Designing A Fighting Robot

Andrew Schreiber
Jul 23, 2017 · 6 min read

So, a few days ago Orlando Maker Faire announced they were holding robot battles. Obviously, I wanted to use this as an excuse to polish off the 15lb lifter I’ve been tinkering with on and off for the last couple months.

Frame for start of a 15 pound bot

Then I made the wise decision to look at both my schedule and my wallet. I just moved to a new place and my fun money budget was lower than I liked. I was estimating about $500–$800 to finish that bot, plus another couple hundred in spares. So I shelved the idea and resigned myself to waiting another year.


Skip forward a few days and I’m talking to a buddy about it, he mentions he has access to one of those fancy Markforged machines that all the MIT bot builder crew loves for their antweights. I dig through my bits and parts drawers and find a handful of little Pololu motors and a matching set of Vex 29 controllers. My Prusa was also looking a wee bit lonely.


Initially I planed on using this as an excuse to learn Fusion360 because the local hackerspace has a fairly active user group. I banged my head against it for a while, after a handful of hard crashes where I lost work I finally ended up with, well, you come up with the word. Mine isn’t sfw.

Initial rough outline of frame

At this point I got frustrated with Fusion being slow and annoying for SIMPLE tasks and also got infuriated with their camera controls for orbit being obnoxious (probably would have been less bad on a traditional mouse but Onshape works fine with a trackpad and keyboard, I’m spoiled) Oh, and the crashes. Did I mention that? What sort of excuse for a CAD program locks up when dimensioning a 2d sketch with no more than 10 lines? Then after work on Friday I was frustrated and needed to focus on something so I cracked open OnShape and did a quick clay model sketch to get a feel for the shape I wanted.

Rough shapes are defined and it already looks more elegant than the last one

So, let’s be clear, I HATE robots that just look like a box. I get that it’s easier to make or machine or fixture but dang it, I am doing this to push myself. Also, I’m 3d printing the dang thing, I don’t care about fixturing, I just need to hit print and magic happens. 3D Printers are like the Walt Disney of fabrication (/s) Anywho, from there I started refining interior details.

Wheels and electronics bay

Well, and some exterior ones, I modeled a reasonable Banebots T40 wheel to use because using a blank cylinder just didn’t feel right. It may not be 100% accurate but the functional parts are there and it’s aesthetically not garbage.

From this early picture you can already start seeing where the electronics bay will go.

Motors have been roughly located

At this point I added in drive motors and started refining the shape of the drum. The original plan was to stuff a 22mm BLDC outrunner motor inside the drum. For a bit I played with the idea of centrally mounting a motor and running a flat belt. Then I realized that looked fugly and I didn’t want to do that. I tend to claim that if something isn’t elegant it isn’t right. So I kept revising and tweaking and looking for motors.

Revision 1 of the frame

So, after a few more refinements the drum shape has changed drastically, the motor mounts have been refined to something that works and doesn’t rely on magic to hold them in place. And the top plate (not pictured) has mounting holes. Oh, on the topic of the mounting holes, before anyone asks if I’m on crack for the size of them… 3d printed parts tend to have issues with tapping them with threads. One common work around is to use heat set threaded inserts. Since I have to remove the top plate anytime to change the battery I decided to use these there. So I’m holding the whole top plate on with a half dozen m3 screws but had to put in a #8 clearance hole.

Total design time — under 6 hours.


The next day rolls around and I go fire up the printer around 4. (I decided to watch Netflix for a while, whatever) I decide to slice the frame at draft resolution (0.2mm layers) because I care about fitting things together not if they are smooth. Honestly, if I could use a bigger layer height without changing my nozzle I would. I did add 4 perimeter layers so I had enough plastic for the inserts to melt into. 4 hours later the frame pops off the bed and I slide the motors into their slots for a perfect fit.

Motor installation was a snap

I decide that the next logical step is to get the drum fit in to make sure it looks sane in the real world (I work better with physical models) I export the STL for the drum with some minor tweaks to it and print out a copy of it (the real one will be metal, but any issues with fit can be solved here cheaper) Once that was done I grabbed a tiny BLDC speed controller that was sitting around to see how wire routing would work. (You can see a small cutout for just that in the above image)

Drum prototype fitted, ESC fitted and wires routed.

So, in a hair over 24 hours I’ve gone from nothing to having a full size model. I have wheels and the weapon motors on order. Once those come in I’ll see about spinning the wheels a bit, I still need a few bearings and screws but I think the general shape is defined.


Takeaways — I’m spoiled with how fast I can get a physical model in my hand for small scale things. I probably could have turned this thing around in under 12 hours from “ideation” to “holding it in my hand” which is absolutely crazy. I don’t have some sort of advanced fabrication facility. I have a $700 printer kit and the free version of Onshape. When folks talk about 3d printing destroying manufacturing and all that I always scoff, but 3d printing (and other rapid prototyping tools like laser cutters) will absolutely have an impact. They lower the cost of iteration. If I had to machine that chassis out of a block of plastic it would have taken hours to do, I have access to a Tormach that could absolutely have done it but I’d be less likely to be willing to approach it as “this is revision 1, it’s going to be on a shelf somewhere next week” if I had to take a ton of time programming the mill, fixturing the block, and, right, paying for a lot of plastic that would become chips.

Anyone wanna sponsor a fighting robot team?

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade