Well, I haven’t written in a while – have been procrastinating until the end of finals and then relaxing for a couple of week but…it seems like I should probably update now, perhaps…a month after the competition :D. First, a post about all the frantic last minute fixes and debugging. Then, I’ll write another post with the results of the competition + reflections 🙂
I left off at the formation of Milestone 7 (vehicle frame check point) and Milestone 8 (making electrical diagrams). The next milestone was to actually complete the electrical system. Near the end, everything blends into a blur of soldering, testing, and taking apart the chain/motor, as we started to operate on a tight timeline to finish before the race.
Wiring: A night of soldering
Aisha and I came in on friday and basically stayed until we’d finished soldering that electrical system, following our electrical system diagrams. I used a lot of the 4mm bullet connectors – replacing some of the ring terminals with them, simply out of taste.
We soldered most of the connections, using XT60 connectors as the main connector from the battery to the rest of the system. (The diagrams are shown in the previous blog post). I also sat down and finished CADing the battery box. I’d previously ordered wood from Ebay, and I set out designing a fingerjointed box to hold the batteries and also a mount for the switch. Unfortunately, the box (and thus, the switch) were behind us – still within hand reach – but we’d just have to learn to instinctively reach behind us to get to the kill switch. (2 switches, one for the precharge resistor, the other the large Hella power switch). The beautiful battery box came out perfectly – t-nuts, fingerjoints, and all! But I had to scrap it – in the CAD, I forgot to put the chain in, and when lasercut, the box conflicted with the chain. I decided to admit defeat, get rid of the offending box sides, and just velcro the damn batteries to the bottom piece of wood.
Hall effect sensors:
Because we used brushless motors, we needed hall sensors to tell the motor controller where the rotor is. First, the hall sensors needed to be calibrated, and we spent many tired hours figuring out the right configuration – at first, nothing worked. This was because the tiny hall sensor board was very difficult to solder precisely. Charles helped us to debug, and we discovered that our hall sensor board was shorted out. I got another board and this time, after careful soldering and calibration, it worked! The hall sensors, however, caused me the most problems in the electronics. Later, the 3D printed hall effect sensor mount broke in the middle of testing, which meant that I spent valuable time taking down the chain/sprocket/motor mount to replace it.
After figuring out the hall effect sensors, we were finally able to drive Phoenixkart around lab. Our first test drive was recorded by Charles with his newly bought Chinese dashcams. With further test-driving, we discovered some problems. First, and foremost, our chain kept falling off. This was largely because our kart “dgonzed“, aka, it didn’t have enough triangular supports to keep the two body pieces of 80/20 horizontal with the front of the vehicle. This meant that when the driver leaned into sharp turns, the kart leaned into them too – during the race, you could hear the kart scraping the ground when I took it around really fast on turns. That was fun.
Back to the problem. The chain kept popping off the sprocket because as we turned corners, the back of the kart would torque, especially because of the way we designed the back wheel supports. Essentially, the back wheel supports are cantilevered, which allows the frame to tilt substantially around the wheel. If I were to design this again, I’d make the wheel axle go through the entire column.We solved this by shortening the chain and perfecting the tensioner. We only finished debugging this the night before the competition. If I were to redesign this kart, I’d make it more stable by having 4 wheels.
Problem #2: Pinning the steering column: the steering column shaft collars that Aisha had cut out just weren’t staying, so she pinned them to the steering column by simply drilling a screw into the shaft collar and the column.
Problem #3: Blowing Fuses: Our battery fuses could only pull 40A, but our motor controller was able to output up to 130A. This meant that when I pressed the throttle too hard continuously, the fuses blew. Charles handily taught us how to program the kelly controller to only output 40A.