TL;DR - Nothing of value. Just project logging my particular failures.
Update of Shame:
- I soldered the three transistors in the top right backwards (correct to the silkscreening), because I missed a note about the silkscreen being wrong.
- I attempt to desolder the transistors to flip them around. I have a hot air SMT rework station, but no, I use the iron from said station to try to desolder the transistors instead. I know the pads lift easy, so I'm careful with it.
- I'm too careful with it, and I keep not using quite enough heat to pull the transistor out. When I do, I either rip or melt off two of the pads on the first transistor anyways.
- It gets worse, I snap the middle leg off of a transistor too, flush with the base (I ordered no spares because "I'm not an idiot, it's not like I'm going to wire them backwards and then break them turning them around", literally what I thought when I was "saving" myself $0.25).
- I use an x-acto knife blade to score around the middle pin on the transistor, then bend up a spare resistor leg into a Z-shape, glob solder on, and hopefully repair the transistor like it's 1957 and they're still precious.
- I test the transistor by diode checking EB, BC, EC, etc on my DMM. Checks good, no continuity when there shouldn't be any, 0.587v drop where there should be.
- I whip up some spare body panel epoxy to give structural hope to the center leg. Repaired?
- I'm too chicken to try to do the coffee straw trick to blow air at the now-filled solder pads when hot to blast the solder out. So I opt to try to just heat all 3 pins at the same time and hope I can push the holes clear when putting them back. Of course this doesn't work.
- I use some resistor leg trimmings to hopefully re-pin the holes. I then cut them flush on the back side, because I'm not thinking ahead.
- I try to resolder the transistors (the correct orientation) to the pins. Of course the pins are so short now that they have no heat sinking and float off their pads on the back side.
- I make it work anyway.
- I test for grounds where they should be, and no grounds where they shouldn't be. Tests okay.
- Possibly ruined board anyways, probably annoying inconsistent bugs in the future with howevermuch power these TO-92s are pushing, flexing the solder as they heat under load.
"I can't fabricate worth a damn but at least when it comes to the electronics it'll be smooth sailing for me" -- MattsAwesomeHubris, 2019.
The good news is that once I've made all the mistakes, only the correct way to do things will remain!
Still haven't programmed the blue pill, might do that next while I try to figure what gets wired to which terminal blocks.