Tank Scooter/Scooter Tank?

It's been a while since I've made an electric vehicle. So, rather than finish one of my numerous unfinished projects like I should, I decided to spend the last month working on something totally new. It wasn't supposed to be design intensive, overly complicated, or even practical, and it generally succeeds in those respects. With (relatively) limited tools over the summer away from college, there would be no pretty waterjet/milled/CNC parts either. This lead to the only reasonable project idea:

I still don't know what to call it yet, but when it is done, it should look something along the lines of this:

SegwayTank?

Its insides are relatively simple too. Two big SLA batteries, an Arduino, a (probably insufficient) SaberTooth 2x60 dual motor controller, and four CIM motors running into a custom 12:1 dual gearbox. Steering will be accomplished by turning the Razor scooter steering post (measured by potentiometer at the base). Throttle control is still not decided yet, but if I can make a handle grip twist system, that will be the way I go. 

Because there were somehow no tracked snowblowers available in June on Craigslist OR eBay for reasonable shipping, I had to buy the tracks and sprockets from a Honda OEM parts retailer for way too much money. Now committed, I had to decide how to power the thing. I'd previously used two CIMs in Not-Segway with reasonable success, so I was fairly confident in their reliability. However, in that project they had been used in conjunction with planetary BaneBots gearboxes, which are expensive. Unwilling to shell out the cash for that solution, I decided to see what I could make using just the ShapeOko and Harbor Freight microlathe.

I knew that four motors could only create more inefficiency and potential problems, but some quick math clearly showed how impractical two would be. I could have upgraded to larger motors, but I thought it was fun to design for four. Shown above is only half of the gearing.

The dual half-shafts (not shown) extend out from the upper axle. The two sides are (obviously) not linked in any way, they just share the same bearing mount plate in the middle. 

Because the tracks still had not arrived, and I couldn't design the chassis until I could measure them, I went straight into gearbox construction. After my recent ShapeOko upgrade, I had a series of bizarre problems which seemed to be originating in the motor drivers or wiring. I tried lots of tests and fixes, none of which really fixed the issue. After getting frustrated and leaving it alone for a week, however, it suddenly began functioning perfectly again. I'm don't know what the moral of the story is, but regardless it meant I could cut out some gearbox plates. 

After previous adventures in ShapeOko aluminum milling, I wasn't too eager to make the plates out of aluminum. Looking around on Amazon, I saw a great price for .5" HDPE cutting boards. I hadn't worked with HDPE, but heard lots of good comments on the ShapeOko forums and seen it in many robots before. I bought two sheets. After dialing in the feed rate with a couple test cuts, I was ready to mill:

The plastic milled beautifully, and is now one of my new favorite materials. I've inserted some of the bearings into the plate in the photo above. I cut three identical copies of the plate for simplicity, then began connecting the motors and machining the shafts and Delrin spacers.

While most of the gearbox went together with spacers and snap rings, the "shaft clips" used above (sold by AndyMark) worked very well on the CIMs.

Turning the snap ring groves into the shafts was actually surprisingly easy. I didn't have a specific cutter for the job, but found that grinding one end of the parting tool to a fine edge was more than adequate. Also shown above is one of the completed middle shafts.

Visible in the finished product above are the threaded rods that hold the sandwich together. Due to a poorly placed hole that I didn't catch in the model, I had to drill out two additional threaded rod holes by hand to make everything fit. Since the rest of my materials had arrived by this time, I took out one of the 12 volt SLAs and gave the assembly its first run. 

I don't have any video of it running at this time, but it is safe to say that it is very loud and powerful. So loud in fact, that ear protection is probably a good idea when running at full power. I actually like the noisiness (it goes well with the absurdness of the vehicle), but hopefully it will lower in volume a little bit when it's mounted. Everything seems to run smoothly and is fully greased, so it looks like it will soon be time to drop this into a chassis. Stay tuned for the next update as I try to finish this project before college begins again.