The Restoration of Overhaul 1: Wait, Why Did We Do It That Way Again?

Last November, I made a trip back up to Boston in order to retrieve some of the heavy things I didn’t bring down when I moved. One of these heavy things I explicitly wanted to get was the hulk of Overhaul 1, which had traveled with me out of MIT, through the Artisan’s Asylum, into the Old New Shop, finally to the New New Shop.

There it is, in the cruft corner of the New New Shop. When the wheel modules went into sadbot back in 2015, the shuffle drive pods were put back in as a visual completion piece. The motors were removed for future other bots – I believe those drive motors might have made it into the Season 2 Road Rash. While they were never used in the Season 1 competition, they were the last piece of the purposeful “Glue 3 designs together” approach we used for Overhaul 1, and the focus of a lot of effort during the build.

Beyond missing a few motors, the bot was exactly in the state it left the Nightmare and Witch Doctor rumble of Season 1.

This was the same trip that I acquired the Benchmaster, Master of Benches on the way up. I somehow managed to fill the back of Coronavan up without even trying. Thus is my life, apparently.

And thus, the conference of heavyweight robots is convened! The still unpainted Overhaul 3 is in the background. As I’ll expound on in its build reports coming some day soon, part of the design mantra was getting back to the roots of what I liked about Overhaul 1. I wanted Overhaul 3 to drive like Sadbot – as a result, I wanted it to drive like OH1. That meant going back to large, bouncy wheels over the old Biohazard inspired 6WD setup of Overhaul 2, and if you recall, 30Haul was made two years ago to explore the same.

My plan for Overhaul 1’s resto was to straighten the frame out so I can easily mount stuff to it again (but not repairing the battle damage!), and then putting some motors back in it. The old battery bay was to become an electronics-and-battery bay since it wouldn’t need enough energy to last a 3 minute match, just to drive around. The actuators for the lift and clamp were in fine enough shape and would just be taken apart for a quick inspection and rebuild if needed.

I began taking the thing apart and assessing what needed to be done. The right side of the frame was caved inwards from Nightmare brushing against it, for instance. This really prevented the shuffle pod on that side from being fully mounted (Its sidewall was also a little caved in, but not enough to matter apparently). Dings, dents, and nibble marks abounded on the rest of the bot.

The “pontoons” in the front warped when welding, so it was already bent anyway, but during the tournament it just ended up bending more. So I also had to figure out how to pull that straight.

I decided to force the frame apart hydraulically from the inside. Doing just enough Big Chuck’s Auto Body to have watched enough repair videos of car and truck body and frame pulls, I was out to try my sense of “understanding how the metal flows” when taking damage. Nightmare pushed the steel inwards, so pull it outwards again to compensate.

Initially, I tried with Mikuvan’s OEM tire jack. While it’s fine and enough for lifting one cheek to change a tire, against the AR400 steel plate and tube weldment, it was just… no.

And so I found myself running to Harbor Freight before closing time to get one of their big 20-ton bottle jacks. With this thing and a cleverly positioned Spool Bus Lifting Tool, I was easily able to force the frame rail back straight again by targeting the upper edge (where it got chewed first). The rest followed without much fuss.

While the tubing is crimped a little on that side now, it doesn’t matter, since all I need is the clearance. The damage is character.

I flipped the frame around to also push out the other side a little. An AR500 plate sits against the bottom of the jack and the recently corrected frame rail in order to boost its rigidity, such that I didn’t just balloon both sides of the frame outwards. I was plenty satisfied with how straightened the whole thing became, really. I didn’t expect it to work out this well!

For pulling the pontoons apart again, I had a creative method in mind. To execute this, I’d first need to weld a pull tab to the end of one of the pontoons. The idea being I’d fixture the pontoon center beam element to something relatively sturdy, and use a come-along or chain binder on…

…Yeah, what was anyone expecting? Dual vantruck metal forming.

I bolted the pontoon center member through one of the former 5th-wheel hitch mounting holes on Spool Bus, suspending it slightly off the bed by using spare Overhaul wedgelets as a spacer. This would allow the beam to deflect the other way as it was pulled. I wrapped the tow chain I keep in Vantruck around the pull tab and joined it up with itself, then attached the other end around the trailer hitch.

I then used a come-along to slowly pull on the length of chain. It looks and sounds far sketchier than it was in real life, but I made sure to use double layered eye/face protection and an few “Anti-kill-yourself” blankets over the chain and cables.

I mean, not that any of that stuff would do much against a potential flying 37 pound pointy steel thing, but it made me feel better about it!

It’s not totally straight (not that it ever was), but it’s better than before for sure. At least it’ll be straight enough to get the bolts started.

With all the frame bashing work I wanted to get done completed, I next moved on to the question of how to put motors back in it. Originally, we had just hung F30-400 Ampflow motors off the sides of the shuffle pods/wheel modules and used some tie rods to secure their back sides. This worked well enough for the time we had. I wanted to execute on an idea we bounced around but did not move on because of the extra complexity.

See, the motors I wanted to use were some XYD-13 24 volt scooter motors that I originally got as a what-if for Overhaul 3. I’m perennially of the opinion that these big ol’ scooter motors are underloved in the U.S. robot fighting scene (but rather popular overseas in the U.K. and Australia, as well as mainland Europe). Uppercut, the team of MIT ducklings from yester-season, also did very well using them for drive.

The plan was to center-mount them in the bot and use flexible couplings to connect them to the shuffle pods. Those seemingly random frame holes near the center of the bot’s wheelbase that were never populated? Well, that’s what they were originally for: Motor mounting.

This is what the arrangement will look like. The motor will drive a very short floating intermediate shaft made of spider couplings, in theory giving them a lot more isolation from the high vibration of the shuffle pods. Now I just needed a way to connect the motors together with themselves.

I decided on the fast, easy, yet effective way – use a Markforged print designed to give a little bit of rotational flex to hang the motors off the two long rails. The motors will be held together with 2 of these and standoffs as a central unit.

Here’s what the design looks like. The three holes are for the motors’s mounting flange, which will bolt through to standoffs.

And this is what it will look like in the design. The former Ampflow bolt pattern will have a small (also 3D printed Onyx) bearing block embedded in it with two flanged 1/2″ bearings to support the drive sprocket.

The design now finalized, after some adjustment of spacings here and there. The motor “pod” itself will be held in place by shaft collars, so I can make everything jiggly at first to do the side-to-side alignment before locking it in.

Fabrication of all this didn’t take too much time. So the next post installment in this “Charles really doesn’t want to start down the rabbit hole of recapping all of the Overhaul 3 content thus far” will be about getting the bot driving again!

Norwalk Havoc May 2021: Scale Model Testing Your BattleBots in Nightmare Mode

It’s time for robots once again! I think what I’m planning on doing from here is, barring one or two more Summer of Ven posts (which is now well into the Fall and Winter of Ven anyhow) I’m going to begin transitioning towards Overhaul, finally.

You see, a major over-winter project of mine was trying to get Overhaul 1 up and running again, at least walking. There was a good bit of fabrication and “What in the everloving fuck did we do here…” reverse engineering your own work from 6-odd years ago. While it’s cool enough in its own right, I’m planning on using that to move towards the long-overdue Overhaul 3 design and build series. I mean, y’all already know what the bot looks like since it’s appeared on here by now, but there was a lot of translation from 30haul’s geometry to the full size bot.

But in the mean time, let’s talk 30haul. Recall the “lessons learned” of the previous March Norwalk post

I’m revising the Snout design to be stiffer on the corners, trading the internal webs for a U-shaped bottom truss shape. The metal is rigid enough on this scale that I don’t think the center webs contributed much.

Next up, after watching enough Battlebots and seeing enough matches at this point, I’m going to make 30Haul an entry in the “fork wars” currently ravaging the combat robot landscape. The joke is that ground-scraping forks grew a foot over the course of the 2020 season, as one of the only ways to ensure a vertical spinner doesn’t get under you, is to get under it first. Keeping the opponent away and under control is just as essential as being able to take (or deliver) the hit…

I’m also beginning to like the Vex wheels less and less, because while they initially offered the bouncy compliance I was looking for, once they start disintegrating, they become more liabilities. The spokes will tear through, leaving the big gaps in the perimeter that then flap in the wind. I decided to try and find some thick foam rubber I can use to make at-scale foamy wheels, not dissimilar to Overhaul’s.

Before any of that can happen, though, I had to have a 30haul that was functional. The damage from March wasn’t that bad in the grand scheme of things, but in the name of thoroughness I decided to completely take the bot apart and remanufacture it to address some of the frame damage and retighten the motor screws, among other tasks.

So here’s 30haul in a pile of itself! I planned to repair the frame as much as I could and keep it around as a backup, because…

Prior to March, I commissioned an entire new chassis to be machined by Wedge Industries. I figured as long as I actually had new 30haul parts, why not just move the bot to the new chassis and keep the old one as the “Just in case…”

To save some expense, I had Alex pound out only the net shape geometry. Since I have my own capabilities, I save some on the machine time and setup when it comes to tapping and countersinking, among other operations. So the first order of business with this new chassis was just chucking the ol’ spiral-flute 1/4-20 into my drill and having a pleasant time.

To the old frame, I repaired the damaged threads in multiple frame rails with threadlocked Helicoil inserts. The rear bulkhead shown here was actually quite bent up from the Ripto match. I used my arbor press to straighten it using some cleverly placed spacers.

Original chassis coming back together with an added weight-shaving pocket that I designed into the one that was sent out for manufacturing. There’s nothing over this area worth hitting really, so no reason to have the weight.

And the original chassis fully repaired and assembled!

The new chassis needed the end-tap holes put in. The way to do this would be to set them up at an angle on the drill press, but I was confident enough in my drill bit rodeo skills such that I just clamped everything together on the bench and hit the front (angled irritatingly – that got designed out of Overhaul 3) bulkhead holes. The rear set I piloted and threaded in one screw before finishing the rest on the drill press.

New chassis fully assembled and ready for population.

Next item on the agenda was to make some wheels. I said earlier that I wanted actual foam rubber wheels for 30haul now to get away from the Vex wheels, as while they did offer the compliance I was looking for, their failure mode was becoming big rubber flaps.

I looked into a few difference sources of “rigid” foam rubber, such as anti-fatigue floor mats and similar. Ultimately McMaster-Carr came to the rescue with a sheet of natural rubber foam – the nice thing was, in the time I had, it just offered a compression/deflection rating in PSI so it was easier to visualize how much give it would have.

To manufacture the wheels, I initially tried to find a hole saw set where I could get a 4 inch and a 1 inch saw on the same arbor. Maybe this was possible in the ol’ glory days of Comedy Central Sports Presents BattleBots™ as I recall a few builders’ reports saying they mounted two hole saws on one arbor to cut washers (e.g. for clutches, spacers, and wheels as well). But these days the consumer systems seem to have evolved and specialized, so the smaller sizes of hole saw usually have direct screw-on mandrels and larger ones have the pin locking ones.

I ended up picking up probably the most horrible Chinesium hole saw set Amazon had to offer in the hopes that it had generic mandrels, but alas, I found that the D-bore of the saws were different sizes as well. Either way, I found out that just hole sawing into the rubber wasn’t bad for alignment anyway. It’ll become round once I do enough burnouts!

I drilled these wheels using the Vex hubs as a template and put the same hole pattern into them.

For now, I made only the two (well, four…) front wheel assemblies to alleviate the worst of the wheel compression issues. What would happen is the Vex wheels would flatten between spokes and then bounce up once it reaches a spoke, and it actually made the bot drive a little bouncy as a result. These are a good deal lighter than the Vex wheels as well.

Hey, no use in spending too much time making wheels if I don’t have the rest of the bot yet. So off we go with populating the new chassis!

While doing so, I re-discovered a problem that might have prevented 30haul from being able to lift much. Recall that I had to step down from a 3-stage P60 gearbox to a 2-stage because the torque of the 3-stage type was rounding off its own output spline. This design stands in contrast with Overhaul 2 (and 3) because 30haul doesn’t have an intermediate gear stage, the lift motor directly fiddles the Big Gear.

I noticed the motor would bind up whenever I spun the gearbox one way. It turns out the output spline on this gearbox was deformed helically in a fashion that positive torque (lift) tended to shove the carrier off the spline, causing it to bind up like a clutch plate against the first stage.

Well, to repair this, I’d just have to replace the output shaft. Trouble with that? To clear the Big Gear and lifter forks, I had to position the motor farther away from the sidewall with the support bearing. Namely, about 2 inches. The stock Banebots P61 gearbox has a 1.5″ shaft.

I was at a juncture where waiting for Banebots shafts to be ordered would mean delaying the bot’s assembly into the week of the event, which I really didn’t feel like doing. So I decided to make a shaft extender, just a turned piece of 7075 rod with a 1/2″ socket on one side and 1/2″ stub on the other, long enough to bridge the gap. A 10-32 screw goes through the middle into the end-tapped P61 shaft.

Here’s what it looks like installed, along with the rest of the motors. Again, the motivating factor here was just to be able to reach that outer support bearing with a stock-length P61 shaft, of which I have plenty of spares.

Wheels loaded on and chassis off the ground now. I like these front wheels already – they don’t cause the front of the bot to sag at all, but are still very bouncy. Boy I wish someone made 3-4″ thick chunks of this stuff!

I finished assembling the bot and put it in “Sportsman Mode” since I didn’t yet know who I was fighting at the time – I’ll change things once I get there. In the time between this and leaving for the event, I put together what I think are the last two 6-fet Brushless Rages left and readied some other spare parts.

You see, by checking out the registration roster, I found out that every other 30lber was a vertical spinner of some sort. Whether disc or drum, this tournament was going to be absolutely fucking brutal. Hence why the title is….

Scale Model Testing Your BattleBots in NIGHTMARE MODE

I teleported up to Norwalk over the course of the day on Friday, May 14th, and got to my (now usual…) Norwalk generic business hotel around 10pm. The event started at 9 the next morning, and it was a full house.

I started 30Haul out in the “sportsman mode” configuration anticipating needing to change out to anti-vert forks. I got “Hyperbite XL” as a first draw. It’s an homage to Deathroll from BattleBots, and luckily was very high off the ground, so I actually decided to roll with these. I was out to test a “Just reach out and grab it” strategy.

Hear me out here – pretty much every vertical weapon worth its mettle takes a few seconds to get up to dangerous speed. I’m content with declaring the clamp disposable and just going up to it and grabbing it. If they try to turn away to prevent this, so much the better.

Well I mean it kinda worked. The only downside of declaring the clamp arm disposable, is you better have a whole lot of them, which I… didn’t. So this was the outcome of “Just grab Hyperbite XL as it’s trying to spin up” – while I was able to keep it at bay, the buckling actually jammed the clamp motor so I couldn’t grab and lift so much as just toss around.

There’s no “stream clip” like some of the other matches, but the match is at 3:12:36 in the main stream recording.

Hyperbite XL managed to machine some of the tips off the forks and snap some of the threaded rods, all from the small amount of contact we made in the match. That’s the downside of fighting a modern vertical spinner – there’s so much power density that the damage slope is very steep. You can really just mess up once.

Nonetheless we were both thrilled enough with the outcome that we decided to call it “Overhaul vs. Death Roll”. See? Battlebots should just be 30lbers.

Great, one match in and I’m already on my spare clamp. I managed to hammer the other side of the first clamp (the half that didn’t become wall art) straight, just in case I had to use it again.

Of course, the next match was against Other Disko, which is a more traditional 4WD vert architecture and multiple-time champion of northeastern events. Great. Luckily, this meant i got to bust out some experimental “Vert Blockers”, extended forks that act as a ground-level push-me stick. There was a lot of fork and counter-fork action at BattleBots this past year, so I decided why not give it a try as well.

This one was fairly brutal and took a long time to recover from. The key takeaway is that Vert-away forks work until they don’t. I mean, the opening gambit was about the best I could hope for, and then it just became pear-shaped.

I think that’s my main beef against the KE meta, really. That because of brushless and lithium batteries, there’s no more compromises that need to be really made on that front. You can have a fast and maneuverable drive, a solid chassis, and an extremely effective weapon, so there’s little point in having anything else.

At one point 30haul must have gotten booped hard enough on the snoot to buckle the leadscrew that drives the clamp. Luckily I decided to actually cut one of these as a spare, or else I might have had to resort to being a lifter only!


(Maybe that would have been better….)

I next faced Marathon, which was an “Overhaul vs. Minotaur” moment, and I was rather pleased by the outcome except that Marathon was counted out under what we think were incorrect conditions. I was still manipulating it when they pulled up the countdown. I was extremely peeved to find that for some reason, perhaps due to accumulated damage and misalignment, that I couldn’t manage a lift after the grab – something was stalling out the lift motor.

(I would find out after the event when I tore the bot down that one of the matches before, likely Other Disko, had bent the lift shaft into a shallow U shape that was binding up if it had much load on it)

By this point, I’d basically run out of fork parts and standoffs, so I was just going into the last match against Stop Hitting Yourself in “With your shield, or upon it” mode. I think it was a little more “upon it” as the dual vertical disks slowly plucked the remaining wheels I had cobbled and assembled out of whatever still was kind of whole.

I mean, in the grand scheme of things, the extremely prototype-y nature of 30haul has been its downfall all along. I was more out to get the shape correct and get a feel for how it’ll drive instead of focusing on functionality, and it still ended up using a few legacy parts from 30hauls and Uberclockers past.

The most important thing was I got a lot (a whole lot, let me emphasize) of recon on how to handle the vertical spinner EDM party that is BattleBots. A lot of the design issues I didn’t like about this 30haul had already been changed on Overhaul well before now, so I got to leave Norwalk this time with some “hmm” lessons for Overhaul, such as….

  • Just grabbing the vert is a valid approach, as long as I can line it up or force them to turn away.
  • The clamp head and forks are considered disposable and will be both treated and duplicated as such.
  • Vert-away sticks work until they don’t, and I’m not sure I’m a judicious enough driver to use them effectively.

I never actually got to test the new “snout” spinner wedge, first because there weren’t any horiontal bar or shell weapons, and second because I just never finished welding it after seeing there weren’t any! To be fair, Stop Hitting Yourself had a horizontal component, which would have been trouble if they decided to use it against me.

Hey, at least a clean picture of 30Haul is on the NHRL hall of fame wall! I think this photo was actually taken during the February one I went to.

With this NHRL, this 30haul’s story comes to a close. Like I said, it well outlived its original purpose and I was just running against opponents with a whole slew of design handicaps. There will be another one in the future, probably not this year, as Overhaul itself is now my focus of attention. And of course the new one will have conversely back-propagated design elements from Overhaul 3!