This episode with Brent Ewald peels back the layers of 3D printing, revealing the intricate dance of design and machine. Listeners will gain an appreciation for the system-based problem-solving that is essential when dealing with the nuanced interplay of a printer's various subsystems and might find themselves inspired by the continuous learning that defines a career in this field.
We open the doors to the nuanced world of CAD and automation, illuminating the potential for prosthetists to streamline their craft with these powerful tools. The chapter sheds light on the precision needed when designing for confined spaces and introduces the concept of auxetic structures, an innovative solution that could eclipse traditional lattices in specific applications. This episode is not just a discussion—it's an invitation for designers to embrace community, share knowledge, and seek guidance in navigating the intricate and fulfilling world of prosthetics and orthotics.
Special thanks to our sponsor: Coyote
This episode with Brent Ewald peels back the layers of 3D printing, revealing the intricate dance of design and machine. Listeners will gain an appreciation for the system-based problem-solving that is essential when dealing with the nuanced interplay of a printer's various subsystems and might find themselves inspired by the continuous learning that defines a career in this field.
We open the doors to the nuanced world of CAD and automation, illuminating the potential for prosthetists to streamline their craft with these powerful tools. The chapter sheds light on the precision needed when designing for confined spaces and introduces the concept of auxetic structures, an innovative solution that could eclipse traditional lattices in specific applications. This episode is not just a discussion—it's an invitation for designers to embrace community, share knowledge, and seek guidance in navigating the intricate and fulfilling world of prosthetics and orthotics.
Special thanks to our sponsor: Coyote
Hey everyone, my name is Joris Peels and this is the prosthetics and orthotics podcast with Brent Wright. How you doing, brent?
Speaker 2:Hey, joris, I'm doing well, man, some some pretty interesting things coming up this this year. There are some new and we call them O codes that are coming down the line, which means these are some things that insurance will now pay for, and it involves robotics so powered orthoses as well as adjustability and I know that you're a big fan of adjustability, and Now it looks like we're gonna start getting, or having the ability to get paid for some of these things, so I think that's pretty cool.
Speaker 2:Okay, that's very exciting, and you know already how much, and then what that will depend on or you don't know yet, or so so yes, on the robotic things and when I, when I say these numbers, just know that these devices are Are also very expensive, so I don't know what the spread is. But there's a couple exoskeleton companies One that most people know of, rewalk that got an L code for their lower extremity, kind of walking suit, for lack of a better term and I want to say it's just north of $90,000 to get reimbursed for that.
Speaker 1:Wow.
Speaker 2:So pretty significant. And then the myo pro, which we've had John naft on before, which is for People that have had strokes, just got a code in January. So January 1st is the first time that you could build these and You're talking about just north of $60,000. So I'm really close to what some of these myoelectric prostheses are Are being reimbursed. So it is a lot of money. But before people think that you know Prosthetics are a greedy money, hungry people just just know that you are fitting the body and that you are on the hook for the care of these patients for at least three years. So it is. It's a bundled code altogether, so if you see this patient hundreds of times, you actually start losing money pretty quick. So there's a While, there is a reward. There's also Very true expenses that go up against something, and then something you know as nuanced as some of this new technology.
Speaker 1:Okay, but it's good to know that there's new stuff possible and maybe, yeah, expand a more 3d print stuff. So that's really really great. My bit of news, by the way the fact that I was reading about a company called Beretta Medical Innovation, something about that. The a Canadian company and they make braces, right, and they have five staff and they make two thousand knee and leg braces a year, which I thought was nice, yeah, as a volume kind of thing, and there's specializes and they are going to Mgf and they're using a completely custom Toolchain like a software toolchain. Let's some customized these things get the files out from this based on the scans then about half an hour. So I and I thought that was very exciting to see that there's a relatively small business doing very well, I think, actually if you do the math on that, and they're now being able to transition to 3d printing in a fairly easy way using automated software. So I thought that was really exciting.
Speaker 2:Yeah, well, and so, and what I love about that is and I heard somebody say, the riches are in the niches, so and I truly believe that now e braces are, you know, fairly commoditized, but if you have a good workflow, especially that moves into digital space there's. There's a lot of white space there for you.
Speaker 1:No, I agree, I agree. I think I braces are my most favorite stuff and I love that niche approach. I think as well, like I love. What I even love is even more niche here, like it's at least post-operative braces and the braces for people that have really different Considerations and really different cases than everyone else. I think that's a beautiful area to expand into.
Speaker 2:Yeah, well, you know what else I love?
Speaker 1:huh, we have a sponsor for the rest of season 7.
Speaker 2:Yes, it's super nice. Yeah, and so our sponsor is coyote design or thought it's a prosthetics and we've had them on. We've always talked about them. They. They carry a range of products, but my favorite product is one that I started with early on and which is kind of their kind of capstone project or product.
Speaker 2:The one that they probably sell the most of is called the coyote airlock and and it's a. It's a locking mechanism. It's a way that a patient keeps their Prosthesis on their leg. But what's neat is, compared to any other lock that's on the market, it's actually airtight. So you do get a little bit of suspension from Suction and and it's funny to hear them talk about it because they sometimes call it suction so it's a little bit of you know how tight the prosthesis in and a little bit of suction, but together it helps suspend the limb so it's not just off the end of their leg and it's a really neat product I highly suggest. If somebody hasn't used it and I'm surprised about how many clinicians don't know About who coyote prosthetics and orthotics is, but hopefully through this we'll be able to get the word out a little bit more- Okay, super cool, I loved having them on and stuff.
Speaker 1:So I really like that and this airlock products, anything that's like that kind of a product, like a complex assembly that really adds value. I'm always gonna be captivated by this and and so I'd like to the sponsors. That's really super good news. And then, what does this mean, the airlock thing? What does that mean for patients, though? I mean, okay, you're suspending it, it's an airlock. You can kind of like put pressure on the, the limit. What does it actually mean day-to-day for a patient, though? You know, have you gone to? If you had a patient where you're like, okay, we had Bob, and then after, with an airlock, bob's life is better. Is that what happens, or is it not that dramatic?
Speaker 2:No, it is, and this is the crazy thing. They have a lot of research around this. So with an airtight lock, you decrease the overall pounding Pressures on the patient's limb, so it's more consistent with every single step. So it, which makes sense, right, if it's kind of this has a suction to your leg, it's going to be consistent every single time, whereas when you do not have it airtight, every single step is different at different forces. So you might have one step that has a high force, one that has a low force, but there's movement that's happening. And this is one of the things that they figured out really early on with the airlock is that it's actually better for the patients because it Normalizes the pressure that's on their limb.
Speaker 1:Okay, man, that's super good. Yeah, super love having the monies nice having, it's always nice having sponsors to make really good products.
Speaker 3:That's right.
Speaker 1:Yeah, it's really nice dude, I like that. It's awesome diet pill where you're like, yeah, try the drain, oh pill, it's nice that we got some people stand is easy to, it's easy. So if they make something useful and then you've already tried and that, yeah, that we were both behind, let's say so that's super good. Okay so, and who's our guest today, then? Right?
Speaker 2:Yeah, well, don't get confused, because we have you all done today.
Speaker 2:He is a principal engineer for out of desk but he also does a ton of innovation, worked for HP for a little while. I'm not gonna hold it against him. He is did go to the University of Michigan and they did something special. I don't really remember recently, but I do know they got beat by Appalachian State when Appalachian went to their Place and they actually paid Appalachian State to go play football and then Appalachian just took the money and beat them anyway. So that was pretty neat. But other than that, you know, brent's been great. I've known him for a very long time and what his Wheelhouse is and this is why we're gonna love talking to him is really workflow specific for multi-jet fusion, because he he does a lot of stuff in the computational design realm and and so while I know that I won't be able to jump in the weeds too much with that, I know that that is super important when we talk about how do you productize a Product or bring it to market where it scales and makes sense for 3d printing.
Speaker 1:Okay, that's super cool. Welcome to the show. What we were, how we're gonna do this. Brent number two.
Speaker 2:Right.
Speaker 1:So so first of like. So how did you first like kind of get involved with with? Well, first like 3d printing, how'd you first get involved with that?
Speaker 3:Right. So while I was at the school that you know was named, I Was in graduate school and you know I was just looking for the next move. I had worked in power generation, so like power plants, pipe supports, heat balance, all sorts of weird stuff there. But then I went to grad school and HP there was an alumni who was at HP who reached out to the university, said any students you know looking for a job, and I Applied. It was. It seemed interesting, even though they were only talking about, you know, 2d printing stuff. The types of problems, even in that space, are really exciting.
Speaker 3:So I went through the interview process, got hired there and Moved out to the West Coast, and this was the end of 2014. And so when I moved out there, about three or four months after I started, they started kicking off the initiative for the 580 and I know Brent writes a big fan of that printer and so they were, you know, not Redividing staff, but you know who's gonna go over to that group, who you know, who's gonna do what, and it was just so exciting to me. So I put my hat in the ring and I got to work on some really cool stuff, starting from, you know, conception to architecture, all the way through to launch, and then, after launch, I Moved over to the Application side, the commercial side of the business, and tried taking it to market, and so I used all the things I learned about how these printers actually work and and tried to help customers use, adopt and, you know, arrive with them Ideally. And that's where I met Brent when I was, when I was over there.
Speaker 1:Okay, that's super cool. And then, and so that's nice to be pioneering, launching essentially a new, well, variant of a technology and new technology in the market always super exciting Kind of thing to do. And what did you learn to this kind of stuff? I mean, because it must have been like for you, must have been quite a learning journey, for you yourself as well, right?
Speaker 3:Absolutely. I mean, first of all, every, every print process is extremely complicated, and that was such a joy to just to embrace you know they talk about. One thing that's commonly you deal with, like HP printers, is you want to balance the heat throughout the bed, and that came up over and over again. It's not only that. That impacts Every aspect of the print process, like powder consumption, local temperature gradients, just and and just learning the complexity and having kind of a native feel for the complexity was so interesting. The other thing that was very important about it was those two things.
Speaker 3:One was system-based thinking, so like you can't only focus on this subsystem, you have to focus on all this you are not focus on, but you have to understand all the subsystems and how they interact. And the second one, and this is something I think we all know I know Brent knows Pierre Kaiser, but he, while we were working on the project, at one point he brought someone in for to give training to the group and the training was basically it was problem-solving training, but there were a few questions they had us repeat that I really loved, which is what are we trying to accomplish? How will we know if a change is an improvement, what changes can we make that will result in an improvement, and that kind of just like explicitly stating what we're trying to do. All that kind of Philosophy was so helpful in understanding system-based processes and outcomes.
Speaker 1:Yeah, I think it's wonderful you brought this up.
Speaker 1:Actually is your main kind of learning things, because we see the so many times and desktop 3d printers, large industrial systems, really in large metal systems, where people get lost in there and in complexity, because there's so many feedback loops and there's so many things that are being changed Either by the user or by the environment, and there's so many things going on that you know it's kind of like one step forward, three steps back and one step sideways and and we've talked about this in a couple of in a very variation in this Podcast one of them is like the settings drift.
Speaker 1:We're both Brent and I are carbs in the fact if you have a desktop 3d printer, you kind of start off with your settings and you kind of do willy-nilly and that they drift over time. Then you don't actually know what you're doing right. That's one of these really simple, understandable examples. But we see from the industrial and especially in the metals world as well, we see all these examples of people developing these really complex machines and not understanding what they're doing and changing things that that also make them go kind of get really lost in the weeds of their own process.
Speaker 3:That's right, and the important thing is to go back and revisit the things you assumed and that is so often missed. Like one example is there you either use a thermo pile or thermal camera. It's something where you can measure temperature most. All these processes require some feedback, like that right, but if you start having problems with things being too cold, you need to check the sensor as well. Like you can't just trust these sensors that are given to you, especially optical ones, because in at least MJF, you're dealing with powders, you know, or aerosol or any number of occlusions that can impact the quality of the measurements. So, absolutely, that was probably the most fun I had.
Speaker 3:I would go back to the lab. I probably spent way too much time there, but the 580 was really a joy to work on. Another thing sorry to think about, though, is there was a new problem that was created because of the 580, which is color. So most of these parts have color on the surface, but you need to have depth in the color in the object, and that's so important or else everything would. Just you wouldn't be able to tell the color realistically. You have to project it down. So there was a ton of computational stuff to be solved as well, not just physical hardware.
Speaker 1:That's super fun. And now you're like Autodesk, so you do like. And how did you end up there? I mean, were you doing the same kind of stuff with them from a software point of view, or is it very different of what you're doing now?
Speaker 3:Man, I've had such a meandering career and I'm still fairly young, so I'm a principal engineer. I'm mostly working on their implicit modeling, which is very cool. So, if anyone's not familiar, there's tons of different subcategories. But generally, let's say, there's explicit modeling, where you define your object, what you want to print, what you want to make as a surface or a collection of surfaces that form a closed volume. And when you're doing explicit modeling, that's your traditional CAD, what everyone thinks of it's like make a cube, your cube is actually just six faces, that kind of thing. And then there's implicit modeling.
Speaker 3:Implicit modeling is you define an object by a volume, and so this is where you have what you can call an arbitrary field. This is actually what in top has made much more mainstream. So they've done a great job there. But we have some features too. But it's where you define it by a volume and you can say a gyroid Everyone uses gyroids, right? That's just an equation that defines a number in space, and so your objects is defined at infinite resolution, which is really exciting. So it's only at the time when you want to use it that you have to convert it into something discrete that you can then print, mesh, whatever, so that's a really exciting class of gyroids.
Speaker 1:And the first thing that's really obvious, of course, is that we don't have to take this drawn out really clunky in the computer thing and then manipulate it right. We can. Just because it's an equation, we can change that thing really really really quickly and really really quite light on the computing power when compared to how it's traditionally done, right.
Speaker 3:Absolutely. I mean, the visualization is going to take a while because you just redefined this thing and so on. Visualization it's actually really easy to do meshes because you just draw the triangle right. If you have a discrete triangle mesh, you just draw the triangles. That's not that hard. But when you do implicit it's, the visualization starts taking longer. But you can define insanely complex geometries at arbitrary precision very, very fast. This is why heat exchangers and all of these products are designed this way, and it's also things that you couldn't really do by hand, and that's what I found very interesting.
Speaker 1:Yeah, I think, if we're going to try and keep it to our audience, I think one of the things is like a foam is random, right, most foams are just random. That's kind of stupid, right. But if you want to make a kind of a directed foam or an engineered foam or something that's a much more intelligent foam, then what we can do is we can define different shapes, different shapes at different levels, different size of these cells, and we can also just change that throughout the thing, depending on what we expect to happen. Right, we could make it softer, or we could make the cells bigger towards the head and smaller towards the outside of the helmet, so we can actually, in a rules-based way, define, well, not just one shape, but a really complex kind of story, if you will. That ends up being a helmet.
Speaker 3:Thank you for that. Yes, I sometimes go a little too technical there, but your point is actually absolutely right in that you can do cells too, so you don't necessarily have to do it as this implicit equation. There's other ways to define that stuff in a rules-based way, which is define your whole space as little boxes and then you put whatever you want into each of the boxes, and this is the computational design stuff Brent was telling you about. That I worked on a lot. I also for a while worked on a footwear effort. So you see mid-soles made out of lattice. Those are just boxes that have rules about how to populate them and fill the space.
Speaker 2:So just curious, I think you've been exposed enough to the prosthetic and orthotic field and definitely the stuff that you did in the footwear side of things. Where do you think? What are some of the bottlenecks that you see our field facing with some of the design side of things, knowing that each one of our things is custom to a patient it's just not in different sizes and all that stuff but where is? Somebody like me would need to look or start thinking about design, so then we can start using some of the tools like what you're talking about to make our design easier in the future, because we're going to start having this data and design data right now, and five years from now there's going to be some probably pretty cool stuff out. Probably in the next year there's going to be some pretty cool stuff out that will make our job easier. How do we need to think about some of this stuff?
Speaker 3:That's a great question. So again I'm going to go back to systems level thinking. So let's say you have a patient. You need information about that patient that enables a custom design, and that could come in the form of a scan, that could come in the form of just some simple measurements to customize something. It could come in. You could name a hundred different ways, but you have to understand what information do I have to do the customization on. So that's your input, your output. You need to know what your output needs to be, and this is the brace that, whatever you're trying to make. And then the question is what other information can I pull in?
Speaker 3:So we've talked a lot about the difference between working with a surface versus working with a volume, versus working with who knows what, but you've got to be aware of what information and what shape it all takes. Once you know that, your design process gets a little simpler, I mean it actually gets very simple. From that point, the key barrier I see to all of this is placing the same feature on each subsequent patient. So when you're talking about prosthetics, the main challenge you have is that, I would think, is that everyone's different, so you might not have all the features that are there. If we keep it to the case of I don't know a hand brace or something, maybe you need to know where the thumb meets. You'd be way better at this. You need to know where certain joints are. You need to have markers there, and the big challenge with that is how do you place the markers, other than having an expert place them, and so that's where I could see a ton of improvement, helping and enabling these workflows to go a lot smoother.
Speaker 1:But isn't it like, isn't it one of the cool things, that we can raise the abstraction level so that maybe I know nothing, I can rely on software written by Brent, who knows like a ton more, who can then rely on other cells or other kind of structures defined by you, so we can actually kind of like, rather than with regular modeling, is everyone's a brick layer right, but here I get to be a brick layer right, but I'm faster. Let's say you get to be more of an architect. So is not that? Raising the abstraction level is not like the biggest advantage for us, is like a really different way of thinking about the whole challenge.
Speaker 3:Absolutely. So let's say, you have a scan and I can write something that could wrap a mesh around or provide padding for whatever the problem is. But I can write something that could do that, given the knowledge of where of the surface of the arm which is given by the scan, and maybe the anchor points, you know where the certain joints are that we want to restrict or guide or whatever it is. Yeah, absolutely, you can start building these layers. So it's really the system of input, labeling, generation and output, and where I come in mostly is in the generation, and then you work with experts to define the other parts.
Speaker 2:So it sounds like really what you're saying is. To be successful is to have a good outline and be able to truly communicate what you want, so you know. For instance, one way to not do it is to say, hey, here's an input and here's an output, but I have no idea how to get there and what features I want, which I would say a lot of people in a prosthetic and orthotic field would do. But rather, if you can do it, say in a manual way, so you can show how you do it currently, that will tell somebody that's helping you create this. You know workflow or what have you. Hey, these are the current steps and these are the important data points for the final product.
Speaker 3:That's a great way of putting it. You know, when you talk about the current way you do, it is very important If you divide it up and then maybe you have a section you want to varoneuilatis on the surface that just kind of holds some structure, maybe provides some stiffness. You know how you get there is very important and, yeah, it's a complete process. I think the most interesting point and me about that whole thing is you learn how to think about the problem and you want to constrain your solutions to the type of geometry you're given. A great example for this is like if you're given a surface and you only have about five millimeters to work with, you know maybe you have to fit some sort of cushioning aspect there. Don't go for a lattice, that's not enough thickness. Maybe you go for, like a leaf spring type design. But it's about you know, design, the way the problem presents itself and the existing workflow matters.
Speaker 1:So what would you recommend? If you're like let's say we have like three levels of like person One is, unless you're like cat has always sound really interesting to me. You know, but I haven't gotten involved yet. What would you say, like if I wanted to harness the power of things like implicit modeling, this generative ways of designing and these powerful tools that are coming out? What would you say? Do I need to learn CAD or do I need to understand CAD? What would you recommend I do? If I now was listening to this I'm a prosthetist, for example, but I know nothing what would you recommend I do?
Speaker 3:Right. So if you're listening to this, you're a prosthetist and you know nothing, I would say what is your end goal? Are you trying to develop your own automated workflow for a specific product? Are you just trying to become strong enough at a concept to do like one-off examples or designs you know for specific patients? You know it really depends. What you do depends heavily on that.
Speaker 3:If you're, if everything is a one-off, you should stick to CAD. I think so because it has all the operations. You don't need to build these massive workflows and you don't need to. You know it's very. There's not much reuse, so there's not much need to do that. Now, if you're looking to essentially become a programmer, slash designer, slash prosthetist or whatever field you're in, what you should do is you should look to things where you can build reusable components. You can still stick to a CAD environment. In fact, you know I work on Fusion 360 and all these CAD packages are essentially they're opening up their API, which is application program mobile interface. I probably watched that but they're opening up APIs where you can control the CAD environment through scripts, and so this is the kind of the repeat operations you do all the time.
Speaker 3:You can start scripting. So a great way to start would be do something and, if you start making your parts in CAD, start scripting little pieces of your workflow, the things you know you do. All the time you do them the same way and it's pretty repeatable. And then, as you start building your script library, you'll start seeing how I can link this to this to this and all of a sudden your workflow start getting shorter and shorter. The work time, the workflows are just as long. In fact they'll probably get longer once you can automate it. And then you start linking them and building them over time and you'll start just saving yourself time. You'll get more time back.
Speaker 1:Is that also because now there's a thing like I could spend 2000 hours learning how to CAD, right, but am I learning how to put the little letters on the manual press? And is somebody else learning like DTP? You know what I mean. Should I instead learn Python and then I'll understand scripting and then I'll be able to use all these other blocks to get a faster thing? Or do I need to kind of know both, at this point at least?
Speaker 3:Right. So this is the beauty of it these APIs are actually written in Python or C++ often, but you can use Python. So if you know Python, you just have to look up there's documentation. You can look online and just say you know Fusion API or whatever I'm just saying that because I work there but you can look up these documentation over these and you can say how do I create a box, how do I create some sketch lines? And there's a programmatic way to do it through Python, where you can create those, and most CAD packages have this capability. Now, is it always Python? No, but Python is probably the lowest common denominator, and I don't mean that in a bad way. I mean everyone has a Python configuration in their package. So, yeah, python is a great way to do it. The only downside I'd say to Python is you can't protect it if you want to share it, meaning Python is a script language, so if you share a script with someone, you're sharing your IP. Just something to think about.
Speaker 1:This is a great thing because Brent is kind of like he's always on this cut. Brent shares a lot of information, he spends an awful lot of time helping other people, essentially, and sometimes these people are nice to him or just nice to everyone in general, and sometimes they just run with it and he'd just say, if someone else is not going to give back a lot of money, so if I could put my intelligence, my workflow, my years of experience in a script, is that a monetizable thing? Is that a protectable thing when at one point, I could actually protect it and sell it maybe, or in some way monetize this?
Speaker 3:Absolutely. Actually, these CAD systems, they have app stores. It's a similar model to a lot of things and if you develop an app that's good and it solves a lot of people's problems, you can charge for it. Now you have to come up with your own mechanism for sale, but that's neither here nor there. Absolutely, you could create a business off of creating tools for or prosthetics and orthotics, and it is very exciting the things you can automate. Most operations you can do in a CAD, in a CAD program. By hand and clicking. You can do with text in Python.
Speaker 1:OK, that's super cool and for you, help us understand a little bit, because one of the most powerful things in additive is the ability, and also in CAD. Now, I think the culmination of that is the ability to kind of like every single part is its own material, if you will. We can give it a different surface, the grip's differently, we can make it different in a gradient type of way across. We give it different properties across every inch of it, and we can also really make a unique way that material will react to someone's leg or foot or something like that. That's super powerful, right, but how do we get started with that, or what are the types of things you can manipulate in that way? Because there's this interplay between what you can do in CAD and what actually works on the machine, and that, to me, is precisely where you work. But it's a very difficult area for a lot of people to understand.
Speaker 3:Absolutely, and I think of it as the fun house mirror of 3D printing. So just because I design it some way, there's a whole processing pipeline between the design I give the printer and then the instructions that the printer executes. Now on FDM or FFF, whichever you want to call it. That's a little simplified because you can give it toolpaths and then you can see explicitly what's happening. But for things like MJF it's complicated. You have to convert this whole mesh model into dots stacked up to the top of the build, and that does take. There are processing steps in there, and so there is a skill to saying this is what I want.
Speaker 3:Sometimes you have to design not for what you want, but these are in more extreme applications where you're like I need to get this exactly precise. As general printing, it's wonderful. But for the very extreme stuff you have to basically manipulate your design to match what the printer is going to produce for your application, and this is true of most printers. You can compensate for warpage, you can compensate for all sorts of things, but that takes printer knowledge and automation on the CAD side. So a great example is Grasshopper. If you create a lattice, you can set the thickness of the lattice to be some function of a distance from an arbitrary point. Why am I saying that? What I'm saying is, if you put an arbitrary point, say you want the thinnest portion to be there, then you can create this kind of lattice that is softer in the middle and stiffer on the outside. So, however you want to do, you can invert. It's an arbitrary function.
Speaker 2:So that's one of the things that I love about talking with people like Brent in the computational design. I'll speak in terms of hey, I want it 4 millimeters thick, and then when you get down to here, it's 2 millimeters, and so people like Brent will be like so that part that's down below is just going to be half the thickness of your overall thickness. So you put it into a mathematical equation or a script of some sort, but it is a little bit of a switch, because how cool is it to take the knowledge that you know of some sort of 3D printing function that you know that works, and as you scale it, then it becomes a function, a numerical function of other features, and I think that is where there is so much power.
Speaker 3:Absolutely, and the other thing I'd say is everything needs to be passed through an expert. I would never say oh, you have your design pipeline set up, just run 1,000 of these braces. No, if they're custom per person, they should be reviewed on every person. I'm sure that's a medical requirement because these systems can go wrong. They're programmed by humans. I would never write a bug, but I'm sure there are people out there that would.
Speaker 2:Right Well, and I think that's a good point that every one does need to be reviewed because each one is so unique. But, in the same sense, what's helped me along the way with designing in this way, kind of this systemized way, is it has started to in my design cycle is I start asking is that really necessary? So can this feature also be featured? On this feature, even though I might have tweaked something a little bit forward based on the patient's diagnosis, Is it truly necessary? Or can I stick with a standard that I know works, so then I don't have to validate another workflow?
Speaker 3:Absolutely. Workflow validation is that's effort to test something thoroughly within the bounds, especially when you start adding. The variability of a printing process on the back end is difficult, so there is value to standardization. Now the nice thing is you can standardize on flexible solutions within some set of bounds. It's not like you only make one size. It's I do this geometry within this surface patch of this scan. Ok, that's something you can regularize. All the things that are complicated, like you have to have trim lines around the outside or something that contains it, that can all be automated if you just know what to look for.
Speaker 2:Yeah, can you share a little bit and you touched on it just slightly, but I think it's worth expanding on a little bit more. This idea of lattice structures, well, it's super cool. You probably see lattice structures on a lot of things that don't need lattice structures, kind of like what you said. But for you know, specifically for the O&P industry, where we're dealing with some things that have a thickness of two millimeters, three millimeters, four and five, you know, if I heard you correctly, you said there's not really a lattice that's going to do much for those thicknesses. Can you just share a little bit about just some design considerations when you want to maybe introduce a lattice structure? Or I know we've talked about oxatic patterns and things of that nature. I think our listeners would really benefit from that.
Speaker 3:Absolutely so. If you're talking about introducing a lattice structure, congratulations. You're trying to take advantage of, you know, 3d printing as it can enable. But there's some things to think about, like do you really need a well, not even need is a lattice structure the right thing to go in here, when you're talking and I'm speaking more about kind of the coarser technologies, even SLS and MJF and FFF, you know, if you start getting below five millimeters, you're starting to get into design space where it's really tricky to design something with any sort of structural integrity that will that you can customize.
Speaker 3:You know everyone wants to talk about meta materials where, oh, if I design it this way, I get this response, this way, I get this response, and it's all based on the same base material. Well, that's wonderful. But if you don't have that much room, you can't vary that much. So. But there's a ton of design options.
Speaker 3:When you're down that low, you can do buckling patterns, you can do waves on a surface so that you know when you start to press it collapses and that would actually expand it laterally, which has some other design impacts which could be beneficial or bad. But generally, when you get that small, start thinking about surfaces or like a uniform surface with leaf springs sticking out it's hard to describe this, but like little lines sticking out diagonally. You can. Auxetic structures are a great way. I mean those are essentially you take a sheet of paper and stamp a pattern into it and this pattern allows the fabric or paper to flow in a certain way. And all of these things are the tools you should look to when you start getting into these narrow spaces. More often than not, you're not going to have a satisfactory lattice for that sort of thing.
Speaker 1:Let me kind of add, anti-laddus propaganda is really welcome here. So I'm glad and it warms my heart, brent, to hear you speak about lattice generally. In our market now it's everybody's like they're for everything. Just throw a lattice at it, let's do it, you know, and so they're just so overused. It seems like really logical to integrate kind of automated or an unpleasant level as well, kind of FEA and all these kind of these kind of testing software right, to have some kind of testing ability where you would just say, well, test, which would be the most comfortable, right, rather than having to do it in a very kind of procedural way and then to end up with a file. But then of course, at the same time you mentioned before, then we have to interface with the machine, and we don't really understand the machine, you know, do you see a lot of progress being made there as like a more base kind of way to really kind of interface with the, with these 3d printers, rather than just using a file?
Speaker 3:When you're using lattices, it's so important to keep in mind the context of the problem you're trying to solve. You know lattices are complex and they're difficult to use and really you're trying to design some macroscopic behavior at a microscopic level and then hope that it scales. This is why you know other design methodologies can really have a lot of benefits rather than lattices as your first stop. As far as you know, interfacing more closely with the printers and handing printers more than just a file. There are tons of companies doing integrations, even with Fusion 360 that I'm on working out with, and HP in particular, you can take your files, connect to your printer through Fusion 3 Fusion and then send the models and customized print parameters and all sorts of things that are really going to should enhance the experience and the control for the user.
Speaker 3:The last thing I'd say is a lot of this stuff with the 3MF Consortium is working to expand its capability to communicate design intent to printers.
Speaker 3:So it's a manufacturing format and the goal here is to, you know, be able to give more information to the people who are to the printer that's going to consume it about how you intended this part to be manufactured. So I've spent this whole time talking about how much work there is and how difficult it is to do so, and I also talked about the value of it. So if you're, if you found yourself in a position where you have a process, a workflow you're doing often enough where it makes sense to automate it, then you know, just be very cognizant of your inputs, your outputs and everything that stays constant, constant. So if you have a shape that you use over and over, maybe you use this bracket over and over you don't have to do a lot of that stuff on the fly. You can pre-compute as much as you want. So again, your inputs, outputs and constants, and then just really keep it focused on, on that. Basic scripting can get you really far in many CAD programs, or, you know, design engines.
Speaker 1:I agree completely. Oh, thank you so much for being here today with us.
Speaker 3:It is. It was a pleasure and this space is rapidly evolving, so it's a lot of fun.
Speaker 1:Well, yeah, thank you so much for being here. The two Brents, I think the other Brent thank you also for being here today.
Speaker 2:Oh, this was good and I think our listeners will be able to take away really a lot specifically on the design and being just super intentional about it, but also know that it's it is hard work and it's okay to get help. So I think those two things really resonated with me.
Speaker 1:Awesome and I hope they resonated as well with you, our listeners. Thank you so much for today and thank you for listening to us and have a great day.