More Weapons than Wheels - An Ant/Beetle Buildlog

Bot Builds & Designs
, ,
1

A few months ago I had an idea for a new 'bot with a single wheel in the middle and two horizontal spinners either side of it. The single wheel provides drive, and the steering comes from changing the angular momentum in the spinners.
This is obviously a very silly idea. Nonetheless, I roped in two friends and set out to pursue it!

image
image998×562 62.5 KB

My original CAD sketch, just to illustrate the idea.

Part One - Centrifugal Farce

I really want this to be a beetle, but it made sense to start with an antweight prototype. Partially to see if this is even feasible, and partially because it’s much easier and cheaper to iterate through design ideas at this scale.

As with many antweigths, the biggest challenge was weight. For want of any real idea of how heavy a spinner needs to be to do damage I bought a pair of the lightest discs in the Monsoon Robotics shop, coming in at 21g each.

PXL_20250125_103909217
PXL_20250125_1039092171230×924 150 KB

Coupling those to a pair of the standard ant weapon motors + ESCs from the BBB shop brought the weapon stack to ~82g. Add in a battery, an N10 + Brushed ESC for the drive motor, and an ESP8266 for control, and we were left with ~24g for the chassis. That’s not a lot of grams.

PXL_20250209_234532767
PXL_20250209_234532767694×924 146 KB

We originally wanted to have one spinner mounted ‘upside-down’ as an undercutter, so that we’d be able to hit at different heights, but we just couldn’t get this shape small and light enough. After our best shot died on the print-bed, we sacrificed our original name choice - Centripetal Farce - to its memory and rebranded as Chinoops.

Even with the much simpler symmetric chassis configuration, it took many iterations and much pondering about component positioning to get below the limit. However after several late nights with Fusion and much resoldering we made it underweight.

PXL_20250319_002853726.MP
PXL_20250319_002853726.MP1643×924 258 KB

3 whole grams underweight infact! (Most of which were immediately spent on googley eyes and decorative/protective duct tape).

Meanwhile the coding had mostly been surprisingly straightforward. Here’s some footage from our first test to see if turning could be achieved:

Yes, my test-box is a reinforced draw… watching this back I realise that there really should be something heavy on top of the 6mm polycarb lid just incase. Lesson learned.

In addition to basic steering and spin-up, we decided there should be a button that would stop both spinners incase the speeds ever got really weird and we wanted to reset. This had the fun effect of undoing all prior steering in a single sharp jolt (as happens at the end of the video).

Unfortunately, the first time I tested this feature the back-emf from the motors killed the ESP, bringing the project’s ESP death-count to 3.

Messenger_creation_1838139357029581
Messenger_creation_18381393570295811920×906 278 KB

Adding a BEC resolved this problem.

Messenger_creation_3935370663350789
Messenger_creation_39353706633507891920×906 303 KB

A few more coding tweaks (including an inverter switch) and Chinoops was ready for a fight in a brewery. Part 2 (learnings from BBBxZeroDegrees) coming soon…

5 Likes
2

Really cute bot and a really interesting idea! Love it!

1 Like
3

Part 2 - Chinoops Learns to Fly

We took Chinoops to a BBB pub event in March.
All fights are in the video link (timestamped Peter Jordan). It was great fun but in this post I’m just going to focus on what we learned, and what we’re planning to change as a result.

What we Learned:

  1. Chinoops has Big Tigger Energy. This is exciting, but not especially sustainable in an Ant arena with hazards in 3 out of 4 corners. I think Craig’s commentary sums it up nicely:
    https://www.youtube.com/watch?v=HGqQQJyY9rs&t=892s
  2. The handling isn’t great. At least some of this is because I totally forgot about the existance of the inverter switch (hence so much anti-wall aggression in the first fight) but the steering is also slower and less reliable than it could be.
  3. The 'bot is surprisingly resillient. The ABS chassis was printed at 20% infill, and in some places is only 1mm thick and perforated. Nothing should be made from 1mm perforated ABS. Fortunately all of the bits that are likely to get hit are 2mm, and these survived many dings in the spinner whiteboard without losing structural integrity. Admittedly we didn’t take any really big hits, but I’m still counting bringing home a still-functioning 'bot as a win.
    PXL_20250405_221324224
    PXL_20250405_2213242241643×924 108 KB
  4. The ‘stop both spinners’ button is actually really useful! There were a couple of times when we got high-sided or otherwise trapped, and were able to unstick by zeroing the spinners to give us a good jolt of torque without needing to have anything in particular touching the ground.

What we’re going to Change:

Steering:

The biggest issue is the handling, so let’s start there. The steering is currently very slow. I could just make the spinner speed change faster to get more torque, but since there’s not much friction to stop us turning once we’ve started, this just causes massive oversteer.
Fortunately, there’s just enough weight and space to add a gyroscope (only 1.5 grams!!) so I can code up a PID loop to increase/decrease the turning rate to match whatever instructions are coming from the controller. I already did something similar for Carbonated Beverage, so hopefully this won’t be too much harder :crossed_fingers:

Bouncing:

The gyroscope also contains an accellerometer, which suggests a solution to the tigger-ing issue: use the gyroscope to automatically detect airbourne-ness, and automatically stop the spinners, so that we don’t just bounce up again when we hit the ground. This could be easier said than done, but I might be able to get it to work.
In any case I want to keep the option to turn it off for melee’s, since it’s too fun to get rid of entirely, and the air’s probably the safest place to be in those fights anyway!

Inverting:

The accelerometer might additionally allow me to automatically detect inversion, so I can’t forget to press the button again. But our existing solution of having teammates yell “Invert!” at whoever’s driving works fine, so I won’t sweat too much over this.

Not attacking the floor:

Speaking of inverting, the handling’s evern worse in undercutter-mode because the spinners interact with the floor. I think I can fix this by adding some stand-offs to the top-plate.

So it’s mostly coding and a bit of CAD. Will post updates as I make progress/discover problems.

p.s. This was my 4th Ant event, and I’ve still never actually won a fight in this weight class! I don’t particularly mind, or I’d make something less obviously ridiculous, but I think it’s a fun challenge to see if I can get at least one win out of this thing!

2 Likes
4

Thank you :heart: Interesting is definitely what I go for :joy: