 Well, thank you very much for coming and welcome. So today, we are going to be talking about the design of an open source modular 5G capable container-based scientific data capture multicopter. Quite the mouthful of a title, but yeah. My name is Mauro Barajero. I'm from South Africa. I'm doing my master's degree in engineering, and this is essentially my master's project. So yeah, let's get into it. So just a shorter title that's a little bit easier to say, and yeah, that's what we call it. The land is in Yali Hexacquad. And it's supposed to be a solution for scientists and for researchers. So let's kind of just dive into it. A little bit of an agenda of today, what we're going to be looking at, what we're going to be talking about. So just kind of diving into the uses of drones and everything and what there is available to people. We're going to talk about a problem identification, seeing what the issue is and how it relates to researchers. Then we're going to talk about our proposed hexacquad. Go through a little bit of a design breakdown, give you some insight into what we were thinking about around our design and why we made our decisions. Talk about some performance that we have so far and then of course some future work and then contributing and the documentation around it too. So without a further due, let's just kind of get right into it. So the uses of drones that become really, really big in our world today and essentially they're these masterful tools that you can really do anything you want with them. It's really limited by what you actually attach to them. So just as some examples over here, we have them for those last mile delivery kind of services, delivering medical supplies and also you have it in smart farming applications where farmers are now kind of using drones to instead of applying a single blanket solution across a crop, you kind of now having localized solutions being like, okay, we only really need fertilizer over here. There's pests over here. How can we deal with that kind of thing? And then you get these more customized drones which researchers are using. So there's kind of picture in the bottom left over there where they're touching a whole array of different kinds of sensors onto it and then performing a whole bunch of different kind of geological studies and everything. Yeah, so just kind of a brief overview of what drones essentially kind of give us and how they benefit us is obviously the applications are far reaching kind of really dependent on what you attach to it, but significantly that high spatial and temporal resolution. So what researchers were working with years ago in terms of working with satellites and low flying aircraft is now becoming more accessible to the individual researcher. It's not, you don't have to be part of these massive research groups anymore like NASA or anything. And then of course, sensors are also becoming a lot more compact and the form factor is decreasing. So you don't need a helicopter to carry your sensor array or everything like that. And then of course, another fantastic thing is transferring risk from people to drones and getting them out of those dangerous situations essentially, so yeah, effectively what drones can provide people with is that high frequency going whenever you like, high quality, and then of course low cost data. You don't need to have a massive budget to do it anymore. So yeah, so let's just talk about some more interaction quickly here. So I just wanted to see just by a show of hands if anybody's actually worked with drones before or used them in their work or anything like that. Okay, cool. And then, yeah, just in terms of that, during your experience, I'm not too sure what you were doing with it, but if you had to modify it or you had to attach something to it, it was kind of quick and simple and easy or a little bit frustrating and yeah, not so good. So frustrating, yes, no? Frustrating, okay. So essentially what the idea kind of ends up around is if you're working with something that's really not designed to be tweaked with, it ends up looking like something like this. So you can see somebody's gone and modified and hacked the whole thing and taken it off and essentially duplicated a whole bunch of hardware just to try and get their function to work, right? So I mean, that works if you know your way around drones and you know everything around that, but if you're, I don't know, for example, a biologist who's been working in the field for however many years to kind of do this is that's where the difficulty comes in, right? So let's just talk about the problem, right? So the problem really lies in and around what you can actually get and what's available to you. So if you're working with something that people are doing in industry already, it's really easy and you've got these fantastic, really, really amazing world-class solutions like those DJIs, especially for farming applications and filming and everything like that. They have an array of sensors which are top-notch and quite form factor has gone really down. But then the other kind of option, if you don't have that access, then you have to look at something else. So you want these massive custom drones. So you've got a massive film camera that you need to attach to it or I don't know, some kind of wind-sensing array. So this really, really caters towards people who are working in industry and have kind of a defined solution essentially and it's really easy for them to access. But as soon as you become a researcher and you're trying to attach these little niche sensors onto it, it's like attaching into a DJI, you have to go and hack it like in the previous image and then all you have to go and spend a lot more money on a custom drone for your tiny little sensor or whatever. So that's essentially what we're looking at and that's kind of the problem that's essentially limiting people around the world today. So this is just a brief of what's kind of the problem. So first of all, if you do manage to mount your sensors and everything onto the drone, how are you gonna reproduce that? How are you gonna be quite exact? How's somebody else gonna do the same thing as you? It's not that easy. There's no, yeah, it's a bit of a hack so it doesn't really work. And then of course, there's limited access to contextual information. So fantastic, you find a way to attach your CO2 sensor to the drone but now you need that other information. So it's fantastic, you have readings but what height, what temperature, what everything else? So that's where you have to get into that duplication. So integration takes a lot of work and expertise going back to that example of simply like a biologist. Maybe they don't have that expertise and it's difficult to integrate, right? So again, it's thinking like, how can you, how are you gonna power it? Where are you gonna store your data? How are you gonna transmit your data? Do you need to add extra equipment on? And then also practically drones crash a lot especially when you're developing them for yourselves and I just have a quick video here just to show you. This was the first time we went flying with my supervisor and I was like, oh, we're not gonna crash. Let's see how it goes. And this was one video. Yeah, so pretty chaotic and propellers come off and it's just a massive domino effect, really terrifying kind of thing. So that was just a drone we were working with and used for, I was using to kind of learn, get around them essentially. So yeah, all right. So what do we propose? We are proposing some kind of modular customizable multi-copter which is intended for researchers to use. It's supposed to be user friendly for them as well. Kind of, it must be accessible for them. I mean, if you have some kind of solution that's really complicated or really difficult to integrate to, it's gonna be, they're not gonna use it anyway either. So let's just talk about some key requirements essentially for us to achieve this kind of goal and get people using drones. So we wanted it to be a simplistic design. Kind of thinking about a flat pack kind of thing. If you get that, anybody can kind of get the components, put it together and it's easy to work with, right? You don't wanna have these fantastically complicated things where you need an engineer or somebody else to go and put it together for you, it's just not practical. We wanted it to be quick to manufacture and acquire parts. So I mean, an example that my supervisor generally talks about is you've gone to some kind of startup drone company or anything like that and they have these fantastic drones and complicated manufacturing processes but what happens when they disappear? If the parts are easy to acquire and you go through these crashes like you saw in the previous video, it's you lose that equipment and you need to then go and explore another solution. So it's impractical. Of course, we wanted to use open source software and we wanted people to be able to integrate and access however they want. We allow them to change things and work with it however they like, right? And of course, user-friendly, fully documented. We're talking about getting people who haven't worked with the drones before to get into working with drones and allow them to use it. So it needs to be fully documented and needs to be really well built essentially so that they can follow along, kind of a Lego follow along style. Here's this piece, this goes there, how do we attach it, everything like that. And then of course, number one, easy to integrate the sensors, all done you. We don't want you to just be using these novel CO2 sensors if you wanna use a camera. Go for it, you need to be able to use it too. And then it all boils down to convenience. If something's not convenient, I can bet you you're not gonna use it. So that's essentially kind of our whole idea and what we wanna go for. Cool. Now designing the Eniala Hexa Quad. So again, just have a couple of things that we're gonna talk about. I'm just gonna try and give you a little bit of insight into the thinking that we went into and behind these different aspects. So yeah, we'll just jump into that quickly now. Okay, so key number one is the structural frame. Okay, sorry, we're gonna talk about specifications first. So when we were talking about this, we were kind of many looking at providing a heavy lift kind of option as well. So we didn't wanna limit it to the small payloads kind of thing. So if you wanna work with a bigger payloads, you have that access to do it. So we're looking at around five kilograms of payload. We yet to test exactly how far we're gonna go with that, but that was just kind of a baseline kind of figure. We thought about having it both a hex and a quad because that gives access to different kinds of people. So if you need that payload with a hex, you get it. But if you only have the budget for a quad, you can have that too. Simple connection methods, we really wanted it to be, like I said, easy to put together again and not have to have these complicated methods. And then expect a fly time 15 to 25 minutes with the payload on. So that's, we feel a bit more of a useful time. If it's no good, you can have five kilogram payload, but you're gonna fly for 10 minutes, right? So moving on. Okay, our frame design, obviously key structural part of the entire drone, everything kind of connects on you and that's, if you're without a structural integrity drone, it's not gonna really work. So we kind of broke this down into three main parts. A central hub, the landing gear, and then the rotor booms or arms. So we're gonna dive into that kind of now. So talking about the central hub, this was an interesting problem to solve because yeah, we went through a whole design iterative process essentially kind of considering how do we want people to interact with this? How do we want them to access it and all that kind of thing? So it started out with this cake tin kind of design in the bottom here where we thought, okay, you can put things in the middle and you can protect them and weather seal it and all that kind of thing. But then that was difficult to reconfigure. If you wanted to change it to a quad or get back into that middle, it's really like, ugh, a bit of a pain, right? So then we kind of shifted towards a modular chassis kind of unit and that was a little bit easier to kind of, if you wanted to be a hex, cool. If you wanted to change it to a quad, a little bit easier to change. Easier to have those cutouts and everything for your quad configuration compared to your hex configuration. So that, yeah, that's kind of developed and it's pushed us toward a modular chassis unit. So this was kind of our first area that we looked at and that's essentially developed and evolved into what our current central hub is today. Just some kind of considerations in what we thought about in the design of this central hub is, of course, weight, strength, shape, surface area, and then mounting strategy. So when it comes to drones, especially multi copters where you are providing the lift with propellers, there's a limited weight budget. You don't want to be contributing a lot of your performance to lifting your frame because then you're losing that on lifting your payloads. And then, of course, strength. Like I said, the central hub basically holds everything else together. If it's too flexible or not rigid enough, things are gonna move and it's gonna be the control systems are gonna get a bit crazy and not work so well. Shape as well, shape was an interesting one. Do we go for a regular hexagon kind of shape? We wanted a bit more elongated because then you can just have these simple box kind of things that somebody can design and attach onto it a little bit easier to attach. And then, of course, surface area as well. Like how do we optimize the surface area? Is it better to be long? Is it better to be round? All those kinds of things. And, of course, using that plate as well for people to mount things onto. So this was kind of guided with some simulations as you can see in the image. So we're doing some stress calculations, checking the regular forces and everything that we'd expect to be applied. And then there was also things that weren't quite straightforward as well. So thinking about is there enough space to hustle the components actually? And is there enough, do our holes even collide or something like that? So it's things that don't come immediately apparent when you're in the initial design but you need to consider it because otherwise you get to cutting and be like, oh, we've made a mistake. So it's the kind of idea of measure twice, cut once, essentially. So yeah, essentially when it came to designing this was kind of balancing all these considerations to have something as practical but also works within our weight and strength kind of budgets. So what we ended up with is a three part kind of central hub. So you can see them divided here where we have a top plate, a bottom plate and then a battery plate. Top plate kind of housing your key components and everything like that. Something that you really need to access a lot of the time. And then within the middle you could have kind of our power distribution essentially something that you need to be able to access but you don't need to have to access it every single time. And then of course a battery payload bay. Those things take up a lot of space. So we needed to have something that's also easy for you to swap a battery in and out of. It mustn't be a mission to get in and out of that bay. Cool, moving on to landing gear. So landing gear was a bit of an interesting kind of problem. Again, also iterative and I mean there are many different kind of existing designs out there. So as you can see on the left hand side here I just kind of had a bit of a sketch idea of what's out there and what people are using and all that kind of thing and kind of evaluating okay what's the pros and cons of each kind of design and everything like that. So we ended up going for something that looks a little bit more like this and that's mainly from experience and from those crashes essentially. So I think the typical thing everybody thinks about when they think about drones is that T-shape kind of design and every single time you crash we break a leg. Every single time, crash, break a leg. So we essentially kind of went for something more like this also backed by some simulations and everything like that to say okay this is what we think is a bit easier, a lot simpler in design and less likely to break as well. Moving on to the rotor arm. So this was a bit of an interesting design problem because it's heavily reliant on the propeller sizes as well. So there's a whole bunch of optimizations again that needed to go into these kind of things. Especially because the rotor booms are part of the drone that your full force of essentially your power system is being applied through, right? So there was again interesting design into shape, dimensions, length and of course strength. So in the next picture here what we have is we, just to show you, we gave a thought into, okay do we go for a square cross section of a tube or a round cross section of a tube? Each having their kind of different benefits. So these graphs were on the left hand side comparing strength to length of the two different cross sections and then on this right hand side was also again which, how they displaced essentially. So for a different length how much would it displace if you applied a singular load at the end of the pole? And that kind of guided us into choosing, okay what would be the optimal solution? What would be the best weight saving with give us the best performance? And then again another interesting thing that we saw is, so our rotor boom arms are a little bit longer than they need to be and the whole idea is because there was a study done by two papers here where the closer the propellers are to each other they more they influence each other. So I'm not too sure if anybody is a Formula One fan but they generally talk about kind of things like aerodynamic things especially like to work in clean linear flowing air. As soon as things become turbulent then they're not so efficient anymore. So that's what's kind of happening in the graph over here where you can see the motor propellers next to each other the closer they are to each other the more they interact and the less efficient they come. So that's kind of guided us to say, okay let's try find an optimal length away where they don't interact too much and they get that better performance. So these are just all bunch of different kinds of considerations that have gone into our design and our length and all that kind of thing. So but there's a very unique part that's missing from the central hub and the frame design and that's kind of these complicated plots that plugs them all together, right? And so you say how do we get rid of or beyond these complex manufacturing techniques from CNC and all that stuff and it's pretty simple actually of printing of course. So there's actually, I think there's about 47 different individual components on the drone and 32 of them are 3D printed. And again that's accessibility. The drones crash a lot. So being able to change those parts is really, really key and you don't wanna have to go and wait a month or two for a manufacturer to get back to you but here's your fancy carbon fiber part or your fancy aluminum machine part kind of thing. And of course 3D prints they give us that versatility. They allow us to create these complex parts in a short amount of time but they also can be functionally loaded which is fantastic. They also allow us to have fast iterations so I went through a couple of different iterations of the motor design. Tweaking things here and there that you don't think of initially and it's a really easy process to follow when you can have that fast iteration. So just to give you a little bit more insight into how they break and everything and how we can design our 3D printed parts is we can design them to be broken as well. So essentially we have these complicated expensive parts in the central hub which are made of carbon fiber and it's expensive. We don't wanna be breaking that and we don't wanna have to be buying more and more. So what we can do is we can design 3D printed parts to act like crumple zones in cars. So we actually had an example of this. We had a motor, the leads came disconnected for some reason and we were, oh, debugging was a bit of a nightmare but what ended up happening was it fell from 20 meters in the sky and it crashed and landed and the landing gear went straight into the ground but what happened is the landing gear sheared and so that's what it's supposed to look like at the top there but you can see they're all kind of, oops, got cut off. And what ended up happening is we didn't actually break anything else not a propeller, not a carbon fiber tube, not a carbon fiber plate, just some 3D printed prints. So 16 hours later we have our prints backing up and ready to go flying again. So that's really quite key, especially if you're thinking you're going out into the field and you're trying to do flight tests, you've been planning this for months, you get there, flying for a week, first crash, first flight you crash and it's just what do you do now? So I mean you could take a printer out with you to do that kind of thing but also if you think about people working in remote areas and Arctic base in the middle of Africa or something like that, if you have something that breaks it's easy to re-manufacture because you have the components there. Cool. Power train, another complicated thing and this is probably the biggest barrier to entry for people who don't know how to build drones essentially. So there's a key components that you need to consider and these all need to agree with each other for you to get the performance that you desire out of the drone. So again, consists of propellers, electric motors, speed controllers and then of course batteries. So the key decision making or design task that you have to do is again propeller and motor matching. So propellers are designed to work at different speeds and provide different kinds of thrust and if that speed does not match the optimal output speed of your motor you can have poor performance. So again, it was a nice iterative design, went through looking at different motors with different propeller sizes, kind of got a ballpark figure from looking at manufacturer data and as well as kind of the theory of propellers. So we were guided here with a bit of simulations but again, this whole process was there's no single way to do it. There's no kind of like plug in here, get out the answer over there. It's really much, very much an iterative process. Yeah, and then of course another big thing that we're looking at as well at the moment is battery considerations. So batteries are the biggest contributor to weight on drones and we really wanna optimize around that and there's two kind of chemistries that we're actually looking at and I'll try and explain why. So we have a lithium ion batteries and we have lithium polymer batteries and both of these have different benefits and kind of performances. So your lithium ion batteries, they've got quite a high energy density. So what that means is you have a high capacity for a lower kind of weight but they don't discharge as much currents. So they don't have that punchy kind of very fast agile movement. It's more of a leisurely kind of endurance flying, essentially, but whereas your polymer, lithium polymer batteries, so they've got a lower energy density, so capacity with more weight, but they give you that nice punchy current allowing you to have that agile, super fast kind of movement in every operation. However, it's a consideration. If you don't need to be racing and flying around, do you need that capacity? So it's again a kind of an optimization deciding what you need. And then flat electronics, the biggest part of it all, without this you won't really fly. So just to give you some insight into what we're using, we decided to go with cube pilots orange cube. It's a fantastic autopilot system. It's got triple redundancies in terms of its IMUs and gyros inside there and a compass as well. And it's also partially open hardware. But most importantly, it's versatile and easy to reconfigure and everything like that to integrate into. So on that cube as well, we're using a autopilot system called RGPilot. There's different versions of RGPilot, so you can use it for rovers, you can use it for submarines, you can use it for fixed wing drones as well. And also really it's open source and it's got a really good community backing it as well. So really easy to integrate and pretty straightforward to follow as well, which is key for what we wanna do. Sensor mounting strategy. So this has been something that we are working on and looking to continue developing. So obviously there's various different types of sensors out there and it's difficult to give you a solution for every single kind of design. So this is just a bit of an insight into what we're kind of thinking. So we're trying to create these simple little mounts essentially and that are easy for people to go and manipulate as well. So all these kind of mounts essentially, so it's particularly the two on the left, are blank kind of mounts essentially, which kind of just fasten onto with an M3 screw. But you can manipulate it and integrate it however you kind of prefer. On the right hand side, we're kind of developing a little bit more of a quick snap on kind of mount as well for ease of clicking on, clicking off, and then we can kind of develop a little bit more. But key to what we're trying to do in the sense of mounting strategy is provide as many different surfaces and options to connect to the drone as possible. So giving you, you can connect to the landing gear, the rotor booms, those tubes are the same kind of diameter. You can connect to the gimbal rails on the top and the bottom, also same diameter, trying to keep things consistent, but as well as trying to connect onto a surface plate as well. So if you, yeah, essentially just trying to give you as many different ways to connect as you possibly can. Cool. And then of course, 5G capable communications. So we're trying to integrate it with a 5G, with a, sorry, a Pi with a 5G hat that's using open interface. And what it's gonna try and do essentially is to port that communication over a regular radio frequency to a 5G mobile network, allowing you to pass over your command and control to that network, as well as being able to do high bandwidth things such as HD, video streaming, all on a single kind of link. So if you do want to know more about that, my colleague gave a talk earlier in the week and it was also at OSS, so I definitely recommend you to go check that out. It's a lot more in detail of what's going on and how we're gonna integrate it as well. Cool, performance. So we're still undergoing performance. I actually kind of started just before I came over here. So we've been having supply chain issues and everything trying to get components to us, but let's give you a brief outline of what we're trying to do and what our tests are gonna look like. So I think you're doing some range testing with telemetry links. So that's gonna do for both the radio link and 5G telemetry link. Active and static kind of endurance. So what I mean by that is hovering versus flying in a grid, for example. And then of course, how are we gonna have endurance with a current drawing sensors as well? So there's no good in being able to say these fantastic numbers with the payload, but you not actually have a sensor that's drawing any current or contributing there. We wanna compare those two batteries as well, like I said, just see what's actually better for our use and of course, doing some kind of payload limiting tests. So what is our actual payload? We know we designed kind of a five kilograms, but what's our max essentially? And then of course, as well with that rotor boom length influence. So is it worthwhile having those longer rotor booms or should we just save on the weight and put them closer to each other? So here's a little bit of a snapshot into what performance we have so far. So it's kind of the simulated results compared to what we've had recorded. Like I said, we haven't really got into it enough, but just to give you a snapshot there, for our simulated flight with a quad configuration, we had about 16.6 minutes of expected flight time, but we're actually achieving 23 and a half minutes with some old batteries. So hopefully when we get some new ones, we're gonna have a lot more, a bigger improvement as well and then get a little bit more comprehensive kind of performance idea. Cool, and then I just wanted to give you a brief view into how we're also looking at performance. So we're using logs to kind of allow us to compare flights here. So in the graph here, what is being compared is two different kinds of propellers. So we can clearly see a nice performance improvement with more expensive propellers compared to cheap propellers. So how that's being displayed there is a lower current use, which essentially will mean we'll be able to fly for longer. So yeah, cool. And then also just to give you a cost overview, because I mean, there's no point in having this if it's way more expensive than the options that they are really available. So for the quad configurations, it's costing around about $2,300. That's for all the components, cuttings and remotes as well. And for the Hex, it's about 27 for everything. And the expense comes in from adding those extra power units. So your propellers, motors and your electric speed controllers, that's where the expense comes from. But just to also give you a bit of perspective on something like the DJI in the video, in the picture earlier, that costs around about 5,950 euros. So I checked earlier today, it's about one to one for dollars. It's about twice as much as we can do. So yeah. And then some future work. So of course, we wanted to continue with our flight testing, get that out and do a detailed review of how we actually gonna perform. And we wanna achieve a reliable kind of hex and quad tune. So what that means is we'll go and tune it, make sure it's flying well. And then it's as simply, as simple as applying those parameters to the same flight controller and getting it to perform as we expect it to perform. Complete that integration again with a 5G and 4G communication. Get it working reliably and see how we can do with that. And then of course, completing our documentation. So I am busy with that at the moment, kind of getting everything documented as well as some processes. We'll get into that now. And then something that's also quite big and important is to be able to get this drone to be fully qualified under FAA. So people can use it for commercial use. It's no good having a drone and you can't really use it for people's work essentially. Contribution and docs. So like I said, we have documentation going at the moment. So it's all hosted off Github and everything like that. And it's incorporating how to guides for people to actually build it. How to kind of insert 3D nuts into prints as well because that's how we, yeah. And then of course, step by step instructions as well. So that's accessible and it's already up and everything like that. And we work on it. But also everything is on Github as well. All our CAD files and everything are there for people to go and change and do whatever they would like with them. And of course, like we love for people to be able to contribute as well and get involved and submit Github issues. So feel free and go explore it and see if you like it and let us know what you think. Yeah. And then of course, licensing as well. So we've talked about going under a certain open hardware license. It's fully permissible, permissive, sorry. And if you'd like to read nor about that license go, it's ready. It's available there. So essentially that's, it brings us to a summary of the Njalehqquad. So yeah, we've developed it kind of through iterative design and it's purposed for researchers essentially trying to get them the perspective that wasn't ready available to them before. And yeah, of course, being accessible and affordable, easy to use no matter where you are, really kind of thing and user-friendly, fully documented. It needs to be convenient for people to use. Hopefully we'll get this performance tested by the end of the year and have the first prototype out and ready for people to try and explore I suppose. But yeah, essentially the whole idea is trying to just give that individual researcher that access to that high quality, high frequency and low cost data essentially. So yeah, essentially that's the end of it. Just want to have some acknowledgments and give a big thanks to the Sloan Foundation for kind of providing us with funding to do this work. The people at Notre Dame University as well for all their help and everything, especially Linux for bringing us out here to talk about this as well. It's quite a cool experience to come out here and actually show people your work. It's quite fun. So yeah, and there's a little contact me. But thank you very much for listening to my speech. If you have anything or any questions you'd like to talk about, let's get into it and we can, yeah, I'm not too sure how long we have. I think we have a little time a little bit. Yeah, thank you. Cool. Yeah. So the question was, do we have other people already working with it or trying to use it and stuff like that. So the first person we actually have working with it is also at the University of Cape Town. They've just started ordering parts and trying to get it to work and everything like that. And so just to give you a bit of insight into what they're trying to do, they're trying to do wireless charging on drones. So they are taking it, our base design and they just change it slightly. But yeah, so that's essentially the main person that we've got working with and we're gonna see how they perform with it. So nobody else is really working with it just yet because we haven't got everything finalized in terms of he has a completely well tuned everything. Yeah, but getting there eventually. So it's still literally just coming out essentially. But yeah. Yeah. Any other questions? We're okay. Yeah. Those are my thanks for a very nice talk. Thank you. Appreciate it. Mm-hmm. Have you also been using open source tools in the design? Like what have you been using for the mechanical drawings and for the simulation? Okay, so that's one snag that I've been using and that's because of what I have access to at the University. So I've been using a proprietary software. It's SOLIDWORKS and it's fantastic for design. It's something that I've learned and used. But what I'm trying to do is I'm going to make sure that those files are also in a format that is usable on open source CAD software as well. So have been using SOLIDWORKS but that's what I've had access to and everything like that. But yeah, essentially. So what will end up happening is we'll have, I suppose three different kinds of files. Have the SOLIDWORKS files that are on there have simple SDL files. And then, yeah, another format that should be able to port between SOLIDWORKS and other CAD softwares as well. Yeah. So for the simulations, we've been just guided by SOLIDWORKS actually has a simulator within it. And it's just a given indication of performance as opposed to you do get these other, they called finite element or analysis tools which give you that stress analysis that kind of blue looking graph. But yeah, essentially just been using the ones integrated into SOLIDWORKS. Yeah. No, no worries. Cool. All right, I think. Is there any questions online or not really? All right, cool. Well, that's I suppose that's it. Yeah. If you have any questions or you wanna talk to me afterwards, I'll be here. But anyway, thanks very much for coming. I really appreciate it. Hope you guys enjoy the last day of OSS. Awesome. Thank you.