1lb - Antweight


Hyperion is my fourth antweight (454g) combat robot.


Hyperion uses 4WD, a compact weapon system to lower COM, indirect drive, a CNC Milled Hub motor system, an additively manufactured poly-amide unibody, titanium for optimal strength/weight where high hardness is needed, and brushless motor technology.

The hub motor:

    Hyperion's hub motor system is similar in idea to what is seen on high-tech electric bikes. What differs in the two hub systems is how small and minimalistic my design is made to be, with a bounding box of 2"*2.2"*1.5" including the shoulder bolt axle and impactor. The construction was made to be simple to minimize failure points, as well as allowing for the largest, heaviest, and most robust bearings possible. To ensure the concentricity of the system, the bearings themselves need to be concentrically sound, even when and after suffering from radial impacts; misalignment is the enemy. The natural shape of the hub motor has been optimized through the use of FEA simulation, and the seven percent increase in production cost (when the quantity is two) is more than counterweighted due to the structural benefits. Three spokes constrain the outer race of the hub motor to the inner race (bearing holders) for the optimal design, which allows for the girth of the spokes. 


    Looking beyond the intricacy of the hub itself, the weapon system, which includes the wedge, works to not only store as much energy as possible but transfer the energy effectively. The weapon stores a high amount of energy for a combat robot of its size (1/2 * (3.217E+04 g mm^2+1.5) * (3035/s)^2=222.3j) with the average KE/lb being 150j, but that isn't what is truly important. What is necessary for high-efficiency energy transfer is the closing speed, the orientation of the prey, as well as the stored energy, and all three main variables are necessary to predict the transfer of energy. One could spin their creations weaponry at 400 MPH (or so is it in theory), but the energy would have an awfully difficult time transferring, only resulting in the creation's own violent gyro forces and cringeworthily slow spin-up—that is on top of the drag created from wind resistance that results in the bell curve of motor current needed to increase the speed to the next tick.


    Though sufficient weapon kinetic energy is 150j, a number higher is preferred. This is due to the energy transfer efficiency never being 100%, so spinning the weapon at a lower speed is bound to increase engagement—increasing efficiency and energy transfer. A weapon-on-weapon hit will increase engagement drastically, as the closing speed increases from the average drive speed of ~8 MPH to a weapon speed of ~150 MPH. This results in the variable of closing speed becoming extraneous, but presents the issues of impactor deformation and yielding of both in the prodigious impact. Perfect efficiency is impossible but getting close to it is not.


    The Impactors: As this bot is modular (Idea expanded upon in Par. 8), I can quickly swap blade impactors in case of excessive wear or swap to a blade better suited for an opponent. This means a thicker blade can be used to better engage soft prey, as energy transfer would be severely diminished if I simply were to cut instead of throw. This also means the blade geometry can be changed: experimental weaponry can be tested without the backlash of high expense.


    In the CAD model, it can be seen that the structural frame components are temporarily bolted together, allowing for disassembly in case a single part is damaged—preventing the need for a total disassembly if severe damage is done in a concentrated area. At least this is for the titanium body as I call it. Under the protection of this body, there's a unibody made of soft and flexible Taulman alloy 645. This unibody—or more technically a base bi-body—is responsible for retaining the electronics and constraining the motors, as well as being the equivalent of a base plate that everything mounts onto. By removing the six bolts on the front, two bolts on the bottom, two bolts on the back, and disconnecting the motor from the ESC, the titanium body can be extricated from the base. This can be done in a matter of a minute, allowing for extensive modularity. Though I can already very easily swap individual parts that are joined to the titanium body, such as the articulating frontal wedgelet (can be swapped to a larger and more protective "full wedge" for deflecting large horizontal spinners) or the weapon disc (can be swapped to a higher inertia weapon at the expense of bite, or swap to a thicker and heavier weapon at the expense of the slicing action and instead of throwing, which could be advantageous in many cases, especially when competing against an ablative opponent.), the entire titanium body assembly can be swapped for different geometry and/or material. This could allow for a far larger weapon, a flipping mechanism instead of the spinner, an ax weapon, and the like. The key takeaway is modularity in the design, something that should be implemented into all robots, as one unchanging design is bound to be ineffective against certain opponents.      


     Following the same observation as fasteners, the combat robot also uses finger joints to ensure proper alignment. This has been incredibly helpful, as the space between the close but observable gap in the remaining space in the female end of the joint signifies where misalignment resides. Next time the female side will have smaller tolerances. By measuring parts, I've conducted a test to find that the laser cutting company of my choice (Sendcutsend.com) has a taper that exerts the material out .064%/.125" of thickness. Using these numbers, I can more accurately and confidently design for tight tolerances.   

    This design uses brass heat inserts extensively in the 3D printed bodies. To minimize wear upon the threads, machine bolts are used. Using bolts as opposed to screws is critical when designing for minimum misalignment, where even .01" can be detrimental to the success of a robot. The other form of female fasteners used is aluminum 4-40 nut strips, which act as compact and lightweight angle brackets. I've used these extensively in all of my antweight combat robots, as they minimize machining time, misalignment, and cost. Indubitably, the largest advantage is the allowance of thin metal plates to fasten to other thin plates with minimal stress along the joint. As one MM titanium cannot be drilled and tapped along the thickness axis (without microscopic bolts), nut strips become the most viable option.   


    My past three antweights, Fiber Wedge, Fiber Blade, and Titan, have all been 2WD. This is because all use unarticulated wedges. Hyperion differs from the past three antweights because of its articulating leading edge. It's wedgelet being hinged is essential to the success of a 4WD combat robot, as there are naturally three points of contact (four is allowed in many combat robots because the wheel with the largest pressure will compress further than the rest, distributing the pressure equally until the material's compression is satisfied, and the compression will allow for the fourth wheel to contact the floor.), the wheels will occupy the threshold and not allow for a static wedge to function as it should; this is why a hinged wedge is necessary. 


     But why would this be desirable? After all, a 2WD bot with a wedge naturally lays more pressure on the wedge. 4WD is advantageous in many aspects, but most notably is the increased mobility and weight distribution. The weight distribution is important as it allows for variable and controlled magnetic downforce. While a 2WD bot may have issues turning, a 4WD robot will not. This is not to say that one is better than the other, as some of the best bots in the world are of both categories; it's simply a question of preference and design goals.    


    In this design, I have a coalesce of plastic and metal, two polar opposites. Many people tend to associate plastic with weakness, flexibility (when unwanted), and an inferior to metal. What many may not know is that high strength plastics such as UHMWPE are far superior to metal for the weight in certain properties. In this design, I use titanium, the highest strength to weight metal, where the thickness is constrained, and nylon and UHMWPE are used where flexibility can be used to my advantage. UHMW is an optimal material for the wheel guards because high stiffness is unimportant. While titanium would permanently deform on an impact, high strength plastic for the same weight would only temporarily deform then proceed to spring back. This is an essential concept in Hyperion's body and wheel guards.


    The orthographic side view of Hyperion shows an array of design aspects, but the angle at which the blade rake begins to visually intersect with the wedge may be one of the most critical aspects of the design. The angle at which Hyperion throws its opponents will allow for optimal energy transfer, as the 28-degree wedge is along a radius to the weapon center, allowing for not only the best angle for deflecting incoming spinners but the best orientation for throwing opponents. The orientation is important, as the goal with a good spinner is to remain planted whilst energy is optimally transferred into the opponent. If the blade pushes the opponent instead of grabbing, both robots will likely have a visual reaction to the impact, reducing control and minimizing energy transfer. If the equal and opposite energy transfer travels through the bot and into the floor, the impact on the opponent will be greatest.


    The orthographic side view also shows the use of round rubber belts. I decided to use these to transfer torque to the front wheels because of two primary reasons: The stretch allows for slight incorrect tensioning, meaning that the distance between the pulleys and the pulley sizes are not strictly constrained by the belts themselves, and also because of the ease of use; only needing to be stretched and slipped over the pulleys without any need for disassembly. And as an added bonus, they come in red.

    Another aspect of Hyperion is its wide track width (The distance of the front wheels) in contrast to the short wheelbase. I design all my robots to be wide as opposed to long to increase the frontal area of the bot, which makes controlling a target easier. The shape is optimal for high-efficiency movement; the wheelbase being short allows for less slippage when pivoting. A large wheelbase robot will require more effort to make a simple Z-axis pivot. In the picture, a red X can be seen. The tips of this X are placed at the contact points of the four wheels, and the intersection shows the point at which the robot pivots. What can also be noted is that each line segment represents the radius of a circle, and this circle is the path at which the wheels will travel upon. This shows that the wheels will undergo a slight lateral movement, resulting in friction and the wheel rotating a distance further than how far the robot is moving (nonoptimal efficiency). In summary, a short wheelbase and a large track width will result in more fluent and efficient movement and is beneficial to control robots.


    Magnetic downforce: The original plan for Hyperion's magnetic downforce solution was as simple as can be; I'd glue them to the bottom face with rubberized glue. As one of the design goals of Hyperion was adaptability, this was a no-go. The second and final solution was bolting on the magnets, which allows for the modularity of adjustment in downforce. Magnets inside the wedgelets proved to be more difficult than the body magnets, as previously designed bolt holes were not present. Implementing magnet hardware (flat head bolt and heat insert) proved to be more difficult than I had originally anticipated it to be, as the .219" large heat insert that restrained the bolt that constrained the magnet was long enough to protrude out of the wedgelets leading face. This would result in the path to the impactor being obstructed, foiling the leading-edge design. The solution was to adjust the geometry, which reduced the wedgelet pressure to a healthier number as well as preventing the obstruction. A lesson that can be learned from this conflict is to never let a feature be an afterthought. I expand on this further in my post How I Innovate.


    Electronics: Hyperion is designed with high power usage motors to reduce acceleration time and prevent the "thing", which is when the robot orients itself in a way that the wheels do not contact the ground and a self-righting mechanism is incapable of correcting the orientation. In Hyperions case, the self-righting mechanism, also called a SriMech, is its flywheel. To allow the weapon system to spin-up with a large amount of physical resistance, a high current yet efficient motor is needed. For my needs, the Emax RS2205 II or the Scorpion M-2205 are the best options, but I've opted for the prior due to cost-efficiency as well as a more detailed downloadable CAD drawing being available. The current draw of the RS 2205 II is 25A stall while I've previously used the D2822, which has a current draw of 12A stall and with lower efficiency. The larger motor needs an ESC (speed controller) to match in size, which is why I upgraded from my previous DYS 20a ESC to a DYS 30a ESC. For the first version of Hyperion, I plan on sticking with the use of brushed drive motors, as the pre-attached gearboxes simplify the drivetrain build process. If I find the drive power is lacking, I will swap to 1106 brushless motors, which after increasing battery size, adding mounting plates, and swapping drive ESCs' for compatibility, the weight savings should be close to nothing, but the performance should be bettered. This will increase the speed as well as torque, allowing the addition of more magnetic downforce. 

   There’s much to do for Hyperion, and I’ll add as I learn. As this blog post is dedicated to the design, a second blog will be created to organize material, which will be called Hyperion - The Build.