I just (2 weeks ago) made my first attempt at BGA reballing.
Yes it sucks, but with experience and decent equipment it may not be that bad.
I ordered most of the equipment on Aliexpress: BGA clamp and universal stencils, extra specific stencils for DDR, flux, Pb/Sn ball assortment.
For heating I have a (precision IR tool) ceramic glass cooktop with spacer and a cheap 858D hot air gun.
I think I have done enough desoldering with this setup to now get the temperatures about right.
Of course the heating and cooling slopes are not going to be as good as in an automated process.
The test was done on a Samsung UE32C6710 TV with random wonky colours, and boot looping.
The problem was thinned down to the DNIe BGA or associated DDR2. On trying to reflow the area I nudged one of the RAM's a little too hard and shorted some balls, after that the options were to give up or attempt a full re-ball.
I used a dedicated DDR2 stencil with a center trench and 0.45mm balls.
The stencil is convenient for placing the balls but then for melting them to the IC it's another problem:
The stencils I have are not very heat resistant, heating them causes a warp/bulge in the middle, then some of the balls move out of position.
Applying a thin layer of flux to the IC before placing the balls then removing the stencil was a little better, most balls stayed in position. Then the problem was getting the hot air temperature, distance and air flow right to melt the solder without the balls moving. I had to do quite a bit of reheating and placing the odd balls by hand...
After resoldering the DDR and reflowing the DNIe the problem was not solved but worse, at least this time it was obvious the weak balls were under the DNIe.
This I had to reball with the universal stencil placing every single ball with tweezers...
Here too there were several corrections to make before all balls would melt correctly to the IC.
For resoldering to the PCB, I used my own flux made of acetone diluted rosin. Preheat took about 7 minutes from underneath until the top side of the PCB was about 110°C, then preheat was cut and hot air applied from above until BGA self positioned.
I checked aspect of outer balls with an endoscope camera.
The final result is a working TV but much too much time was put into this. It was more a training approach than financial one.