Hello and good day to you from episode 28 of our podcast series, Project Breakaway. A metaphorical and literal time in the day when we here at Predator Cycling, take some time away from working in the back shop to come and share with our listeners what we're doing, how we're doing it, what it takes to do it, our ideas, our innovative success stories and even our missteps and failures. If you find yourself with an interest in bicycles, composite manufacturing, out of the box design or even curiosities beyond, I encourage you to stick with us, settle in and learn a little. I'm Courtney B, co-owner and project manager of Predator Cycling. I'm here with my partner Arm Goan, the other co-owner, CEO, lead designer and engineer and unofficial physics specialist. How's it going, Arm? It's going good. I was going to say physicist, but that would not be a lie. That would be a lie. Also unofficial physicist. Yes. Yeah. Physics. Okay. So, welcome back. We just uh did a discussion in our last episode with Kurt Chan from Ansis. And we discussed uh optimized topology and simulation. Mhm. And it even though it was frustrating because our house audio didn't work that day, but Kurt sounded great. Yes, he did. Uh, it was a good conversation. Yeah, no, it's super interesting. I mean, it's something that we've been talking about a lot and we've been working a lot on. Um, and a lot of the tools that we've been using. I mean, obviously we're Ansis people here, so we use uh Ansis Discovery and Ansis Mechanical for all of our topology optimization needs. But, um, it's been a super powerful tool for us to be able to leverage um 3D print into production. Right. So, it's timely. Okay, so our discussion was about simulation and new age and at the end of the discussion we talked about additive 3D printing titanium metals, blah, blah, blah. So, it's timely because of an unfortunate event that happened during the team pursuit at the Olympics two days ago. Yeah. Uh, during a Velodrome event. Yeah. So, just quickly to talk about the team pursuit. It's on the Velodrome. So, it's the circle track and it's four guys on the same team just doing a time. Yes, so uh track just real quick is a You're a this is your event. I love track. Track is my event. I love track, track is my event. But you're not a pursuit guy. I did do pursuit. I did do team pursuit. Um, uh, yeah, I love the pursuit. I I ended up later um towards the end of my racing. I did uh mostly the the Karen and the Kilo. Were my event, but the track is a 250 meter um basically NASCAR track, oval track. Um, it's banked at about 45 degrees. ish. Um, in the corners, each track's a little bit different, but that's the general thing. There's not like an Olympic standard. There there is Olympic standards, but I don't really want to go too far. It's a rabbit hole. But all the Olympic tracks. they're all 250 indoor tracks. Um, and track bikes are special because there are no brakes. Um, there is a fixed gear, so you pedal backwards, you go backwards, you pedal forward, you go forward. Um, and then like one gear. Oh, you mean like. Like an old school bike. No. Wait, no. It's not good. You pedal backwards and that was your break. Right. So this is you pedal backwards, you go backwards. It's a track bike. It's a fixie. Um, that's what is used to race on the track. Um, and the team pursuit is a special event within on the track. Um, where you have a team of four, uh, men or women. Um, and you have two teams, well, you have multiple teams where you seat them, but then when you're racing, you have a team on either side of the track. So, um, you basically kind of like try to catch each other. Um, but you're racing for time. So you have four riders. So there's a team on the other side of the track. Yeah. And the okay, I'm jumping ahead. But I didn't. But yeah, so there's basically a team on either side of the track. And you're basically, uh, racing for time and then once you make it to the finals, uh, you are racing to beat the other person. And, uh, you typically, uh, you go from a standing start. It's 4K. Um, so that's. Um, that's typically 16 laps if it's on a 250 track. There's also a 333. Um, but uh, yeah. So it's basically 16 laps, 4K. And they're going really fast. They're going over 60 K an hour average speed. It's about 65 K. They're averaging, so they're flying. What is that in miles per hour? I'm doing that really quick in my head. That's like around 30 miles an hour. A little over 30 miles an hour. Right? Yeah. That doesn't seem that fast. Right? For four minutes. Oh, I don't know. I mean, they're moving. That's for me. Anyway, so, uh. The what happened quickly, I'll tell you and then you can tell me more. Uh, so the rider in the fourth position on the Australian team during the team pursuit, uh, he was literally just going around the corner. Yeah. And uh, his bars just snapped off. Yeah. And he crashed. Yeah, my heart sunk. I was watching it live. You're like the only person watching it live at 3:00 in the morning. I think yeah, it was 3:30 or something in the morning. Um, yeah, I literally was like, oh, I think I screamed in the shop. Because I saw it. Um, it was terrifying. Um, I mean, he's luckily, he's okay. But it was just like, there was just like, he's going. And then all of a sudden he wasn't going. Right, so I I just watched the video. Because you told me about it. But I didn't see it because I was sleeping at 3:00 in the morning. Yeah. One of us has to sleep. Yeah, so, uh. I just watched the video and um, yeah, the bars literally just like, I mean, they didn't even like. When you told me they snapped, I thought, oh, they broke in half. Well, they they broke off from where it joins to the frame. It's a special frame. It's I mean, this is all like, I mean, we just from footage and what people have shown pictures of. It's it's inconclusive of exactly what happened to cause the break. They're still investigating. It's investigating. But um, and this is also not a thing to like point fingers of any kind. happened to anyone. Yeah, and it just it's super relevant to what we were talking about last podcast. Right. So like to bring this back up is really. So basically Team Australia is utilizing new tech. Right. and 3D printing. Yes. Uh, titanium, which we discussed with Kurt. Yes. Um, it's slowly becoming more mainstream and that's what we are eventually going to start diving into here in our shop. Yeah. I think. Yes. Not maybe bars, but titanium parts, components. Yes, we're looking at metal, we're definitely looking into metal. Uh, we're very uh into the techy side and simulating first, which is why we talk to Kurt and NASA all the time because we're into simulation for it. Yes. And I I think, you know, I I know we've talked about these things a lot on this, especially on this podcast. Um, but um, this concept of um, well, okay, so with titanium 3D printing or any sort of 3D printing, you there's no penalty for making things super light. So if you're adhering to that kind of concept, the lighter a part is in the form of being less material, um, the less it costs to make in 3D printing. Where in subtractive manufacturing, you start off with a massive block of titanium that costs a lot of money. And then you. And then you. You cut it out. You have to cut it out, so the more material you remove, it costs you more because one, you bought a bigger piece of material and two, you had to spend more time cutting it. You wear out tools, you wear out machines. But that's for mass manufacturing, right? No, just for anything. Then why are these 3D titanium printed parts so expensive? Just because it's new technology and the printers obviously just are. So like. So some of some of these 3D printers, I mean, when you're talking about laser 3D printing, you could be the the very basic entry level machines cost, you know, about $150,000 ish and they go up to about, you know, three uh uh $750,000. Um, so you're talking about very, very expensive machines. Um, and when you're talking about 3D printing like this and the the types of tech that most of these people are using, um, it doesn't print fast. It prints pretty slow. So you're going to go through like maybe one build chamber like a day. Sometimes it takes two days to go through a build chamber. So you could have this part on it, let's say a half a million dollar printer, um, and you can only cycle maybe 300 prints, um, a year. So. But people are going this route because they can make cooler designs, different angles, stuff that a CNC machine couldn't cut out. Mhm. from a piece. Mhm. of metal. Yes, absolutely, and there's it's very interesting. I mean, that's why we're exploring it for some of the some of our parts that we've done like the the genius bottle cage and um some of our cleat adapters and parts. Um, so absolutely, it's very, very interesting. Um, but when you start talking about um, um, 3D printing in in materials, one of the things that I don't want to say limitation, one of the the side effects of uh of going down the 3D print path is is time. Um, it's usually not the quickest process and that inherently will drive your cost up significantly. You buy a half a million dollar printer, um, and you can only do 300 prints, let's say a year. Like you do the math out, it's a very expensive just to recoup your cost on the machine. Forget the the material cost that's really expensive and sometimes running the machines can be expensive. So, But it's like the cool new thing. It's the cool new tech. Because there's not I mean, not just this specific company and there's other companies that have release parts and components that are 3D printed. Yeah. It's very cool. It's very cool and and, you know, especially I know we've talked a ton about topology optimization and optimization of of materials and and and and lattice structures and generative stuff. And and 3D print really lets you benefit from that. I mean, a lot. Like you can really benefit from that material stuff. Now, the the, you know, we talked. Physics is only good as the data you give it. Solvers are only as good as the data you give them. So they're optimizing for the data points that you give it. So. So I want to talk about physics. Yes. Uh, design and simulation physics versus live live testing of the physics. Right. I am a physics skeptic. Skepticism. Am I saying that right? Yes. Skeptic. Of physics. Yes. Now, uh, I'm not a scientist. No. I didn't actually skip physics in high school because it wasn't a requirement. But, uh, what usually happens here at the shop is arm will simulate something. And he'll say, uh, well, I did all the math. And it it math shows this. Mhm. And this will happen. Mhm. And I say, okay. And then we make said part. Uh-huh. In reality. And then and and I know you know, like, oh, there's all sorts of different things that could happen. Well, for sure. Uh, but it never works. Like, I just feel like the physics never works. Well, I mean, it depends. Something like this, I was with Kurt with skeptic, I don't know why this brain fart. Uh, where it is. Anyway, I was a skeptic of 3D print of titanium. Because to me that seems dangerous. And I feel like I was proven right. Yeah. So I mean, it's new tech. So, um. Okay. So again. It comes back to this idea. Um. And yes. And so, okay. Well, two things. One, a lot of the things that we test out when we're testing out prototypes are literally taking things to the absolute extreme. So, um. When we do that, yes, we have we have a higher fail rate than we when we don't push it to that absolute extreme. But you got to know where your limits are. So you can kind of know how far you can take something. I think for to me, an old school mentality is I've got a big block of metal. Yeah. And it's sturdy as hell and I'm cutting from it. And that should be a fail safe. Like that. And to me, it's like I'm creating this thing out of metal. And I might have thin spots in certain places. Mhm. And that seems unsafe to me. You have a point and also it's it it well. have a couple points there that I think are really relevant, especially in the 3D print realm. Um, I I come back to my original point of topology optimization. Simulation is only as good as the data you give it. Um, and what I mean by that is that the compute like software and simulation cannot um predict how uh it's a topology optimized part will do. If you're giving it loads that are not known to it. So, for instance, um handle bar. The even the software can fail. I mean if it doesn't If it doesn't know. Well, I I wouldn't call it a failure on the software. I would call it a failure on the user not fully understanding the loads in the situation. But the software can't know everything. It the software It only knows as much as they've made it. No. Well, when you're talking about forces and loads and especially stuff like answers and stuff, um, it it will account for all forces and loads that you apply. Now, the problem that you have here, you have, let's say, let's let's take this handle bar situation for an example. You have um ISO testing protocols and I know that this company that built these bars tests to those protocols. And exceeds those protocols. Um, which is great. And they do it in physical testing. In simulation realm, um, you have, let's let's think about the handle bar. You have a person with their elbows supported on the bars. When they're arrow bars. So you have, let's say, 100 pounds approximately static load there. You have the rider at an angle going into a corner. Um, at, you know, 65 kilometers an hour. So you have, you know, this centripical force pushing against the outer. You have inertia. You have gravity. You have all of these different forces acting on that part. And then, of course, you have the mounting hardware. And the mounting how that's mounted and connected to the bike and the deflection and the sturdy of that mounting. And how that's affecting the part. You know, all of these add-ons, then you have the whole thing of like modulus and fatigue. Um, and then like harmonics and frequencies like from the track. So there's a lot of outside forces. And you, you know, when you simulate, you typically don't simulate for all of those forces and all of those conditions. Because that would take forever. Do we? We we simulate. Um. For a lot of it. A lot. So when we run the genius bottle cage for instance, we have. We have about, I think about 12, we have 12 loads that are for the use case of the part. And then we have another like 15 different simulations we run to test for um repetitive use, fatigue life. Um, vibration, um, all kinds of other tests that we run as well. To to test for the longevity of the part. Um, and then we actually cycle test them out. That's something simulation can't do cycle testing, right? It can, it can. You can do cycle testing. You can just repetitively in simulation like put a load on a. Well, how would that? Uh, yeah. You can. Ansis can actually. So, but the thing is, okay. So here's an example. So this is all great. And. I mean, this kind of goes down. I think what you want to talk about a little bit later. But like, you can do all of that stuff. So if you're, okay, in 3D printing, right? You design the part. In, um, in software. And then you basically send it out to a proprietary, typically a proprietary software that's made for the printer. So that the printer can understand what you're trying to make. And then it makes it. Um, and then you have to take that part out. And there is a curing process or post process that you have to do for titanium 3D printing stuff mostly. Um, and you have to clean it up. You have to sand blast. So there's a billion fail points throughout the process. I don't want to say a billion fail points. It's pretty digital, pretty far along the process compared to older manufacturing protocols. But you're talking about like a finishing process. There's a finishing process and curing and that's where all the, that's where all the shit goes wrong. That's right. Yeah. So like when you take a belt. That's where all the shit goes wrong. So like if you take a belt sander and you start polishing out titanium bars. Um, and if you have wall thicknesses that are, you know, a millimeter. Half a millimeter. I'd go right through it. You could go right through it pretty quickly. Um, and. The the problem that you get is is that the simulation and the data that you did was not accounting. For for shaving off that little piece of material there. And goes back to your point of when you machine something from a block. Um, you typically don't get into some of those small tiny details and those wall thicknesses because you can't. It takes too much time. It's too much effort. The tooling gets too crazy, it gets exponentially more and more difficult to remove that material. So you don't remove it. You just leave it. So you have these areas of like high stress that end up having blocks of solid material. Right. Well, that's the risk involved with manufacturing that I fear on on our end when we release products. Mhm. Because we can simulate all day. Right. And you. You trust the simulation. Yeah. I trust no one. This is true. But I. Yeah. And so when we get like writer. And then we get like for the genius water bottle cages, we get a customer feedback. About, you know, oh, the material's like stretchy or this bends or this wobbles. And that wasn't accounted for in your simulation. No, it was. It on the genius water bottle cage it was. Um, the one thing that we don't. Or I'm sorry, the answers didn't, you didn't show like a wobble. It does. Oh. Yeah. We have that. Why were you surprised about it when they said it? Well, I wasn't. Surprised. So, oh, so you're talking about is the torsional twist on the cage. Yes. Yes. So we test for the forces and act. But we didn't account for. What I didn't. I take that back. I didn't realize. That when the part, when the cage, when you put in the cage from the side and that angle and the weight of the bottle cage in it. It would have more than what I had anticipated in side to side flex. Now, there's no issue. On it from a stress point or uh uh deformation from having the part in and out or fatigue life of it. it just gives you this perception that it's loose or it's. There's just like a real-time comfort issue. Yeah, it's just, you know, it's just a user input thing. And it it hasn't shown to be an issue at all and it's it's exceeded every one of our tasks. But uh we saw it very quickly when we were doing, you know, our our small our cycle testing here of the physical product. Uh we we figured that out quickly. Um we've tried to strengthen that torsional twist. But you know, we're have to add more material to do it. Um and there's no it doesn't do anything. So. Right. Um. I think that's just my fear. Uh adding 3D additive prints going forward with our products. Yeah, for sure. I mean, it's always something you need to be concerned about. I mean, you always have to think about these things. And also need to be thinking about like, you know, real life conditions. And how people are really going to use it. Like, for instance, something that we didn't simulate for exactly. Um we tested the the um fatigue on the bottles with a vibration involved. But like we only tested like two types of vibration. So like we didn't test for like, for instance, if someone used it on a downhill mountain bike. Um, you know, course. We didn't test the bottle cage for that. Um. What vibrations did we? We tested for like what we thought would be a normal. Um um uh road, you know, road and like cross country cyclocross type um impacts and and and loads is what we tested for. Consistent. Yes. Not like. Yeah, yeah, consistent. Just like, you know, like. Yeah. Yeah, like kind of like just like a bumpy road, just like bouncing vibrations. Is what we were testing for. Consistent writing. Yes. Uh, okay. Uh, so then uh we wanted to talk about how we keep our simulations digital longer during the design process. We keep things digital throughout the process. And that's what makes a huge difference for us. Because one of the things that we do is is we obviously design everything on the computer. And um and then simulate everything digitally. And then we make molds or we do 3D printing directly from that. Um we don't on our 3D prints do post processing. Um we're very careful on how we do it so that the actual part that we produce. Is the part that we simulated for. Um and then we have a a kind of like a closed loop system. So we basically uh test our prints till failure. Um in simulation and we do a bunch of different types of simulation for it. And then what we actually do is we actually print it, make the part, process it, and then we do that same failure test on our testing machine. So that we can actually see what happens. And then make sure that what we see um the failure that we get on the part. Um we actually that's how simulation predicted it to be. And and if it didn't, then we know that something went wrong. Like we knew that something in the manufacturing process went wrong, something we didn't account for. Um or you know, we didn't set up our simulation correctly. So. because you kept it digital longer, your theory is that if there is a failure, you would be able to better identify the failure. in the manufacturing process. Well, both. It's a closed loop system. It's like a validation type system. It's like someone second, you know, checking your work. Okay. So like we designed the part, we simulate the part. Um, you know, did we account for this? Did we account for that? Like what happened? We assume that we were making the part like this. Well, then you actually manufacture the part. You're like, well, you know, we sanded it, we did this, we post-processed it. We ended up milling this one part off. You know, the part became different than the part that we actually designed. Yeah. Um, keeping that net part correct. Um, we're not correct. There's a long word, but keeping that net part. The part that was the simulated part is really important. Especially when you're getting into topology optimizations. Um, 3D print, any sort of like high strength to weight ratio parts, like that matters a lot. Mhm. And then one of the things that we do is is because we're able to test. And we're able to test in house to a level of accuracy. That's um, you know, very close to what we're simulating at. And we can validate that process and then come up with our own testing protocols. To be able to test production parts. Um, is really important. And we can validate that. And so if something goes wrong in our in our manufacturing process. We can catch it. Really fast. Like we can be like, oh, well, why? Why did that all of a sudden, you know, why did that deflection change? Um, you know, by 2%. Like it's it's never been out of that zone. So, let's play a not so fun uh imagination game. Uh, we have manufactured a part in the Olympics and it has failed. Why are you playing this game? Because I want to know. How would we react, respond, check our tech, use simulation to identify a problem? Because what the current situation is in the Olympics is that it took quite a while for the manufacturer to reply and respond. Yeah. To a failure. And the whole podcast is basically how we would discuss our missteps and failures. Yeah. And I'm going to future proof myself. Yeah, I mean, well. I mean, we have a part that fails. Okay, horrible. It happens to every manufacturer probably. I mean, it happens. You might not hear about it. Yeah. I'm I'm yeah. And it's in this giant public setting. And this you know, your brand is dependent on how you fix the situation. Right. What so so we've collected, we've used answers. We've collected simulations, blah, blah, blah. We've found, you know, it's been, you know, less than 12 hours. We're we're diving into our simulations, right? Yeah, I mean, the first thing I would, I mean, I mean, I probably the first thing I would do is just hide. I mean, make sure that the whoever's racing on it's okay. Well, obviously. But I'm talking about from a manufacturer. Engineering reverse engineering perspective. For reverse engineering perspective, the first thing. I would do is figure out what part it was that failed, the serial number of that part, which you should be able to know because you'd know which rider is writing what part. Um and then quickly look up test results and fatigue testing of the part that we have on in the data of our database. Um see if there's any other part, if there's anything that was out of whack or, you know, um anything that didn't match. And then if something did and I saw a fingerprint of something that was weird in that part and see if any other part that we manufactured had that same issue. And quickly notify those people to not write on that part. Right, because it's about survival and, you know, Well, also like I just, you know, I'm going to I'm going to make an assumption, there's there's three ways a part like that could possibly fail. Physics. One would be um installation, something a crash, I mean, somebody crashing on it previously. And having fatigue or a crack in it somewhere that wasn't identified. Oh, you know. Um the other one would be um. Improper installation, um if you would if you were accounting for depending on how the the clamping system and everything connects. It not being correctly torqued or spec or something like that. Right. That that could be another one. Um and then the other the other the processing part that I could see. Is something in manufacturing process. I would hope that our our our our testing would have caught anything that was improperly manufactured. Um or like an internal flaw. Um or a post processing problem, like we were talking about like sanding, removing, machining, something that was done towards the very end of it. That maybe didn't get caught in in the if there's a process that didn't get caught before testing. But somewhere there. It could go wrong. So the first thing I would do is go back and try and figure out everything that's in our control. And that we have access to on data and go through everything. Right. I mean, that's the first thing I would do. Nothing is a sure thing in exploration of new digital design and, you know, 3D print. Yeah, and then the other thing I would also do is I would go ahead and say, hey, this is where it broke, I'm going to run a new simulation based on everything I know. And should it have broken there? Right. That would be my next thing I would do. Um that way I would know like. Like, you know, that's weird, like you. Depending on where the crack or the break happens. You know, you can know a lot. Like. Oh, is it were we pushing that a little too far? Like was that like the thin height that was a high stress point and we tried to save 10 grams by reducing the mass. And it ended up being higher stress than we thought it was going to be. And it broke. Or was it, you know, a part was over clamped or. Yeah, that's what I would start doing. I'd start diving into all of the data that I have access to. Until I can actually get my hands on the part. And also know if like, you know, we have an issue with something that we assumed. There's an assumption that we made that was untrue. Oh, I'm just highlighting the importance of simulation. Yes, there's there's there's a lot of importance. And all of this, you know, crazy. But also like 3D printing, metal. Yeah, but also the big thing too to think about too is like all these things that happen, right, that this obviously is, I mean, it's it's horrible. But like, you know, it's also a learning experience on your product and knowing your stress points and and failure points. Um, and building a better product. So, um, there's a lot to learn, um, from from things like this. Yep. Uh, is there anything else you want to talk about, uh, in terms of simulation, 3D printing? No, I mean, I think like I I've already highlighted it. But I think one of the big things that's one of the things that's taken us a long time to get to get just right and to, um, uh, why we've we haven't brought out a lot of parts. Um, in the last years is is we've really been working on this workflow of keeping things digital, of validation, of testing in house. Of a lot of this kind of stuff is stuff that we've been working on really hard. Um, and it's it's really difficult. That's why a lot of people don't do it. Um, but yeah, it's it's a process we've been working on really hard. And it it helps us make our simulations better. It helps us understand how loads and forces are acting on things better. Um, and it helps us each time we cycle a part to make it better. Um, and safer. That's the goal. That's the goal. Okay, uh, well, things to mention real quick before we head out, uh, the Kio and Delta. 3D printed cleat adapter wedges are being packaged this week. Um, they're on their way to Amazon for fulfillment next week. Keep an eye out on our Amazon store site as well as predatorscycling.com for their release. We're also experimenting with the release of a speedplay cleat protector shim plate. These are a stainless steel plate that is meant to protect the speedplay wedge from the base of your carbon shoe from damage of wear and tear. Mhm. Um, and then our available thickness will be 0.8 mm thick. Yeah. So they're um thicker than um ones that you that speedplay used to make them. They've stopped making them now. But um they're thicker than those, um they're a little more um durable and they also have a a nice textured finish. So they take the wear better. Um, so yeah, this is kind of an experiment for us to see if this is something that we want to go down the road of. Um. Well, it's an experiment for your feet and not for Olympics. Yeah. Um, but they're designed around the new Wahoo cleat system. Wahoo is a company that purchased speedplay. Yeah. Recently. So. Yeah, so this is designed around the new cleat, they made a couple updates to the pedal and a couple updates to the cleat and this is for the new um Wahoo cleat. Great. So, uh, check that out. And then, uh, I think that's it. Wrap it up, Olympics are coming to an end here. Yeah. More cycling. There's some more, yeah, there's more track still. More track to go. Yep. More fun times. Yep. Okay, we thank you for choosing to take some time with us. And we look forward to. for the future breakaways. Like for us on Instagram and LinkedIn, Facebook, Twitter and in person here in Tennessee. We ask our listeners to please share, like and subscribe. We're available on all major streaming platforms. Thanks for listening, have a good one and find some time to break away.
EpisodeAug 5, 2021 · 30:29
← Podcast
Project Breakaway with Predator Cycling
28: Olympic Oopsies, The Simulation & 3D Print Frontier, Ep. 28
Hosts Courtney B and Arm Goan follow up on their previous discussion about simulation and topology optimization, connecting it to the potential of 3D printing in cycling production. This topic gains timely relevance as they analyze a recent "Olympic Oopsie" where an Australian team pursuit rider's 3D-printed titanium handlebars snapped mid-race. The incident prompts a deeper dive into the benefits and current challenges of advanced additive manufacturing for high-performance sports equipment.
Loading player…