This thread is dedicated to all the little things which we've been able to repair rather than simply trash and replace. I think it is sadly a dying skill set (and even mindset in many cases), but here is your chance to post your proud moments of saying "Not today!" to good ol' Murphy's Law.
My most recent accomplishment in this vein is fixing a garage door opener which suddenly refused to open one morning...
Wife was headed to work and goes into the garage after I hand off her morning coffee, as per usual, but then instead of the sound of the garage door going up, I hear a muffled "Oh no". I meet her in the garage and after a brief double-check that the wall switch and car remotes only result in a relay-like clicking noise from the unit and no further movement, I pop the door off the carriage and manually lift the door for her to get out. Now the investigation begins...
LED on the back of the unit is making 5 flashes and then a few three other struggling barely-flashes before starting the sequence over again. Pull up the manual online and find the troubleshooting step for 5 flashes says the unit might be overheated (unlikely - hasn't been used yet today and it is normal temperature in the garage) or the RPM sensor is having a fault. Manual suggests unplugging the machine to reset and try again. Easy enough, but no dice. Last instruction on the 5 flashes troubleshooting matches the description for 6 flashes: Replace the logic board assembly. Sounds fun and/or expensive.
For the sake of thoroughness and funsies, I look up what a logic board assembly costs, and compare that to an entirely new unit as well. Turns out that the logic board surprisingly actually costs less than a whole new unit, at around $75 USD and $200 USD, respectively. Still, I know from experience that oftentimes you can find the failed component with a little work, and it is often a few-cent passive component rather than a proprietary chip or motor etc.
The manual didn't include disassembly instructions, so I faff about for a bit trying to decipher how to pull the cover apart. I've removed all the screws I think hold the cover on, and it still refuses to budge, but ultimately it turns out that some old rubber damping pads inside had sort-of fused it all together and you just have to pull harder than I expected.

First checks: No fuses are blown, and none of the capacitors have spewed their guts, so it won't be quite that easy. Next, scanning for any obvious scorch marks on the board: there are a handful of resistors and a diode that have some brown haze around them, strangely enough. I put each of them to the meter and they seem fine though. I look more closely at the back of the board, and after staring at all the tiny SMDs for a couple minutes, I find that a tiny little SOT-23 has blown half of its lid off. No black marks around it, but I'm pretty sure a culprit has been found. A diode check on the pads confirms that one of the legs is no longer connected at all.
I get out a loupe to inspect what remains of the package, and 75% of the part number is blown off, but luckily a handful of other SOT-23s nearby matching the first two (only remaining) characters of the blown device's part number suggest that it is a "K1G S3"-something. A few datasheets later and it looks like this is a NPN signal transistor (https://www.jameco.com/Jameco/Products/ ... 540800.pdf). I am certain that I don't have any "K1G S3"s laying around, but I do have a handful of "SS8050"s (https://www.mouser.com/ds/2/149/SS8050-117753.pdf), which might fit the bill.*
*Closer examination of the specifics between these parts reveals that my replacement transistor has a lower maximum voltage rating, but a higher overall power rating. I am not actually sure what the nominal voltage is on this particular trace, or what transients it might be exposed to, but I strongly suspect that it is still within the design parameters for the board.
De-solder the destroyed component and dead-bug style solder the replacement TO-92 package onto the old pads and we're ready to plug it all back into the garage door opener again. It works! An hour of work saves me at least $75 - not including what time difference getting the replacement board etc. would have been.
Notably, I have not determined what caused the original transistor to fail, although I suspect a more-or-less random occurrence. If some other undetected failure mechanism still exists then this replacement may not last very long, but it has worked for at least a handful of cycles so far.
Repair Win:

Explanation for non-electronics nerds: The replacement part is the "towering" (only ~5 mm tall) black semi-cylinder component, the two black rectangles directly to the right of it are the same type of component it replaced.
Other nice "Repair Wins" that come to mind (but which I won't bore you with the details as above):
- Fixing a neighbor's laptop charging port which had broken off of the motherboard inside. Value = a new laptop, or at least pricey repair bill?
- Bypassing broken house thermostat "auto"/"on" switch - permanently shorted to "auto" now. Value = new thermostat and/or cost of technician visit.
- Sewing ripped backpack/bag straps etc. Value = new backpack to invaluable if performed on the trail (yes, I keep a sewing kit in my camping gear).
EDIT 1: OK, I remembered another thread which touches on some of the same things as this one (viewtopic.php?f=42&t=121040), but it was a little more specific towards what we keep to repair electronics, and we could leave this more open to all kinds of repair.
EDIT 2: Image added.