Robot Ruckus 2021 didn’t go very well. Let’s get that out of the way first.
And unfortunately I don’t really have enough footage to make a proper video, so I’m going to skip it for this competition. Sorry everyone! Instead I’ll try to sprinkle more media in here than I normally would.
I’m going to spend this post talking about what we did going into the event, what went wrong, and what we’re going to do going forward.
What we did different
CnC Manufacturing
Pierce now has access to a nice CnC machine, and this has upped our manufacturing game tremendously.
The Halo chassis, partially machined
With a CnC in hand, we can add more complex geometry to Halo’s chassis. We no long use wheel weights to balance the bot, instead adding material to the ring to ensure balance by design. Flats were added across from holes to make the holes easier to punch. And the need for motor mount plates was removed, we can bolt the motors directly to flats on the inside of the ring.
For Hit ‘N Spin it made a big difference. We now machine Hit ‘N Spin’s chassis out of UHMW blanks. This should prevent Hit ‘N Spin from exploding on contact like last time.
Hit ‘N Spin chassis being machined
Hit ‘N Spin not exploding
Halo Electronics Placement
I wanted to simplify the wire routing in Halo by moving the circuit board near the battery. My thought was the circuit card could double as power distribution, such that the ESCs route to dedicated pads for each. No more Y splices.
Here’s what we came up with:
This new placement saved a lot of weight, and made our “squishy bits” a smaller target.
After seeing Project Liftoff get taken out by Sawmuri, we invested in some better top and bottom armor for your battery box. Seen above, both sides have a UHMW/Nomex (fiberglass)/PLA sandwich to deal with a variety of incoming attacks.
Halo Electronics
A new board was spun for the new placement in the chassis, but no major changes were made. In fact, a lot of extra pads and unpopulated features had to be removed to make room (it was a layout nightmare).
Unified Codedbase
I spent a good amount of time combining Halo and Hit ‘N Spin into a single codebase. This ended up more than paying off, as any improvements in one automatically payed off to the other bot. I also took the time to completely remove ST’s HAL libraries, which I’ve been suspicious of from the beginning. Now all of the code that is running has been written and verified by me.
The Handheld Controller
To support a new experiment I’m working on, a new handheld controller was made from scratch. I’ll detail the experiment in a later post once I have it working, but I’ll briefly list the specs here:
Raspberry Pi 4 with Touchscreen Hat
Thumbstick and deadman’s switch
Xbee radio
Picamera module
USB power bank
The controller software was running off of a python application. The screen showed battery information and RPM, which was nice to have. But the touch screen wasn’t as reliable as a mechanical switch, and caused some problems during a Halo match.
What Went Wrong
Both bots had individual problems.
Hit ‘N Spin didn’t have traction
Going into RR2021 I had the highest hopes for Hit ‘N Spin. It seemed nigh indestructible, and it’s unusual geometry is hard to fight in antweights. But ultimately, it was defeated by the floor.
This is close to the floor used
Losing a match because of the floor seems like a right of passage in battlebots, and I can now say I lost a tournament because of it. They used a plastic backsplash panel as the floor for the tourney. It was enough grip to get spun up, so it took a lot of experimentation to figure out what was going on. Turns out, it was not enough grip to translate. Hit ‘N Spin spun up, then sat there until it got counted out.
Thankfully, Halo maintained enough traction to translate, but…
Halo had motor problems
The untimely discontinuation of our beloved AX-2810Q motors caused us to source a replacement model. The new motors were a rollercoaster. They boosted our acceleration and top speed appreciably, which was great to see. But two days before the event we realized they increased wheel wear dramatically.
And finally, when our first hit landed, the bearing in one of them exploded.
For the second match we were able to mount the old motor on instead, but incredibly that motor lost its bearing too!
Though even if the motors had held, there were issues in the battery box.
Fracture points circled in red
The battery box was cracking in multiple places, from the shock of impacts alone. It was still together by the end, but we knew it was only a matter of time before it exploded.
The impactor on Halo is about a half inch longer than the first revision. And it is much more fortified, such that it does not yield on impact. As a result, Halo hits way harder than it did at first, and its internals get shredded in the process.
We’re playing a game of finding the next weakest link. Eventually, we’ll find them all. Until then I expect us to continue losing more than we win!
What We’re Going to Do
Customized motors
Clearly stock motors aren’t cutting it. We need to replace the bearings at a minimum, but we’re looking to go further and convert the propdrive motors into hubmotors. This should hopefully put motor reliability issues to bed.
No more PLA
We used PLA for the battery box, because that’s the only thing our printers can do. We knew PLA was brittle, and used careful design to mitigate issues. I then hit a brick a couple times to double check.
But clearly in the heat of battle, PLA can’t handle the shock. It will be for prototyping only. We will be getting printed parts made professionally and mailed in from now on.
Figure out how to translate on “weak” floors
I suspect smarter motor control could have saved Hit ‘N Spin. So I plan to acquire the same flooring and get it to run reliably.
Driver Change
Pierce will be taking over driver responsibilities going forward. After so many events, we’ve agreed that I’m better at running the media side of things. When these bots are working they basically drive themselves, and Pierce has a ton of RC plane and drone experience, so we think he’ll do great.
As for me, I got an enormous amount of practice with my Chronos 1.4 during the event. I’m now at a point where I can reliably keep the action in focus, so it makes sense to have me do camera work full time.
Build a 30lb Hit ‘N Spin
Okay, this isn’t really a lesson learned. In fact, we probably shouldn’t. But it just sounds too fun, and we’ve got ideas we’ve been dying to put to the test. So we’ve started designing our next bot, which we’re calling “Bat out of Hell”. Hopefully we can get it together for this year’s Norwalk Havoc Season!
Chronos 1.4 Captures the Glory
To wrap up this post, here’s some footage I caught while we were sitting around being losers.
If anyone was at the event and I might have gotten footage of their bot, drop a contact form with a link to a bot photo and I’ll send you what I have.