← Outside of the Box
Outside of the Box
Robot Archetype Series Deep Dive: Melty Brains, Shell Spinners & Ring Spinners
01:04:05
Loading summary
Transcript155 lines
- [00:01]
Jacob
Christine's talking all things that spin as she talks to a few different teams about their melty brain shell spinner and ring spinner archetype robots. Any meltibrain can be built with just strong engineering foundations. People should not be as scared of them as they are.
- [00:17]
Nathaniel
If you build your rodent spinner well, which of course is the hardest part, it can really start to wear down the weapon assembly.
- [00:28]
Christine
Hello everyone, and welcome to another episode of Outside of the Box. And we are kind of, I'd say, kicking off my series on archetypes, which the, the episode that I already came out with was kind of the kickoff, but this is more the kickoff of the deep dive into specific archetypes and especially ones that are very similar. And the, the reason why I want to do that is because some of these similar archetypes are also very different from each other. And I want to highlight both those similarities and differences and talk about, you know, like, what is involved in making these robots and all of those things. So for this first episode, I'm going to focus on the archetypes that are really less common, I think, and I think are some of the most interesting and can be complex even, which is, as you can see, highlighted by my background, the more circular type of robots. So specifically shell spinners, ring spinners and melty brains. So to kick off this episode, I have Jacob here from the greatest challenge and SRI representing the Chunky, Chunky, the Georgia Tech's, you know, wonderful, beautiful 30 pound shell spitter here to talk with with me. So welcome both of you to this episode.
- [02:01]
Sri
Yeah, thanks for having me.
- [02:03]
Jacob
Thank you.
- [02:04]
Christine
Absolutely. Well, I'm really excited to get into this because probably one of the biggest things for me when I watch competitions and I see maybe people in chat or hear people talking is when they see something that is circular and looks similar to a hockey puck and kind of think that they're all the same and they're, they're very much not. So I want to talk about, you know, the differences and, you know, all of that. So I'll go ahead and actually start with you, Jacob, because I think melty brains are one of the most, like, confusing things for people because they see it and don't necessarily know like, what that is and how that's different from any other circular shaped robot. So just give me like the brief summary of what the difference is.
- [02:56]
Jacob
Yeah, that's a great place to start. So what a melty brain fundamentally is, is it's very similar to a Thwackbot, if you've Heard of those? So a thwackbot is just about the most simple robot you can build. It's a battery, it's two wheels, and then maybe a hammer on front. And thwackbots drive to the center of the arena and they spin up and what they do is they wait for their opponents to come and engage with them. So a multibrain and a thwackbot are really pretty similar mechanically and even electrically. But the key is, what takes a thwackbot to a meltibrain is software. It's programming its sensors. So a meltibrain has electronics in it that can feel how fast the robot's spinning. It uses what it feels to know where it's pointing, and then the driver can instruct the motors to pulse based on where the robot is pointing hundreds of times a second. And on average, the robot can go and attack. It can drive in a direction towards its opponents, and we call that translational drift. So that is what a multibrain is.
- [04:14]
Henley
Yeah, that.
- [04:15]
Christine
I mean that, that definitely is a good description of what it is. Now, sri, as far as shell spinners, other than the obvious, like when you see a shell spinner, I, I think most people know what that is because of the aforementioned shell. But how functionally is it different than ring spinners or melty brains?
- [04:41]
Sri
So compared to a melty brain, it's very different because usually the drive and the weapon is fully independent of each other. Where meltes use essentially their drive train to fully spin the entire bot and use the full weight class. So for greatest challenge, for example, the full 3 pounds as its weapon, a shell is usually fully independent. The weapon is different, similar to any other archetype where the weapon is independent. Shells and rings are actually pretty similar. The main difference is usually how the weapon itself is driven. The easy way to explain it is you can say it's a shell, kind of like a turtle shell. Another way to explain a shell spinner is like imagine you have your traditional drive base, say on a four wheel drive, you know, robot, or like an RC car even. And then imagine you put a bowl on top of it and the entire bowl spins. That's an okay way to say it because usually we like to call chonky at times a big blue walk. And that's how like the origin of some shell spinners started, right? Like Ziggo. Not Ziggo, Blendo was essentially a big walk that then they spun up with a big engine really fast.
- [05:53]
Christine
Yeah, yeah, that's a, that's a good way of putting it. Now for, for those, like I said, the reason, big reason I wanted to talk about this first is because there's so, you know, many questions as far as differences and that certainly like highlights what those differences are. I guess number one question for both of you before we dive into more specifics is why do each of you think that more people don't build these type of robots? Because they're pretty uncommon as opposed to other archetypes. And you know, like I said, I think that they're pretty interesting. But just hearing from your perspectives why you think people don't build them more
- [06:34]
Jacob
often, I can jump in and say from melty brains, it even starts right away. People can be scared to build the archetype even because of the name. They say melty brains get their name because it's so complicated to build one. It will melt their brain. And I just think this isn't true. So yeah, it's the people here, code and sensors and they say it's not for me. But I think that any melty brain can be built with just strong engineering foundations and people should not be as scared of them as they are.
- [07:19]
Henley
Yeah.
- [07:19]
Christine
What about for you sri?
- [07:21]
Sri
For shells, I think a major barrier is cost and or complexity. In the lower weight classes, like three pounds, it is very difficult to make a rigid shell that doesn't touch your own frame. Like if you want to make a TPU based shell, what ends up happening is your shell ovals over time and oscillates while you're trying to spin up. So you have to find some way to keep it rigid. So I've seen all methods of doing that, whether it's like carbon fiber stiffeners or aluminum or whatever. But that adds complexity, adds cost, adds weight. And then in the bigger weight classes, in something like gigabyte with, there's a, they have a big rolled shell that's then welded together that's very expensive, very difficult to make chunky. It's a billet shell. So incredibly expensive to do it the way that we do it. But there's a reason for it because it's very strong. It's a common thing to call it the shell tax. Like I would say most shells compared to their, you know, other comparative bots of their weight class are easily 50%, you know, 40% more expensive. And then the complexity behind it, you want to use like really big motors for shells. And the entire goal of a shell is also be really flat and low to the ground to keep your center of mass as low as possible and stable for that you either need a pancake style motor, so a large diameter flat motor, which usually it's hard to find that in a easy form factor that has a lot of torque, or you need to do some sort of alternate methods of transmission. So something like a bevel drive or a tangent drive or a lot of different ways. And that's more mechanical complexity, more difficult. So it's not the easiest archetype to at least start off with.
- [09:03]
Christine
Yeah, yeah, for sure. For somebody who's less experienced. So I want to deep dive more into each of these and the interesting thing is I think for, for both of you that meltes tend to skew towards the lower weight classes and shell spinners tend to skew towards the larger weight classes. So I think that that's, you know, difference there. So, Jacob, for Maltese, give me kind of your deep dive into, you know, the. If somebody were to decide to build a multi kind of, where do they start? You know, how much like, what does that look like in terms of potential cost, potential complexity, you know, things like that? If somebody would want to make one.
- [09:55]
Jacob
Yeah, absolutely. So for me, kind of the bible of Meltibrain building is Spencer's hardware blog. The team behind Halo. That's how my teammates and I got started in 2019. They have a great blog about how to do each step of the melty brain algorithm, how everything works. So start there. And what we did is we didn't start by just jumping into CAD and building, designing the robot and buying $1,000 worth of parts. We built successive prototypes to understand the challenge. So our very first multibrain you could just hold in its hand. It was one motor taped to the table that would spin an Arduino and an accelerometer and an led. And we just coded that. We wanted to see that as this little robot spun that the LED would know where it was facing even though the whole thing was spinning. And we worked our way up from successive prototypes. We went from that to a 3D printed 1 pound ring and then a metal 1 pounder and then we worked our way up to the 3 pound weight class.
- [11:09]
Sri
Yeah.
- [11:10]
Jacob
Start with Spencer's hardware blog. And iteration is the key to bringing a project like this to life.
- [11:19]
Christine
Yeah, it sounds like a good starting point for sure. Kind of a similar question to you, sri. I mean, obviously like something like chonky is going to be, you know, quite a bit of scale and investment to put something into building because like you said, it's already more expensive. So talk about the process for building something like that. And I mean like you said, it's, it's probably not even as possible to do it at a smaller weight class without a lot of thought put into
- [11:50]
Sri
it for something like a chonk of. I mean it took what, three and a half, four years of iteration for it to get to where it was. Doing it at a lower weight class is definitely possible. It is difficult, but it is possible. Doing it the exact same way, like all the methods that we use definitely don't scale down very well because we rely on bigger batteries, bigger motors and actually packaging. You know, the shufflers don't scale down very well. It is hard to get our style of shufflers to scale down, but it definitely is possible. I mean if you look at shell spinners, like why you yu is incredible. And that is a labor of love. Like if you, if you look at how that thing is laid out and all the work that Tim Bowens has put into it, it is definitely possible to make something similar. Maybe not to the level of a Tim Bowens, but it is possible. There's bots like, like rage bait in the 12 pound class they just recently competed. Those are first time builders in the, the University of Cincinnati Capstone stuff like they, they were able to start out with a relatively successful like an actual spinning and driving 12 pound shuffler shell. That's a lot of, you know, qualifiers and they were able to do a lot better than, you know, the first iterations of Chonky which were made by very experienced people. So the actual start of it is very similar to other bots. Like there's not cut and dry like, like a blog or anything but. Or if you have experience building bots, like if you've made a mid cutter before, the process of actually prototyping and starting isn't much different. But getting it fully dialed in, I won't say will happen first go around. A good amount of iteration I think is definitely necessary to fully optimize everything as with any robot, really.
- [13:48]
Christine
Yeah, I mean that is true for certain. Now I want to talk about like your specific like bots and things like that because I'm curious Jacob, with the, the greatest challenge as far as the way that it's actually designed. It's a little bit different at least than what I've seen from other meltes and things like that. Talk about the decisions that went into the designing it like that. If you had, I'm sure you mentioned Halo, you know, inspiration for, you know, doing a bot like that.
- [14:25]
Jacob
Yeah. So for us Halo was a huge inspiration in doing the Ring melty design, where all the mass is around the edge of the robot, as opposed to maybe the most popular melty brain these days right now is the Cashmere's Project Liftoff, which is we would call like a disk melty. And the advantage of the ring design is you get higher moment of inertia. It's like a ballerina spinning on ice, holding its arms in versus you throw them out. You can store more energy at the same speed when all that mass is around the ring. And we found that appealing for dealing more damage at lower speeds. But as we've seen, either design can be very effective with robots like Project Liftoff. So really there's. There's trade offs, and either design can work. We just like this one.
- [15:32]
Christine
Yeah, yeah, for sure. And I mean, both. Both have, you know, definitely had some success. I mean, the greatest challenge, I believe that you've made the semifinals twice. So, you know, it. It certainly works pretty well. Now SRI for. Because I know that, like, we. I've talked to, you know, to you and your team and things like that about Chonky, but I don't know that I've ever asked, like, the question of why a shell spinner and if there was any direct inspiration behind that specific design, because obviously, like, it's been very successful throughout its iterations and things like that. So I'm kind of curious about that.
- [16:17]
Sri
So I joined Chonky during Chonk3 a little over two years ago. So I wasn't there during the Genesis. Right. But I do know the story behind it very well. So unrelated to Shell, this started off at Modo 2023 or 2022, when the original team members, you know, Brian Epstein and me here in Agarage, they. They saw a knockoff weight for the first time in the test box, and they heard it shuffling for the first time on a steel floor, and it deafened the entire room, essentially. And they're like, hey, we want to do that. We want to be able to essentially draw the entire room just by moving, not even firing a weapon. You know, when you go to Norwalk, you see the test box, when you see a big beater spin up or a big drum spin up, the entire pits hear it? And it essentially draws a crowd of 100 people within a couple minutes. Right. So essentially, can you do that while just driving around, you know, how much impact can you have on the crowd? So they saw that and, like, we want to do that. We want to make a big dual cam shuffler. And how do you put the biggest weapon On a dual cam shuffler, naturally the progression was let's make a big shell. And then the name came from that let's make a chunky shell. There we go. From there they, they took inspiration. So Robojackets has a long history of shufflers, you know, dating back 2016, 2018. So they took a lot of design inspiration from those in the past. Took some design inspiration from Knockoff White. If you look at the shufflers between Knockoff White and Chunky, there's a lot of, a lot of similarities. Even though we did not look at their designs at all. I mean you arrive at some similar decisions as you go on. Design wise. I've been able to look at gigabyte CAD in the past and funnily enough, chonky's almost like frame size, everything almost exactly a 50% scale down gigabyte we have never looked at. Again never looked at the bot separately other than oh they have a good self writing portal. Okay, we'll just yoink, take that. But honestly, no direct influence there and just you, you arrive at similar ideas or similar scale just by existing. Really?
- [18:35]
Christine
Yeah, I, I think engineering works like that sometimes. You know, similar minds think and kind of reach the same conclusion. Now I want to ask because I think this is important like from both of you for your specific archetype. What do you think is like the biggest piece of advice or knowledge that you would tell somebody who's looking to build that type of robot? So I'll go ahead and start with you Jacob. Like if somebody was like I want to build a multibrain other than like hey, look at this guide. Like what's the big biggest piece of knowledge you could give somebody for that?
- [19:17]
Jacob
Where do I start? I guess the single most important thing you could do is join the Hockey Pucks of Death Discord server. It's got dozens of passionate meltibrain builders who will answer any question you have about their designs about how to solve problems with your designs. So yeah, get involved with the melty community is like the best place to start as far as like, like specific, like technical advice. I think I'll say two things. The first is that yeah, you just want to, you want to jump into the, the code and get testing. Ideally it's best to, it's great to work with a partner or a team to build one of these robots. So like when my team was starting out, I did the hardware and electronics and my teammate did the software. And that's division of labor is often what can even make a multibrain Build possible. And the second thing I'll say is people like to focus on the code and the electronics and the difficulty of that. But you can't underestimate the mechanical aspect either. These robots, they're experiencing at their radius, 400 g's of acceleration, which would turn you to jelly. Like a fire. A fighter pilot blacks out at 15, you get turned to jelly at 400 GS. So the mechanical strength is, is paramount. Not, not to mention the shock waves of energy that travel through the robots as they ricochet around the arena. So that's kind of. You got a three part answer there, but that's fine.
- [21:08]
Christine
Yeah, that's my response. That's great. That's great. Yeah, the lengthier the better for that. That's what I want is, you know, if there's someone out there watching this right now and it's like, I want to build that type of robot, I want them to have as much information as they can get for sure. Same question for you, sri. Other than like the best advice being don't assume you're going to build a shell spinner and go win two golden dumpsters, because that may happen, but, you know, not, not necessarily. So what, what advice do you have?
- [21:42]
Sri
Shells are first and foremost a packaging problem. Getting as much motor as possible while also having enough battery to sustain spinning and driving to an extent is incredibly difficult. Like if you look at any shell layout, if you see even the smallest bit of wasted space, it's probably a bad shell spinner. Like easily you have to be able to properly lay everything out. It's why bots like chonky have a very unique layout. Like people see and they're like, oh, this makes sense. It's laid out in a very specific way. We have to do everything incredibly carefully. Everything is no like square centimeter is wasted and it's very difficult to do that. And that takes iteration again. That's why, you know, old versions were kind of bad because there was wasted space part of it, you know, and then there were a lot of kinks to, to iron out. Other thing is, traction is incredibly big. Chonk of really benefits from the shuffler bonus in this way where because we are a dual cam shuffler, we have two big feet on the ground at all times. Each of those feet with what, six cleats. So in reality we have 12 spikes in the ground at any given time versus a wheeled bot. Even with cleats, if it's four wheel drive, you have four points of contact on the ground sunk in to an extent. So when we spin up and we still do it, but when we spin up, we stay planted. The body of the bot isn't rotating because usually what will happen is the body with the lower moment of inertia will try to spin first. It's why with Gigabyte, even with the amount of magnets they use and everything, when they do a hard spin up, their body rotates while they're also trying to spin up. And that one is very hard to drive because you can't spin up both your body and the shell and try to drive forward because then you'll just run into a wall. And you see us do that a lot. Anyways, the other thing, a very big thing that was helpful for us, and this was a suggestion from Derek, one of the drivers of Gigabyte was having a separate weapon operator is really valuable because the concentration to really regulate a shell, the spinning of the shell while also driving, is a lot because you absolutely cannot turn up the weapon while you're in the air. Or I think people call it coining. So when the shell is kind of not fully stable and playing on the ground, if you try to spin up while that happens, just going to fling yourself in a random spot, you're not going to spin up. And that, you know, leaves you vulnerable to get little lashed and then, you know, stuck on a wall, not so fun. So if you're fully stable, spun up, you're going to do great. I don't think anything has properly stopped us if we are actually fully spun up. There isn't a bot in the world that has properly killed chonk of. And I think that stays true for almost any shell because you got what, 50 of your whole mass dedicated to the weapon. There's not a lot of bots that can stop, you know, in the three pound class, one and a half pounds, you know, that's what, 600 grams of spinning mass, it's difficult to stop that. So properly getting planted and spun up is, is a lot.
- [25:10]
Christine
Yeah, for sure. Well, that's, that's great advice.
- [25:15]
Jacob
And if I, if I could, I'd like to echo something Sri said about shell spinners. That's true for Meltis, too. And that is to not expect to take home, you know, golden dumpsters at your first event. Absolutely. These designs take iteration and learning through testing and learning through failure. I think we went.02 at our multiple events before getting even one win. But if you, yeah, solve things iteratively, you'll see the bot last longer in matches and start taking matches and Going further. So that's true for Melty's too?
- [25:54]
Christine
Yeah, yeah, a hundred percent. Well, this was extremely insightful. And we'll have more on this episode, um, to continue with Event Horizon top talking about ring spinners, but I really hope that this helped give some knowledge to anybody who's seeking to learn more about these types of robots. And we'll, you know, we'll continue on. So thank you, Jacob and Sri, and if you're watching, keep watching this episode. All right, so just to continue the discussion for this episode, I have the team behind Event Horizon and members of Singularity, you know, both very similar ring spinners. But I think, you know, as I've said, the big focus of this episode is talking about how these robots are different from these similar looking robots and kind of dive into that. So I have Lucas, Nathaniel, Henley and Max here with me to talk about ring spinners and circular spinny robots. So welcome to the show.
- [27:11]
Nathaniel
Thanks.
- [27:13]
Christine
Yeah, absolutely, absolutely. So when I was kind of conceiving this episode, I think several people that I know pointed to Event Horizon as like, hey, that's a pretty successful ring spinner. I think really the first thing that I want to talk about is how, like, ring spinners are challenging. You know, you don't see a lot of them out there that are, you know, getting into the tops and the finals and qualifying and stuff like that, just because it is a, I think, a fun design, but it can also be complicated. Even though they look like a really simple robot. So. So I guess talk about, like, the fact that they are difficult to, like, make in a competitive way, even though they look very basic, if you're just, like looking at it and don't know anything about the robot.
- [28:07]
Nathaniel
Yeah, I guess I can start off. So, like, the main challenge with a ring spinner is holding the weapon. Compared to something like an overhead spinner or a shell spinner. There's no central axle, so you have to hold the blade on the outside rim. And because that part is much closer to the impact, it gets a lot tougher to hold. The way it's typically done is with a whole load of ball bearings, typically on the top and bottom and then on the inside. And so you can do that, but once you do that, you run into the problem of how do I drive this as well. So having all those things come together in a durable manner is where things really, really get tough. So I guess we can talk about some of the first ways I tried or my team tried to drive this system. The first thing we did was friction drive, which is basically where you take some sort of rubber wheel and you press it up against the inside ring of the blade. That does work, but you run into problems where as you spin, the rubber can wear away. And if you lose all your rubber, you no longer make proper contact the blade and it will stop spinning. Other issues with that is you need to spring load it somehow. A lot of people do like a segmented part where the whole motor can actually like get pressed out against the blade. But we ran with it and the way we did it was just by having the TPU of the chassis kind of flex the motor into the ring of the robot. And it worked for a while, but it's just such a pain to deal with and like fix and all that sort of stuff. So for a while we wanted to switch to a gear drive. But you can imagine if you had a gear like gear teeth need to mesh. And if you're trying to support this surface with gear teeth, you can't really run a bearing against it. And so that's one of the other big challenges. So Lucas started, wanted to get his hand in machining and were able to do some cooler 3D geometries. If you want to talk about that Lucas, how he supported blade.
- [30:05]
Lucas
Yeah, for sure. So the, the solution to that problem of like, okay, so now you have a gear profile, but you also need to support the axle somehow because like the, the problem with using gears a lot of the time in like large impact loading situations is that gears can the gear just break or alternatively the, you know, the wear surfaces are really small because it's just the tips of the surface of the. The tooth. So the way that we solved that. Which was very painful, unfortunately. Which was painful for you? Yeah, painful for me.
- [30:40]
Nathaniel
A few buttons.
- [30:40]
Lucas
For me it was like really easy to cat. I'm sure
- [30:45]
Jacob
the.
- [30:47]
Lucas
So we had a gear profile that. It's always hard to explain this with, with no visuals, but I do have
- [30:54]
Nathaniel
CAD if you want me to pull
- [30:55]
Lucas
that up that you can pull up. Okay, that's good. But just for just as a, as a, as a primer. It's basically a gear profile where if you put a plate that covers the gear profile on the bottom and by doing that you're able to ride on the plate instead of the gear teeth. But the gear can just be. But the gear on the motor can just be spaced up so that it meshes with the blade and it doesn't mesh with the plate contacts, other things. And that was. And that works really well. And that's what we're using now. But the main challenge with that is actually making that. Because the issue is, is that in the weight class that we're at, to get sufficiently strong blades that they don't break, which we've had problems with before. Four, you can't really like 2D geometry, water jet, a gear profile, and then put a plate on the bottom and then like bolt them together because there's just so much extra flex in the blade up and down. It's like substantially weaker. It's like nearly. I think it's like nearly 70% weaker or something like that. As a result, in bending by breaking the beam into two segments. We checked a little while ago, but it was. It wasn't good. And so you kind of have to make it all of one part. But to make it out of all one part, it, you need to mill the gear teeth, which literally, like professional machinists will look at the part that I made and be like, why? Why did you do that? Yeah, because like, they know how bad and annoying it was to do this. And like, I have literally spent, like, I want to say the better part of a year and a half perfecting how to do that. And it was very much not a simple process to figure out. Like, how do you hold the stock? How do you hold something that's so big inside a machine that's relatively small? How do you make sure that your end mills don't break when you're going into the gear teeth? Because the gear teeth are basically full slotting, but you're using an end mill. It's basically the same size as the gear teeth at the very end. And that whole process is super non trivial. And it's something that, like, even if someone were to ask me, like, how do you do it? I couldn't. There's no secret. Like, I could. It would take like an hour to explain everything that goes into it.
- [33:23]
Nathaniel
Yeah. But we've made quite a few blades now, so you've got a lot of chances to practice. Awesome.
- [33:28]
Lucas
Yeah, I. I've made like nine at this point. Each one, I think took close to 40 hours of machine time.
- [33:36]
Henley
Making the blades is like its own thing unto itself.
- [33:40]
Lucas
Yeah.
- [33:42]
Henley
Even if you've already run all the stuff in Solidworks and just have to punch it in for it to do it, you still need to watch it. And the pure hours it takes, not even counting the techniques that Lucas has had to make to do it efficiently, is just a full week of going into the machine shop and, and making sure it's running right and, and just, just making the blade like I I.
- [34:09]
Lucas
The most recent set that we made, which was the most complicated we want, we made by far because it had not only a gear profile but also an undercut ball bearing raceway on the other side. All in the same part still.
- [34:24]
Nathaniel
That actually ties into a good spot like so ring spinners as well. They have the huge advantage of 360 degree both attack and defense, which is amazing. But with that comes a big problem. We try and spin up for the first time in a match. Your whole chassis wants to counter spin. And what Lucas was referring to is putting a ball bearing inside the blade. And the reason we've done that is because we actually have started working on singularity, which is a ring spinner. There's two blades that counter rotate. And that completely solves a huge portion of the problems that ring spinners have, which is they like to coin so you can't drive them once they're doing that. And they like to be completely uncontrollable because you try and turn just a tiny bit and we can talk about this for sure. You try and turn and the blade will just take you and make you turn way faster than you expect. But having two that counter rotate gets rid of all the gyroscopic forces and counter torque reaction forces.
- [35:16]
Lucas
Yeah. Also I want to say that I suspect unfortunately with Event Horizon being one of the most, I think, combat robots of all time. But in addition, because of the, because of the sheer viewing, but because of that, I feel like I understand why people are confused between melty brains and ring spinners specifically because of Event Horizon to some extent.
- [35:40]
Nathaniel
Because we do look like a melty if we're uncontrolled.
- [35:43]
Lucas
Yeah. Like the robot is spinning while the weapon is spinning. So it's like really not obvious that the robot isn't designed to spin. Like the whole chassis isn't designed to spin.
- [35:56]
Nathaniel
And that's common for a lot of ring spinners too. You'll see when they start, they'll like the count. The center will start spinning the opposite direction. I think probably in the previous episode you went over it. But Miltzi has two wheels and it spins them in opposite directions to get the whole chassis and blade all as one piece to spin. And the entire thing acts as the weapon. Ring spinners have a dedicated weapon assembly and a dedicated drive assembly.
- [36:20]
Christine
Yeah, yeah. One other thing, because you, you mentioned about like that comparison to melty is now I don't think anybody necessarily would confuse a ring spinner. With a shell spinner, because appearance way wise, they look very different. But something that we talked about in the first part of this episode is that functionally, they're closer to being the same versus like a melty. So, you know, maybe talk a little bit about that too, because like I said, they look very different. But, like, functionally, they're doing similar things.
- [36:53]
Nathaniel
Yeah, like, so shells, they have similar to ring spinner. They have a large weapon around the whole robot electronics and drive, and that's what spins and the center stays still. They look different typically because that the entire shell is spinning, yet it stays planted. But the thing that makes them very similar is the way they will drive and act, where you still have that counter rotation of the chassis versus the blade when you try and start spinning. And you still have that issue driving where trying to go straight becomes a lot more tough With a weapon that has its center of rotation around the center of rotation of the robot. But yeah, the main difference there is how you hold the weapon. A lot of shells, actually, I think definitionally shells tend to hold the weapon by a central axle in the middle of the robot that kind of comes out like an umbrella. Although there's one notable thing I believe. I think one of the ones Adam Savage or Jamie made. Blendo. Yeah, Blendo is actually a ring spinner with a cap connected to the ring. So it behaves more like a shell spinner.
- [37:56]
Lucas
Yes.
- [37:57]
Nathaniel
They don't have a central axle, which I thought was really, really fascinating. Yeah. But there's a lot of, like, interplay between them. It's just shells. Like, they have that top that also spins with their weapon versus us, where we can be flipped over. That is a huge advantage of rings, is you can be flipped and it behaves exactly the same because our wheels poke through the top and bottom. But as we saw recently. What was that, Lucas?
- [38:20]
Lucas
You also don't have to have that like, weird, irregularly shaped bar that you need to self. Right. But you need to have on the top of the shell spinners.
- [38:28]
Nathaniel
Yeah, for sure. For sure. I was going to say we ran into a problem with that recently. Or we went up against. Shoot. Who was it? It was the wall, I believe. The wall, part two.
- [38:36]
Henley
They.
- [38:37]
Lucas
Yeah.
- [38:38]
Nathaniel
Our robot has only ever fought in the past, has only ever fought verts and horizontals. And we usually. We have perfect defense against verts because it's a horizontal blade. And they can't really ever hit the top of our robot. So we've been fine against them. Horizontals have always been a bit more Tricky because they can slide under the. The ring. And we fought the first horizontal we ever fought. What was it? It was, I believe, Spartan. First collision, they sliced straight through our chassis and damaged a weapon motor. We went down. So since then, we've been like, oh, how do we do better against horizontals? And we've kind of come up with a few different things. And what we've kind of settled on now has been a shield where the. The ring itself has a. Basically a bar that sticks up and down so as it spins, no other spinning weapon can get into the side. And so we were all excited, like, oh, this is perfect. We have defense against horizontals and verts now. And then, of course, last event, we go up against a. An overhead cutter, and they just stab straight through the middle of our robot.
- [39:37]
Lucas
So it was honestly like, if. If you had to show somebody, like, exactly. How does a ring spinner or how does a hammersaw counter a ring spinner? It was like a. Literally a picture perfect one hit, like, sliced a battery wire, like, perfectly in the center. It was hard to watch.
- [39:55]
Henley
It turned us into a YouTube short.
- [39:57]
Lucas
They did. It was really unfortunate.
- [40:00]
Nathaniel
Event Resident and Singularity both have a lot of, like, really, really big impacts. And that's one of the reasons why we are so gung ho on getting these designs, because they have so much potential. They just have so many little areas, though, where it's like that one small thing will kill the entire robot, because if your weapon goes down, you're done. Like, you don't have any other thing. Like, as a ring spinner, if your weapon dies, you're done.
- [40:24]
Christine
So, yeah.
- [40:26]
Lucas
Oh, sorry.
- [40:28]
Christine
No, I was just gonna say that is very true. I mean, obviously, like, hammer saws would be a very not ideal robot to fight. From your perspective, is there any archetype that you think ring spinners are uniquely designed to do well against?
- [40:46]
Nathaniel
So a lot of people don't think this, but I think they do great against beater bars.
- [40:52]
Lucas
Yeah.
- [40:52]
Nathaniel
So a lot of people are like, oh, yeah, beater bar will spin faster and they'll just pop the ring spinner up. And while true, a lot of collisions will go to them. The amount of force that a ring spinner can impart, having such a large moment of inertia in that weapon, it really starts. If you build your ring spinner well, which, of course is the hardest part, it can really start to wear down the weapon. Assembly of. Of beaters. We've had a few instances now where Singularity, especially because it's got that. That Pinch point. And we don't think about initially. It also has a pull where if one lands and the other lands will actually rip open the robot. So we fought against multiple verts, multiple drums and beaters. Now where we've landed one on an inner brace that holds their weapon and the other on the edge of their weapon and just pull it open. It's really, really, really effective when that happens. Yeah, horizontals are still tough, but I
- [41:46]
Henley
think it was mothership that. I think it was mothership that was the beater bar that we like dislodged their axle or something.
- [41:55]
Nathaniel
Was that at the last event?
- [41:57]
Henley
Yeah, yeah, the.
- [41:58]
Nathaniel
There was.
- [41:59]
Henley
It was.
- [42:00]
Nathaniel
There's been a few.
- [42:01]
Lucas
There's been a few. There was also that. There was the. The European British champion. The guy with the. The red drum.
- [42:10]
Henley
Yeah, yeah.
- [42:12]
Nathaniel
I'd have to look it up. I'm so bad with names.
- [42:14]
Lucas
I have no idea what. I just remember talking to him and I was like, wow, what? He's telling me a story. He's like a helicopter mechanic
- [42:22]
Christine
with.
- [42:23]
Henley
With how we do against those, you see. And also to some extent with the. What was it? The.
- [42:30]
Jacob
The wall.
- [42:32]
Henley
You see just the chunks that get taken out of the metal, or you see the actual axles that hold these things in place. Those get warped or they just get like chunks taken out of them. You can see from the walls when it's. When it's a singularity hit, it has the distinct curls in opposite direction. And that's not something you get in
- [42:53]
Christine
a lot of bots.
- [42:55]
Henley
They just don't usually have two counter rotating blades, I don't think.
- [42:58]
Nathaniel
I think no one else actually has done a counter rotating horizontal yet. Definitely not a ring spinner. But I'm also pretty sure no one's come with a horizontal, but there's a lot of advantages that come with it. I think, though, when it comes to counter rotating, I think ring spinners are actually uniquely suited to it because if you have like two bars, like, if you ever use scissors, right, and you try and cut something that's not really cuttable, sometimes those blades will pop open and you don't actually get a good shear. They just kind of slice between. And that's like a huge concern. If you have your pivot point really far away from the contact point, because you've got a lot of leverage there. You can just crank those guys open. But on a ring, you're holding the weapon right next to where it actually impacts. And so we hold the two blades together a lot better. So I think rings Are really, really well suited for doing this.
- [43:47]
Lucas
Unfortunately, with the rotating horizontals, you're prone to doing this. I mean, think about it. If you. If you hit and try to scissor, you're not. The blades won't span. You'll do this. Yes.
- [44:02]
Nathaniel
Yes. You flip sideways. Yeah, that's fun. A lot of people confuse Event horizon Singularity. Event horizon had its streak of fame where it got stuck in the wall a while ago. And I think that's where a lot of people actually know it from. Because I think that clip, it made the rounds for sure. Funnily enough, that was the first iteration of Event horizon that actually used ears. And when we switched to a gear profile, we were able to spin the blade a lot faster. We jumped to a much more powerful 2550 KV motors and a 4S battery. So we were spinning near, I think, almost 280 miles an hour tip speed, which is. Rings tend to spin slower than most other weapons do just because there's so much more mass there. But, like, we were very excited about that. And despite that clip having no drive, just it shows how effective a weapon like that can be. And that's like, again, why we've been pursuing it so hard, because we really think that, like, if you can get everything lined up, it can go really deep. But so far, we haven't quite figured that out yet.
- [45:08]
Christine
Yeah, I mean, when you. There's certain robots that, you know, ones that we've seen that to look at them, you know, maybe you don't think that they're going to be able to do as much or what have you, but it's all about iterating on that design and like, also driving well obviously is a big part of it too. I'm curious because I know, like, with one thing I've been talking about on this episode, for anybody who wants to make a robot like that, in terms of making a ring spinner, is it something that is costly to make because of the way that it has to be designed, or does it have more to do with, like, having the experience and time to put into it? Because you definitely did a lot of iterating on Event horizon for sure. And it seems like it took some time to get it to where you wanted it to be. So is it like more that or is it also a cost thing? Because again, when you look at it, they're very compact and everything. You know, I don't know. I've never looked into what it would take to make something like that.
- [46:12]
Lucas
Personally, I think that I don't want to say that like we've reached the pinnacles of what rings runners can do, right. I'm not trying to say we have not. We have not. Right? But at the same time, I think that as we realized that, we started to go deeper and we started to realize that material constraints are actually becoming a problem now. And like we're actually like physically fully shearing our weapon apart. Like, like a full like quarter inch by quarter inch cross section. I think bigger, like half inch by quarter, like shearing apart. It unfortunately makes the realization that without really, really durable materials or some geometry that I'm. I honestly don't know what you would do, but I think there is a solution somewhere. You. To make a ring spinner that can survive its collisions, like long term against robots that can hit really hard, you do need like high quality materials that we got for relatively cheap. But when I say relatively cheap, I mean we got 10 blades for like $860 of stock. Of stock.
- [47:26]
Nathaniel
Plus access to CNC machines. And someone knows how to cnc so.
- [47:29]
Lucas
Exactly. Unfortunately, to make this blade, because there are some men sometimes where I was just like, oh man, I don't want to make this anymore. I want someone else to make it out of house. Professionals can do it better than me for sure. Four grand, I think was what we got the quote for.
- [47:46]
Nathaniel
That is specifically for our design. You can make a cheap ring spinner. It is not so prohibitively expensive. Like you could get an AR500 blade. The problem with doing something like that though and just using like, you know, AR 500, AR 600 is they tend to bend a bit more. So you'll need to buy more and have more spares and that tends to get more expensive as well. I would say, of course people say,
- [48:08]
Lucas
oh, sorry, go ahead.
- [48:09]
Nathaniel
Yeah, I was gonna say people will say just make the blade bigger. But you have to fit all of your stuff inside the blade as well. And as you make your weapon, as you make your whole robot like slightly larger in diameter, the weapon weight increases a lot with a small, small increase. So we've kind of hit the sweet spot at right around 200 millimeters diameter. But trying to make that blade like any, any wider, any taller adds a lot of weight.
- [48:34]
Lucas
I would also say, like to say that I think, I feel, I believe they're called bevelstaff. I think their archetype, at least material constraint wise, has legs in the sense that they, they go for less motors, only one draw, only one weapon motor drive. And they. But they use a much thicker blade than ours, but the blade is, the diameter is much smaller. And from that perspective you could do that from a much more economical perspective because you don't need like esoteric materials that we use. You can use like just really thick AR500 and you get the benefit of just beam bending because the thing is much thicker and that is much cheaper to, to approach. And I think like, I think in principle you can make a ring spinner that has like a thicker blade, like three, eight or maybe even half. I'm not sure.
- [49:22]
Nathaniel
I think that's mooncake actually. That is very similar to that.
- [49:24]
Henley
Yeah, yeah, yeah. Because it's got the one central one that's, that's big and then it's kind of like it's really dense.
- [49:29]
Nathaniel
Yeah, they run friction though, which is some like you can get friction to work. It's just a pain.
- [49:35]
Henley
It was, it got so much material inside the chassis that we had to clean out. There's so many air canisters.
- [49:42]
Nathaniel
Yeah. So essentially if the inside of your race of your, your weapon isn't smooth, like if it's water jet for instance, or laser cut, your O rings will wear away way faster than if it was like a CNC surface. And unfortunately the cheaper options are like laser cutting and water jet and so those will eat away at your O rings much faster than any other thing will. And we should poke mooncake again to see how he, how he does that. But yeah, it's, it's tough. The other issue too with, with O rings is the like imperfect torque transfer which makes your startups a little slower and kind of, it does kind of limit your top speed of what we've noticed, which is why we're able to start going way faster when we got, when we went to switch to gears.
- [50:29]
Henley
Cost and material wise. We also like heat treat our blades, so there's that as well. And for our armor, because our blade takes so much weight, we have like a carbon fiber top plate that fits into it. So it's not like it's one solid print either. So it does. Yeah, it used to be. We also used to take like pliers and stuff and manually carve out new plates mid competition.
- [50:54]
Lucas
But a lot of things when you're
- [50:58]
Nathaniel
in the pits and you need a part, you know, anything goes. It might be sacrilege to like take this wonderful like two day long print and just take flush cutters to insert cutting proper holes into it. But you know, you're at the event. Yeah, you make do. It's. That's part of the Fun of, of combat robotics too is you know that
- [51:14]
Lucas
to, to, to further answer your cost question, unfortunately into to the point of carbon fiber. Specifically carbon fiber is really good for what we're using it for. And it's like exactly what you want to use it for because it's high stiffness and but it's not being shock loaded like in, in any significant way that would cause it to like delaminate or whatever, as long as it's not being hit on the side, which ideally it doesn't. And so in that, in that sense, carbon fiber in this particular instance is like pretty much as best as you can do for the stiffness that you need. And so as a result of that, you kind of need to buy it. And carbon fiber, unfortunately, if you're not going to water jet it and you need 3D geometry pockets, you have to get it CNC'd. And CNC. Even if you have access to CNC yourself, doing it safely is like impossible unless you have like really high quality machines, like hundred thousand dollar machines I think to do it right.
- [52:10]
Nathaniel
And water baths.
- [52:11]
Lucas
Yeah, and water baths and stuff. And so you kind of have to do it out of house with CNC Madness or other people. And unfortunately that adds that basically tax on like a $300 just there.
- [52:23]
Nathaniel
Yeah, you don't need to have pockets in your carbon fiber. And it's better if you don't because carbon fiber is a lot. It's better if it's monolithic as opposed to having multiple depths cut into it just due to the nature of a fibrous material. But it's a lot sleeker if you have nice countersinks for your, your T nuts, which is what we use for holding a robot together. And of course because we're using T nuts, we actually do want them recessed because the essentially I should probably talk about this. Event horizon started as being two pieces of TPU that were just screwed together around the blade. Well actually, sorry, good point. It started actually as two pieces of.
- [53:03]
Lucas
Oh yes, that's true.
- [53:04]
Nathaniel
It did start as tpu. Yeah, we started with a nylon blend which we had done all the research, like oh, this is going to be super, super impact resistant. And it is only if it's water like moisture treated. And the first event we went to, we didn't realize that really. And I think it was our second impact, the two halves just ripped apart. It was a very fun video.
- [53:30]
Lucas
We have a great video that we can maybe send you after to put and post. But it's like frame one, the Bots together, frame two. There's three parts. There's the blade, the chassis, and the chassis.
- [53:42]
Henley
It also made it to our second event when they were announcing robots. They're like, didn't that one explode last time?
- [53:49]
Lucas
Oh, yeah, yeah, that was.
- [53:54]
Nathaniel
First iterations usually don't go well anyways. Yeah.
- [53:57]
Lucas
After deciding how to. After realizing you need to moisture treat it, be like, oh, wow, we should have done that. Then we had this whole episode of trying to learn how to moisture treat nylon and also retain dimensional accuracy, which is impossible. If anybody's watching this, do not try it. It is not easy.
- [54:16]
Jacob
You can do it in your oven.
- [54:17]
Nathaniel
You can just fine. Our parts are, again, they're 200 millimeters in diameter. So if you try and moisture treat them and basically all it is, you leave it in a plastic bag with, like, some water, and it'll absorb it. It actually expands like a sponge. And so our chassis went from being nice and flat to being very like, like party popper shaped. So, yeah, it was. It was a bit of. At least on my end. It was. It was quite disastrous because I was like, oh, God, the events in like a day. How do I fix this problem that you just fit. The two sides don't fit together. I ended up, like, putting in the oven, trying to dry it back out.
- [54:54]
Henley
That was a fun call.
- [54:55]
Nathaniel
Switch to TPU after that, and it's been wonderful ever since.
- [54:59]
Lucas
Yeah. We're not doing research on it. And unfortunately, the correct way to moisture treat nylon requires, like, weeks of. In a. In a temperate, humid environment like a rainforest. That is the correct way to do it.
- [55:11]
Nathaniel
That would still cause it to expand.
- [55:13]
Henley
Yeah.
- [55:13]
Lucas
Like a greenhouse might work. We didn't have both time and a greenhouse, but, like, maybe, I don't know.
- [55:20]
Henley
Ring spinners are not the easiest beginner project. They are not the most budget friendly. It was cool. They're not the most budget friendly. They're not the easiest to assemble. You need special materials. It is a little bit of a logistics nightmare. But if you build one, the. The great part is that once you've built it, it's cheaper the second time. Surely you could just steal electronics from your previous iterations. They don't want you to.
- [55:51]
Nathaniel
It is true. We rarely have. We rarely have electronics break inside a robot. Like, ring spinners are really, really good at keeping their insides inside and safe.
- [56:01]
Lucas
Yeah.
- [56:01]
Nathaniel
So you save cost on that. Yeah. So once you have one that works, the upkeep is pretty much just maybe a new blade here and there. If last one didn't get Trashed. Also, I will say what we do is kind of, we have the budget through the school to do this and we tend to change our design just, just enough where we need a lot of new parts every single time. If you're a new builder and you're building rings, you don't have to do that. You can make incremental changes that don't require that much remodification. Although with a ring, with how compact everything is, if you want to move a motor a little, it tends to shift everything else a bit.
- [56:38]
Christine
But yeah, good, good advice. I guess my last question, because I talked a little bit about this in the earlier part of the episode too, but from a different perspective point of view is that ring spinners are a robot that you typically see at the lower weight classes and not so much at the higher weight classes. Kind of talk about why like a 30 pound ring spinner would be a big challenge.
- [57:10]
Nathaniel
Prohibitively expensive almost is like the main thing. The engineering gets tougher as you get bigger as well, where it just. You have a lot more area of blade to support, which increases friction and makes. It makes spinning them fast enough to really do damage much more challenging. Like I think Ring Master, I believe, is one of the, like the battle bots, the large ones. And you can see that they have a really tough time getting their weapon up to speed to do proper damage to opponents. Once it's up, it can do it. But like, there's just so many caveats of making something like 30 pounds or even higher with a ring spinner that makes it. If you don't have a really, really solid foundation, I think would make it prohibitively expensive to really try because you just don't know if it's really going to be effective until you build something like that.
- [58:00]
Lucas
I will also say that I believe the root square law governs that. The fact that, that if, if Event Horizon was scaled up, I don't think the cages can handle it. I'm gonna be real.
- [58:15]
Nathaniel
There's a claim for sure,
- [58:20]
Lucas
they're not really great at handling it now. And so if we, if we actively decided to like limit the tip length so that we couldn't penetrate all the way through, it might be okay if there's somebody who just didn't know this was a problem. Perhaps, but I don't know.
- [58:40]
Nathaniel
I think in £30 you'd have a really tough time getting your weapon up to the speeds required to do that.
- [58:45]
Lucas
Maybe, maybe, I don't know, it's just, I mean, maybe we can scale how
- [58:49]
Sri
many Motors we use.
- [58:51]
Nathaniel
Yeah. We have been talking about making a 12 pound version of maybe Event Horizon or Singularity and we think 12 is probably something that might be reasonable because once you go to that class, we could probably stop using a CNC blade. We could probably do a multi layer design because it's thick. Yeah, we've been looking into it, but it is, I think we've done like what we've done eight iterations of event horizon and we're on our fourth iteration of singularity right now. So 12 practice robots before even thinking about going to, to 12 pounds. So yeah, it's. Yeah.
- [59:35]
Lucas
I would actually major a guess that the reason that there are no 12 and 30 pound ring spinners is not because they're perhaps that they're not as unfeasible as we're making them sound, but that people just don't go for them and don't try them at the three pound weight class to get enough experience to make it at the 12 and 30s. Like I, I suspect, I suspect that that might also be what's going on because like again, there are like four ring spinners at NHRL.
- [60:03]
Nathaniel
Yeah.
- [60:04]
Lucas
In the 3,000, like there's really, there really isn't any builders other than like us, Mooncake double stuff and like one other. There was the one more that we
- [60:14]
Nathaniel
did a rumble with.
- [60:15]
Lucas
There's the one more that we did the rumble with. I forget what their name is. Florida State, I think. I'm sorry, we're forgetting your name. But I think it's actually that people haven't gone there yet. It's not that it's probably, it might be prohibitively expensive, but it might not be prohibitive. The engineer. I think it's more that just people haven't gone there yet.
- [60:40]
Nathaniel
I feel like ring spinners do kind of have that. They have this like aura of like, oh, it's a ring spinner, it's cool. But it doesn't really ever do well. And when people are going to spend the money to make a 12 pound, they want it to be competitive and do well. And so when you do look into the higher weight classes, unless it's you know, like a robot that's like going in for, for comedy, they do tend to be really, really a lot more like, oh, this is going to be like we're putting a lot of effort into making this like a meta robot and ring spinners aren't the meta. And so it's just, you don't see it when it requires a bit more
- [61:12]
Henley
money to make ring Spinners are also more delicate. So if you're dealing with more forces at high, like, higher weight classes, you could also have those issues. I know before our new, like, bearing cage thing, like before our groove that has the bearings in it, we had issues where it was just a groove between blades that held the bearings or the ball bearings, ball bearings versus bearings, whatever. And they'd get wedged. So you also have to. How, how do you hold those? How do you, how do you deal with hits and stuff? Because it is, it's very tightly packed. Even at, I guess, a larger scale, you'd still. It's a lot of moving parts. It's. It's a lot of forces rubbing against each other, more so than just a usual singular blade. It's expensive, it needs special materials. So I think it's, it has trouble scaling. And like Nate said, if you're putting all this money into it, you probably want something that you either you've already tested or is proved to work. And I think that's why you get a lot more variety in the three pound, like, arena, and you get a little more wackier designs because it's not as much of a financial commitment. Still a commitment, but not as crazy.
- [62:35]
Christine
For sure. For sure. Well, I mean, that's, that's a lot of really good information, I think. And I guess, like, I, I'll throw the challenge out there to builders. You know, let's, let's see more ring spinners. Get, get, get creative. Get creative.
- [62:52]
Lucas
Yeah, we wanted, we want to see it. It's just. No one else did it.
- [62:55]
Nathaniel
Yeah, yeah, we'll do, we will do
- [62:57]
Lucas
a rumble with all the ring spinners.
- [62:59]
Jacob
Again.
- [63:00]
Lucas
There'll be so much energy in the box, it'll be unimaginable.
- [63:05]
Henley
Another Beyblade match.
- [63:07]
Lucas
It was great the first time.
- [63:11]
Christine
Yeah, I, I'd love to see it. Well, yeah, that's, that's a lot of good stuff right there. So if you're watching this episode. Hey there. There you go. There's some, some archetypes that we've talked about that we don't see as much of and it would be cool to see more of. So make sure that you, like, subscribe, Share Comment, Comment. If you love ring spinners, like, please, please do let me know and we'll see everybody next time on the show.