

Oh yeah, so that's a good example there. I'm just gonna say this is a good example of a part that we're working on with a large aerospace manufacturer on, and it may seem ridiculous, but a lot of companies, they'll start with — this is like a seven kilogram block and end up with a 1/2 kilogram part through machining. There's a lot of parts like that on a plane.

And in terms of time savings to think about, CNC machining, you not only have to develop a program to make the final part to control that end mill, but you also have to develop some pretty complex tooling to hold that part while you're machining it. And so it's sort of a two-step process, and it's a lot of programming, it's a lot of design, it's a lot of just waiting to receive materials, waiting for materials to be machined. It's kind of like a chicken or egg thing, you have to machine a fixture to hold your part before you can machine the actual part. And so it's not an efficient process. It is a process that can get you to a very accurate, valuable part at the end of the process, and it's used in conjunction with a lot of processes because it can produce very smooth surfaces and accurate features. But yeah.

Student: [Question about titanium wire availability — inaudible on captions.]

Alex: Good question. So I guess I haven't really got into our process yet, but titanium wire, thanks to the welding industry, is something that is what we call commercial off-the-shelf, right. It's used by a lot of manufacturers currently to do welding and joining and repairing operations.

Exactly. Yeah, so you can buy big spools of this stuff from hundreds of different manufacturers. Yeah, it is more expensive than billet. If you think about the process to make a big sheet of titanium, it is less energy intensive, less resource intensive than drawing something into a thin wire. But it's normally about — we're seeing it's about two times more expensive to buy wire than to buy like billet. If you look at metal powders, which are used by a lot of metal printing processes, that's often four or five, sometimes even ten times the cost of plate.

Any questions on CNC machining and how it compares to additive manufacturing? Yeah. So casting — well, there's a lot of different types of casting. I guess some of the most popular forms of casting, like sand casting, die casting, if you're looking for a high throughput sort of process, it involves tooling. In the case of die casting, you can normally only do die casting for lower — correct me if I'm wrong — melting point sort of metals, usually like aluminum, magnesium. I don't think you can usually die cast like steel or titanium. But where you can do die casting, of course, there's a tooling step where it takes a long time to design, manufacture the tooling to do it. In lower volume casting, typically like sand casting, where you'll make a pattern and then you'll form sand around it and then you'll use that as sort of the negative to push the liquid metal into, it's a lot slower of a process, but it's relatively low cost. You can produce some complex shapes, but not the same complexity that you can get out of additive manufacturing. And at the end of the day, it's also sort of a near net shape process, like 3D printing — you guys familiar with that term, near net shape? So you end up getting something close to the final shape of your part. You're still gonna have to do secondary operations to smooth surfaces to meet tolerances.

And so you end up forming a case — it's another step in the process, but a lot of parts in aerospace are near net shape cast, but then you're not done. And I think that's part of the story here, that you can additively manufacture something, but you're not done until you end up with final surface finish, final part — final surface finish, final part shape, okay, surface roughness and things like that. There are a lot of steps in additive manufacturing. It's just one of those steps on the way to the final part. Yeah.

I guess there's some other examples of near net shape processes. And near net shape is sort of a little bit subjective. I mean you can look at this process like sand casting and say that 99% of sand casted parts are gonna need some sort of finishing step. But there are examples that don't, like a manhole cover. If you can deal with a surface around a whole part that looks like it's basically the texture of sand, then that's fine, then it's — in that case it's a net shape process. But for most processes, you're getting something closer to the final shape, but still any sort of functional surfaces, sealing surfaces, bearing surfaces, you need them to be very smooth and very tightly toleranced. So machine steps, polishing steps, or grinding steps, or whatever it may be, are used after that process. So here's some examples of other near-net processes that are typically near net shape. Yeah. I mean feel free to stop me at any point.

Tom: [On chemical milling / etching processes] Very corrosion-resistant, so the acids they use are pretty aggressive acids. And then that means like it makes for the office like, right, if you leave it in there for an hour or several hours in the acid. And so you're not getting final precision until you machine the final precision. You just chemically etch away. Are you aware that, because it's a complex — a part that's a complex part — it wasn't just a rectangular parallelepiped that you just cut off the surface, but you got contours and everything else. So these are complex parts. You look at that practice that we had that was up there a little while ago, that's too complex to go just machining the surface plus the outer cases.

Alex: So I guess let's talk a little bit about some of the processes and sort of the thermal management component of it. This is a huge thing in metal — I guess is a huge thing in any manufacturing, but especially in 3D printing. When you're melting metal, it expands, and when you cool it, contracts. And when you have one part, one section of your part that's very hot, and one section of your part that's very cold, you start to deal with a lot of internal stresses. And parts manufactured with a lot of additive manufacturing processes will literally tear themselves in half, especially along layer lines where there may not be great fusion and strength. So this is an example that's not atypical in 3D printing, where you'll have a powder bed build that will tear itself apart. And so this is a large factor when you're thinking about how to design parts for the process, when you're thinking about how to orient parts within a build, when you're thinking about how to support them and how to optimize parameters for a build. And so different metal manufacturing processes have different amounts of sort of heat loads and residual stresses like this that they create.

Another factor with thermal management is also the effect on the grain structure and the crystals in the final metal part. A lot of really high energy processes, we'll start to see what is called columnar grains, where a lot of heat penetrating into the part will actually create anisotropic crystal structures, which most industries do not want to see. Someone like Boeing wants a part that is what you call isotropic — equal strength, equal properties in all directions, because it's easier to design for, it's easier to validate. So this is just something to keep your mind on, that it's one of these sort of dirty secrets about 3D printing that a lot of people don't talk about, but it's something that people that use the technology are very aware of and need to be pretty experienced with. Yeah.

I mean, I'm not an expert from a metallurgical perspective on some materials, but there's a reason why a lot of materials just have not been printed and probably maybe cannot be printed with a lot of processes. When you look at aluminum, right, one of the most widely [used] materials — powder bed processes have really only been able to process casting grades of aluminum, which are difficult or [low?] strength. They haven't been able to process your typical aerospace grade high-strength aluminum because of issues with cracking and this kind of thing. Do you have anything to add to that?

Tom: It gets very specific. Steels are not a problem. Those steels are not going to be the parts unless you're getting to high alloy steels. In fact, you might want to mention something about — here, the titanium is fine, certainly a niche in the medical business and the aerospace business. Nickel alloys that are high value-added, $200 a pound anyway, and they're going into high temperature applications. So you really have to look at the metal — beautiful thing before you start. That's one of the problems people are trying to — carbon steels, or brass eyeglass hinges, I mean...

Alex: Yeah, so why don't I use that as a good segue to kind of explain what we're doing, how it works, and how we're thinking about the applications and industry and adoption of this technology. So, why Digital Alloys. We've created a new metal manufacturing process called Joule Printing. We call it Joule Printing because Joule, unit of energy, because of the translation from Joule heating, basically resistive heating, just like a coil in a toaster, right. We take a welding wire, which we discussed earlier, which is great because you can find it in a lot of different metals, a lot of people manufacture it, you can get high quality, low cost compared to powders, which is how most metal additive manufacturing is done. It's a lot safer. A lot of companies don't like dealing with powdered metal because it's a risk of inhalation, it's a risk of flammability because it's a really high surface area, it's a risk — well, it's prone to oxidation.

Yeah, this is titanium. Yeah, so what you don't see is sort of the bigger printer behind this is in argon. So we have a chamber, we purge it with argon.

Student: Would it be on fire?

Alex: Yeah, it would be heavily oxidized. You'd lose a lot of strength in the material, it would look funky colors. I don't know if it would catch on fire, I don't know, but maybe if it was powder.

Student: [Question about long narrow parts and thermal management.]

Alex: A part — I think you could have some super long narrow part that it stays super hot. Yeah, it's something we are — it's a big part of our development right now, is figuring out thermal management. A lot of times it's figuring out — what you see here is what we call printing on a cold plate, so it's just room temperature ambient environment. It's really hard to get good fusion to a cold plate, you know, with molten metal. So we're playing with ideas like, do you heat up the plate, do you heat up the whole chamber, to maintain that thermal environment and to try to reduce the thermal gradients. And frankly it's something that we're still figuring out, and it's very geometry specific.

Tom: Like you said, when they started doing this about a year, fifteen months ago, the big problem was they — yeah, that's keeping the heat in the metal. Has too much thermal, you know. And so they weren't getting the diffusion, then we're getting porosity in the final layers. And it turns out, it amazes me that you can have a machine that sloughs that fast, but they just sped up the machine to go faster, and they solved the problem. But again, you're gonna have some problems with residual stresses eventually. They could not make that as a six inch by six inch part without getting crack.

Alex: Yeah, not today, not without really heating up the whole environment and the plate. Yeah, we have, I guess, it's a good problem compared to the opposite, which is that the faster we print, the easier it is to print because we're printing on hotter metal essentially. With most manufacturing processes, right, the faster you try to go, the more complications arise. With our process, it's very hard to print slow because it's hard to print on a cold surface, basically. It's hard to get good fusion, it's hard to dump that much metal into the process — so that much energy into the process — without destabilizing it essentially. So that's why we like materials like titanium, because they do — what's the right term — they have a low — heat — they're low conductivity of heat, yeah.

Student: [Question about toolpath optimization and software.]

Alex: Yeah, good question. So our long-term vision is that our customers don't have to be experts on sort of the thermal intricacies of our process and have to develop these really optimized raster patterns to get a good part. We want software to be able to do that, and so we're trying to develop some pretty advanced software that will understand some of these physics and be predictive and develop sort of a tool path that makes sense. At first, of course, we're going to have to give a lot of customizability to users because it won't be perfect, but that's the sort of end goal. Yes, there is a lot of science and thought to how to — what is the best tool path strategy, we call it. A lot of deposition type technologies will often just kind of trace the outside surface because they want to get a good exterior kind of surface, and then they'll sort of fill it in like you're coloring in something. But we found the most success right now with this sort of back and forth, like polarized kind of rastering, partly because the point in our tool path that is potentially the lowest quality is when we have to do a turn — that's sort of the most complex part. And so why not put that on the outside of the part that you're gonna machine off anyway. I'll pass around just — this is just a section of a coupon like this.

Tom: [On microstructure] That transformation gives you a grain crystal structures. With different alloys you get different things. We call widmanstätten, and those are not as harmful in titanium as they might be in steel. Not that they're terribly powerful in steel, but it's not the optimal microstructure. This is a nice cast microstructure, but actually...

Alex: Yeah, and so we also like titanium because, I think for that sort of reason, it absorbs any oxygen that's in the chamber when you're printing. Steels — the oxides, I guess, remain on the surface, and then it's harder to get good fusion to the next layer. So that's another reason we kind of like titanium right now.

Student: [Question about orienting stock to get different properties along grain directions.]

Alex: Yeah, and so I guess any other material, living [rolling] structure — okay, it can — in a rolled structure, okay. So you do have to worry about anisotropy in the titanium.

Tom: I think at an early [unclear].

Alex: And this is a good segue to instead of — how we're working with customers today, they want something that is repeatable and sort of as simple as possible to qualify, because any new manufacturing process is a risk. And looking at being able to say a process is isotropic and extremely consistent and repeatable makes it a lot easier for a company to adopt and design for and qualify. There could be advantages if it's sort of anisotropic — think about what people do with carbon fiber, right, composites and things like that. That can definitely be some great design benefits to that, but also with that comes a lot of complexity. So right now we're trying to focus on just a consistent isotropic part. What we are going to do in the future is actually start experimenting more with multi metal parts, where we can — we have sort of a treed system in this printer where we can retract one metal wire and insert another while we're printing. And so you can start to imagine how you could produce some very optimized mechanical, electrical, thermal designs by using different metals in different places. And so that's sort of, to a lot of engineers, that's sort of the Holy Grail of where you can go with manufacturing. That said, you've got to figure out how to print a single metal part really well before you start distracting yourselves with the complexities of multi metal. So that's sort of future roadmap stuff. Yeah.

Student: [Question about wrong geometry and temperature gradients.]

Alex: The wrong geometry could, and a big enough temperature gradient could affect — but we're also making sure we're getting really good fusion between layers, and so that's a key part of it. Our process is actually a lot lower energy than a lot of other processes. To give you guys a couple points of reference, we're printing right now, like this part here, we're printing titanium at about two kilograms an hour. We're printing at a couple — about two volts, so I think like an AAA battery voltage, but a lot of amps — about 300 amps. So voltage times amps is power, 600 watts. It's only about 300 watt hours per kilogram to print titanium. If you look at most welding processes or most powder bed laser processes, it's about an order of magnitude less energy input to get a kilogram part, which I think will really help us in terms of residual stresses and heat cycling and these kind of factors. And that's what we're finding so far. I mean we're printing parts on a plate that is typically a lot thinner than you would use. I mean even this, I think, is quite a bit thicker than you would use in typical additive manufacturing when you're making parts in laser powder bed — if you guys have seen that process — you're using like a multi inch thick plate because otherwise the thing will literally just potato chip.

So any questions on the process before I go more to the business side of things? I think we have another seven minutes — that right?

Tom: Oh we have until the end of the hour.

Alex: Oh, you have it. Okay, fifty-five, okay, cool. Any questions, I guess, on the process before I go further?

Okay, so next question. We have this exciting new technology, what on earth do we do with it, right? If we stay in the lab for five years tinkering around and never make a dollar, who's going to continue to fund that? So on the business side of things, early on we looked at this process, and we said, okay, it's high speed relative to other metal 3D printing processes, we're able to use off-the-shelf material, so it's a relatively low cost wire. And of course it's a near net shape process. If you guys saw the little part I passed around, it's kind of like similar surface to a sand casting. That surface is not going to be acceptable for the majority of engineered surfaces. So it's a near net shape process, basically, to get as close to that final shape part. Low-cost material, fast process, where does this make sense in the business world, where could manufacturers really use this, and where could it drive value?

So we said, well, first of all, if we think about machined parts, right, let's look at parts that are — one, the material is very high cost, the material is hard to machine, so it's expensive and slow to remove material. And that kind of got us to our first two target applications, one of which being titanium. And the industry that uses the most titanium is aerospace. Actually the company that uses the most titanium is Boeing, who is one of our early investors, which is great. We have a strategic partnership there. And then we looked at tool steel, because tool steel is not as expensive as titanium, but it's still expensive compared to most steels, and it's very hard to machine, it's very slow. A lot of cost is actually eating up tool — just in the tool wear from machining tool steel, just the end mill has to be replaced a lot, and those are expensive.

So that led us to tooling applications, where you're making molds and dies for molding, for die casting, for stamping, and so forth. And a lot of that industry is in automotive because of their volumes, because the sizes of their parts. There's also a fair amount of that in like consumer products, but to be frank, a lot of that's over in China, and as a US company we said what companies can we partner with here that are manufacturing parts here. A lot of the tooling in the US is automotive. So we are working with a lot of the leading companies in the aerospace sector, in the automotive sector, on these titanium sort of components, and tool steel for tooling. And it's not all the conformal cooling applications like we talked about, even in just producing an insert that goes into a die that — let's say forms like — so I was at Ford a few months ago and saw this massive die that the whole hood of a vehicle starts — is just a thin plate of steel or aluminum, and then they put it into this massive die and it stamps it right into the shape of a hood. Inside of that die there are what they call sort of wear components — they're the parts around the edges that see the most force, that have the most intricate geometries, that have to bend the metal and cut the metal. So we look at a lot of those parts that they're machining today. A lot of them look like — kind of look like those brackets, you know, they're fairly complex shapes where they're really inefficient to machine. And so we're printing — so a lot of parts that we don't even need to do the complexity of conformal cooling, just printing simple-ish tool steel parts that are inefficient to make using machining right now, for companies like Ford.

So if it's interesting, I'll go into a little bit about just sort of how an aerospace company thinks about a technology like Joule Printing, or any sort of new manufacturing process, and how an automotive company does too. And stop me again if you have any questions.

Tom: [On the value of automotive tooling] The automotive — the value of the factory for parts that are going to go in the automobile — talk about that — are going to go into the dies, after the parts they go into the automobile. A die — you have to see one of these dies, believe it, they weigh 200 tons, they stand about the height of the desk, it's a piece of block of solid steel with contours, they're highly polished. And those guys can cost — a set of dies for — and if one of those cracks or wears out, all of a sudden you're out of production. You don't have a spare die to make that. Okay, and if you're out of production for six months, you can't sell that car for six months. You have to get in there. It's the speed of getting back into service on the die, not a part that's going to go to the customer, what's going to keep you in production. You'll pay anything to get back into production with that die within a week, because otherwise — yes.

Alex: The dream for a company like Ford is to have this machine right in their stamping factory, next to a CNC machine, and if a part breaks, like you mentioned, they can very quickly create that near net shape, machine it, get it into a die. Or just to produce that first die potentially, they can produce it faster, they can get to a part faster, they can potentially launch a vehicle faster. I think it's something like — a completely new designed car, like think of like a Tesla, like starting completely new, right, with a totally new chassis and body design and architecture, I think it's something like a billion dollars to create all tooling for a completely new vehicle. And it's crazy to think that sort of investment goes in before you've even made the first car and sold the first car, right. You're putting that sort of investment into tooling. So any way to improve the efficiency, the speed of producing it, the cost of producing it, it's hugely valuable to these companies.

And we're especially looking at large parts because automotive companies have outsourced almost everything now — all the small plastics and components and even small metal parts, they're looking to suppliers, a lot of it's from overseas. But the big parts, they still want to produce in-house. I mean, the people know why you'd want to produce a big part in house first — produce it in China? Yeah, just the cost the shipping in. So a lot of these large metal heavy big parts they want to produce in-house, and they want to do it efficiently. So this is a tool that they're excited about to help do that.

We also love tooling because when you look at putting a part in a car or in a plane, there's so much testing that goes into proving how robust it is, how strong it is, how it ages, how it deals with different chemicals it may see, with UV light, with thermal cycling. There's endless months and months of testing to validate a new manufacturing process, a new material, potentially to go into a vehicle. And it's even longer for aerospace, as you'd imagine. Getting into tooling, they're a lot less risk-averse, if they have another machined part sitting there right behind it that they could swap in if the part breaks, it's not the end of the world. So you can move a lot faster from a business perspective, you can get into real sort of revenue and production a lot faster than working with a company like Boeing on titanium parts. It may be five years before that part ends up in a plane, and probably for good reason, right. We all get on two [these] planes and we expect a certain level of safety and vetting to be done to any structural component on that plane. And so it takes a really long time for them to do that with a new manufacturing process. So we're kind of doing two things at once because we know that titanium is really exciting and Boeing is really exciting as an investor, but if we just work on titanium for Boeing as a startup, we're gonna run out of money. So that's why we're doing automotive tooling and sort of titanium aerospace as our two kind of launch applications from a business perspective. Yeah.

Student: [Question about hybrid CNC + printing machines.]

Alex: You know, it may make sense to integrate our technology into like a big CNC machine or something like that. Some companies are doing that, basically what you have sort of like two heads, right, one has a mill, one has a printing head. The problem we see with that is a few fold. So one, you're kind of building two expensive machines into one, right. Lasers are expensive or print heads are expensive, or CNC controls and equipment — you're kind of putting two expensive machines into one with sort of inherently half utilization of each, right, you can only do printing or machining at the same time. So there's sort of an economics challenge there with that architecture. Another is that we mentioned all the different thermal dynamics going on in a printing process. A lot of times you have to do a final machining step to get to really accurate tolerances, so doing machining to a somewhat hot part, right — you could have a molten metal part, so you have to wait for it to cool to a certain degree until you can machine it. But you're still dealing with all these thermal stresses in the process, trying to machine in process. Then you get to the final part, you're probably gonna have to stress relieve, when you stress relieve it's gonna move, and you're gonna have to machine it again. So it kind of defeats a lot — I think it defeats a lot of the purpose of doing it in a hybrid approach like that. The place where I think makes a lot of sense with hybrid is if you want to produce really intricate, highly tolerant, smooth internal features you couldn't access after the part is printed. In the sense of like conformal cooling like we talked about, if you needed those conformal cooling channels to be super smooth, the only way you can really do that is to machine it while you're printing it, otherwise once you've formed it you can't access these non-straight channels. So that's my take on at least —

Yeah, any other questions? How do you — just put in an oven, I mean there's lots of different stress relieving programs.

Tom: [On a different process — possibly binder jet] So now it's got a nice pot on that. Of different processing — binder jet and whatever — in this size of the part, and it goes up to about a meter in size. And the meter inside, you're basically looking down — well, not pixels — so you got pixels the size of half a centimeter, okay. You can have very fine pixels in it. Yeah, there this season one on the bottom where you have not fallen — that small soccer ball. You gotta be — here, they're so different. Procedure — every process has its own strengths. There is no one universal additive manufacturing.

Alex: Yeah, we think about just like conventional processes, right. There's a reason that we have hundreds of different manufacturing processes today. It's not this sort of winner-takes-all market, right. Each one is a tool that has its own benefits and has its own challenges in different applications. So when we think about all the different metal processes, some of these are some of the key categories. There is for sure some overlap, but we think for the most part there are different tools for different jobs. Right now, our first printer, we think makes sense for parts coming between the size of a tennis ball and a beach ball. You look at a company like Desktop Metal and their binder jetting, and we think it makes sense for much smaller parts, much thinner parts, because you've got to remove all that binder and then get it down to density. And then you have your big, more welding type processes, where you're laying down like a hot dog size bead of metal in some cases, where it's only going to make sense for big, simple, large kind of structures, mostly aerospace structures, industrial structures. Yeah.

We think we can even do — with a current wire, it's just like, you have to start somewhere, right. We had to build a machine starting somewhere. We chose a size that we thought was kind of in a nice sweet spot that wasn't well addressed with current technologies. And we're also solving a lot of the thermal management issues around how do you maintain heat in that part. I mentioned we use an argon environment — if you try to build a really big chamber, a big machine, it's gonna be really expensive to purge that with argon. So you have to start to think about local shielding, or are you printing in a vacuum, factors like that too. So eventually, yeah, we think it'll scale, but it's gonna take time. Yeah, I'll stick around if anyone has questions.