The Overhaul 2018 Design and Build Series, Part 2: Where Everything Gets Easier

Hey! There’s a robot-related TV show premiering on May 11th you might be interested in. There are robots on it, and they do stuff. They might even tell you about how the robots were made or about who made them! I might even be on it occasionally (But for sure not the first episode: The new format of the show was filmed in fairly cleanly episode-divided chunks, and I’m not quite at liberty to say which episode(s) Overhaul stars in)

Guess what? It’s finally after BattleBots. This means my life has finally returned to roughly normal (whatever that…. is), and most importantly, I can actually finish these damn build reports. Remember back in the day when this site was more hardcore, where I posted basically daily about what I made that day? It turns out “real life” is a class you can’t skip too many times a week. Build everything in college and ditch your classes, kids! I mean, uhh, be a responsible young adult and remain engaged in your education. Yeah. That’s the right message to send! Something something public facing role model…

We return to the design stage of Overhaul by picking up after the most imperative task – redoing the steel frontal parts of the bot – was finished. In fact, I left this post half-finished before I dove into making sure everything was done and had spares, etc.

Everything else honestly seemed easy by comparison, because I already determined what was going on with the other aspects of the bot beforehand, and it really only needed to be pounded through. The next two priorities after the new forks and clamp actuator were to finish designing the drive wheels so I could immediately start 3D printing cores and molds for production, and retrofitting the bot with Brushless Rages.

 

The wheel technology I wanted to use on the bot was pretty well developed by prototyping it with Überclocker last year for the Franklin and Motorama events. I essentially just scaled it up and kept the “scooter wheel” style molding features.

 

Something cool you can do with 3D printing easily is make fully interior voids that have no opening to the outside world. I didn’t want to waste material and time by printing a huge wheel which is mostly hollow anyway, and wanted more material perimeters near the highly-stressed hub area. But a fully spoked design would have been extra fragile in my mind.

Solution? I enclosed the spokes with endcaps that have 45-degree chamfered lead-ins so it can print without support. This way, I get the concentrated materal perimeter in the center and the outer regions, as well as two relatively solid endcaps. You can’t see these from the outside at all – they look like blank wheels.

The molds are constructed the exact same as Überclocker’s, too. I’m hurrying on the wheels first because I wanted to test the viability of the “twist to unlock” demolding strategy that I piloted with the smaller wheels. As you can see, I designed in giant wrench flats (or perhaps vising flats) so I can hold the mold in something. Up until this point, I was completely unsure if twist-to-unlock was even going to begin to work!

Parametric generation made designing the 3″ front wheels super easy! To really do parametric modeling well, you have to pay a lot of attention to the order that your features were made in. I’ve practiced using parametric-CAD for its actual parametric properties more in the past few years with consulting jobs, and Überclocker’s wheels were the first multi-variant parametric part of this complexity I’ve done and had gotten it to generate correctly on the first try. I haven’t even dared touch fully parametric assemblies.

The parameters were essentially related to wheel diameter/feature thicknesses, number of thru-slots, and suppression of the interior spokes of the larger one.

(Useful side note – the continuation of that article series about horizontal modeling is something that experienced CAD users all do subconsciously. I learned it the hard way through many of my models exploding, and watching friends with bad CAD habits having entire assemblies made of parts that are exploding. If you look back through how I generate Overhaul’s relatively complex wedge facets, that’s probably the best example I have visible of horizontal modeling concepts)

I imported the wheel assemblies and also added new 12-tooth drive sprockets. I’ve described many times how Overhaul was very under-geared with a design top speed of 18-19mph and could not use nearly all of its velocity space in the arena, coupled with limited traction (hopefully less an issue this time). My experiences with Clocker at Motorama with its new 10mph top speed showed that it felt a lot less squirrely and linear to drive despite not having the best traction.

Going to 12-tooth motor sprockets from the 15 tooth ones would bring that down to 14mph, which was historically a “sweet spot” speed for the 48ft BattleBox.

The liftgear remains pretty much the exact same as last time, but the gearboxes are now the BaneBots BB220 series. I got to test drive these in some of my recent consulting projects after talking with BaneBots post-Season 2. The problem with the P80s was the Double-D coupling inside starting to round off under high-torque loads. The BB220 shares a gear pitch with the P80s, so all my spare purchased gearsets are still useable, but have output stage carriers that are twice as thick and connected using a 12mm hex bore and not a 10mm DD.

I only had to design a different mounting plate to adapt these – the ratios are otherwise the same. BaneBots only sells 4:1 stages for this gearbox right now, but with the ring gear being the same gear pitch and tooth counts as the P80, you COULD fiddle a 3:1 stage anywhere but the output.

Next item on my agenda was the “Anti-Cobalting System” for the outer frame rails. I stewed pretty hard on how to implement these. The ideal solution would have been to box off the top and bottom of the rails with an intermediate tying member, or try to do it Clocker style with a thicker single spanning piece.

Problem is, there is a lot going on in that area – on one side, all the liftgear intermediate bearings are built into the frame rail, and the front drive chain also snakes around there. There’s also not much space to attach an upper brace plate on the inside frame rails without making it fully service-dependent on removing the arm towers (and hence the top half of the bot) for any kind of access to the drivetrain from the top.

I didn’t want to sacrifice that serviceability, and I was also much LESS concerned about “Cobalting” the rails save for a direct side hit because of DETHPLOW now tying both sides together with wubbie isolation. So the ACS became a single bridge plate which spanned the entire unsupported length between the center and front axles. I decided to make it from left over 4mm titanium stitched in through its entire length by 1/4-20 Grade 8 screws.

In a realistic direct hit to the frame side, that plate is still going to buckle and likely pop a few screws. Generally though, it takes transferring a relatively minusule amount of energy to the inner frame rail to prevent buckling. If I had more material and time, I would actually have made an entire width-of-bot bridging piece to act as a huge gusset for this whole area.

But I don’t! So here we are.

That’s actually….. it. There’s not much else going on in this bot which is substantially different this year. Electrically, though, it’s a different story. I decided to drastically refactor and simplify the electrical deck. Last season’s mantra was designing the E-deck and battery as two modules which are replaced wholesale in event of failure, then we figure out what’s wrong with the broken one later.

I really consider that system over-engineered now, and especially with DETHPLOW mode, I needed a lot of that weight back first. With the ESC choice being standardized, there wasn’t a need to make a whole rack of them removable at a time.

I also thought about the number of times I swapped a battery out to charge it and replaced it with a freshly charged one: 0

Every lithium battery worth using in a robot nowadays can charge at 2-5C rates. That means a full charge in 30 minutes or less, and matches at BattleBots will not occur that quickly. Overhaul is also not a bot which is so strenous on batteries that it will roll through an entire charge in one match – Overhaul 1 took up about half of is battery nominal capacity, and OH2 was even worse at like 1/3rd per match.

Therefore, I settled for keeping the Brushless Rages on a single plate accessible from the top for individual removal if needed, and batteries considered now non-removable and better armored within the bot.

So here’s what’s going on! My HobbyKing sponsorship was renewed around now, and they finally had the high C-rating Graphene packs in stock and ready to fire (heh) over to me. I was interested in these last season, but they had been very recently introduced then and the larger sizes were not yet in production.

I am not going to harp on the potential upsides and/or downsides of graphene battery marketing (bad sponsoree…. bad!), but 65C lipos are 65C lipos. Technically Overhaul would be just fine running 2 of them, but I had space for all 4.

The batteries form a single layer in the bot instead of being double-stacked near the back now.

A little hole-patterning later, and the new e-deck unit is basically done. The whole assembly is now wubbie-suspended within the bot, with the batteries (in real life) double-sided taped together into a brick and then sandwiched between the aluminum plate and a lower either-metal-or-Garolite plate, depending on available weight.

This is the assembly by itself. I found some space to squeeze in the 7th Brushless Rage to handle the clamp drive. This whole stack is around 3.5″ tall, so it leaves about 1/2 of air gap between the top plate and my ESCs. That miiiiiiiiiiiiiight be enough?

Seen in faint outline in the e-deck installation photo is a new top plate. I decided to do away with all the fancy cutouts and vents since the ESCs have a giant heat sink for a home. It will exist in two versions – a titanium 4mm one for DETHPLOW mode which trades about 3 pounds I can use, and one made of 4mm AR400/500 steel that weighs more for wedge fight mode (which is looking more and more like it’ll need ballasting)

Finally, the completing modification….. is moving the master switches to somewhere else that isn’t directly accessible on the top of the bot by wayward hammers. Hey, if someone reaches all the way back there (last season), we’re fucked anyway, right? Well guess what – someone did reach all the way there, and we were fucked.

The new location is accessible with the same tool, and with the activator still standing off to the side. The switches face 45-degrees upward directly under the arm tubes and sunk into the frame rail cubby – formerly occupied by Overhaul’s well-meaning but ineffective server fan exhaust port.

So here we go! The two master configurations for Overhaul this time:

General purpose match mode -wedge fights and vertical type weapons alike get the long arms and Limited Liability wedges, with exact positioning depending on who. The heavy top plate is in play. The configuration weight here is 230 pounds only, so I have a lot of wiggle room for silly accessories, minibots, and customized countermeasures.

The anti-KE DETHPLOW mode is specifically for horizontal bar and disc spinners. This mode is actually questionable against higher-hitting bar weapons like Icewave, but I’ve also not had to face such a thing yet, so hell if I know what happens!

And that’s it! The fabrication of everything obviously had to move quickly, so the build reports for Overhaul this year will be a little short. Stay tuned!