AM_F2019_05

Next Thursday, we're actually — will be, if all goes well and something doesn't happen to me — will be my last lecture. Actually it's not my lecture, it's actually an extra lecture, it's the seventh lecture in the series, and it will be Alex Huckstep, who wrote these blogs, these eleven blogs on Digital Alloys. He'll be here, I've invited him, he's excited about coming, and hopefully you'll have some questions for him, okay. And I think I said in the beginning, one of the reasons for doing it — well, people are always asking me about additive manufacturing, because you see it in the news all the time, and I found his blogs and I thought it was a balanced presentation for someone who's selling additive manufacturing. He's not overselling it, okay. I think he has a fair approach of, this is the strength in this area, and this is the strength in that area, and this has the advantages over here, but he is still selling it. So, and I actually sit on their scientific advisory board, so you don't have to tell him that I've said some of the stuff is crap. I don't actually think Digital Alloys' approach is crap.

And what I'd like to do next Wednesday is talk to you about what strategies make sense. I mean, hopefully you might have learned something now about all these different companies, and they have different processes, and the processes have different approaches, and they're focusing on different materials. And what's the best approach? They're focusing on different industries, and I think I've already said to you what industries make sense. Aerospace. I'm sorry folks, automotive and shipbuilding don't make sense, for — well, not shipbuilding, maybe ship repair. I told you the US Navy is spending hundreds of millions of dollars, because if you've got a billion-dollar ship out in the middle of the Atlantic Ocean and a part breaks, you will pay anything to get a new part, okay. It has not — it doesn't have to do with manufacturing, it has to do with maintenance, it's okay. And I mean in the military, losing a capital ship, okay, when you're in the middle of a war or whatever, is not a good thing.

And in fact they did, during the Second Gulf War, they had a cruiser — and the Second Gulf War was supposedly only about three days long, right, when they did their final assault — and right in the middle of it, one of the cruisers, they have 3,000 psi airlines that run almost everything on the ship, brass elbow failed, and their whole 3,000 psi airline went down. Well if you lost all the electricity in your home, what can you do, right? You can't run your — you can't, you know, in the old days you couldn't use the landline telephone necessarily, your furnace doesn't work, okay, you can't run your computer, you can't watch TV, everything's built around the infrastructure of the electricity. Well on the ship, that 3,000 psi airline is basically the same thing — everything runs off the compressed air, and all the ship had to come offline, and you're losing one of your major ships in the middle of an assault, okay. It's like, you know, it's like someone hit it with a torpedo and wiped it out. They weren't happy. And it turns out someone had — the brass alloy was the wrong brass, and it failed by stress corrosion cracking. Simple little thing, but it wiped out a whole ship in terms of its utility. But you know, if you could replace it quickly, you've got some advantage.

Okay, so additive manufacturing. If you don't have any questions, hopefully you will have some questions to ask Alex next week. We've gone through parts 1 and 2 sort of, and he listed all these different companies and different processes — directed energy deposition, powder bed fusion, powder binder, and then other over here. And look at all these companies, these are not little bit players. Hewlett-Packard, GE, Trumpf, Laser Coherent Radiation, Lincoln Electric, the biggest welding company in the world, okay. And then you got these little companies like Digital Alloys and Desktop Metals, which have very close ties to this department, okay.

Now let's go through what I skipped yesterday, a real-life example, okay. So this is a blog basically that you can find on the web. There's this guy Don Nelson, called the Additive Report. He writes newsletters just about additive manufacturing, okay. And this is just, what, ten months ago or something. And he writes as — he's a science writer — he learned that eyeglass hinges are a good additive manufacturing candidate. And in a recent telephone interview with the CTO of this additive manufacturing company — and here is a quarter, and there's an eyeglass hinge made by powder bed fusion with a binder, okay — and he says they can make forty-five thousand pre-assembled twelve by six by — twelve by five by six millimeter eyewear hinges in a four-hour shift. Wow.

So what are these things made out of? I mean look at them. Well these eyeglasses are titanium, but you tend not to — in the powder bed fusion you tend not to use titanium powder. A, it's horrendously expensive. B, titanium compared to steel powder is pretty cheap, but — if it's not titanium frames, it's probably brass, some sort of brass, could be stainless steel, okay. And powder bed with the binder is probably fine for stainless, it might be okay for brass and whatnot. But we'll talk a little bit about what this is competing with. But there's the part. Can you see the surface properties of this? And they're talking about, you can take three pieces, this hinge and that hinge and the hinge pin which goes in — and we've all had problems with the screws in the eyeglasses if you wear glasses — and they said we can make them all in one part, and so we reduced the number of parts, which is a very important design rule in manufacturing, to reduce the number of parts to simplify things and save money. But do you think people really want a hinge that has that surface roughness? It's got to be polished up, particularly if one of the hinge pins is this thing. You don't want the thing flopping around, right, and if you got surface roughness like that, you're not going to have the tight clearances you need for an eyeglass.

So aside from that, but anyway, let's do a cost estimate. To make forty-five thousand in an hour, uses a multi-head printer which costs a million dollars and probably is half the size — I don't know, am I sized this room. And if you just do the amortization on a four-hour shift, that's about five hundred dollars a day for acquisition, buying the million-dollar machine and maintenance, and that works out to about a third of a million a year. And if you start looking at them, a mechanical machine requires a fair amount of maintenance, mechanical parts wear, okay, and this thing's moving pretty fast. The materials, if we use three grams of fine metal powder per part, that's going to work out to about three thousand dollars for a four-hour shift. The net cost is — I worked this out, it's about eight cents per hinge, okay. The fallacies are the surface finish, and the fact that — well, I don't really care, I actually am going to make the frames in three parts: this leg, that leg, and the frame for the lenses. And then I'm gonna assemble it, so I'm gonna do the assembly anyway. I don't really need to make three parts together — at least not these three parts. In fact, I might want to disassemble that. How do I repair that if the thing broke? Now if I lost a screw, I could just replace the screw, which is what — okay, so it doesn't really make much sense to me.

Okay, plus, how do we make these hinges today? Well, if you go down here to Attleboro, Massachusetts, there's a company called Bleach and Garner [Leach & Garner], which is the reason Attleboro is the gold capital of the world. They make — they have what they call a fancy wire division. Instead of making round brass or gold-filled wires, which are brass with gold on the outside, they have, if you will, they have a room that's the size of this room that has all these dies, which instead of a round wire, they have all kinds of shapes that look like maybe this, okay. And then if you put a hole in that, that's a big long wire like that, but you basically slice it like a piece of bologna, okay, in a shear machine, and you got your hinge — one of your hinges, right. And in fact if you do it right, you got two hinges, because you put two of them, you got one on each side, right. And then you just use the screw, and screws are cheap. A hinge like that for your eyeglasses, probably less than a penny.

So this is only about ten times the cost, but it's all one piece integrated with a lousy surface finish. And you're telling me that after two hundred million dollars and three years of effort, this is the best example that this company can come up with for an additive manufactured — it's not what they tell you, it's what they don't tell you. You got to learn to fill in the blanks, folks. You got to be critical and say, well, does this make sense? When I learned about this from someone else in the department, they said, you know what they got on their website — well it turns out it's not their website, all you have to do is look up the company and the name of the company and eyeglass hinges, and you'll pop up this Additive Report, okay. That's all in the last year. And they've raised four hundred and twenty million dollars of venture capital. General Electric put in eighty million, Google put in a hundred million I think. And now the best thing, after two years, they can tell me about, is eyeglass hinges.

Oh, rapid prototyping. Do I need forty-five thousand pieces? Okay, well sometimes maybe I do, but not usually in rapid prototyping. So there's all kinds of dimensions: how many parts do I need, what type of material, what quality of surface, and things like that. And that's where these eleven blogs kind of march through some of these. And one of the blogs on surface finish was just done like three weeks ago, and I bet that you've — Alex, what his next blog is going to be — and if you actually walk through the blogs, you're gonna see, where do you get a lot of his information? He goes to conferences on additive manufacturing, or he talks to his customer, okay, and his customer tells him what he needs. And I don't know if any of you heard it, but when Dave Hill was here, he said a lot of these things — he wasn't talking just about additive manufacturing, but a lot of these things are solutions looking for a problem. The eyeglass hinges, here is a solution looking for a problem. There isn't a problem in making eyeglass hinges inexpensively right now. But someone there thought, oh, we could make small parts and we can integrate three into one and no one else could make an integrated one.

Student: [asks whether the existing method makes the hinge in three different pieces]

He's part of three different pieces, right? That's right. The answer is, so what, can you do it economically? The answer is no, okay. Can you do it with a decent surface finish? No. Is it repairable when you actually get it in service, so that your optician can repair it or you can repair it? I used to repair my own, okay. You know, and you start looking at all the attributes that you want in a manufactured product. This is not the best example you should be able to come up with after two hundred million dollars in three years of working on additive manufacturing — at least, I don't think so.

But you know, the other telling thing about this — that machine there is a Desktop Metal machine. You walk down the hall twenty yards, and there's one sitting there right in front of the door, and it's on top of a desk. Mike Tarkanian put it on top of a real desk, okay. And you know how many of their own metal parts are in that machine? After three years and being the manufacturing revolution, how many parts do they use in their equipment that they're selling? Excuse me. That's critical thinking, okay. And if you don't like the answer, ask them, okay.

Now, part of the step in any binder process is, you start with a design. And you know, in the early days, thirty years ago, of desktop 3D printing, the big deal was the software to take a 3D object in the computer and slice it into layers in the computer, and then change that to the rastering for the robot that's doing the 3D printing, okay. That was what everybody from 1990 to 1995 was working on, that software. Well that software you can buy from twenty different places now, because there are writers working on it. And you make your green part — and I passed around a green part the first day of class and it came back in two pieces, because it doesn't have a lot of strength, does a green part, okay. It's just sort of like a piece of candy that you can snap, because it's got frosting and the binder's not all that strong. Then you debind it, you get rid of all the extra powder and stuff, and you put it in a furnace at a lower temperature and you boil out the impurities of the binder, the adhesive that's holding it together. That used to be called binder burnout, and when Mike Cima became a young assistant professor here, he changed that to binder pyrolysis, because "binder burnout" sounded sort of non-techie, okay. But now it's called binder pyrolysis, because you're burning it off.

And then you have to put it in an oven, because whenever you put powders together — I don't care what you do, if you study powders, and the ceramicists more than anyone else, they've been studying it scientifically for over a hundred years — if I take sugar, or salt, or you take your powder and you pour it into a vessel and you measure the density, it's half the density. It's 50% dense, half of it's air. Doesn't matter if it's spheres, it doesn't matter if it's irregularly shaped, unless it's like fibers, okay. But for any kind of sort of equiaxed but angular or spherical or a smooth powder, 50% porosity. So you burn out the binder and you still got 50% porosity. What kind of strength you kind of have on 50% porosity? It's even weaker after you burn out the binder than it was when you had the glue in there holding it together. But you're going to be careful. And then you sinter it. And what happens when you sinter something that's 50% voids? How many voids do you have left? About 3%, okay. You will never — any metallurgist or ceramicist will tell you this, you always get a little bit of porosity left over. It has to do with grain boundary motion and things like that.

There was a faculty member here when I was a student, Bob Coble was — he had been at General Electric as a ceramicist, he came out of this department, went there, and he developed a process to take an aluminum oxide powder, and he could make it basically a hundred percent dense by putting in beryllia impurities that had impeded the grain boundary motion. And when they sintered that aluminum oxide tube, they could get a translucent tube, and it was called Lucalox — that was the General Electric name. And all the light bulbs around the country, particularly the outdoor light bulbs, are high-pressure sodium vapor lamps because Bob Coble invented Lucalox, okay. Which all had to do with grain boundary motion and sintering. And people didn't care if it wasn't a perfect cylinder, it still held the sodium vapor inside the lamp.

Student: [question about 50% porosity]

Yeah. By fifty percent. Well that's the way it works in the service. Now I've got a piece here — this is a piece made on the, I think, was made on this 3D printer at the MIT Institute, and it's now been sintered, so I could drop this on the floor, it's nice and hard. But look for how many right angles that are still right angles that you can find on this. But you'll find that the precision is not so great, and some of the detail on the surfaces — well, maybe the powder wanted to sinter into a ball rather than stay in a straight line, okay. So there's certain surface tension effects in metals, and remember, metals have higher surface energies than any other material — we can go through that — but they do, by about a factor of three. And those surface tension effects will lead to all kinds of defects in the sintered part on the surface, okay.

Student: [question about uniformity of defects]

They won't be even, in a complex part. If I just made a cylinder or a sphere or a rectangular parallelepiped, yeah they're fairly uniform. But if I start making complex parts, some areas will start to sinter before other areas, because you put them in the furnace and the smaller areas start to sinter first. And it does it, and they've never solved that problem. And we knew they never would, when they invested the 420 million dollars — when they signed those papers before they ever did the experiment, anyone who had ever worked on powder metallurgy or powder ceramics knew that part of the process was an Achilles heel, okay. They have no solution for that. Their solution is, it's a slow process to make binder powder — binder parts — but we're gonna have multiple print heads. Oh, what a wonderful idea. But you start looking at the other steps in the process. Yeah, you're right. 50% porosity, you're never gonna get a precision part.

Student: [question, possibly about laser/electron beam powder bed]

No, they — depending on whether it's electron beam or laser powder bed, because if you go back to his little chart of all the companies he kind of — and if you go to his website, he actually has a video of laser powder bed or maybe this electron beam, where you've got the powders, and you can see the process in a video, and you see that it's a very violent process. You're spitting out powders as you're melting powders, and you say, well what kind of defects is that going to produce. But go look on the blog, and you'll see that it actually — because you're melting the metal, you get a better surface finish than a lot of things in the powder bed fusion. That's not a bad surface finish, better than most additive manufacturing parts. But the precision goes to hell because of the shrinkage. And if you melt by directed energy deposition, which is what he calls what you're talking about, or bed — this surface actually should be better, except you got all this little spatter of molten metal drops that land on the surface that end up introducing defects. Should it be a hundred percent? You can get close to a hundred percent density if you go slow enough. And if you try to go fast, you start getting spatter all over the place, okay, molten metal spatter, that's a problem, okay. But if you go slow enough, you can get 100 percent dense. Pratt & Whitney showed that in 1975 when they tried to make the laser disk, right, this is called laser glazing, and they just went around and around.

Anyway, so does that answer the question? None of these — there is no process out of all those companies with all these different processes, there's no process that has all the good attributes of anything. And that's why there's so many different processes. Everybody says, oh I can get rid of this problem, or I can get rid of these problems, but they introduce other problems. So when you're actually thinking what am I going to use this for, you have to say, well what is my application, what attributes do I need in my finished part, what is the problem I'm trying to solve, okay. If you're just saying, I'm a solution looking for a problem, the odds are you're going to go out of business, okay. People don't buy solutions looking for problems. People hire you to help them solve their problems. If you're not going to solve a problem, they're not going to give you any of their money, okay.

So anyway, directed energy deposition has some of its problems, but it also has to compete — and this is not part of Alex's stuff, this is something I put in on competitive metal forming techniques. So if I'm trying to make something that's sheet metal, I could stamp out parts in what we call progressive stamping. And this is actually parts that will be cut up and interfaced together to make a heat exchanger, okay. And there's just two different ones here, and you'll see it starts out with the sheet and it forms little domes and deeper domes, and then those form cylinders, and eventually you punch out a hole, and you slip those together and stack them up, and you got a heat exchanger, and these are the tubes in the heat exchanger. That's progressive stamping. If you've ever been in a shop where they're doing progressive stamping, you see the stamping press going chunka chunka chunka, and you're doing a whole row at a time. That's pretty fast, okay.

Another technique is deep drawing. This is the way we make — we use 40% of the aluminum in the world goes into deep drawing. You start out with a sheet, and you make a cup, there's a little cup, then you take that cup and you make a deeper, taller cup, you take that cup, you make it taller still, and you take that cup and you make it taller still. What's this starting to look like, if I blew it up to the size of a Coke can? Looks like a Coke can, right? Exactly, very good. That's not what they were making. They wanted to make this. I got these from Alva Kaufman, who was the faculty member who taught me as a sophomore in this department. And he was retiring at age 55 the day I came as a young professor, and I ended up sharing my office with his graduate students. These are samples that he had collected over the years.

He was going off to become a rich man selling Christmas trees. He had a farm in New Hampshire, and he decided — he lived up in Marblehead, and he was decided he was fed up with MIT, he was gonna retire, he was going to put his Marblehead home, this is 1976, on the market for $90,000 — ridiculous price — and if anyone offered it, he was going to sell it. First person came by, offered him $90,000. So he moved to his farm in New Hampshire. And his farm in New Hampshire, he had 10,000 Christmas trees. In the spring, he would go out and he would trim the trees, good exercise and nice cool, okay. And he could sell those Christmas trees — he would sell about a third of them a year, and they were perfect pyramidal shape because he had — or, he trimmed them each year. And he could sell those for about thirty bucks apiece back then, which is a lot of money — Christmas trees today, he would be selling them for $150 a piece. He was making enough to live on handsomely by just spending a few weeks trimming trees in the spring, and enjoying living on his farm the rest of the year. But he had another business on the side. He would go to yard sales with his wife in New Hampshire, and they'd pick up junk, they'd take the junk to New York City and sell it as antiques.

Student: Markups about ten to a hundred, right?

What's one person's junk is another person's treasure, or whatever the saying goes. Anyway, they were trying to make these little cans, which are rectangular cans, which they made from that other thing. And they made these because when you had vacuum tubes in the old TVs, you sometimes wanted to put a can around it. This is the can for the vacuum tube. There are one two three four five six parts here, which you can pass around — you know, either that — I have some others that are bigger and heavier, but these are easier to pass around, okay. [Tom passes samples around.]

There's another process, and I actually, back when I was a young assistant professor, I went to Pittsburgh once and I don't remember why, but I went through a plant that was doing cold heading. Now, cold heading was a process developed after World War Two when someone learned that if I put a phosphate coating on steel — and how do you put a phosphate coating on steel, you just dunk it in phosphoric acid, okay. Phosphoric acid, is that hazardous? Well, you know, it's Coca-Cola, right? Coca-Cola is just phosphoric acid. So it's not all that hazardous. If you had — too many. The other day I had a classmate here who used to go through a case of Coca-Cola a day, and then he ended up in the infirmary with anemia, because he lived on Coca-Cola and M&Ms. Couldn't figure out why he had anemia, okay. But if you phosphate-coat this — you can take a wire, this is 3/8 inch diameter wire, phosphate coated, and then put it in a stearate bath. Stearate, what's that? It's soap, okay, same type of thing. You feel stearic acid, it's a stearate.

That's the sheared blank. You first just square it up, and you put a little indent on it. The next one, you start to make a hole in it. The next one, you make a deeper hole — you'll have these — done. Next one, you go deeper, and you put a flange on the bottom. The next one, you actually punch out the bottom of the hole, so now this is the only piece of scrap, and I'll leave the tape on here, but it's just a little piece of metal. Then you put a knurl on it. Then you thread the inside. Then you plate it, to give it some corrosion — galvanize it. And in those steps, you've gone and made a metal insert that you can put into a piece of wood, drill a hole, hammer it in, and now you can put a machine screw into the piece of wood to assemble furniture. This is a fastener. They make that in about — little over one second, because that machine goes seven times a second, hammer blows bang, and you have never been through a plant any louder than that plant, okay. Just machines hammering away. Called co— cold heading. Almost near net shape — some of it is near net shape.

The only other cold heading plant I've been through is a plant that was supplying steering gears for General Motors. These were parts, some of which weighed five or ten pounds, and some of those were warm headed. They would heat them up to three or four hundred degrees, because to try to do them cold — now you wouldn't want to touch one of those when it came out of the machine, friction had made it a little hotter, okay. But it was the ability to lubricate that surface that allowed them to start cold heading machines. All the nails you've ever seen and most of the screws — if you go to the website, if you go to Google, you'll find a company called Fastenal in Rockford, Illinois, and they take you on a little tour of their plant, okay, a video tour, very good tour. You're gonna make those parts by additive manufacturing? No, I don't think so. That's been optimized over 70 years, you're not. That's called heading, okay.

But stamping, progressive stamping, all the deep drawing — that's the deep drawing, or drawing and ironing is actually the beer cans or the Coke cans — all these different processes of making things, they're a lot faster. Pixel-by-pixel manufacturing cannot compete unless you have a functional advantage to your part, okay. If it's just a paperweight or a little fastener that holds a piece of wood furniture with a machine screw, completely weakened, or an eyeglass hinge, okay, you're not going to compete. You've got to have a functional advantage. And you'll see later in things like General Electric — and General Electric in their aircraft engines, they're saving tens of thousands of dollars on a part that used to be twenty parts that were welded together, and now they can make it in one part with all these little nozzle designs, and they'll get more efficient flow in their nozzles and better fuel efficiency. Well that's worth millions of dollars, okay. So yes, additive manufacturing makes sense, and you can do it, and you can pay five thousand dollars for a little part the size of your fist, okay. But you can't make an automobile part okay, or some pixel-by-pixel.

If you go to what's called near-net-shape manufacturing, the Air Force in the 80s and 90s spent billions of dollars trying to make things near net shape, and we'll talk about that a little bit later. And these — if you just looked it up, I think, on Wikipedia — these are some of the technologies. Cold forming is the type of cold heading. Super plastic forming. If you want to know about these forming things, I actually give a module on deformation processing. But again, pixel-by-pixel manufacturing cannot compete without a functional advantage.

Remember I said there's structural materials and there's functional materials. Structural materials are made in very large volumes. You're not going to make steel I-beams by additive manufacturing. You can make them by rolling. That's how Bethlehem Steel — has now defunct, that we're [were] first company that hired me — they became the second largest steel company in the world, because they learned how to roll I-beams, and they built New York City. The tallest building before Bethlehem could roll I-beams was about ten stories, because you had to make riveted I-beams and things. When they learned to roll them on the mill, all of a sudden a cheap process to make I-beams, efficient use of steel, you can go up 100 stories, okay. And so some people consider that rolling of structural shapes was the technology that — Bethlehem [had] two technologies that built this second largest steel company in the world. One was structural rolling, structural beams. The other, they came up with a type of tool steel that was better machining than anything else, very high hardness and stuff, and they got a patent on that.

Student: [question about additive concrete construction]

Yeah, I actually, I think I have seen it, there. They're just squirting out concrete, okay. I'm sure you can do it, but, you know, I think I can pour concrete into forms cheaper, okay. Right. One of the ways I think of it — I like to think of it is, if I have a plane, and I am trying to pass material from an unformed shape to a formed shape, I could do it by making liquid metal and pouring it into a mold. What's the flux of atoms across this plane as I'm pouring the metal into the mold? And if you take my casting module, the first class, I compare that to vapor deposition, because for whatever reason ten years ago, everybody said we're gonna use molecular beam epitaxy to grow things and blah. The fastest you can grow something by vapor phase deposition is about a millimeter an hour. It's actually about a half a millimeter. You can go through the kinetic theory of gases, it's only so fast that you can bring atoms to the surface, and if you try to bring them to the surface faster than a millimeter an hour, it's not a vapor anymore, okay, it's gonna start to liquefy, okay. And if it's a liquid, well, you might as well just pour it, okay. But by vapor deposition the fastest I can grow something is about a millimeter an hour, so you're not going to make very thick coatings, in general, by vapor deposition.

What Giancarlo is talking about, okay, I can have a little pixel of cement that maybe the size of a golf ball, and I can squirt that down pixel by pixel. But what if I actually had a form, and I poured the concrete? And you're gonna have to have a form, because the concrete that you do pixel by pixel still has to have some support, if you put it in there very fast, right.

Student: [follow-up — apparently noting the additive concrete process doesn't use a form]

So they don't need a form, but I bet the surface roughness is — piece of, you got to come by and you got to give it a skim coat. Yeah, it's a horrible thing. So there's advantages and disadvantages. If you're going to build one house of a unique design, that concrete process might make sense. But you know, Thomas Edison once started to market concrete homes, and on one of my websites, son's, on materials selection, I have pictures of Thomas Edison, and he built some of these two-story houses out of concrete. He was trying to come up with a cheaper way to build homes. I think some of these homes still exist down in New Jersey, okay. So there's again, you can do it, and you know, the people who invent that process, they'll go out and they'll try to find some Additive Report — die — some reporter, and he's gonna say, oh look at this great technology, okay. Where's the critical thinking behind that article? Yeah, it is — it's great for building a one-of-a-kind home, okay. Well I guess people want one-of-a-kind homes, okay.

Other questions? So another thing — I'm not trying to rush this necessarily, but part four, design rules. Well, one of the design rules, there are limitations on the size of what you can build, because most of these processes — not all of them, but they build up residual stresses. And you can see what happens with the residual stresses if you're trying to electron beam or laser glaze or melt the surface, you're gonna end up cracking, okay. That 1922 January 1922 Welding Journal article, residual stresses and cracking, okay. And later Alex goes through and he has a nice little plot — may not be his plot, he probably collected a lot of this stuff from various sources — and it shows the maximum size when you get to blogs 9 or 10.

Now he has — I like quotes — this is by Antoine [de Saint-Exupéry], saw an expert or — who is he? La Petite Prince [Le Petit Prince], The Little Prince. He was a writer in France. His book translated into 300 languages, sold a hundred and forty million copies. It's about this little prince that visited different planets in the universe, and one was a planet where he learned about loneliness, and another he learned about hope. Anyway, but Saint-Exupéry was a male pilot, he flew airmail planes before World War Two. And so he had some engineering knowledge. He says: "The designer knows he has achieved perfection not when there's nothing left to add, but when there's nothing left to take away." Well, of course, anybody who's working on airplane parts knows that you're looking for the lightest weight, and your perfect design is when you've shimmed the way — you whittle away every little piece of extraneous part to get the lowest weight for still adequate strength, okay. And so he has this nice little quote. He's got a number of — people like to quote him. He was a very articulate airline pilot.

Okay, so here's the economics, this is part five. And what he's done is he's looking at titanium. Why is he looking at titanium? Well, that's their bread and butter. Oh, I passed around my little World Trade Center thing — virtually every class, right? That's their part, they're making titanium parts for Boeing, or they hope to sell parts to Boeing. He says cost is the largest issue gating adoption of metal additive manufacturing. That's true for most things, okay. I remember a very prominent faculty member of this department, when he was an untenured associate professor, we were walking across campus over towards the Sloan School, and he says to me, "Did you see Joe Clark's article in the Journal of Metals?" I said yeah. He said, "You know, I never thought about it before, but cost really is important, isn't it." I thought, woo-hoo, okay. I said, yeah it is. But you know, now he's one of the most successful people, because he's learned that cost is important, okay. He's started many companies now and stuff.

But they're looking at titanium here. And so he's broken out laser powder bed, electron beam powder, directed energy wire, directed energy binder jetting, and their process of Joule printing, and lo and behold, their process is cheapest, okay. But it's dollars for cc, okay. If you change it to dollars per kilogram, $1,500 a kilogram, eleven hundred dollars a kilogram, of course there's cheap. This is just the same thing plotted on a different axis. And I did the cost ratios — I showed them to you yesterday, I cleaned them up a little bit yesterday after class — and so here's titanium, really about nine hundred thousand dollars per cubic meter in the wrought form, thirty-two hundred — and steel five thousand per cubic meter. Additive manufacturing steel is cheaper powder than titanium, and the ratios are still five to one and six hundred twenty-five to one. I think this tells you why they want to focus on titanium. They only have to come up with an advantage. They get sort of the five to one disadvantage, and you'll see later in his blogs, they've really done it, okay. Boeing's really done it, okay, when you look at the overall system cost.

And then he's got a plot — we won't dwell on this — but cost comparison of the different things per unit printing cost. This is just for printing, and of course there's not the cheapest, binder jetting Desktop Metals is the cheapest potentially. But this is the printer utilization time. In the binder jetting things, these only have about a thousand hours per eight-hour shift for a year, which is about half time for the eight-hour shift. That's because you have — these are mechanical machines working very rapidly, there's a lot of maintenance time, okay. These are not machines like — you have a steel caster that operates for three years without shutting it down, okay, for repairs, major repairs.

Now if you start looking at the wasted material — this is actually the final part made out of titanium, I believe, this is in his article, and it's some sort of bracket. And they found that they can start with a plate of the same material, build up some things on the top, and then start to machine them away and end up with this part. And you say, well gee, that's a lot of work — except if you look at what's called the "buy to fly" ratio, and this is if we're talking about aerospace parts, this is a very important parameter. The Air Force has been calling this the buy-to-fly ratio. Buy-to-fly ratio, dollars per kilogram. They're claiming Joule printing, their process, is flat depending on the buy-to-fly ratio. That's not really true, but you know they're going to present it the best way. But if you talk about hogging the whole thing out by CNC machining, the buy-to-fly ratio raises the price to three thousand dollars a kilogram at thirty-to-one buy-to-fly ratio. There are a lot of parts in an aircraft, particularly military aircraft, that are thirty-two to one buy-to-fly ratio. That's why the Air Force spent billions of dollars in the 80s and 90s on net-shape manufacturing processes, which didn't include additive manufacturing, okay.

Now here's a buy-to-fly ratio of a real part. Titanium 6-4, this alloy we made over here at Watertown Arsenal after World War Two, the most common high-strength titanium alloy in existence right now. Here's the final part, it's a hinge or something. It's a T-shaped part. In the old days, you would make this out of a solid block of titanium, and you'd have a seventeen-to-one buy-to-fly ratio. 95-point-four percent of your material goes as machining chips. Yes?

Student: [asks what "buy to fly" means]

Buy-to-fly is how much material do you buy from your material supplier, to how much weight do you fly on the aircraft. That's why they call it buy-to-fly. Sorry — no, thanks for asking that. Probably wasn't clear to some other people. To me it's second nature because I've been using it for thirty years. But what's the weight of the part you buy from your material supplier? So they buy a titanium plate from TIMET. Titania — TIMET. To give you a difference in the scale of these industries, TIMET is the world's largest titanium producer, and they can't afford to build — to buy a rolling mill to roll titanium. We don't make that much titanium. Or they can't sell that — we could make it, but they can't sell it at the prices that it goes for, they would. They purchase one day on the Republic Steel rolling mill per month to roll all of their titanium for that month, okay. And given the fact there are probably 12 plate rolling mills in the United States just for steel, you can figure out that ratio, okay. One month — or one out of 30 days, because these things run 24/7 at a steel mill. When you invest a billion dollars in a rolling mill, you keep it busy, okay. But all of the time, all of the titanium production in the United States, titanium plate production, can be made in one day in one month in one mill, okay. So when we talk about scale, you got to remember there's huge differences in scale of things, okay.

And this is from Leo Christodoulou. I didn't know until yesterday, when I was actually looking at the blog — Leo's wife Julie, doctor, both of them are doctors, he used to be heading a lot of the materials work at DARPA, the Defense Advanced Research Projects Agency, giving away hundreds of millions of dollars on advanced manufacturing. I learned that he's now chief technology for additive manufacturing at Boeing, and his wife used to give me research too at the Office of Naval Research, okay. "Historic aerospace buy-to-fly ratios could range from 15 to 1 to 30 to 1 on the high end depending on the shape of the part and the manufacturing process. In the next few years we expect additive manufacturing to significantly reduce these ratios while also being cost competitive." He is not an idiot, okay. He used to give away hundreds of millions of dollars on advanced manufacturing, and now he's working with Boeing to figure out how Boeing can make use of it, and tomorrow we'll talk about some of that.

But just to give you, in the last couple of minutes, here's a picture of what Desktop Metals — not Desktop Metals, but who is it, we're [it's] Digital Alloys is doing. This is their process, not making little tray towers, but this is probably sped up a little bit but not too much, but they're making some complex part that they're going to machine down to another part. But they can make that in minutes, okay. And the amount of titanium they're buying is a lot less. And they go through the economics here, and it wasn't Alex Huckstep that did this, this is Boeing has done the cost-value analysis on how they make this part currently. Oops, I didn't hand around the sign-in sheet, sorry. Anyway, so tomorrow we'll talk some more about this. And they show that you can make the part in eighteen days versus forty days for CNC machining, and a reduction in cost of a factor of two and a half when you look at the whole system. So additive manufacturing works for these parts, doesn't work for high-class hinges though, okay.