We injected voltage on "purple line" which is the U4 input and observed the voltage on the output after the coil which is the 3.3v line.
Around 10V the output came alive with around 1V on it.
as I said earlier: "if you are super double sure there is no voltage on purple and no short to ground then inject 5V there, either way set current limit to 100-200mA and go from there"
what made you inject >5 V??? "go from there" doesnt mean keep injecting more voltage
, it means continue diagnosing. There is no situation possible where 3.3 DC-DC converter doesnt work at 5V, but starts working at 10V (other than it being faulty)
Around 20V the output
and you kept on going
At around 24V the output was just under 2V and then something happened and the voltage was brought down while the amperes spiked to around 0.5A.
And that was it. Now the 3.3V line is short to the ground.
just as I wrote: "injecting 5V might go bad without precautions, for example U4 can be bad, pass 5V to microcontroller and fry it"
tldr either:
-you blew U4 and it passed 24V to the MCU
-or there was always short to ground on 3.3V rail
As a last resort we cut off the output trace of U4-to-coil and tried injecting some amperes to 3.3V line directly. All we got in return was CPU getting hotter the higher we went. Not sure why but it appears to be fried now.
That was supposed to be the _first_ thing to try
At least it was a learning experience for you.