The Design of the Stance Stance Revolution 2!

chorlLeave a comment

Alright! Picking up from the last episode of stance, I was beginning to embark on the hyperfixation stage of designing. Basically, spitting things into a CAD model isn’t the part that really takes me forever (unless the model is extremely detailed or need simulation, I suppose) but the initial ideation and concept stone-turning.

A lot of this is just done by staring at the ceiling blankly for a while, or mulling over while I drive some place – much of the reason I seemingly enjoy driving up and down the entire Eastern Seaboard. When I taught 2.00Gokart, I called this “Grounding the design” in my lectures. It’s always easier once things condense because you start having actual dimensions and requirements and such!

For Stance, I started with a rudimentary driveshaft design for the worm gears to get some dimensions to work with. Next, I started hollowing the “block” out in a way that I can adjust the placement as the needs evolve. This took a few back and forths of moving the worm gear socket around (and changing its depth) as I dug up bearings I could use.

One of the first headaches from the brainstorming and cartoon model poking was the wheels. As mentioned previously, if I actually wanted to make 45 degree conical wheels, they had to get large. This to me seemed like a questionable benefit at best and a weight penalty pretty much by default.

To press on with the design, I decided to make a set of O-ring wheels like I have done for many bots before. These wheels will be used with soft (50A durometer) O-rings so they have some semblance of traction and shock absorption. If I found out I had weight or space later, maybe I’ll revise this to be a set of cut foam wheels.

Unusually, the O-ring groove was biased towards one side because they’re going to be touching the ground on the 45 degree. The “outside” was heavily flanged and the “inside” not so much. I usually specify O-ring wheels with a roughly 10% ‘stretch” so they don’t just flop off – this means specifying, for instance, a 2-3/4″ center circle O-ring for a 3 inch center circle wheel (which might end up around 3-3/16″ or so on the outside.

The bore was modeled as a hex shape with corner knockouts to prevent material bunching from the 3D print process affecting how the shape fits on the shaft.

i’m adding more sockets, caves, and spider holes for everything else here. The motor will sit in the middle between the two wheels and I’ll need to buy some shaft stock to make an extra-extra long dual shaft. No biggie there. On the ends with the worm screws, the shafts will ride in 3mm bearings embedded into the print.

With 3mm, 6mm, 8mm, and 10mm bearings already specified and it doesn’t even look like a robot yet, this is looking like it’ll become another Roll Cake or something. I can no longer design a bot that doesn’t use 57 tiny little ball bearings.

I’m choosing to retain the wheels with some “Poodle clips” or wide-shoulder E-clips, instead of a screw or nut and bolt setup. This will let me pop the wheels off quickly if need be, or… THEY’LL POP OFF QUICKLY IF NEED BE.

At this point, I was still entertaining the thought of the bot being made as two fully independent halves, which will be retained by some method. I started down the road of a dual dovetail joint retained by a pin doing down the longitudinal centerline, almost like a big Anderson PowerPole connector.

I’ve modeled stand-in side plates for mounting the weapon hubs here. They’ll change as needed later, and were for now just defined using the chassis “block” geometry.

I had enough of the design together to now make some comparisons to original Stance Stance. It occupies the same rough footprint, but the weapons are heavier and smaller in diameter (largely as a response to the modern meta of compact vertical spinners) and it has a lot more invested in the drivetrain. Stance Stance Revolution 1 used only two Fingertech Silver Spark motors for drive. How it was even able to move, I have no real idea.

Check out the preliminary weapon mount design as well. Two blocks, sandwiching the sidewall in the middle, with a single large shoulder bolt or similar running through it. The blocks will be attached to both sidewalls for rigidity.

It was also time to de-cartoon the blades, which were at first made for the visual only. I kept the basic shape, but I designed the counterweight side more functionally and used it to tune the center of mass.

Realistically, once you get within the 0.01″ mark…. I don’t think you’re even going to notice the benefits. Manufacturing tolerances alone will make any more precision not matter, and definitely the first few hits will ruin whatever you had going for you anyway. I got lucky and within a few “iterations” of changing the counterweight radius and angle, got the center axis to with 0.002″ of the coordinate frame.

I cooked up a hub to hold two 10mm bore type 6200 bearings. Severely overkill, but I already owned several. I probably could have saved a bit of weight here and used the smaller 6000 type bearing. The drive engagement is with 4 dowel pins on the inside of the blade bore (not visible here).

The best part? The weapon retainment is also with a snap ring. A very large, beefy one, mind you, at 2 inches in diameter.

The CAD has reached the “Hard to look at without vomiting ” stage at this point. I mushed a few dimensions around to optimize for both weight, which was becoming a concern, and serviceability (such as tool access, how to drop a part without also dropping 5 other parts, etc).

It was around here that I decided to backtrack significantly and get rid of the “Must be able to split in half in the funniest way” design. The center wall took up very precious space in a small bot and I was unable to find a battery I found satisfactory to fit in the awkward remaining triangles. I simply backed the model up all the way to the beginning after the motor and gear sockets and removed the big angled cut. Some very very angry assembly mates later…

I blew up everything more recent than the cut and started by emptying out the center of the chassis. It was kind of a manual Shell operation, because the actual Shell operation was too complex to drive all the custom thicknesses and skip/ignore-this-area demands.

As it was going to be 3D printed anyways, I wasn’t extremely concerned with excessive solid areas but did want to try and expose as much “perimeter” as I could, as well as get ahead on making space to mount internal parts.

More details going into the frame now. I added a center cage to hold the battery I had in mind, a 850mAh 4S that I had one of and bought another of. I also added a serviceability change, which was ensuring the motor and worm assembly can be removed as one.

The support bearing sockets were turned into a mild snap-in-place socket shape, such that the flexibility of the plastic let me push the bearings past them but the socket shape otherwise keeps them braced against the gear driving load. The motor mounting flange was “cut open” and I’ll only use two screws to hang onto the motor. This was done so I could avoid needing to untighten the worms in order to dismantle the drivetrain – I could, in principle, make 4 or 5 of the same motor-and-dual-worm assembly and swap any in and out.

Other features added by now include fastening holes for the side plates, which will simply be attached using Plastite (thread-forming) screws. No heat-set inserts, no backing nuts, no helicoils. Just big threads.

Weight was well beyond a concern by this stage. I had to change the thicknesses and dimensions of several parts a few times.

Originally designed to be 3/16″ thick, the discs had to become 1/8″ thick AR500 steel. I changed weapon motors to 2830 sizes because the two 2836 size 450 helicopter motors were going to be too heavy by almost an ounce apiece.

The battery mentioned earlier was revised from 1000mAh to 850mAh for the same reason, and I ended up buying titanium bolts for the weapon shafts instead of steel.

im not going to spend money on this robot

The weapon drive itself was kind of a detail left for last, because there was really only 1 choice given the placement of everything. That was gear drive. Gear driven weapons are relatively rare because most people can locate the motors sensibly, but when you are this close….

Now, belt drive does have its advantages still even at this close-in distance because of shock absorption. To counter that, simply make the gears B I G G E. I designed these as metric Module 1.5 (or about 16 DP) in case I wanted to use a pre-made pinion. The big ring gear would be 3D printed from nylon and fastened to the blade hub’s back using a bolt pattern, shown above.

One of the “blocks” that held the weapon shaft bolt was turned into more of a spoon shape with an integrated motor mount flange. I decided this was going to be more rigid and position-determinate versus having the motor flange cut into the side plates and the weapon shaft bolt blocks being completely separate.

While specifying the gear stock I was going to make the pinions from, I also decided to make a quick dumb version that was to be 3D printed from nylon as well, but had a set screw hub using two backing nuts stuffed into a socket.

I wasn’t sure if this was going to make it into the final version, since I think I’d rather have some solid metal in a weapon drive hub. But it would allow me to just bang the weapon drive out quickly and test it. And it wouldn’t hurt to have around as spares anyway!

After the weapon drive gears were finalized, I massaged the mounting bracket a little more to allow for easier service access.

As a last touch, because I did this many, MANY times during the design process… I indicated which way on the bot was the front. Many patterns and mirrors were lost to doing it on the wrong end! I’ve also done a bit of weight reduction here, cutting away more areas of the material which will be covered up by the side plates anyway so I didn’t need any more plastic there. Alignment cutouts for the bottom and top plates can also be seen.

And finally, the completed design!

Those side and top & bottom cover plates are going to be made from carbon fiber made by CNCMadness. The discs will be from SendCutSend. I just about don’t have to lift a damn finger on this bot except to make the aluminum weapon hubs and the driveshafts!

From here, it was “Watch my Ender 3/5s do all the work”! On the next episode of Stance, the Buildening…

That’s It, We’re Bringing The Stance Back! Introducing Stance Stance Revolution 2

chorl4 Comments

Yeah, yeah…. I know, I know, the SSL certificate expired again. I’m glad like 5 of you still e-mail me about site bugs. I changed my hosting options in January and I guess just I had to log in to click the button again, but had been too lazy to actually do so. Not like anyone reads websites any more, they just want me to stream working on stuff or make short form videos. All I ever do now is log into the admin panel and delete spam comments anyways!

Anyways, Operation IDIocracy and the New Robot Trap House basically ate my entire 2023. With that said, by about December 2023, I had unpacked enough of my tools and machinery to the point where making something was plausible again. I figured a new bot build would be a good stress test and will force me to unpack and organize more things as I needed them.

With that being said, I was also broke, so no new 30lbers! The idea was to see what I could bang together for Motorama using just my vast trove of random parts, with maybe a McMaster order thrown in. As expected, that lasted all of maybe 15 minutes.

Welcome Back to Stance

Almost a decade ago on this very website (scary), I built the world’s first, and somehow still only, dual 45-degree angled spinner called Stance Stance Revolution. It was built as a bit of an inside joke based on Counter Revolution from BattleBots’ 2015 reboot season (and RoboGames before that) and Plan X, which caused some Internet Consternation because its (reversible) weapon usually spun downwards. So let’s make a bot where you have no idea which way it’s spinning, but it’s alright because I don’t either!

It somehow kicked ass. While it never won anything outright, it had a terrifyingly good run at some of the old MassDestruction events held at the Artisan’s Asylum (then in Somerville) makerspace. After these events were done, I kind of just put the husk away, swearing I’d rebuild it at some point. Like 27 vans and 5 BattleBots season later, IT WAS TIME.

The plotting for SSR2 actually went back a long time to around late 2020, after I’d moved to the Old Robot Trap House, rallied the troops to build the first chassis of Overhaul 3, and then decided to skip BattleBots’ 2020 season due to my team being dispersed all over the country and travel being very difficult. So I was basically thinking of other things to build (the same thought turnovers lead to random projects such as the Omnibot reboot)

Here’s the first concept of the new Stance Stance Revolution I made around then. It’s just a solid blob model with no details, but it conveyed the vibe of version 2. The biggest change is to remedy a major problem that version 1 had: It didn’t even have stanced wheels. Basically I was thinking of making conical wheels, whether by a hot wire foam cutting jig or 3D printing little TPU cones or something, so the whole drivetrain sat at 45 degrees, in-line with the weapon disc on each side. Extreme camber.

I basically didn’t even think about or look at the concept until late last year, when I was considering my options for Motorama. I picked it back up again and decided to take it more seriously. One of the challenges was of course how to even drive the wheels?! I brainstormed and chicken scratched out some ideas, including putting motors parallel with the wheel face (so they, in the global coordinate frame, were 45 degrees angled up) and also using bevel gears built into the side of each wheel mating with a pinion embedded in the frame.

From there, I turned my attention to an old friend, worm gears.

Motors parallel to the wheels presented a packaging problem because they would stick out a lot, especially motors with gearboxes, and I’d likely have to raise them fairly high to fit them in the frame. The bevel gear idea had more merit, but I’d still need to run a geared motor and design something to split the power to both wheels, like a layshaft in the middle. Ideally I’d be able to arrange a motor long-ways in the robot and have it drive both wheels using a single shaft.

So why not strip two holes with one impact driver…. and handle both the 90 degree + 45 degree power transmission turnaround and the gear reduction in one? Open worm gears aren’t super commonly seen in robot fighting, I think more due to the need to precisely align two gears. The general lack of backdriving capability also could make your drivetrain more vulnerable to sudden torque loads, which could shear off the teeth. Nonetheless, I think this approach is under-loved and therefore the solution had more appeal to me.

Above, I’ve made a few worm gear toy models to play with both in CAD and IRL. I wanted to get a feel for what my sensitive variables are when it came to making the mounts for these. The bronze gear was ordered from SDP-SI and its matching worm gear was just a generated profile by Autodesk Inventor. It sits in a little frame to be made from whatever material I had loaded in the Ender flock at the time.

The motor is a random nose hair shaver motor or something that I got in big sacks off AliExpress. One of my more recent habits is just going to AliExpress and searching for abject parts – sort price by cheapest and find the you-pull-it junkyard part equivalent. There’s a lot of sellers who sell harvested parts like from some broken Xiaomi drone. That’s how I scored a bag of drone motors for about $1.90 each, by searching “brushless motor” and sorting by price increasing. Trashcopter and its friends are basically the outcome of that search!

These motors were built like regular 1806 class R/C outrunner motors, but had an extra long shaft which extended out both sides. It seemed almost ideal to use for the drivetrain, so I decided to design around it. I think they were like $1.50 each or something! Clearly removed from some device, with wires of inconsistent lengths and color orders. They could certainly have less exciting retirements.

The assembly was helpful in deciding if I wanted to make a fully integrated “print-in-place” hub plus gear, such as out of nylon, or order properly manufactured gears and then machine them. This decision was also dependent on how the wheels will actually get mounted to the gears (or shafts).

For instance, if I could position the gear very close to the wheel, then it might be worth designing the wheel hub with a worm gear stump on it. If it HAS to go through a shaft, then I’d rather buy molded or machined gears and eliminate one source of precision loss (3D printed parts are never perfect, you can only design around the inherent unevenness).

To answer this, I had to think more about the chassis layout and see how real part dimensions will completely destroy my blob model and make me start from scratch.

I began a new sketch model where I imported the parts in question and just rotated and clicked & dragged everything into place for a visual. The chassis was defined from a rectangular block that had chunks successively slice off each end until I liked how it looked.

This immediately showed me that the biggest problem I’ll face is the diameter of the worm gear is going to forcefully drive how big the wheels are… and my goodness did they have to get BIG.

If I tried to make the wheels smaller, the worm gears ride so low as to be basically outside the bot. The worm gears I had in mind were 30 teeth; I clearly had to go to 20 teeth or even smaller in order to make this fit with wheels that weren’t 5.5″ in diameter.

And yes, I briefly did a design study on What if I gave up on life, a.k.a just used big foamy chunks like Susquehanna Boxcar but with straight wheels. This obviously gave me plenty of volume for everything and was perfectly reasonable, if I wanted to give up on life.

I decided to back down to a 20 tooth worm gear instead. This is kind of on the edge of what ratio I thought would make the bot driveable instead of twitchy or squirrely-fast, but it did allow the gear to be packaged fully inside the chassis while retaining reasonable ground clearance.

The earlier question of “Could I 3D print the worm gear and an integrated wheel hub” was answered and it turns out no I could not because of the need to clear the motor diameter. A dead shaft in this case would actually have fairly little support on its bottom side as well, so rigidity would be horrendous. I decided that it was better to keep the worm drive and motor enclosed inside the bot as well, for alignment and cleanliness reasons. This basically made decision to use a live (driven) shaft hung from two bearings.

So, with this in mind, I went ahead and placed an order for a big sack of worms and worm gears from AliExpress. I bought a metric set of trial parts to begin with, so it was easy to find substitutes on the Chinesium Market. Dimensions for these gear families are pretty standardized with only minor differences, so while the order was arriving (about 2 weeks) I went ahead and pushed the design itself, leaving some slop space in case part dimensions had to change.

To make the shafts, I was going to be lazy and use 3/8″ aluminum hex stock with one side machined into a partial rounded polygon at a 8mm diameter. This lets me use 8mm bore bearings which are super common, but also retains flat surfaces for fastening. The step from rounded to sharp will be the seat for the bearing.

With this approach in mind, it was time to start the detail design. I had enough information here using its positioning in the chassis mockup to drill deep into making the driveshaft assembly comprising the worm gear, shaft, means of attachment, and bearings. The general idea was to seat the motor and worm gears in pockets cut out of the chassis “block”, such that a plate that is mounted to the angled face keeps the gears in place and also doubles as the structure to mount the blade hub. If I remove the weapon and the side plate, then I can pull the gear assemblies out quickly.

In the next episode of Stance, designing the rest of the chassis and bringing it to a state where I’m ready to attack fabrication!