Having nothing to do in my workshop, I decided to practise my skills a bit. I digged up a scrapped graphics card from the junk box and decided to try to desolder the RAM-chips (BGA) after seeing how "easy" it looks when Louis Rossman does it.
I applied flux around, launched the hot air station, and started to heat. Realizing after a few minutes that nothing at all had happened, I tried a combo of 1) other nozzles, 2) higher temperature and 3) more airflow.
At the last point I had 400 degrees celsius and 90% airflow. Zero reaction. Even heated on the back side, no reaction.
Finally I gave up and simply pried off the chip to see how the solder balls were laid out, so I could use that info for the next chip (which went just as badly).
Then I tried the 400C / 90% setting straight upon the solder balls of the pried off chip, but the solder did not even melt. My next approach was to use the soldering iron at 350C straight on the balls, with and without a wick, but not even that melted the solder.
What I had to do was to apply a large blob of fresh solder to the iron tip, drown the solder balls in it, and then – finally – I was able to remove some of the balls with the wick. Note: some of the balls, not all of them because they did not melt.
What the hell is this BGA-ball kind of solder anyway, that does not melt?
Best Answer
Thermal inertia is playing against you. Also take into account that lead-free solder needs temperatures in excess of 220°C to melt down (compared to 180°C for tin-lead solder), so the thermal gradient will be quite high to begin with.
Because of this, I would recommend preheating the board to 120°C by using one of the following methods:
Then apply hot air to desolder the BGA chip.