Summary

We drove to the Museum parking lot in order to start for the morning, which we were told supposed to open up for us at 8:00 am the previous day. Unfortunately, this was not the case, and we waited there for about 10 minutes before being told by the organizers that it was probably okay to start on the opposite side of the road. There was a bit of a scramble as we ran through our sanity checks, but ultimately we managed to start on time.

We made a pit stop in the McDonald's for iced coffee and a washroom break, where we briefly debated whether or not to take the car through the drive-thru. Ultimately, we decided against it, and instead it turned into a PR session as some curious members of the National Guard came by and took some photos with us. After spending at least 30 minutes there as we waited for people, we decided it was about time to get going. Then we struggled for a bit to get the car started again, where it seemed like Driver Controls was hard-faulting. We ended up fixing this by rebuilding and reflashing Driver Controls with Link-Time-Optimizations disabled.

We drove for a little while longer, before running into the same Motor Controller issues that we were seeing the previous day. Our driver, Micah, noticed that it the issue seemed to happen consistently when the motors controller transitions from the six-step square-wave to the sensorless control mode, which uses a space-vector modulated sine-wave output. This allowed us to rule out an issue with Driver Controls and the Motor Controller Interface. We tried debugging the motor controller CAN bus, but then realized that somehow one of the connections to our CAN-USB dongle was damaged and needed to be resoldered. One of the residents came out, and we chatted for a while before borrowing an outlet in his garage to set up a soldering iron.

We managed to start moving again, but the convoy decided that we needed to recreate the issue in order to have a shot at understanding what was going on. We likely weren't going to be able to do so at our stage stop, and so it was probably going to be somewhere around here. The convoy pulled into an empty church parking lot, and we proceeded to drive around in circles as we attempted to replicate the issue and debug. We managed to get it into that transition phase at 10km/h, and verified that everything on the System side seemed to be okay. It was likely a Power issue, or the Motor Controllers themselves. However, we were also racing against time in order to make it to the Stage Stop, and so we collectively decided to trailer and make use of the extra time to debug that night. At this point, we didn't really have a plan, and we were just hoping for the best.

We arrived at Scott's Bluff National Monument, and all we saw were a few team trailers. We were a little confused by that, since it was almost the end of the regulation raycing day. As it turns out, we had missed a route update, and the stage end wasn't at Scott's Bluff National Monument, but rather, at the Oregon Trail Museum and Visitor Center. So we had to drive another 10 minutes to the actual stage end location.

At dinner, we sat with PrISUm and they offered to help out in whatever way they could, since they were out and had begun coining themselves as "Display Class". It turns out that we were at the same campground that they were staying at, and so they came by afterwards. It's hard to explain an electrical problem and solve it, since there's so many ways things can go wrong and typically that only works when the problem is trivial or there's something obviously wrong, which hopefully we've already resolved.

We jacked the car up, and attempted to replicate the issue that way. Unfortunately, we weren't able to do so, probably because there wasn't enough friction on the wheels in order for the motors to draw enough current.

Anyways, at this point, Karl noticed one of the observers (Spencer Berglund) hanging out by our car, and since he wasn't our observer at the time, Karl asked if he knew anything about Motor Controllers. Turns out, that was probably the smartest thing that we did all race. He took a look at our Tritium WaveSculptor 20 configuration, and immediately, he pointed out that the RMS Peak motor current maximum on them was rated for 100A, with a peak current of 122A, but we had set it to 130A. Motor impedance will limit the current at higher speeds, which limits the bus current. He suspected that although we weren't pulling anything near that, flashing a value outside of the valid range of operating values was causing them to fault, since they would hit the hard limit first.

Spencer also suggested that we run PhasorSense and ParamExtract, and since we didn't have anything to lose, we did. It was a lot of effort in order to plug in our CAN-USB adapter into the Motor Controllers, since we needed to unmount them and then attach the dongle, which took a good 30 minutes at least.

The plan was to find a parking lot and drive around, and see if we could reproduce the issue with these new changes. However, we first attempted to try and clarify a regulation that stated that we were not to drive after the end of the raycing day to figure out if that was restricted to driving the route, or just driving in general. We didn't want to be driving around on gravel, as not only would traction be severely reduced, but we would probably wake up everyone on the campground as our motors transitioned from the six step square wave to a sinusoidal wave. In the end, we decided that the regulations were vague enough that we could probably argue our way out if they decided to penalize us, and so the best course of action was probably to just notify our observer about what we were planning on doing. At this point, our observer had fallen asleep, and so we talked to the observer for Onda Solare, who was totally cool with keeping an eye on us (since they had impounded their battery). He was super helpful, and read over the regulations before unofficially endorsing what we were about to do.

However, at this point, it was already past midnight, and we still needed to charge (which would take at least 5 hours), so we decided to call it a night and hope that we fixed everything.