Doing one final post after uploading the scraps of documentation I was able to salvage. There’s a bit of a story here, I’m not expecting anyone to read it, but it feels good to vent…

History

This was a passion project & capstone for my electrical engineering degree. The project was about 3 months, from conception to final deliverables (demo, documentation, and formal-ish presentation). I’ll be blunt… this timeline is incredibly idiotic, and the department should have made it 2-terms. My partner worked on the firmware, and I managed schematic, layout, and test. I also had some help from professors and lab technicians in the department who were kind enough to participate in design reviews.

For the most part my partner and I worked in parallel (he had a dev board), and I had LTSpice and enough of a home lab to get by. When it came time to integrate the system, it was HELL!!! I soldered the first board by hand, and flashed it. All was working great then the STM32 stopped talking. I desoldered the MCU and tried again…. After about 30minutes I had another dead STM32.

After way too much debugging I found it was caused by bad part selection. My 3V3 switcher was meant to be a fixed 2V5 output. The impedance in my voltage divider was obliterating the transient performance of the device. In desperation I tried lowering the impedance of the divider (pissing away power as a result), and this helped for some time but eventually another STM32 fell victim to my bad design. In the end we cut our losses, and demo’d the systems separately. FW using the dev board fly wired to the display of our case, and HW a newly populated PCBA and some of my lab equipment (sans 3v3 switcher). [See Demonstration Day – Log]

I should also mention that this integration hell happened the night before the final demo. We were already working on a tight deadline and hadn’t budgeted in any hiccups. My friend and I were debugging until 2am, and crawled back to university the next morning for demo day. Here’s (I think) a funny picture showing the state of my “lab” at the time, the morning after the debugging hell. (cold coffee and all!)

Demo

Never before seen demo! The thing did actually work, albeit this was just my chopper circuit operating w/o the STM32 at the time.