Kicking the CAN #4 — Hunting the Bug (STM32)

2026-03-05

Claude Code very quickly adopted the stance of "I did everything correctly" and then came to the conclusion that it must be a hardware problem. In between, an unsuccessful attempt was made to swap sender and receiver, to make the LED blink differently or not at all, and when none of that helped, Claude Code started chasing a fixed idea.

I know this from earlier experiments a few months ago — it happened quite often back then and was usually accompanied by confused hallucinations.

After Claude instructed me to measure the voltages at RX and TX, and these didn't match expectations (TX was near 0V), Claude was convinced it must be TX, specifically because it was connected to USB. It asked me to look for a solder bridge that it believed must exist on my Nucleo board. There was supposedly only one way to get the expected voltages, and that was to physically disconnect USB from RX.

Since I already knew this state of fixed ideas from earlier experiments, I ended the session to clear the context and continue the hunt the next day.