ALEX HATTORI

View Original

Uppercut CAD Reorganization

I never made a post about how Uppercut (250lb robot for the tv show Battlebots) was designed due to the time crunch of build season (~6 weeks from when we found out we were accepted until we had to have a finished robot packed in a crate). Overall I was very happy with how it turned out. Mason and I did the majority of the CAD and integration with others jumping in and many many group design reviews. We all used Solidworks 2017 at the time (2018 now) and GrabCAD for version control.

The physical robot we built worked about as well as we expected, and in competition managed to make it to the top 16 where it lost to the eventual champion, Bite Force.

Back to CAD. Unfortunately because Mason and I have different CADing styles, the main design file is a little temperamental. I focused on making everything into the smallest subassemblies possible and then build up my main assembly with many modules, whereas Mason preferred to design a lot of parts in place in the master assembly. Now that we’re considering working on a few new design changes to the robot remotely (its locked up in its crate at MIT) during this quarantine, I thought it’d be a good time to clean the CAD up. Overall I am very particular about never leaving an assembly with any overconstraints or unsolvable constraints but its getting hard to maintain as our design becomes more complex.

For example, some parts have hole patterns that are defined in assemblies but now that we’ve already manufactured the parts I made copies of the parts and fixed all the holes so we don’t accidentally modify the CAD to be different from the real life hardware. I started doing this because we switched from using socket head bolts to flanged hex head bolts (because of the integrated washers) and when I changed that in the butt assembly, everything in the main assembly broke.

Similarly a lot of welded assemblies were designed in the main assembly but now that the parts are locked in and already welded I made everything more modular. In the above assembly, all the welded steel parts are assembled in that assembly, but now I have subassemblies that consist of welded structures that actually are connected.

When we were first designing the robot, we didn’t know how configurable it would be. Now we know we can choose to have many different types of armor, self righting mechanisms, top plates, bottom plates, motor controllers, motors and plenty of other attachments and so I decided a complete overhaul of the CAD was necessary.

I started by making a new assembly. I wanted to leave the original design assembly intact in case we need to design new parts in a similar manner, and so I also made copies of many files, labeling them as “fixed” files.

Now, the new Configuration Assembly starts off as the completely open and exposed butt of the robot. In the butt assembly we can choose between using Arc200s, 6 Fet Vescs and 12 Fet Vescs (motor controllers). I might add another layer for battery configurations which we may need due to requirements of the new motors discussed later. The nice thing is that all this nesting of assemblies makes it super easy to add another configuration.

The next level we can choose any combination of aluminum or steel top and bottom plates.

Then any combination of front or back welded butt armor.

Then we can choose which motors we want to use to drive the weapon. This was the main driving feature behind doing this reorganization. We are trying to remotely design/build/test a new set of weapon motors that are IPMs (internal permanent magnet motor). These motors are unique because they can field weaken especially well, meaning for the same kv motor, we can gear it up more, and still reach the same top speed. Using this motor setup is risky though because it involves changing our physical motor assembly (and adding a spur gearbox between the motor and the pulley), making a custom motor controller, and changing our battery setup (we need ~200V for these), all while we are scattered around the country with limited tools and the robot is locked up at MIT. In order to mitigate risk we are trying to make everything modular so we can build them separately during quarantine, and then bring them to the competition, put them in the robot and test.

After choosing the motor, we can choose which self righting attachments, tubes and/or horns, we want. This is important because with the IPMs and twice the torque, we may be able to flip ourselves over with the torque from the blade and thus won’t need self righting attachments.

Next we can choose which pulley armor we want.

Followed by what attachments (forks/wedge)

And lastly the blade (fist, bar, hoop).

Overall the assembly is very comprehensive now with almost every bolt in place. Additionally I have measured weights of a good chunk of the pieces from last year so we can substitute those into assemblies to account for welds. The configuration assembly will be useful for estimating what armor configurations we can use, and also is a lot cleaner than our original design assembly.