Pictured above is the bizarre metal creation that holds the whole vehicle together. It isn't a good structural design, but it is exactly the correct size to fit everything that needs to be held between the tracks. It will begin to make more sense in conjunction with the rest of the components. But first:
Now painted blue and with the custom gearbox mounted. The drive shaft just barely visible below the motors will be supported with a HDPE/Delrin bearing block attached to the frame. The same system will be used at the rear for supporting the track's idler wheels. The shaft is mounted to the sprockets through a custom aluminum adapter and shear bolt. This system isn't ideal, but it's about the best I can do with the budget and tools available.
Now everything hopefully makes more sense. This will be the scooter's final configuration (obviously with the bearings, batteries, and electronics not pictured). Unfortunately, I don't have many more pictures after this stage. To summarize, the SLA batteries were mounted to the baseplate and soldered together in parallel. The Sabertoother 2x60 was configured correctly for tank drive, and wired up to the motors and an RC receiver for testing. That's when things stopped working.
I've had problems with the Sabertooth 2x60 speed controller before on Not-Segway, but it thankfully kicked back to life with some gentle tapping with a mallet right before I had to present the vehicle. I assumed that glitch was just a random occurrence, because I otherwise really liked the Sabertooth. After this project though, they've pretty much lost my business forever. After correctly connecting everything up, I encountered the strange error of only one motor side working. I spend hours troubleshooting, including switching it out with Not-Segway's controller, which worked perfectly. I finally found that the only way to get it to function on both channels was to switch the settings from RC-mode to Microcontroller mode. After this, it worked normally and was very responsive. My theory is that the lost-signal sensing feature on RC mode is tuned way too high on the non-working channel, leading to the Sabertooth constantly keep it locked down. This is frustrating, but not a mission critical feature for my purposes because I was planning to have all input sources go through an Arduino anyways (on which I will write my own signal-loss code). It's not worth the trouble of returning the unit, but I'll be keeping an eye out for other controllers with similar features and cost.
With the electronics working and a freshly cut aluminum deck in place, I was able to take it out for a quick spin on RC mode before packing up.
This is tank scooter's state as of now, partially assembled but sitting in a corner until this winter when I come back. Check back for steering installation and some driving video then.
