Designing in Color – Greg Paulsen on 3D Printing

macro gives you more electronics manufacturing options so you can get to market faster now with one platform you have access to several factories in North America for all your manufacturing needs. Skip the line by using factories with an immediate manufacturing capacity for PCB assembly and system integration. That's also called Buildbox. Go from prototype to high volume production with one manufacturer macro fab and take advantage of our lower costs through automated platform driven processes and relationships with key part distributors and manufacturers Learn more at macro fab.com.

Welcome to the macro fab engineering podcast. I am your guest, Greg Paulson.

And we are your hosts Parker Dolman.

And Stephen Craig.

This is episode 197.

Greg is the leader of the application engineering team at xometry. An online instant quoting platform for custom manufacturing projects, which utilizes a professional network of 1000s of manufacturers. His background is in applied additive manufacturing, having experience as an engineer running machines and ultimately using the parts. Greg was last on the podcast for Episode 181, where we talked xometry pizza and digital manufacturing.

Welcome back, Greg.

Happy to be here. Ben been listening even more since the last podcast, I'm a few episodes behind. But I think we are. We are moss buddies like manufacturing as a service, like enthusiast, and I found kinship with macro fab.

I've not heard that before or heard it said that way moss buddies, I like that,

like a taco bell live Moss, you know, we can just go on and on.

So this time, we're gonna be talking about 3d printing. And kind of like the design aspect of 3d printing in you know, it's not just like machines and, and tools and stuff. But like actually, like, the specifications that go behind, like, building your thing out of 3d printing processes.

Yeah, it's exciting to talk about. And that's really what it's about, right? So all these things are a means to end. So a lot of times we focus up in the sky on processes, and I've even seen in the past, like, oh, well, what's difference between FDM and SLS? And really, the question is, what the hell is FDM? Or SLS? You know, what are these acronyms that stand for? So I think comes down to kind of knowing which, what these processes do, right, you know, and what does the parts look like? And then we can actually go back and figure out, you know, how to actually design well, for each one of these processes and when to use what

you imagine since they're basically these processes are different 3d printing topographies would be a good word for it, I think. Maybe they call them

like family, sometimes families TM, yeah. Methods. And yeah, so I mean, it all in all, like 3d printing, I think by by definition, and I kind of use a terry Wohlers definition, he's an analyst has been the industry for forever, but by definition, it is the CAD defined generation of object, typically layer by layer, by carrying or fusing a lesser material material together. So fusing a filament, fusing a powder with, with laser, for example, curing liquid resins, you're taking something that is kind of this a morphic material. And you are using some sort of energy Asian to bind it together, and kind of stack it up on itself and monitor itself. So we are talking about it's not just giving us layer side by side to glue to glue it together, it's actually to fuse underneath that gives you your third dimension, hence 3d And, and build these parts up. And what's unique about that, versus you know, milling or injection molding is that, you know, milling I'm taking from a larger stock, right? I'm taking a, you know, a piece of billet material, a rod stock, and I'm forming, shaping, bending using the dye and sometimes and I'm reducing it down to a shape. And that means that my mindset when I'm doing something like machining is almost the mindset of a, you know, like thinking in cylinders, like, what can I make my drill do like how can I go and buttered up to the side of this part to make a feature? How do I like how can I access this by making something that has almost a linear projection, and everything that we know? Like everything saying that your your paradigm is around right now is around this idea of directional approaches to manufacturing. So even when we look at molded parts, if you're doing an injection molded enclosure, for example, injection molds are CNC machined. So the basic principles are the same from a, you know, from standpoint of design for manufacturability has to do with access, how can I access this with a tool that's protruding from something and get into the build as features. So when I build complexity, with traditional manufacturing, CNC, machining, molding, or like, I'm adding axes, I'm adding specialized tooling, I'm adding angles. But when I put them on the end and say I have this a morphic material, you know, I have a, you know, my 3d printer, I'm actually able to think more about it's almost like a routed method like, like tree roots, tree branches, like where, how can I stack and build and grow out. And it has a completely different mindset for building features. So it's something that is almost similar by tribal knowledge, a lot of software's catching up. But it's a very, very different mindset, from approach to building especially designed if you're designing specifically to build parts and additive for the for its full life.

Something though, that is interesting, like, yeah, for sure, when it comes to subtractive manufacturing, like CNC, you do have to think of everything as how's my bit actually going to get in there and do it but but there is still directionality with 3d printing, because the bed or however the the the actual layers are deposited, still comes from like, particular direction. You still I mean, you do have to think about it in a in a in a different way. And and certainly what I've seen in the past is, there, there almost seems to be like an optimal way to play something in a bed, something that you get, you get different prints based off of different directions, which is sort of different from subtractive. Right?

Yeah, absolutely. And like, so a lot of these traditional approaches for the vast majority of these processes, like what I think about, you know, FDM, which is very akin to what most desktop 3d printers are, and that's where I have my spool of plastic filament, you know, I'm melting it out and extruding out with you know, you know, a small nozzle, that nozzle was basically being controlled. To do its best interpretation of what your cat is. And this will go to tolerances in a second here, but it's the the machines trying his darndest to make that shape. And, yeah, you're right, you're working in a world of gravity, you know, you're working with a world of constraints, and you're working with a thermoplastic. So plastics themselves, they behave differently as a cool, right. So just like injection molding, where you can, if you have uneven wall thicknesses and things like that, the actual cooling of the part can make your make your features curl up and warp and sync used to have some of those buttons more like kind of a micro level. So like, it's almost like doing a bunch of little micro welds when you're building together. So with FDM, so that Fused Deposition are sometimes called fused filament fabrication, you are fusing a to a build table. Because one gravity exists, you know, that's, that's the thing there. But to your you're holding that, that structure and actually needs a almost like a a datum to reference off of. And when you're building, building parts up from that you're building on a plane by plane basis. So like, we always call them slices, which, which I'm taking my 3d model, and I'm chopping it into these little 2d images based off the layer heights of my layer height is 10 1000s. Like, every 10 1000s, I have a 2d image cross section of that part. And, in general, like when I think in 3d, like for 3d printing, and 3d setups, a lot of times my mind instantly tilts that part. It instantly tilts a tilted to a 45 degree angle, too, because sometimes to your point, it's, there's their optimal pathways in order to get as many features to build what we call natural, which is unsupported. versus building, if a lot of times, if you have a safer design, and you just set a flat on the ground, it's going to have a lot of areas under under shadow, and that means I have to actually build sacrificial structure underneath that support structure to compensate for them.

And it takes longer to do it that way.

It takes longer sometimes because I'm doing the support structure. And yeah, it's and so I can be a little bit fancier by mitigating the supports on this. Now on a desktop, it's more of a kind of like a pain in the butt. And, and also it's also a challenge because I have usually cosmetic differences because I actually have to kind of peel off that structure and that structure is made of the base material I'm already using. So if I'm making my part out of let's say like, you know, nylon or Yes, I'm building my supports out of that. And then I got to kind of snip them off and say that down, I have some cosmetic issue. So sometimes I like to mitigate support just because of that. But ultimately, with, you know, with industrial platforms was my, my experience, you know, 12 years, you know, working has always been in industrial, manufacturing additive manufacturing platforms, you were angling that to build natural, again, like to mitigate that support structure. And, and also, we find that just adding an angle to any of these three dimensional parts, helps give the properties of the part itself a little bit more how I say it, I'll say a little bit less extreme differences. Because if I take something that is built vertically, like a pencil, if I if I set that pencil vertically, it can snap very easily, because the between layer adhesion is just that little circle, going through. But if I build it at a diagonal, sometimes that layer adhesion is a lot stronger, because now you have this kind of like oblong oval shape at the cross section. So sometimes it's not just about, hey, you know, what's most efficient? It's about, hey, what's the function of this? Like, I'm actually using this part like are, you know, you know, obviously, I work for xometry, and our engineers who set out to do these build setups, they are looking at, and they're looking at feature sets kind of thinking, what is, you know, what is the customer mentality? And they're actually oriented? It's not just for what's most efficient, or saves the most cost? It's actually more about what's the function of this? Like, are there snap tabs? Are there the features and the learning? Orient based off? What's going to resolve the best for that?

You know, actually, as a little bit of a side tangent, Parker and I were discussing this a little bit earlier. I'm curious. So as an engineer, if I knew I had a part that I wanted to be 3d printed, and let's say let's go with the pencil example, how do I specify to my manufacturer don't print it vertically? Print it horizontally, or at an angle or something like that? What's the best way to get that point across? If I have the secret sauce of what direction it should be printed?

So I mean, are

expand on that? Sorry, Greg, expand on that would be like, you know, like, let's say it was a brackets and you know, the voting force is going to be the certain direction. Yeah, how do you explain that to a, a shop to get that made.

So I used to do this. So my background before xometry, I worked in product development. And I was the guy making drawings. And it was actually very interesting, because the way that we worked with 3d printers, they always treat them like a net shape. So the actual part like if you're used to running the ERP, the actual part ID was a part that had post finishing, like was drilled, milled, sanded down and to its final form. And then we had a sub component, which was the STL file, which I would send out to the 3d printing. So like, the demand for that part will require one 3d print of this and we treat it as a single item bomb is a single item bomb, but it was the only cheat I could do with ERP to make 3d printing makes sense. And, and I would wish that it had its own individual drawing. And the drawing usually had a kind of a Z, like little z with up arrow pointing for z direction. And that was actually a pretty clear cut way of communicating what's important and even like funny and important dimensions, critical dimensions, which is something that again, it's for reference more than anything else, but it's something to know like what's important for this, a really good example is again, we're, we're talking about something that's moving in the x y axis, and that it has this Z travel, whenever I lean things, there can be a little bit of wobble, although these machines are actually really precise and really repeatable. But just just, let's just say that if I build a cylinder on a diagonal, or build a cylinder, inside or below, build a cylinder vertical, honestly, the vertical is always going to come out more consistent with the roundness of the shape. So sometimes you have features, especially with additive manufacturing, where I have multiple angles or directions of stuff happening. So I take a multi part assembly, right, so a lot of times you can take a look at assemblies and say you have a little call it that gets tightened up with a, you know, with a threaded insert in the screw in one direction that's holding a camera and then you have, you know, kind of an off angle feature that is, you know, say a venting port for example, and, and you have a you know, another thing that actually hooks up to the main chassis of your body that's going you know, perpendicular to everything else. All of a sudden, it's very important to communicate to your, to your vendor or to whomever you're working with. This is the most important thing, because I may actually change my whole orientation around that most important thing to make it print vertical, so that it's been scanned in that kind of the xy direction, which has the most stable control to it. Sometimes again, it's inherent like sometimes you just like, when I look at a part in CNC milling, I can pretty much tell you if it's going to be a lathe part or if it's going to be a CNC milled part just because all sudden, if I start seeing lots of trends of axial directions, then I'm probably gonna say, Hey, this is a turn part and I know which direction is going to be. But being able to be able to communicate that tribal knowledge is sometimes more more difficult. And that's where a quick drawing or just a little note saying, this thing is critical. Like, that's all you need to do, because most of time, these texts, they've made more parts than you ever make in your life. And they kind of know the best way to orient. I'll tell you that there are exceptions to this. So this is, say I'm doing an FTM. And say I'm making something that's the size of a dinner plate. But it just so happens that the dinner plate has a cap to a, you know, another dinner dinner plate piece, and it has snap tabs. So this is where we run into trouble. Because like I said, if I build something that's thin, like a thin and cantilever vertically, in a lot of our processes, not all these process, but for example, that fused filament, it's going to be a lot weaker, and that's where you kind of have to make this trade off. Because building the dinner plate vertically will probably raise the price of it by an order of magnitude. Because all of a sudden, you have, hey, you guys taking 12 inches of my Z and and if I'm running at, you know, 10s out layers, that's a lot of layer, layer start stops a lot of cleaning for the nozzle, and the most, most expensive thing is not material. It's time. Yeah, these machines are often you know, several $100,000 from the service level. And you're talking about, you know, an hourly overhead rate that will easily consume any material costs that you're concerned about, on these on these platforms. So one of the things that we kind of raise our hand on if you sold me that that piece, and then you're like this is important. Also, these snap tabs don't break won't break. That's where I raised my hand and say, Listen, I, I have six other 3d printing processes that may help you out here. And that's why I switch you over. So when I think about FDM I think about parts that are designed that work well for CNC machining, but the second that we start going into enclosures, housings, things that have more organic shapes, I start to lean you towards things that have this more like more of a morphic material, like selective laser sintering, or Multi Jet Fusion, which basically which is based off like a powder bed system. So there's also design cues that may change change over that way. So FDM when you're designing you have like I said, you have a very stable x y, you can build long broad flat features, you can we are proceedings can build up to 36 inches, which is you know, it's a problem like industry big for 3d printers. And, but But it's weak, its weak on the small pieces, like it's weak on the small stuff, the small detail features, you know, cantilever things, I call them the god pins where you have this beautiful, you know, party cost $1,000 And then you have a, you know, quarter inch long pin that protrudes you know, three quarter inches vertical, and snaps up immediately. And at that point, I'm like, that's when I kind of sit back and I'm like, if your part will if the gravity like literally the weight of your part will break a feature on your part, then you may want to look at make that feature replaceable, maybe design it as a whole and you know, buy a 20 pack of pins from McMaster that you can stick in there and you know, sacrificial pens all day long. But the you know, there's there's things that as FTM is really good at. And then there's things where all of a sudden, it really really starts poking to another process. I'm an SLS fanboy like I have that. And that's because when I when I was running, running machines myself, I was running a selective laser sintering machine. I actually got exposed to that in grad school at James Madison University. I think that machine ended up at Virginia Tech now but at that time, we had some we had it there. And I just fell in love with this stuff.

So SLS doesn't require support structure, it it is actually a laser that's fusing nylon powder together to make those shapes. But instead of having support is kind of self supporting its own heated material. So if you think about this powder, like like a flour, almost everything is heated up in that chamber to close around 140 Celsius. And this chamber is it's it's nitrogen controlled. It's inert gas, like this whole thing is just a giant, temperature controlled sensitive oven. And you're going layer by layer, but when those little selective cross sections are being melted by a laser, it's not like this full on blast melt. It says gentle nudge from a, you know, from a powder state to a melted state.

So it's already almost at the transition Yeah, because it's

usually melting right around that it's like kind of like 160 s or so by temperature wise, that Celsius and, and so you have this, you have this powder powder bed. And since it's so hot already, the material doesn't kind of flex up. So say I had a completely cool powder, like completely cool room temperature bed that does the same thing with a laser, what you'll see is they're almost like wax kind of peeling up, like curling up on itself, because the second it hits it with laser expands and then wants to contract. And when it starts to contract it, it's kind of like you know, doing a curl up, you know, it just starts peeling up on itself. And we actually call it peeling, by the way in industry, because it can give you a very bad evening, because it'll jam your rollers and things. But the. But by keeping everything near melting temperature, it just kind of chills it, it stays, stays in the same place. And as my Z layers get stacked on and on, I'm able to continually build. The other beautiful thing about this is since I don't use a Porsche structure, I'm not constrained by that, you know, gravity, if you will, that you will have most of these processes, you're not just sticking it to a part bed. So now an SLS and Multi Jet Fusion works in similar way different melt way different paths and melting those parts, but very similar process. And even the base material is exactly the same. They are kind of floating in space. So I can actually nest multiple parts together. So SLS tends to be cheapest of all the 3d printing process from an industrial standpoint, because it's not that the process is cheap itself. It's not that the machines are cheap, it's that I could just build so many parts of actually get scale with it, I get an economy of scale with those processes. But I like it because I have that freedom of design, it tends to be more forgiving in that z direction, tend to have more within isotropic results. It's not perfect, but it's better, it's a lot better, especially compared to FDM, which just has these extreme swings, depending on what direction of travel your, your designs going. And, and also, because I've done this, you know, you're able to kind of treasure hunt to get the parts out. So it's also it's kind of fun and almost therapeutic to clean out in the morning. So that's other things I just I remember so fondly I'm sure a time has held because you're talking about it even even as it's cooling down. Sometimes it's the material is like near boiling temperature, when you take out this giant block of what we call like cake, it looks like a big rectangle of white powder. And I remember in the morning, we just happen to have one of these at our engineering firm, we would take out this powder and and I just put two, two nitrile gloves on. And just hope I didn't scald my hands because I just thought like, Okay, I put on my earbuds start listening to like audio book, put on my cans, put on my respirator and just go to town and you stick your hand and near boiling hot material as your as you start to like dig through for your parts. In reality, like if you're a service bureau, you usually lead a day to be like, like 18 hours of cool and do all that. But I was working in product development. So you know, every every hour counted. So we were a little bit more rushed on the jobs. also lower the height. So it wasn't like insane, but it was uh but yeah, it's it's, it was very cool because it gives us quickly quickly iterate designs, and since it was so cheap on a per part because you're able to stuff the machine full of parts, you can start thinking in shotgun approaches of design. So this was designed for injection molded a lot of times that we're using that Celeste as a surrogate for for rapid prototyping, but we would try six configurations, just put them all at once. And then give them to our stakeholder and be like choose to or choose one and we'll get we'll start working for that. So it was our own genetic algorithm of figuring that out, you know, trade off, right and running an evolutionary pattern, you know, calling five of them going on this one, six more configurations, go and go. Yeah, if you have

if you have the bed space, and then that totally works out. Yeah, sure. So

yeah, so the big difference between least from a design standpoint, from SLS to FDM, is you don't have to worry too much about overhangs with SLS. Correct.

Yeah, you don't have to worry about overhangs. Now. Let's go back to

here's the but yeah.

That being said, it's all it's all fun and games until the everything starts cooling down. So here's what can happen. Say I Say I'm using money in typical platform for SLS machines, although there are like largers that kind of like double size platforms. The ones that I prefer tend to be about 13 by 13 by 23 inches. And the biggest reason why is you could stuff it full and still be done in about 26 hours, plenty And from a from a process workflow standpoint, if you're running it as a as a service industry that allows you to essentially have, you know, one set of one one builder parts being made, what these check these trades can be changed out, one of these trades can be cooling and the one of them can be broken out. So you could actually have a perfect little three day schedule with one machine parts being made, cooling broken out. So like you could have very consistent three day lead times with these type of parameters. But the so I start printing and say I have a part, I'm 13 by 13 inches, so you have a 14 inch part. And if I am building a 14 inch part, I'm building a vertical. And I'm building this. So again, this build could be you know, when I first start that first layer, the first four and a half 1000s layer of that part, it means that final part may not be complete till 18 hours later. And a lot of weird thermal stuff can happen during that time. So if the part is designed in a way that's self supporting, so say it has kind of like a lattice structure or like kind of a C shape to it, or S shape where it where it has kind of features that if you if it was the same shape and paper and you sit on a table, it wouldn't fall over, then it may be more thermally stable. But say it's just kind of the, you know, a long, thin feature, all of a sudden, I get very strange stressing and crosslinking across these parts, and especially once the parts started cooling out or outside the build, I could get a lot of weird stress and I started to see warping. So warping occurs, and it's very, it's probably the most frequent complaint about the SLS platform is going to be warping of long, broad flat parts. And, and again, like a good rule of thumb is if I can grab a feature of my of my part and the rest of the part sags, then probably this this feature is going to be prone to warping. Because it's not it's not supporting itself. But in general, sometimes it's just something of unavoidable. Because I have this, you know, I have this unconstrained it's not fit by support structure, it's not, you know, essentially bound down to a build table. So I have I have it kind of de stressing and cooling in a fried environment. One of the things that I kind of talked about when I do webinars and other series with 3d printing processes is SLS, I talked about the rule of the fist, which is the best size part for SLS, even though I have this 13 by 13 by 23. Build area is basically parts around the size of the fist, tolerances tend to be very stable. For the for the processes, your feature detail tends to be, you know, generally good. And I can stack a bunch of them in a machine at once. So my scalability, my price is good as well. So you don't always want to kind of focus on the like, hey, what's the biggest footprint? That may not be the right question. It's you know more about like, what your project is, you know, what your budget is and what your goals are. And like it, you know, sometimes it makes sense, even like if you're building a jig or fixture, sometimes it makes sense, just parse it out into smaller, smaller pieces, just because it actually tends to be more economical and sometimes prints faster because you don't have this giant Z constraint behind it.

But I just want to note on Multi Jet Fusion, because it's in no way is an underdog it's actually a fantastic performing machine. It's an evolution of what SLS has been SLS is was kind of invented in the late 80s. And it's been around for it's been around for quite a while. And it really has been perfected in a lot of ways. Multi Jet Fusion, which is on by HP has taken a kind of concept of high speed sintering, which is a different way of fusing materials together. And it actually is adding a it does a single pass with kind of this large longer inkjet or if you will, and like an inkjet printer, it inks where the parts are going to be so instead of a laser scanning and making that cross section like you're doing SLS, this inkjet is going across and essentially in one continuous movement in key in the cross section. So if you look at this, you see white powder and say you're making a bunch of doughnut shapes, you see a white powder, then you see a bunch of circles, you know you you'll you'll see these show up, and then it does a second pass with a heat bar. Essentially all that black inked part areas are going to absorb more of that heat and create a center effect while the other material where there wasn't any end game there is not gonna absorb as much heat so it remains powder based. So the biggest difference there is that on the part by part basis, okay, no big deal, but 20 parts 40 parts, my throughput is higher because each layer takes about seven seconds less. And we're dealing with four and a half hour layers. Seven seconds left This could be like getting my build out, you know, six hours sooner. And if I have a really high scan time, like so that laser is just one little beam going back and forth. It's moving super fast. Like, as humans proceeding, it's really fast. But it's still not doing all that stuff at once, like so like, if you have a consistent layer time, like Multi Jet Fusion has. The advantage is throughput. So, when I start working in production, using Multijet is my lean for that. So it's surpassing SLS, by a production standard, because it just can do more parts faster. So one to one, like one part one part, you don't see the difference. But if you're doing 2020, or 40, or 40, or you know, 300, or 300, it's a huge difference.

So is there any disadvantages to that type of system or

Multijet your your parts do come out Gray, or and oftentimes, or die black. But color is coming out with HP right now. So they're looking into color. And it's pretty mature, it's good in writing like early Z Corp color 3d printers, back in kind of like 2007 ish or so where the whites are not quite white, but they're pretty good because you haven't had color 3d printing like this before in your life. So your your, you'll take anything, it's technical. And it's nylon. So like I, you know, call it 3d printing exists like the Stratasys, J 750, which is a poly jet kind of this, this little injector UV cured machine. It's really good for making these, these shapes that can be used as medical models or industrial design models. But as an engineer, I haven't had a lot of use for it. Because PolyJet breaks very, very quickly. And also engineers, they don't know how to design a VM or all files, which is another problem. But you know, that's another episode there. But, you know, engineers don't design and color right now is a problem right now. So full color, 3d printing is fantastic, except no one actually knows how to make the export for it. But I think in the future, we're going to have more and more of these color options available. So I think the HP is actually going to be leading that. And I'm really excited to be able to add some stuff or add a logo or do a par mark or something directly with the print. So you

might have you can just like put your QR code of what that part is, like on the bottom of it.

Yeah, I mean, it's super powerful. And, and, I mean, right now, SLS, I mean, SLS is not like oh, I have all these you know, fantastic benefits I it does have trade offs, like the surface finish of both these processes is is grainy, almost like a sugar cube

Excuse me. And the call as I get here, I'm just going to clear my throat

and the also you have to die the material so like that's, that's the other thing is like he it's usually it's stark wide and you're kind of stuck with with dying to a mono mono color. So it's still the perception is not as strong as some other materials are for like cosmetic look, but a lot of people are getting used to it just because of how cheap it is and how available is right now,

are there any impacts on the functionality and strength of the material based off of the type of manufacturing?

So it's their industry, there are strengths differences, like so of like, like I said, FDM I can build more boutique materials, right. So I have like abs as a all terms, they have Woodfill they have all these type of things, especially in the desktop market. So I can get different accoutrements to enhance my print by materials. But I still do have a directional challenge, right. So like, vertical thin features tend to be weaker, you know, have the an isotropic features to it. processes like SLS, Moesha, fusion, even even some of those photopolymer processes, especially carbon, I have less challenges with with direction. So I can focus more on those materials. But the materials may be limited, like like SLS is great, but it's bulk great, like I have this material and this material only. So you are you're really focused on nylon for those. For for some of these newer processes like carbon DLs, they kind of were looking at this like, hey, what if we didn't think thermoplastics? Like what if I'm thinking just what can I do with a 3d printer. And so they have these compounded resins that are actually getting a different strength property that act more like a urethane where it has a second post cure to it. And they're they've actually been looking at this very different way for you know, for 3d printing. So it's, instead of saying hey, everybody runs, you know, abs injection molding so we better be better get the best ABS out here for our 3d printers the same. What can I do for the material proper These with, you know, photo curing and how can I optimize that. So carbon has come out with digital light synthesis, it's similar to SLA, which is another UV cured resin process. But essentially, you know, if you think SLS is a powder bed, these, these UV cured resin processes are a liquid bed. I'm putting a liquid in, I'm curing it, usually with a UV or a digital light projector. And I'm building these parts and I'm either drawing them under, or I'm drawing them up. So So carbon DLs actually pulls the parts out by projecting underneath. So underneath that transparent window, at the bottom of this liquid liquid resin chamber, and they have some cool IP around it that allows them to continuously pool so they have this, you know, full continuous growing process to make these parts. But my excitement has actually been more about thinking about the material as what material is really good for additive, not just what material is really good for. I already know ABS injection molding, and carbon has these a urethane based rigid elastic, last America urethanes. A urethane base silicone, a super high 10, a high detail material concidered Ester, which we found very popular, because before that all we have for high templates, altom, which is fantastic. But it's FDM, which sweats the small, small stuff. So now we have people doing really small channel detail features using this new printing material. And I'm more excited about that, like, you know, anything that gets me a first print is also the final print for production is a really good example of the future of 3d printing.

While the I'm looking at a picture of the form three L I haven't seen that before that is monstrous. For SLA printing.

Yeah. Yeah, for for SLA, like a lot of our platforms are actually 26 inches, no problem. But for a prosumer, mid range, 3d printer. The form three was awesome. And they have some really cool technologies that they call it like low force SLA. What it really is, is it is like the carbon DLs printers, is projecting LiDAR, it's curing from underneath and then building the part up. And the panel that's carry underneath just have to be flexible. So usually you have to kind of pluck the SLA part away from that panel. And that'll sometimes mean I need really beefy support structures, because they have to be stronger than the panel adhesion that's happening every single layer. Well, now with their basically by adding flexibility to it, it kind of just gives them a full peel off and you do a lot less support structures, and you end up end up getting less like jerkiness between layers. Because even on these liquid photopolymer bases, you still have a little bit of a kind of a zigzag rigidity, like like are like layer line stepping, which is really apparent FTM less apparent in SLA, you know, kind of apparent in FFLs. But, you know, you, the more you do to make that transition przy height, smoother, the smoother the outside results will be as well.

Yeah, that makes sense.

So one thing I want to talk about, and this kind of goes back to, you're talking about a designing a print that had a post in it. And you know, instead of designing the post in it, like you're using like as locating peg, like for fixture, instead of designing it that way you would design it as a whole and then just by fixturing pins or dowels basically and then put them into your print. I want to talk more about that and stuff like, like printing threads, or using just put up a hole there. So use a self tapper screw or do use inserts and trade offs of those, those types of systems and, and, and the processes as well. Yeah, so

I am the biggest believer in commercial off the shelf items ever, like MacMasters must be dial. So just just understand that not everything needs to be custom. And, you know, I once had a customer where the least homey design for a potentiometer knob. And I showed them a website to like a guitar website where they sold hundreds of versions of potentiometer knobs because I'm like, sometimes you could buy this stuff for $5 and it doesn't need to be custom manufacturing. Let's let's let's see how we get help here. My general rule for 3d printing threads is don't like unless a really coarse threads. You're going to run into interference and resolution issues. Because when I'm using these, these processes Um, each one has a little tiny new and they get really annoying on small details. So fused filaments, right I am, when I'm moving that layer V up, I have to kind of start somewhere on the next layer. And a lot of times that'll leave a little bow bulbous feature that we call a zipper. And if that if that feature just happens to be generated on your thread, all the time you have a you have a bad day, you have the interference right there. In laser powder, bed fusion, you have first off, you don't have that smooth surfaces, right, you have that matte surface finish to it. But also, I'm using a heat to bond these features together. So I'm using like I'm keeping solid 140 Celsius, I'm thinking with a laser. And in the small holes as they have a small threaded hole, what's going to happen is that heat from one side of that hole is going to radiate to the other side of the hole. Well, all the heat from all this senses, it's a circle is radii in all different directions, that actually centers more. So most holes tend to be about 10 tau smaller than what you expect them to be. So again, if you're making a female female thread, you're gonna have a very tight thread if you don't offset correctly, and then you have to deal with offset and tuning your parts was iterative process. Meanwhile, if you just created the drill diameter for a tapped hole, you know, SLS, nylon, HP motors have fused in materials, even a lot of our FMLA materials and carbon DLs materials tap like butter. So I can go in and just tap that feature out and you have perfectly smooth machine like threads to them. I also I will say for FDM because you have that layer WISe process. I like I like inserts, so again designed for our brass, you know screw to expand our press to fit inserts. And those can easily be inserted in you don't always need to heat steak or ultrasonic weld inserts in a lot of the times the screws that expand is good enough. But even even adding something like a like a heat fake insert is pretty simple, you do it exactly the same way you would with an injection molded piece where you can use either a special tool tip or if you're lazy, like me use the one of the screwdriver style tips on your on your soldering iron, and just got to be gentle with it. And you know, once once the heat gets in those inserts, again, it'll just flow smoothly right in place. My advice and again, like, it's just from experiences, if I'm designing a part and it has threads to it, I'm very likely either tapping or adding inserts as a default, like I I never think about designing threads, unless it's extremely like a custom thread like, you know, like a double helix or something like that, where it's just not off the shelf. Or if it's a very coarse thread where I know, hey, if I offset this surface by three sow, I'd probably be good enough thing, and I'll be able to print it out, knowing I may need to make two. Yeah,

it's also when I look at projects on online and stuff. And people will print threads and stuff. And I'm like, inserts are so inexpensive, and they just work better, structurally stronger, and all that stuff. But it's also same thing with bearings, is people will print bearings, it's like you can get a skateboard bearing for 10 cents 10 cents, and it works so much better than anything you could ever print.

Well, bearings are beautiful demonstrators are the freedoms that you can do with complexity and 3d printing, right. So like in SLS, I can I can print free floating objects as long as they have about half a millimeter gap in between them. And again, I'm gonna say you know, with a bunch Asterix, beside that, but a bearing is a good example. You see that all the time and like 3d printing experts where someone prints it out? And it is it's a cool example. But you're absolutely right. If I'm practically using a bearing in my project, I will very likely purchase that bearing and use the use of 3d printing process to make the custom part of my feature.

Exactly, yeah.

At the same time, press fit. threaded inserts are easily obtainable and easy to work with. I mean, you don't have to have crazy machinery if you're doing prototyping with them. It's really easy to work with.

Yeah, I have um, so Well, now our additive team has it but I'm a big fan of these Stanley color yellow cases. They're, they're about two inches tall, and they have a bunch of little compartments to them. And you can take the compartments out as you need them. And, and we used to just have basically in stock like every single type of every single McMaster insert, we just bought a pack of them and you put a label on and if someone asked for, you know that quarter 20 You just pull that out, you've said about your project, take a drill, you drill out the holes, and then you put in the inserts, you know, put it back in but it's really it's you know, from a production standpoint, it's also really easy to organize like it's really easy to to add custom with stock components in those ways. So I'm a big fan of again, pins inserts, even. I know we do a lot of fixturing for BMW, BMW, the they use the 3d printing aspects for the contoured and the custom features, they also have a lot of machining parts, because sometimes they just need something that's super stiff, it's super reliable. And they, you know, they live with machine components. But if they're doing a handle, or if they're doing a clamp feature, they're buying that it's just a McMaster feature to it. They're not they're not trying to reinvent the handle. So there it is, you know, when I think about 3d printing, at least, the way that I use it, when I'm looking for practical like, it's this 3d printed part is designed for 3d printing, it's, I look at it more the same way I look like, as a machine shop part, you know, it's going to be something that is designed for that process. And 3d printer just happens to be another tool to make apart. It's not, it's not special, but it's also not, you know, weak, it has really unique opportunities, when it comes to design and freedoms that you have.

Yeah, this reminds me is back at back at Mac, when Steven was there, we were designing a whole bunch of fixtures for holding PCB panels and stuff. And we designed it out and we made it so that you could it was seen see not 3d printed, but we made it so you can make it make it flat stock. So you didn't have to actually like machine features, because we had like, pegs and stuff, where so if you just like handed a 3d model of someone that had to make it out like a half inch piece of aluminum, instead of like, you know, a 16th inch piece of aluminum, and then we just your press the Tenzan and stuff like that. So it's one of those using using off shelf parts with your custom glue, so to speak. Which is your your 3d printed part. In this case?

Yeah, we do it so often. You know, on our on our website, we have, we have seven 3d printing processes, the only process that you can't install inserts or get tap is PolyJet. But every other project that we have, you can like you can just select five inserts and type in which ones you want. And just make sure that your design has a proper hole diameter like that manhole diameter for that, but we take care of the rest. And it is just so common to do that. And like and again, the beauty of that is that means that all you do is open the box and then you just assemble your parts. You don't need to think about all those little things like even to your example, like with sheet metal, the same thing like he if you want to plug a few inserts in like we could do that work on the side, because it's it's assembly, but it's not really like it's still continuous what the part so we kind of call it part manufacturing. Yeah.

So now your bomb has two part numbers on it. Yeah.

The challenge is that a lot of people design the inserts in as a body in the solid model. So maybe like a solid part file, but it has like extra bodies are some of those inserts. And that throws our our quoting engine for a loop, it usually throws out an error, or says hey, this is assembly No way, you know, we're blocking and they're like, why isn't my part uploading. And, and it's usually if you just suppress those features, because we kind of you know, if we kind of get it to a drawing or if you have an image, that's enough for us to work off of, for for incorporating inserts into your design.

Oh, cool. I think one of the last topics I kind of want to talk about is kind of like the pros and cons of let's say running your own 3d printer versus say using a service like xometry to order things because I know a lot of people on our podcast or their makers are self starting engineers. And so they have their own 3d printer. It's like why would they use a service?

Yeah, and my I think my answer is Why not both, you know, if I if I was in the shoes, and I was starting up as a as a company, and I had like hours mattered, like you're you can't walk away from from a growing business, you can take time off from it. And the beauty of a lot of these desktop 3d printers is they tend to be relatively low cost. They may be a pain in the butt, which is something from a service and I don't want to deal with I'll never use one for my service, but they maybe maybe enough of a positive in your workflow that you can get some parts out that can have get you kind of to a good enough state to move on to your next iteration of project. What services offer is access to millions of dollars of equipment and infrastructure without any purchase other than just the part you need. So you can get the material from an industrial printer and use printers are often hundreds of 1000s of dollars to purchase in the first place. And they will usually require their own work sells personal protective equipment. You know postprocess the equipment person Prior to protective equipment for the post processing equipment, you know, it's it's a, it is an entire factory. Yeah, that's

actually one thing people don't think about is like an FDM printer that's printing PLA or ABS, it's outgassing into the environment like like, I always think it's really funny, where it was like, Oh, this was MakerBot big thing was like, every engineer at your company could have one on their desk. And it's like, sure, that's great, except it's off gassing all these noxious VOCs. Right next to your engineers. It's exactly what you want. Like,

molten plastic. Yeah. It's by the ways, but you know, when I, when I was, you know, more running the machines than anything else. So my wife, you say, like, you smell like nylon? Well, I guess. The other thing is talking about is it's also cleanliness, like I always say, like, if you're running a powder bed platform, so whether that is the SLS a single shift fusion, metal 3d printing, like DMLS, what you really are, is you are a maid, you are cleaning constantly. Like, there's, there's a couple of things I always had with me, you know, when I was when I was running these machines, you know, I have my little pen live for inspection, and it's just in there, my computer, and then I have my shop vac, these things are always around, because there's always powder, there's always stuff that is making a mess. And so you clean it up. And so working with a service, you're you're getting the benefits of the infrastructure, you're getting capacity that is crazy big, like the amount of work that we can do. And the amount of parts that we can produce, in days, dwarfs anything that's possible with, you know, with internal engineering capacity, like I want, I, you know, I want our customers to have their own manufacturing, whether it's 3d printing or other things, because it actually makes them better designers, because they're working with us iterations and working with understanding of what these processes can do really well. But ultimately, when it comes down to a certain material that may not hit the temperature range, a part size that may just not fit on that printer, or, Hey, you just need this many parts, and it will take four weeks for your single printer to do, you can get it back in like three days or less, you know, I think our FDM lead times now are that down to two days, for, you know, through through a service company. So there's a lot of no brainers that come around come round with that. But in no way is it cannibalized in either direction, because it both serve a very, very specific and very important need in product development.

You know, I would also say something to add on top of that is you also get somewhat of the guaranteed specifications that go along with those machines. So if you want a very known tolerance, you can get that if you're purchasing it, if you're doing it on your own, then it's up to your calibration, right.

And in your calibration can go out, you know, you change the material, you change something else and you get a deviation. One of the things that we look for when we vet for what what machines or classifications add next is, Can I do a print now and be satisfactory? And then can they order you know, 30 of them three months from now and get the exact same part at these industrial platforms. Again, mechanically, the deposition of material and stuff may be very similar to what you have on these, you know, on these, you know, sub $5,000 machines or even some sub $500 machine, sometimes the mechanics are very similar, but the consistency and repeatability is what a service industry works for, because we cannot afford downtime, we can't afford to do an engineering effort on every single part. We really want that digital, you know, digital path of CAD file, you know, CAD file, uploaded, interpreted, quoted, purchased, and then sent to machine to make, and we want to we want to keep that very, you know, continuous throughout the process, because that's, that's throughput, you know, that's throughput, that's reliability for the customer sake, and consistency, they know what to expect when they press go. But I think the the takeaway that I always have is there's, there's, there's still a lot of wild west going on with additive. So it's not like the CNC industry where it's like, oh, I just want a T six temper of my aluminum 60s 61. And you know, that you could buy that from like 90 different suppliers and you know, have a reasonable expectation, that's the same material. There, there are a lot of question marks and the more and more you work up to from as a production facility, the more and more you have to kind of play it safe. And that's, that's why, you know, like xometry is a great example because it's something I know very well you're working with, you know, OEM equipment run on OEM parameters with OEM supplied material, because you have a digital stability like you have a process stability, there that you can actually convey to your customer. The more I work down towards an engineering I've also, I'll even say when I ran SLS, for this, for the engineering company that I worked for, I was tweaking numbers all the time, because I kind of just needed the shape. And I was able to kind of look at it. And I was like, Okay, I need to be more watertight. So I'm just going to crank up by two degrees Celsius. And I was able to do some more experimental work, like I was able to throw different powders in, make my own custom mixes to see if it work. But that's because I had that freedom because I was working, like I was an employee of the company in which I was working to make this parts for so I had, I had more of the r&d, you know, mentality for it. So it just really depends on you know, where that you know, where you are, like on what you're able to do with these printers on why you're doing that in the first place, because I couldn't give that inconsistency to a customer. But I could take it for myself, because, you know, it's my machine. I'm running it and you get the parts I make. But you know, as a service. You you think more and more in scale.

I totally makes sense.

So you guys have 3d printers, though each one of you?

I do not actually

know I've got one. Yeah, it's pain in the butt.

That's sad. Like, when was the last time you ran it? So it's

actually yesterday, actually, I was, I was gonna bring it up when we were talking about warping stuff, because it's FTM. But I only print. I'm kind of like you, I guess I don't, I don't care about the color. I print in just because I do a lot of automotive parts for my project. So I do like polycarbonate PC. And I'm trying to do this print right now, which which I'm getting some layer of separation.

I'll describe it for the listeners. So he printed something where the internal cavity if you imagine a tube on the side was printed horizontally. And what happened was he has a overthink base to his to his part. That base was a the base actually was probably better adhered to the to the base plate. And as the internal stresses a polycarbonate was polycarbonate is it's great tensile strength, which also means that when that material cools is going to pull a lot harder, though he has a thin area in the middle of that horizontal cylinder that just tore itself in half, and he ripped his part in half. Yep, I've done that with metal parts. I've ripped stainless steel parts in half by design. So you

can fix this with superglue.

So we didn't touch on metal 3d printing too much. But metal 3d printing is actually closer to FDM. And a lot of ways, because you're strong arming this metal down to a metal build plate, the whole thing is made in room temperature. So the power the laser is doing is micro welds, you're building the cedars from the mountaintop. And it's literally a continuous battle of the the metal distressing against that build plate and who's going to win this inch and a half thick build plate, or the part if the part wins and appeals up, which means like it pulls off from its support from the base, or it literally pulls itself off, which is which we'll call the crack and tear down down the side. Like you know, you lose like it's it's a bad day for everybody. And you're talking a very high overhead rate and just you know pain in the butt because it's like, oh, let me just peel this off my hands. You know, like let me get this little wire EDM or the bandsaw and then when we get this to, to a mill at a plane off the face of my support plate, and not to mention you're dealing with you know, metal powder, which itself is dangerous. So, yeah, yeah, I've been there. I've done that a metal it's, it's it's Daniel, Dad news bears, but we learned from that and that's why we don't print hockey pucks and metal folks. Yeah.

Oh, fantastic.

Oh, cool. Steven, you got anything else? Greg? No,

I think I'm good. It was really great to have you on again, Greg. We appreciate you coming on. It's always fantastic.

Yeah, thank you so much, Greg, for coming on to the podcast again and talk 3d printing with us. Yeah, happy

to talk and obviously like I you know, I like to stuff so anytime you guys need us. You know, let me know happy to reach out and talk some more and even pick your brains if you're trying to troubleshoot your polycarbonate printers

might might be I'm running through some dedication problems right now. So yeah, so with that, uh, Greg, you want to sign us out?

This was a macro fab engineering podcast. I was your guest Greg Paulson.

And we are your hosts Parker Dolman and Steven Craig. Later everyone take it easy thank you yes, you our listener for downloading our show. If you have a cool idea, project or topic let Stephen I know Tweet us at Mac fab at Longhorn engineer with no O's or at analog E and G or emails at podcast at Mac fab.com. Also check out our Slack channel. If you're not just grab to the podcast yet click that subscribe button that way you get the latest episode right when new releases and please review us wherever you listen as it helps the show stay visible and helps new listeners find us