Why does my DIY clock module reboot the second I plug in USB?

And keep pulling the ribbon while flashing if you want, that's still good habit even with the diode fitted, just for peace of mind while you're poking at firmware.

I've seen this happen with other little electronics I've built. It feels like when the USB connection starts up and tries to talk to the board, it suddenly needs a lot of power, right? Maybe that sudden change in how much electricity is being drawn is what is causing the reset, even if the voltage stays fine. It is not just the power going in, but how quickly the chip demands it. Does the process of flashing firmware pull a lot of current that the rest of the circuit cannot handle?

The problem is almost certainly related to current draw rather than voltage polarity, which is why the flashing resets it. When the firmware loader initializes over USB, there is a massive transient current spike. My last board failure happened when I pushed a module drawing 400 milliamps onto a power rail rated for 300 milliamps. The voltage simply dropped too much to keep the microcontroller stable, even for a second. You need a rail capable of handling that sudden surge during the flash process.

I had a wierd thing happen on a diffrent little board I put together and it was similar, but not the same issue. I thought it was more about noise than the power spikes, which is a guess. I tried just dropping in a big capacitor near the USB connector on my end and it seemed to settle things down when the firmware was flashing. It felt like whatever is happening on the data lines is also causing a sort of internal hiccup during the reset phase.