This video provides a clear, practical look at the iterative design process, proving that real-world failure is the most honest teacher for any engineer. It effectively shows how competitive pressure forces the kind of rapid innovation that lab testing simply cannot replicate.
Deep Dive
Prerequisite Knowledge
- No data available.
Where to go next
- No data available.
Deep Dive
Something New? (Bell Brawls)Added:
[music] [music] [music] [music] [music] >> Damn, when they said time flies as you get older they were not joking. The two months between Robot Wars and now have thankfully been a lot less busy than the ones immediately preceding it. Allowing me to spend more time on my personal passion project.
Oh, [ __ ] What has two months of hard work got me? A brand spanking new version of a brand spanking new version of Helton.
That's right. I brought my very first robot back from the grave, gave it a long overdue upgrade, and competed with it at the BCRC Battle Royale.
I really enjoyed putting all my design skills to the test while making this robot. Enjoyed it so much in fact, I actually designed the majority of this robot in one weekend. Using practice I'd gotten with Occam's Razor and a generous amount of advice from professional bot builders, I was able to build a robot that solved almost all the problems older versions of Helton had.
For example, although older versions of Helton had a decent drive system, the damage output was very limited by its small weapon.
All of my victories up to this point have been victories of control and attrition, usually lasting long enough to make it through a judges decision and winning that way. I never landed any critical hits. Any significant damage I caused to another robot was due to fatigue from being hit many times, with most knockouts being caused by electronic failures.
But now, with a much more optimized weapon equipped, version 4 finally has the damage output to shred unlucky opponents on its first hit, while sacrificing uh almost none of the control that Hell on already had. I've also sort of learned a lot when it comes to physically building the robots. I actually only broke one thing while building this robot, that being this speed controller. I shorted the two ground wires and I can't figure out how to undo it.
And even though I wired a capacitor backwards at one point, my smoke stopper saved me. So yeah, although this version does take a much longer time to build than previous versions, I was able to do it in a way that doesn't break $200 in speed controller. Oh boy, let's get into the design changes. This will take a while.
Although most of the inspiration for this robot came from previous versions of Hell on, I also took a large amount of inspiration from a robot called Fragment X.
>> FRAGMENT X OH MY GOD!
>> OH MY GOD!
>> [cheering] >> FRAGMENT X IS LAUNCHED INTO THE ARENA WALL!
>> CURRENTLY THE TOP ranked beetleweight according to the RCE. The creator, Zach, actually mentored me during the creation of this robot. Many subtle design details were lifted directly from this robot.
Out of all the things that are instantly recognizable as different from previous versions, I'd say the most beneficial change was the new weapon system.
Older versions used quarter inch TPU solid in by the chassis as uprights, with a shoulder bolt screwed into them.
They they were not very good.
The shoulder bolt got pulled through, the chassis bolts got ripped out, the whole weapon system was bent backwards and hit my own robot. There are so many things was with how this old weapon was mounted. I originally did it this way cuz I didn't like how the weapon phase wires were exposed in many robots, but the chances of them being hit is so small compared to the absolute nightmare that was this mounting system caused me.
Please never do this.
Version 4 now use a pair of proper ultra-high molecular weight plastic uprights screwed in place from the bottom and the front. This system is a thousand times more reliable than the old version.
The ears are now incorporated into the uprights instead of being their own thing and they give the weapon 6 mm of clearance with the ground instead of only 1.5 to stop the issue of balancing on one ear upside down. I was a little worried about the ears breaking off, but UHMW does not fracture easily with robots only being able to pick small chunks off of them.
The weapon was also significantly changed.
I actually had a guy ask me a while back for the CAD file for version 3's weapon and I was like, "Thanks, but are you sure?"
This weapon is garbage. It had a diameter of only 95 mm but still weighed 190 g. The new weapon has a diameter of 125 mm and weighs only 150 g, which is a much more efficient use of energy and weight.
Finally, instead of using a shoulder bolt to mount the weapon to the uprights, I used a threaded titanium rod with screws on each side, which saves weight [music] compared to a steel shoulder bolt.
Moving to the bottom of the robot, Pelt Tong is now four-wheel drive. The drive system is based off the Akum's Razor drive system that I successfully tested at Robotronica and it translated well to three-pound robots. The front wheel is only half the size of the back wheel.
This is done to save weight and improve rear wheel grip.
TPU pulleys are keyed into the wheels linked by a 74 tooth belt, which is coincidentally the same belt that the SSP uses. Very convenient.
And also to prevent this from happening, I stopped using 95A TPU for the wheels and started using 95A TPU. I was afraid that this would make the wheels too squishy, but they worked fine.
Helton now fully relies on carbon fiber plates for structure with only a thin TPU frame to hold everything together.
There are two large plates on the top and bottom for stability on the XY plane and two smaller drive plates on the sides to maintain belt tension and Z axis stability.
The magnets have been made smaller because of weight and because the bottom plate is so recessed, they have to have little platforms on them to keep them close to the ground.
In order to access the drive system and improve repairability, the side armor has been made removable. Three screws directly attach the armor to the chassis and two more use the dead axle and the standoff for some extra support.
The armor has also been made thicker as well to provide a little more space between an opponent and my drive system.
Now for the internals. Helton version 4 now uses a six cell battery. All the electronics have been retrofitted to operate on 22 volts.
But other than the TBS cap combos and the link switch that did not arrive on time, the electronics are the same as older versions.
The motors on the other hand have changed. I've actually downsized the drive to run on the smaller 2006 motors.
This had to be done to save weight and decrease the thickness of the robot.
They still work just fine though.
Although the weapon motor is the same model as the previous version, this robot uses one with a lower KV than previous versions because I did not want the weapon to explode under the higher voltage and weapon size.
And now for the config mounting system.
It sticks out of the robot rather than being recessed into it. It uses two M5 bolts to hold in the configs instead of an M8 bolt. And there are multiple ways to mount the configs with three notches being on each side.
Helton version 4 has a lot of configs.
Currently mounted are the main forks.
They are a little bit longer than the previous forks and have a steeper angle of attack.
I now also have proper drum forks. They are shorter, thicker, and have a heavy fillet on them to help deflect either weapons.
I also have these super long titanium forks. They're not very strong, so you'll only be seeing them against hammer saws and the drill bots where the possibility of them being damaged is low.
I also made this attachment. This can go on any one of these forks and provides a larger control surface area.
Although you'll probably only ever see it with the long forks.
Moving away from the forks, we have the wedge. And yes, I still am using the PBK wedge because bending titanium sucks ass. King for also has a belly which is what I used successfully with Hell Tongue in the past and Alkane's Razor.
Not all my configs use this mounting system, however. I also made a smaller, thicker weapon with its own special uprights. This was made for use against horizontals and beaters as a smaller weapon that spins more times per minute works better against those robots.
To make up for the smaller size, I used a motor with a higher KV, but after testing, I realized that this made the weapon a tad fast. I also made this anti-hammer saw configuration with a massive TPU turtle shell, a titanium bottom plate, and a weapon with a very high rake angle. I had to remove the side armor before this to be within weight and it looks absolutely ridiculous, but it works.
And finally, there are two copies of this robot I could use.
There there was two copies of this robot I could use. The design stage was relatively painless, but once I had to bring the robot into the real world, I encountered many problems. Thankfully, due to aggressive overestimation, my robot was actually super under weight when I first built it and I was able to put some more weight back into the chassis for some extra strength. The biggest problems I had had to do with the drive and weapon clearance.
I previously heard somewhere that you generally want to make your weapon as close to the ground as possible. With this in mind, I gave this robot a weapon ground clearance of 1 mm.
This is ideal for beaters and drums, but for a slightly more flexible robot like this one, it caused the weapon to hit the ground all the time, which isn't good.
To solve this, I had to add impactors in the front which constantly dragged on the floor.
This preventing the weapon from hitting the ground for the most part, but it also had the consequence of lifting my entire drive system off the ground whenever my robot tilted forward even a little bit. Putting way too much upward travel on the forks also didn't help.
These factors impacted the driveability of this robot, causing it to turn often when I did not want it to. Having a high center of gravity and not putting tolerances on any of the parts didn't help much either.
I tried to fix these problems as best as I could and magnets helped a little bit, but I still had issues during the tournament. Nevertheless, I had a functioning robot and off I went to San Jose for its first competition.
Hey, continuity error. I hated the first time I tried to do this, so I'm doing it again.
Before I get into the videos, I'd like to talk about my sponsor, the Kinship Center. The Kinship Center serves as a specialized social center specifically curated for neurodivergent youth and young adults. Unlike social groups of the past, it fosters connection in a more natural and approachable way for people that usually struggle with it, making it a highly effective space for those who haven't thrived in traditional environments. Anyone is welcome and if you or someone you know might benefit from this program, you can learn more about their mission and their services by visiting their official website.
All right, time to show you what version 4 can do.
>> [music] >> 5 4 3 [music] 2 1 >> In this tournament, I had 16 opponents competing against me, many of them being very skilled.
I was definitely very nervous about this tournament, especially considering my last performance at the Bell Brawl.
But also because I was essentially bringing in an entirely new robot, different from a design that I know that works to an extent. I may have had some experience doing thorough research and got help from the best. But, there was no true way of me for knowing if this robot would work or explode into a million pieces. My first opponent was Swordsman of the Underworld, a wide vertical spinner that entirely relies on a baling for its ground game.
I was well aware that at this point that metal forks work well against the baling, but I was a little worried considering his weapon was an inch larger than mine.
The fight started, and we made contact, and both of our robots immediately died.
Then, our robot woke up, I got into a better position, and hit him from the back so hard it evicted the battery from the robot.
Not too much to say about this fight.
[clears throat] It was most certainly the record for my fastest knockout, and it will most likely remain that way for a long time. I'm unsure exactly why my robot shut off. It happened a few times during testing, and it was not ideal.
I eventually came to the conclusion that it was due to the receiver being right next to one of the magnets. I'm not sure if this is actually the cause, but I moved it, and it never happened again.
I went into the next match hoping it would be more of a fight, and my next opponent was Riptide, a dual disc spinner.
Now, I was actually a little bit more confident against this robot. Its weapon was smaller, and so were its forks. But, I believe it's important to never let your guard down. Always prepare for the worst against any opponent.
The fight started, and I botched the charge. My turning is subpar here, and I actually got right in front of him, but he was unable to capitalize. I destabilized him, and then launched him across the arena. With him damaged on his head, I hit him again, ripping off his wheel. The opponent then tapped out.
A very brief, but satisfying match that showcases only some of the massive upgrades that version 4 has received.
The massive range increase greatly increases the chance of heavy contact, which greatly increases the chance of crippling an opponent.
However, the next match I knew would put version 4 to the test because my next opponent was a rematch against Vanquisher.
It has been 8 months since my victory against him at SCAR, and since then he has gotten a pretty big upgrade.
Vanquisher now has much longer forks, magnets, and a modified hub motor specifically designed for its weapon.
Before the match, I actually installed this mini Bane thing because I thought he still was using his smaller forks.
But, this actually didn't change the outcome of the fight. I botched the charge again and turned too soon.
Vanquisher was able to push me into the wall and hit my fork inside armor really hard.
This caused my fork to crack, obliterated my front wheel, and bent my armor under my back wheel high centering me.
And there wasn't really much I could do at this point. I tried to get him to hit my armor to fix it, but it didn't work, and so I just tapped out.
More design problems for me to fix. Yay.
My armor was far too floppy and does not have enough support at the ends. I will fix this by increasing the number of print walls and adding a second standoff in between the wheels.
They also caused my drive plate to separate from the front of the chassis.
I'll fix this by adding more support at this area and will add washers at the screws at the front to help hold it in.
Also, my drive plate cracked, and I can't really do much about that. What really sucks is that I had only made two of everything in order to prevent waste if I wanted to change the design, but this also comes with the consequence of making one of the copies unusable.
Thankfully, I had the other copy to work with, and it was coincidentally already in the correct configuration for the opponent I was to face next.
My next opponent was Squirm Blob, a drum spinner robot.
This fight was thankfully much more interesting than the first three.
The fight started and I screwed up the charge again and ran at the robot.
Skrimblo was able to get into a good position and charge me before I could respond. Thankfully, Skrimblo missed the charge and he drove right into my weapon, causing me to roof him. I charged him, and even though I missed, trying to respond caused him to destabilize himself.
Due to the small weapons particularly high gyro, turning from you was hard. I was able to get a lucky hit on him though. The driver of Skrimblo was very skilled and was often able to face me during my charges and hit my forks, which did not work as well as I would have hoped.
And even though he wasn't very stable, the gyro of my weapon made it hard to capitalize on opportunities.
But eventually, Skrimblo blinked, turning into my weapon and causing me to roof him hard.
This disabled Skrimblo's weapon, allowing me to go in for another attack.
Skrimblo's drive started to suffer, and he had no way to defend himself against the heavy hits I could dish out.
Eventually, I landed a devastating blow that ripped off Skrimblo's forks and crippled his drive further. It was at this point that Skrimblo finally stopped moving.
>> And 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, knockout. Helltongue is your winner.
>> My fight with Skrimblo was an intense one that showed even more of Helltongue's new strengths, such as its vastly improved ability to self-right.
However, the fight with Skrimblo revealed a few issues as well. The drum forks I had made got completely chewed up.
When I was grinding them down, I never cooled them down. This caused the forks to become annealed, which made the steel very soft. This isn't too hard to fix, but I will need some new forks.
Also, at some point during the fight, Skrimblo broke my magnet ring for my weapon.
This becomes important later.
My next opponent was Stardust, a long thin robot with a gargantuanly thick disc spinner.
In order to make sure my forks would pick them up, I moved them closer to the middle, and that was kind of a nightmare, but I made it work.
This robot did worry me. It looked fast, and the driver had some experience.
I knew that any fight against this opponent would be tough.
The fight started, and I was actually able to avoid an undesirable encounter.
Stardust's massive weapon caused him to gyro over, and I was able to nick his weapon. This happened a second time when he turned to face me to charge.
Stardust failed to capitalize on a few moments when I was vulnerable, even though he managed to nick my fork.
Stardust in general is a very unstable bot and turns slowly, allowing me to outmaneuver him quite easily. I missed a few charges and made a few mistakes that could have ended badly, but eventually, I was able to stop Stardust from getting back on its feet for the most part.
Eventually, the accumulated damage was too much for Stardust to handle.
The robot stopped moving, and the driver tapped out.
>> Tapped out. Hal Tong is your winner.
>> Another great victory, but even in this fight, I encountered a problem. Hal Tong does not like to go backwards. It often veers to the left or right when I try to do so.
This became important at this point when Stardust flipped himself back over while I was charging. I stopped and tried to go backwards several times to get in a safer position, but couldn't. Stardust could have capitalized on that situation, and that would have not been ideal.
Again, I'm hoping fixing the impactors in the weapon slots will fix this. Now, onto my next fight. I knew it was going to be a good one. I was to fight the robot Kamikaze, a beater robot built by the same guy that built Zephyr.
Now, due to the aforementioned magnet ring crack, I was unable to use my beater weapon.
This sucks because I didn't realize this had happened until after I installed the weapon. But thankfully, I had no time to switch back. I installed my beater forks in the middle and loaded myself into the arena.
The fight started, and Kamikaze launched me into the wall, but the force of this caused his receiver to malfunction, leaving him stranded. This allowed me to recover, and I got into a better position and got a hit on him. But, I oversteered and missed the next charge.
He was able to hit my forks several times and sent me flying. But, thankfully, this caused no lasting damage.
Kamikaze was very good at charging, and I had to play my cards carefully, knowing when to attack and when not to.
Suddenly, Kamikaze stopped moving again.
But, I was unable to capitalize, however, as I oversteered again and couldn't take a damage.
But, eventually, my beater forks finally worked as intended, and I was able to send him to the roof and rip off his fork. And then, after a particularly bad charge, I launched him into the roof again and disabled the camera.
The cord actually jammed Kamikaze's weapon, and the referee decided to pause the match.
The match resumed, and despite Kamikaze's weapon not being tangled anymore, it never turned on again. This gave me a big advantage, and I was able to get some good hits on him. I left him with a scraping before a little bit, but I didn't think much of it.
Unfortunately, something terrible happened. My weapon started to slow down.
I had realized that I had accidentally put in a half-dead battery in my robot.
I tried to conserve energy after this, but it was a fruitless endeavor. I eventually was left stranded in the arena.
>> Shut down, restart. Shut down, restart.
There.
Come on, Jake.
Unstick.
>> This is the second time in a row that disorganization has caused me to be knocked out of a tournament. And this time I was actually winning. This loss will surely haunt me for years to come.
In this tournament, I had gotten fifth place, the third time in a row.
However, this one was better because there were 16 other opponents to compete against.
This was a great performance from a new design, and there are lots of opportunities for improvement, as well.
The next time you'll see Hell Tongue, many of them will be fixed, and I'll be even scarier.
As for those changes, apart from the ones I've already explained previously, I'm going to explain that in my next video. But, I'll flash the ones I'm thinking about on screen right now.
Thank you so much for watching, and I'll see you all next time.
>> [music]
Related Videos
Beyond Robotics | European Rover Challenge 2026
beyondrobotics
189 views•2026-06-01
Beatbot Sora70: JetPulse Technology and AI obstacle avoidance and navigation!
DroidModderX
26K views•2026-06-02
Tesla FSD 14.3.3 Hits Phoenix Streets - FIRST LOOK
anthonystesla
114 views•2026-05-29
Elon Musk Just Revealed Fremont Line for Optimus Gen 3 Mass Production
TheAINexusOfficial
180 views•2026-05-30
人機一体「零式人機 ver.2」 子ども企画【おもしろ発見!モビリティー】 #乗り物 #automobile #robot #shorts
KyodoNews
1K views•2026-05-28
China’s New Luna AI Robot Looks Shockingly Human...
NextGenHumanoids
850 views•2026-05-28
Reachy Mini: the $300 open source robot you can actually hack — Andres Marafioti, Hugging Face
aiDotEngineer
662 views•2026-05-29
柔軟指×AI画像処理食品の仕分け作業システム!#柔軟指 #ロボット #自動化 #製造業をもっと盛り上げたい
KiQ_Robotics_Corp.
113 views•2026-05-28











