I just found out today that my robot, Portable Apocalypse, made it onto the cover of Servo Magazine! There is also an accompanying article describing the development of the robot similar to what I have posted in a few previous articles here on my blog.
Since the descriptions and pictures are spread out overa fewposts at this point, I thought it might be a good time to put together a quick and dirty photo post summarizing what goes into the construction of this robot.
The general construction is described in detail in my original post about Portable Apocalypse V1. I had previously described my plans for the V2 build that would compete at Motorama 2020, but the gallery above is the first time I have shown the pictures here
And now that I have learned a few lessons from Motorama 2020, there is Portable Apocalypse V3, where the main differences are switching back from an aluminum to a titanium frame and learning to better pad the electronics.
Thanks to the fast turnaround times from SendCutSend I was able to take V3 from finished design to fully built in only a few days! Now I just have to wait for this whole pandemic thing to blow over so I can compete with Portable Apocalypse V3!
Bottom Feeder managed to qualify for NHRL’s December finals back in May, I won’t have any opportunities to compete between now and then, and after the end of August my free time for building will be severely limited. So that means I had about 3 months to figure out design upgrades for Bottom Feeder, get parts ordered/fabricated, and get everything built and tested. Plus I already had the new iteration of Apex Predator all built from the beginning of the year, so I’ve given it a few tweaks and flourishes and I plan to bring it along for some grudge matches.
Bottom Feeder Upgrades
The most obvious issue that Bottom Feeder faced in May was that its two losses stemmed from the disk being knocked off by vertical spinners. I had an aluminum gear for mounting the disks to, but it was damaged in the match against KillaJoule, so the rest of the tournament I was stuck using my backup UHMW gears which were not nearly so resistant to the disk screws pulling out. This time around I have aluminum gears, upgraded to larger fasteners, and have plenty of spares, so the disk should stay solidly connected this time around.
A less noticeable improvement is the shape of the gear teeth themselves. I had previously designed my gears using Solidworks’s gears toolbox. What I later discovered was that all the Solidworks toolbox can generate is basically placeholder gears with teeth that are actually fairly misshapen. This helps to explain why the gears on Bottom Feeder were noisier and had much more backlash than I expected, and both of those issues disappear after switching to proper involute tooth profiles.
While not as obvious as a disk that goes flying across the arena, the previous iteration was pretty rough on the motors driving the disk. All of the motors lost their retaining ring to hold the rotor in axially, which probably allowed them to float a bit on big hits. With the motors being located directly over the spinning disk, this may have contributed to how all of my motors were rendered inoperable by the end of the tournament with their bearings destroyed.
The first place I found to improve on this was to use better motors. The original PropDrive motors contained a small and a medium bearing for the rotor to spin with. The new Cobra motors have two medium and one large bearings, and thus should be able to take much higher forces. They also come with shaft collars, which should work much better for retaining the rotors than the dinky clips. They also have a few other design/manufacturing features that should make them more rugged, like having the magnets surrounded with epoxy from the factory (free battle hardening!), and having a more substantial interface between the rotor cap and the magnet ring. Hopefully these will last through the tournament.
The other thing I am doing to try to protect my motors is changing up the material I am using for the motor gears. The last iteration used UHMW for the gears since I could machine it myself and it is low friction. The unfortunate part is that I believe that it was still stiff enough to transfer a lot of the impact forces back into the motors and contribute to their failure. This time around I am trying out TPU gears, with the thought that their flexibility will help to smooth out the shock loads seen by the motors.
In addition to the same disk design that I used in May, I also have a couple of other disk options I can use depending on what opponent I am facing. A longer option should be good against other horizontals to help out-reach them, and a thicker disk should be good against verticals to help the disk deflect less on weapon-to-weapon hits and prevent any resulting self damage.
The frame and drive systems held up great in May so there wasn’t much that need major improvements, but I made a few tweaks to improve how the drive axles are mounted. Previously the axles were just shoulder bolts threading directly into the side walls of the frame. This allowed the threads to be bent to where one of the wheels was rubbing enough to make driving straight nearly impossible, and with enough abuse one of the bolts eventually snapped off at the thinner neck of the threads. The new frame has the shoulder of the bolts go through the frame wall with a tight slip fit and thread into mounting points further inside the chassis. This should eliminate the bolt neck as a likely failure point and reduce the chance of tweaking the shafts into binding up the drive system.
Minibot – Remora
At the May NHRL event my father pointed out that many other teams were taking advantage of NHRL’s unique multibot weight bonus for minibots which could actually have a decent impact on the outcome of a match, and that he would be willing to drive one for me. Since he has exactly zero experience with combat robotics besides watching me at a few competitions, I decided to keep it simple and give him a 4WD wedge bot that he can use to get under my opponents and let me line up for a better hit.
I wanted something in the vein of the D2 kits, but with a drive train that (hopefully) has a better chance of surviving going into the ring with my 12lb opponents. The frame is designed to be the same build style as I have for Bottom Feeder with interlocking aluminum walls, CF base plates, and a bit of UHMW armor, but brought down to the 3lb scale. Building this first actually allowed me to use this build as a testbed for some of the frame tweaks for Bottom Feeder. I don’t have the best history with making titanium wedges myself, so just throw an off-the-shelf D2 wedge on the front and I’ve hopefully got a bot that can take a hit from a 12lb opponent (or quite possibly my own 12lb bot) and keep running.
The drive train is also like a shrunk down version of the 12lb bot, with brushless outrunner motors through planetary gearboxes (Fingertech Mega Sparks), XL belts to drive the wheels, bolts through the frame walls for the axles, and BaneBots compliant wheels. In my testing at 4S without the wedge in place it has enough drive power to wheelie itself into cartwheels from a standing start if I hit the throttle, but I can make adjustments in my transmitter to try to keep the wheels on the ground, and I hope that means it has enough power to wedge itself under a 12lb opponent.
Another fun fact, three generations of my family will be involved with getting this minibot to the finals: I Designed it, my daughter helped build it, and my father will be the one to drive it.
Apex Predator Upgrades
Despite some people having strong opinions to the contrary, Apex Predator has a few key differences from Omnivore, namely that it does not have the capability to run as an undercutter and as a result has more weight available for heavier and more substantial armor. The vertical spinner configuration of Omnivore from May is much the same as Apex Predator’s setup, so there were a few takeaways that I have implemented before I bring it along for grudge matches.
The new tooth plates for Omnivore worked pretty well and held up much better than the standard flat head screw teeth, so I carried the concept back over to Apex Predator, but this time with the aesthetics tweaked to look more like fangs to better fit the theme. The 2L belts worked well to drive the bar, but there were no off-the-shelf options for the motor pulley, and the limited selection of belt sizes meant I had to use a fairly large pulley on the bar, which I think might have contributed to Shifty being able to snipe the belt/pulley on his first hit. Consequently I am trying out an XL belt with a smaller pulley, but I am printing it from TPU to let it have some little bit of give since it can’t slip like the V-belts. Finally, I’m trying to speed up the bar a bit to help me continue to come out on top in weapon-to-weapon hits.
Bottom Feeder is basically ready for the finals, I just have to solder up some spare electronics, tune my transmitter settings, and put it on a shelf to wait for the day to come. I’m a little sad that I can’t formally compete with Apex Predator in the tournament, but hopefully bringing it along for grudge matches will give me an opportunity to test out some of these new tweaks before bringing it back another year. I also have Portable Apocalypse and Omnivore sitting on the shelf in the state they were at the end of the November/May tournaments. They are a little worse for wear, but mostly just need a fresh belt/wheel/receiver to be combat ready again, so they may get chucked in a box and dragged along for grudge matches as well just so I can make the most out of my trip to Connecticut.
In thinking back on my build history, I have built a reasonably successful series of 2WD undercutters ranging from 150g up to 12lbs before transitioning to 4WD for what I feel have been my best undercutters to date in Bottom Feeder and Omnivore (hard to argue with a place in the finals and a record setting knockout). I also have a successful 3lb 4WD vert design in Apex Predator, and I now have a bit of experience in the 12lb weight class, so maybe next I will take a crack at a 12lb vert design.
NHRL’s May event has come and gone, so it’s time to introduce and review the performance of the two bots I built for the event.
New 3lb – Omnivore
Omnivore is my new 3lb bot as the next iteration of the design I started last year with Apex Predator. Since this event was restricted to new bots only, this design was made to have interchangeable weapon systems to be able to better counter a variety of opponents. I only managed to take it to 9th place out of 60 competitors in the weight class, but I will take solace in the fact that the bots that knocked me out went very far in the tournament, with one going undefeated all the way through the winner’s bracket and being crowned as the most destructive bot in the tournament.
The central frame that both weapon systems mount into retains the same overall shape, but every part had to be re-engineered to make the modular weaponry possible, as well as implementing a few upgrades.
The overall shape of the design remained the same, but all of the UHMW parts went from being solid in Apex Predator to being highly pocketed in Omnivore, reducing their total weight by over 30% to free up weight for the modular components.
The portions of the uprights where the bar mounts to the frame were upgraded from 3/8″ thick to 1/2″ thick to better take the force of the upgraded weapon.
The “bunny ears” at the front of the frame which hold the weapon off the ground while inverted were one of the parts most likely to be damaged enough to require replacement, so they were upgraded to be a separate replaceable part to obviate the need to replace an entire upright if this small section was damaged.
The portions of the carbon fiber top and bottom plates that connect the wheel guards to the main frame were cut down to allow larger wheels to be used for the undercutter configuration.
The weapon uprights were modified to allow the two plates for the undercutter to key into the frame for mounting, and to provide an additional set of mounting points for the larger front wheels to fit in.
The drive gear motors were upgraded to a lower gear ratio to increase the top speed, and uprights were improved to have the motors mount directly to them to make them easier to replace if damaged.
The vertical configuration still uses the same Fingertech beater bar at its core, but with the majority of the components upgraded.
The tooth screws were replaced with asymmetrical AR500 steel tooth plates, increasing the inertia of the bar as well as the durability.
The timing belts that I have repeatedly lost in matches before were replaced with more reliable 2L V-belts and custom printed pulleys.
A TPU belt guard was added to attempt to shield the belt from some damage
The grade 8 bolt that the bar spins on, which has a tendency to bend under strong hits, was replaced with stronger alloy steel bolts.
The stock motor and ESC were upgraded to more powerful versions.
The battery was upgraded from a 850mAh LiPo to an 1100mAh LiHv battery to give me a higher voltage as well as enough capacity to run at full throttle for the entire 3 minutes without running out of power.
The forks were switched from AR500 to titanium for weight, and their mountings were adjusted to prevent them from tucking under and propping my wheels up like in the last event.
The printed wedges intended to help combat opponents with horizontal spinners were upgraded to also guard higher areas as well
The undercutter configuration took it’s initial inspiration from my older 3lb undercutter design, Portable Apocalypse. Like Portable Apocalypse, the weapon shaft setup mounted to two metal plates with the end of an inverted brushless motor being used to drive the weapon. Unlike Portable Apocalypse, the limited space for the weapon module (and limited selection of small 2L belt sizes) did not leave me with enough space to use a belt reduction to drive the bar spinner this time, so I made my first attempt at driving my weapon with gears instead. The additional space needed under the frame for the undercutter was more than was available in the design for Apex Predator, so I also had to go to making some larger wheels to give me the needed height. Because I didn’t want the bigger wheels to drastically change how the bot drove from the other configuration, I also increased the tooth count on the wheel pulleys to increase the belt reduction ratio and keep the top speed about the same.
Based on my two matches against two opponents armed with the now ubiquitous Fingertech beater bar, at least one effective way of countering the widely used weapon is to build a better beater bar yourself. I was consistently able to win the weapon-on-weapon engagements, and the improved bar configuration now has enough power to bounce the opponents off the ceiling on good hits.
I need to make absolutely certain that I loctite EVERYTHING that I can. I initially forgot to loctite the threads on the bolts that serve as the axles for my wheels, and it almost cost me my first match after one of the axles simply fell out of its own accord simply from being backed out by the rotating of the wheels.
The TPU belt guard served to protect against grazing hits from two beater bar opponents, but failed to stop a direct hit from an opponent with a single disk from cutting straight through the belt.
I need to better pad my receiver, as the same hit that severed my belt somehow managed to jar my receiver (all the way at the other end of the bot) hard enough that it seems to have sheared the processor chip off of the receiver, fully disabling the bot.
I need to figure out a way of protecting against hammersaws, as one opponent was able to slice directly through my top plate and break my receiver in half to disable me in one hit.
I may need to look into a mini-bot to accompany the main bot, as they seem to be very effective in harassing opponents, and the NHRL weight bonus makes them easy to add in.
The 4wd drive system, frame, electronics, and weapon systems seem to be working very well. I feel like I may need to redistribute weight somewhere to beef up the defenses in some places to make the bot harder to knock out with a lucky hit. I also want to improve on my interchangeable weapon modules to be able to swap them out more quickly and with less disassembly necessary. I also think that I will be looking into making a hammer saw module as another weapon option.
New 12lb – Bottom Feedeer
Bottom Feeder was my new 12lb undercutter bot, and the spiritual successor to my last 12 pounder, Antisocial Distancing. I managed to take it all the way to 4th place in the competition, qualifying me for the invitation only finals event in December where I could compete for the grand prize of $12,000!
Three of the problems I had seen with Antisocial Distancing were an unsatisfactory amount of driving control with the 2wd setup, inadequate systems to power the weapon, and a frame that wasn’t holding up as well as I hoped. The driving I improved by going to a larger version of the belt driven 4wd setup that had worked so well in Apex Predator last year. The weapon power I improved by switching to two smaller motors with larger ESCs to drive everything, and large toothed gears to transfer the torque. The frame I completely redesigned with durability in mind. It has laser cut aluminum side and back walls, thick UHMW rear armor, carbon fiber top and bottom plates, and a billet aluminum front bulkhead meant to take the punishment from both my opponents as well as the weapon mounted to it.
I need to figure out the best way to mount the weapon to the gears, or at least bring more spares. I had an aluminum gear to securely mount the disk to, but a bunch of the teeth got taken off in a direct hit from another horizontal spinner. My spare gears were all UHMW plastic, and while they spun things just fine , any hit from a vertical spinner was enough to rip the screws out of the plastic and send the disk flying.
I need to find a better way of mounting or protecting the weapon motors. All 3 motors I brought wound up destroyed, with the forces I placed on the motor cans being too much for them to survive for long.
I can’t rely on the motor retaining clips. This probably fed into the previous issue, but all of the motors quickly lost their retaining clips on the shafts, leaving them free to drop down into the spinning disk/bar and further expose themselves to impacts.
One motor was actually enough to drive the disk in a pinch, but having two for redundancy and spreading the load was nice.
Next time when I am taking the motors apart for epoxying, I should also take the time to drill out and tap the holes on the end of the can for larger fasteners.
I may actually have the drive setup too fast for my current driving skill level, so I might tone it down a bit for next time around until I get more experience.
The frame and drive were extremely durable and don’t have much need for improvement. In my last match I quickly lost my disk again without a spare aluminum gear, but the rest of the bot just kept on running and coming back for more punishment for the full 3 minutes. Give me a spare aluminum gear and another spare weapon motor and I could have had it ready to go for another match.
Bottom Feeder will be back, and looking much the same as it did at this event. The weapon mounting and weapon motor mounting could use a bit of improving, but other than that I am very happy with how this bot held up. It’s nice to take a leap on a bunch of design choices that are new to you and largely have things work out exactly as planned.
After the good first showing of Apex Predator in November’s Norwalk event, I started to want to make something similar for the 1lb category as well, since it has now been a few years since I had really made a new 1lb design. Since this is in many ways a smaller version of the 3lb I decided to name it Mesopredator, which is a predator lower on the food chain than an Apex Predator.
As with Apex Predator I wanted to make it a vertical spinner, but I didn’t want to start out by sinking a lot of time and money into making my own weapon from scratch to start. This time around, I chose to go with the KitBots Saifu drum rather than Fingertech’s new 1lb beater bar setup. This choice was largely due to weight, since the Saifu drum setup is the lighter option and would give me more breathing room to fit in the 4wd system that I wanted, but it also has the advantage of eliminating the driving belt as a possible point of failure.
The frame will again be my current favorite construction of CNC cut UHMW and carbon fiber parts, with the wheel guards pulling double duty to also act as supports for the far end of the drive shafts. The weight is tight enough that I only gave it a pair of tiny titanium forks and some questionable mountings to get under opponents, so we’ll have to see how well those hold up and perform.
I went for a belt driven 4wd setup kinda like in Apex Predator, including my custom treaded foam wheels and combined pulley/hub setup. Unfortunately, with the much tighter weight constraints on the 1lb class I had to drive the back wheels directly off of the silver spark motors, which may increase the likelihood of motor related issues. I also found that the bearing blocks were too heavy for this design, and the flat plate bearing mounts were too big for the tiny frame, so I wound up making my own DIY flat plate bearing mount from a CNC cut carbon fiber plate and a printed spacer. Hopefully this all winds up working well and being durable enough.
The frame is mostly complete now, I just have to cram the electronics into it and replace the printed test fit plates with permanent carbon fiber ones, then it should be ready to go. I would love to debut this bot at Bayside Robotics’ March First State Fights event, but I won’t be able to attend due to scheduling conflicts, so we’ll see when it actually gets to fight. I don’t plan to retire my 1lb undercutter, Someone Else’s Problem, just yet since since it has been doing quite well and won the last event undefeated, so whenever I can get them both into an event next could decide if I keep going with both bots or pick one to try to iterate the designs.
The design of the new version of Apex Predator is now complete, and it is now 2/3 built, with the electrical work and the carbon fiber plates being the remaining work to do before it would be ready to compete. Since the first iteration performed fairly well back at November’s Norwalk event, the design is mostly the same but with a bunch of minor improvements spread throughout.
The previous battery seemed to only have enough juice to run everything flat out for about 2-1/2 minutes, which I found out when I didn’t even have enough power left to drive back to the door at the end of a brutal 3 minute match. The new battery adds 30% to the storage capacity, so it should give me plenty to last through my matches.
While the Fingertech beater bar is a solid off-the-shelf option, it isn’t going to be the most powerful weapon in the brackets, so I wanted to see what I could do in that regard. Adding grub screws to the unused teeth holes ups the spinning mass a bit, and a new motor with a slightly higher KV rating combined with the new battery being the high voltage variety gives me an extra 2k rpms at full throttle. The napkin math says that all together this should increase the stored kinetic energy by about 1/3.
The drive motors are also getting swapped out for a lower gear ratio version, allowing for a higher top speed.
Having experienced the repairs in the pits showed me that it was not very easy to work on the weapon system without having to halfway disassemble the frame, so a few minor design tweaks now make it much simpler and less time consuming to remove the weapon components for repairs.
After seeing how my front fork-lets dragged the floor and flipped up when reversing as well as flipping under and holding my wheels off the ground, I knew this needed to be addressed in the next version. The new forks have a bit more ground clearance and can no longer leave me stranded if they hit a divot in the floor.
The center rails of the frame may be the single most time consuming part to replace after getting damaged since everything else mounts to them, and I had one of the bunny ears get chopped off of them twice last time. To try to prevent this requiring a full replacement, I made the bunny ears replaceable by themselves.
While not a direct upgrade to the design of the bot itself, I hit on the idea of 3D printing jigs to perfectly place all of the screw holes that I would need to be drilling by hand, making the frame parts much faster and easier to produce.
So now with a bigger battery, faster drive, more powerful weapon, and easier repairs I feel like I should be ready to shoot for the podium next time around!
I finally made it out to November’s Norwalk Havoc event, and it was definitely a different experience. I talked with some of the competitors there who had also competed on the BattleBots TV show who said that the production value for NHRL was about on par with the TV show, and I believe it.
Since I had both two beetles I had (re)built over the pandemic lull as well as a last minute 12lb, I made the slightly masochistic decision to bring all three bots to the competition. This kept me hopping the entire time, as I was constantly charging batteries and assessing damage on the last bot that fought, then running up to the green room as I was being called up to fight another bot. I feel like 2 bots per event is a more ideal load, so I may have to be more selective on what to bring next time.
Apex Predator was my dark horse of this event. It may be sporting a standard Fingertech beater bar as its weapon, but it managed to make it to 9th place out of a field of 90 beetleweights, and it could have gone farther if not for my negligence in the pits.
My first match was against a first time builder with Buzzer Beater, their take on the bog standard 2wd Fingertech beater design, and it ended with me scattering their bot’s guts across the inside of the box. (Sorry guys!) Apex Predator then went on to rack up three more victories against Tothik, Diamondback, and Spartan.
The match with Diamondback was the first time this bot had needed to go the full 3 minutes, and it was here that I discovered that when running the weapon at 100% the whole time my current battery only has enough juice to go for about 2 minutes and 45 seconds before things start crapping out. There was a lipo out in the box from the destroyed mini bot that both the builder and the production crew wanted me to destroy for spectacle after the fight was over, but I simply didn’t have enough left in the battery to even drive over to it!
The next fight against Quarantine was a back and forth slug-fest where I had the upper hand for the latter half of the match, but a judge’s decision didn’t go my way in the end. Because I didn’t see any major external damage I just charged the batteries and assumed I was good for the next match. My last match with Dread Hades proved my folly, when I got to the box and discovered that one side of drive was completely locked up, leaving me all but immobile and an easy target for Tyler’s typical excellent driving, knocking me out of the tournament. I had expected that this would have meant a damaged or disconnected gear box, but a post-mortem after the event revealed that I had actually managed to pop the back of the motor off and locked up the motor. At least my double-geared gearboxes proved to be quite durable, but it looks like I will have to do some additional battle hardening for next time.
Inspect everything – I have always said that winning these competitive events requires paying attention to all of the small details, and while I had done fairly well on this in my design and preparations, I was slipping by the end of the day. After being up and running around for 16 hours straight I was more than a little tired, but if I had properly inspected/tested the bot between matches I could have caught the motor issue and had the match with Dread Hades go quite differently.
Bring more spares – After less than stellar results with my last attempt at a vertical spinner I hadn’t expected much out of this bot for a first outing, so I hadn’t expected to make it so far in the tournament and I had only brought a minimum of spare parts. Another kind builder give me spare weapon belt when I ran out, but now I know to bring more!
Battery life – I turned down my weapon speed to make it through the rest of the competition without blacking out, but I would like to find the weight/space for a larger battery so I can just go all out.
Horizontal wedgelets – My solid TPU anti-horizontal wedgelets finally got to see action against Spartan and performed well, but seeing the simple design of Lynx’s new titanium wedgelets has given me some ideas for possible improvements. The one issue with going in that direction is that I have historically had significant difficulty with countersinking screw holes for titanium wedges, and subsequently keeping them on the bot.
Battle hardening – I had already put loctite on all of the screws for my drive motors to keep the gearbox together, but now it looks like I need to work on keeping the motor itself together as well.
Custom weapon – Now that I have seen how well this design has performed, I think that I will start toying around with ideas for a custom made weapon to give it even more kick. I will probably stick to the Fingertech beater bar for version 1.1 of the design and just work on refining some of the other aspects, but it will be on the back burner for a version 2.0 down the road.
While the design seems to be quite durable and destructive, Portable Apocalypse continues to suffer from a streak of bad luck that has prevented it from going deeper into the brackets. It seems that the changes I made for the kits and this current build finally managed to tame the electrical gremlins that had plagued the bot in its last two outings, but other things cropped up to go wrong instead. On the bright side, I can be happy that the Jolt! kits have performed well, with one of them at this competition making it farther than I did and finishing in the top 1/3 of the field, but I’m a little annoyed that I can’t seem to manage a bit more success for myself with this design.
For my personal bot I had elected to make a disk that sacrificed the reversible tooth for a more aggressive tooth styled to hopefully better counter vertical spinners. I got to test this out in its first match against Mantella. While it was another great back and forth match, the judges ultimately decided against me. A quick inspection in the pits revealed that one of the weapon-on-weapon hits with Mantella had managed to break off the front half of the disk’s tooth. Without a reversible tooth on the other side it was now virtually useless to me, so I signed the disk and gave it to Mantella’s driver as a trophy and switched to my standard backup for the remainder of the event.
The bot was then technically recorded to have two wins in a row, but the first one had the opponent forfeit, and the second one against Silent Spring ended in seconds when the first couple small hits of the match caused whatever electrical issues Jamo was experiencing that day to flare up again, and the match ended after his bot blacked out. Going into this match I would have loved to be able to say that I knocked Silent Spring out of the tournament, but I don’t feel like this really counts in anything but the most technical sense.
The final match against Ti is where my luck fully gave out. The fight started out well enough with Portable Apocalypse’s typical chaotic bouncing around the box between trading blows, but then the glue on one of my custom made wheels gave out, followed a few seconds later by Ti landing a direct hit to my weapon belt. With the weapon disabled and one of the wheels missing while my opponent was still doing just fine, I had little choice but to tap out.
So technically on paper Portable Apocalypse maintains its ongoing even win/loss record with a 2-2 performance at the event, I don’t feel like it truly counts since the two wins were from a forfeit and the opponent randomly blacking out in the opening seconds.
Belt guard – I had previously gotten lucky and never faced another horizontal spinner that was at the proper height to get past my disk and hit my belt. Perhaps it is time to work out a proper belt guard to prevent it happening again.
Custom wheels – I had decided to try my hand at custom EVA foam wheels and printed hubs this time around. The wheels worked out fine, but the glue to attach them to the hubs didn’t hold out. I may continue using the custom wheels, but it is probably back to the Fingertech hubs for me.
Redesign? – Portable Apocalypse was my first bot that I felt really proud of building and I loved hearing that opponents were nervous of facing its chaotic wrath, so I don’t want to just give up on it, but it is hard to ignore that I have not had the most success with it (even while others have used the design to win competitions). I think that some type of changes will be necessary to get better results. While Portable Apocalypse has always been all about the damage and aggression categories, it can be difficult to demonstrate control when you are repeatedly flinging yourself across the box, so perhaps it is time to reign things in to a more controllable level. Apex Predator aptly demonstrated the good performance of the new 4wd drive train that I cooked up, so I think I will look into the concept of converting the design to a 4wd over/under cutter as a first direction to iterate.
Antisocial Distancing gave a pretty good performance and qualified for the December final event, but it was plagued by some issues at the end.
In the week leading up to the competition, I had finally gotten the bot assembled, only to discover that my chosen weapon ESCs just didn’t have the oomph to get the weapon started. (I also suspect that my using 4S batteries with the motor rated for 5S-6S batteries is involved, but that is a topic for another time.) I found that I had an old 60A Aerostar reversible ESC that was able to get it spinning, and I found online what looked like it should be a rebranded version of the same ESC without the reversing option that could be delivered in time for me to drive to Norwalk. To minimize the wire clutter from the extra lead for reversing, I elected to put the new ESC into the bot, and off we went.
Upon getting to Norwalk and running through safety I discovered that despite my assumption that the two ESCs would function the same, the new ESC was also not able to start my disk up from a dead stop. Thankfully, I found that I could get the disk spinning a little by doing a 360 with the drive, and that was enough for the ESC to take over and do its job.
I didn’t get a chance to immediately swap out the ESCs, and that may have actually been somewhat beneficial. My first opponent forfeited, then I managed to notch KOs against Carmen and Smeezus, just using my little spin move to get the disk going whenever needed. The match against Smeezus was my longest yet for this new bot, and the result was that the built up heat from the ESC was enough to melt through the shrink wrap that was protecting it and holding its heat sink on. As a result I decided to finally take the time to put the original 60A reversible ESC back into the bot before moving into the winner’s semifinal match.
The match against Hot Leaf Juice started out well. I was finally able to spin the weapon up without resorting to robotic ballet moves, we started exchanging a few blows in the opening seconds, and I managed to tear one of their wheels off without taking any damage myself. Then my bot just stopped moving. After tapping out and getting the bot back to my pit table, I found that the radio receiver was just dead. I swapped in an identical backup receiver and tested that the wheels were working, so it seemed like I was ready to go again.
My loser’s semifinal match was against Rip and Tear, a bot encased in large quantities of ablative foam armor. I got my bot into the box, the doors closed, I drove to my corner, I tried to twitch my weapon to check it before the match started, and the bot died again, unable to even drive back to the door. Mercifully, they let me take it back to get it running again. After finding the radio receiver dead again, I swapped in a different type of receiver this time, and it was back to the box to try things again. This time once the doors closed I found that my drive still worked, but my weapon was still non-functional. With no more spare ESCs, there was nothing to do but press on. Despite my aggressive driving, a judge’s decision for Rip and Tear served to knock me out of the tournament, but left me with 3rd place and an invitation to the December final.
New weapon power system – My weapon power system wound up being what I could scrape together in the final days before the competition, I did not have an adequate chance to test or improve it, and the evidence points to my ESC eventually trying to destroy my receivers and/or itself to knock me out of the tournament. I need to find better components that won’t let me down for next time, and give myself time to test them out.
Test between matches – Again i was faced with issues in the box that I potentially could have fixed in the pits between matches if I had taken the time to test and find them. I need to make having a full test between matches be my SOP going forward.
Frame improvements – While the frame overall held up well from the battering it took, by the end I was seeing cracks forming in the bottom plate where the weapon shaft went through it. I should find some way to beef things up there.
Disk improvements – The disk held up throughout the event, but by the end the tooth was starting to look a bit bent and blunt on one side, so a few tweaks to the design are in order if I expect it to make it through a larger bracket.
Overall I would say that my first trip to Norwalk went fairly well, and I had a blast the whole time! Apex Predator pleasantly surprised me, and I definitely want to iterate on it to see where it can take me. Many of the issues that lead to my defeats were a least partially self inflicted, but I can learn from them and improve my designs to come back even stronger in the future.
While I may have qualified for the December final in the 12lb category, I unfortunately don’t think that I will be going this year. With the event just 4 weeks away and all of the other plans that I have keeping me busy this holiday season, I would feel massively rushed just to try to get a bot ready, and with the bracket for the final holding twice as many bots as I just faced I would probably need another bot worth of spare parts if not a whole spare bot if I wanted a chance at the grand prize. For the moment I will choose to save my sanity, rebuild and upgrade in my own time, and come back for another attempt some other time.
With only a few weeks between deciding to finally step up to the next weight class and build a 12lb and actually needing to have it ready to fight, I had to get moving to get everything finished in time. The CAD wasn’t even 100% done, but I had to start ordering parts to make sure that it would all be here in time. Nothing like a ticking clock to motivate you!
Some of the first components to arrive were the motors and speed controllers. Unlike the lower weight classes where I can just slap them into the frame and expect them to be fine as long as they don’t take a direct hit, 12lbs is the weight class where the G-forces alone start to be enough to break motors. To improve their odds of survival I applied epoxy to both the stator windings and rotor magnets to hopefully keep everything in place when the sparks fly.
Because I chose to keep this design as a fairly direct scale up from my 3lb design (Portable Apocalypse and the Jolt! kits), the motor pulley is going to be mounted directly onto the end of the can again. This probably isn’t the best idea for the longevity of the motors, particularly when the weapon disk is going to be spinning just a few millimeters beneath the pulley, so I decided to add what reinforcement I could. One shipment of bearings from Amazon and laser cut parts from SendCutSend later, and I have a nice beefy support for the far end of the motor.
While I originally designed around continuing to use a 3D printed V-belt pulley to drive the weapon, I decided that it was time to finally make the leap on getting actual machined parts made. The new aluminum pulley came out beautifully, and it will make me less worried about the pulley melting or de-laminating during a match.
The drive motors also deserve a little TLC to keep them in top condition. A printed shield will help keep the loose electronics from entangling or bogging down the rotor, and a bearing at the end of the can will help to keep everything happy.
The overall design is again largely a scaling-up of my 3lb design. The frame is entirely made from 1/8″ 6061 aluminum plates held together by standoffs. As I realize that this alone probably isn’t enough to stand up to the rigors of 12lb combat, the entire perimeter is reinforced with 1/2″ thick UHMW walls that the standoffs are pressed into. Hopefully this composite sandwich is tough enough to hold up. The front section where the weapon shaft mounts gets some additional reinforcement, with another 6 aluminum plates stacked up to increase the thickness, and a thick aluminum spacer to bridge the remaining distance. The result is solid(ish) aluminum from top to bottom for the shaft to slot into and clamp down on. Finally, some thick strips of UHMW are bent and bolted onto the frame to form the armor.
Now I just have to finish getting it wired up and I should be ready for its debut at Norwalk Havoc!
This past weekend I attended Bayside Robotics’ second event, and it certainly proved to be an interesting day.
1lb – Someone Else’s Problem
This was the event where Someone Else’s Problem finally came into its own. It originally started its life as a 3D printed design that was hardly able to stand up the the force of its own weapon, went through two design iterations that I never actually competed with, and finally reached its current configuration with its carbon fiber frame. I finally got to take it for a spin back in July, but a bit of bad luck getting stuck on the “pit”, followed by my receiver plugs getting disconnected to fully knock me out of the tournament. Since that last tournament I shrink wrapped the connectors to prevent a repeat of the disconnection, and I charged the batteries, but it was otherwise already perfectly ready to go.
The first two matches went alright, eventually going to judges decisions that went my way, but the unique combination of centrifugal/gyroscopic forces with my weapon still haven’t been tamed and left me more often than not bouncing around the arena with little control. After this, I decided to try turning my max weapon speed down as an attempt to get better control. (I have the weapon set up on a 3-way switch which only gives me off-1/2-full speed control.)
I didn’t immediately get to try the adjusted tactic out though, as my next opponent forfeited the match. What is notable about this is that it wasn’t a typical forfeit where he was unable to get the bot working before the start of the match, he had just come out of a victory relatively unscathed. He told me that he believed that my bot would cause more damage to his bot than he would be able to repair for the next round, so he strategically chose to work his way up through the loser’s bracket rather than risk being more permanently knocked out of the tournament by facing me. I think that this marks the first time that I have won a match based purely on intimidation.
My next match placed me against Sleep Deprived, the very well driven Viper kit that knocked me out of the previous tournament. The rematch went my way this time, with the decreased weapon throttle giving me much better control, sending me to the finals. Sleep Deprived then worked its way up through the loser’s bracket to face me again for the final match. I managed to disable one side of his drive again, but then pit maintenance oversights caught up with me.
After my second match of the day, I noticed a chip had been taken out of my motor pulley. Rather than risk it becoming an issue, I decided to replace it, but I had forgotten to bring loctite with me to the event. After the next match I should have checked that the screws were tight again, but since I didn’t it resulted in 3 of the 4 screws falling out and remaining one hanging on by a thread. The result was that the pulley came loose and allowed the belt to fall off, so I had to resort to pushing him around the box for the rest of the match until it went to a decision in my favor. 5-1/2 years after building my first antweight I had finally made it onto the podium!
While the bot is looking a bit rougher in a couple spots after the tournament, I would be happy to throw it right into another one with just fresh batteries and replacing those lost screws. There isn’t a lot that I see as easy opportunities for improvement at the moment. I’m a little disappointed that I have to turn the destructive capability down a notch to get things a bit more controllable, but I still think that is an improvement overall.
The one other failure I had in the tournament was the gearbox on one of my Silver Spark drive motors. Both now and in July I had a gearbox fail by having the first pinion in the gearbox start spinning on the shaft rather than driving the next gear. I suspect that there might be something batch related there, because I have yet to see similar issues with motors that I purchased at earlier or later dates than those two, and I thought that the more commonly seen failure modes were seeing the gear teeth get chewed up or something get bent. There aren’t a lot of more durable options at this weight class than the Spark motors with bearing blocks, but I now have a couple 16mm planetary motors coming to try out as a potential avenue for durability improvement.
1lb Plastic – Backlash
Backlash did about as well as I expected, seeing as it was a first attempt at the class, and a fairly rushed last minute attempt as well. The first match started out okay, I got a few hits in, but then the weapon broke, followed shortly by the frame. Between the poor performance of the design, the need for a replacement drive motor for Someone Else’s Problem, and the challenge of keeping up with trying to get everything charged and repaired for 3 bots, I decided to drop Backlash out of the tournament and promote it to spare parts.
I may one day return to the 1lb plastic class, but I think that it is different enough from the design constraints that I am more accustomed to that I have much to learn before I can hope to reach any similar level of success.
150g – Pocket Apocalypse
While Pocket Apocalypse came away with the first place in the competition, unfortunately it only amounts to a bit more than a participation prize. While there were several other fairyweights that signed up for the event, only one showed up, and that one only ever got one side of its drive working in the box. While I’ll take the trophy to add to my collection for showing up with a working bot, I hope that I can have a good competition to really earn it next time.
The time has come for me to step outside my comfort zone and push my boundaries again. Allow me to introduce my first hobbyweight robot (in progress) – Antisocial Distancing, now registered to debut at the November Norwalk Havoc event!
This design is basically going to be a scaling up to take my undercutter design from Portable Apocalypse from 3lbs up to the 12lb class. I’m starting out by taking the same basic overall shape and making the frame out of 1/8″ 6061 aluminum plates sandwiching UHMW walls, with pressed in inserts to make bolting it all together simple. The drive will be the bog standard 3536 brushless/BaneBots gearbox combo. I’m hoping that the compliant wheels will give me a similar shock absorbing effect to protect the gearboxes like I get from the Fingertech foam wheels in the lower weight classes, because I expect this bot will wind up every bit as acrobatic as it’s predecessors.
I’m trying out simply scaling up the weapon drive by hanging the motor from the top plate and mounting the pulley onto the end of the can again, but I’m going to try supporting the motor with a bearing around the can to help it survive the larger impacts. It has worked for me at the 3lb class, so I’m more confident in making that work than trying to drive a pulley with just a set screw. I’m also trying out another first by having the disk pulley properly machined out of aluminum rather than printing it.
There are so many question marks and unknowns in my mind with this build, since I have never built anything on this scale before (or even given the 12lb class nearly as much attention as the smaller classes). I need to learn to solder pinions, grease gearboxes, epoxy motors, bend both aluminum and UHMW parts, and do more manual fabrication than I have grown accustomed to recently (fully CNC part construction techniques have spoiled me). I have almost 8 weeks from when I ordered most of the parts until the event, some of the parts may take at least half of that time just to get into my hands, and I still need to take measurements from the off-the-shelf parts after they arrive before I can order the final laser cut parts. It will be a bit of a race to finish and get there, but I think it is doable.
I have now officially added the Spark! 150g robot kits to my shop! These kits have now gone undefeated to take first place in the first two competitions they have entered, and now you can own one too! The carbon fiber frame is tough, the titanium disk is destructive, and with only a few screws to attach and 3 wires to solder it couldn’t get much easier to build.
The first two kits are ready to ship immediately. Further kits ordered in the short term may be delayed by several days as I am waiting for additional drive motors to arrive. All orders will be shipped as soon as all required parts are on hand on a first-come-first-served basis.