[sdiy] Reflow revisited and a bit OT, sorry

[Tom Corbitt]

BGA's aren't that bad, the only real issue with them is you need an
x-ray to be sure what's wrong with it and the layer densities required
to pull out everything on high pin devices pushes the cost out of
hobby world.  I've done some hobby bga boards (low ball count) It's
really neat to watch a BGA chip self align when it hits the molten
phase. I've had good success with a hot air rework station. With FPGA
devices, as long as you don't short power and gnd pins, it's trivial
to write a self test to check your banks for shorts after you've
soldered it down.

The problem with the xbox is that microsoft skipped the underfill. The
thermal expansion differences in the cpu/pcb material stress the balls
and over time you get the ROD.

All the large count bga work I do professionally gets x-rayed (usually
about $50 per board depending on volume). Some assembly houses will
shoot the whole board, it's really cool to see the internals of chips
and passives (wire wound resistors are always cool)

Tom C.

On Sun, Aug 29, 2010 at 4:37 PM, Travis Shire <tshire at charter.net> wrote:
> Those console games use a Ball Grid Array system for the large chips.
> There's a "right" way to rework boards with BGA that involves a special flux
> and temperture regulated reflow oven that can get to the correct temperture
> for lead free solder. Its not something I'd want to do without the right
> equipment unless its just for experimentation on something that was headed
> for the trash anyway. A friend of mine has a shop that does a *lot* of those
> console repairs and he's invested a ton into the gear to make it happen.
> And....its a semi permanent fix. There *will* be more problems down the
> road.
> http://jdelectronicsinc.com/index.htm
>
> So....I wouldn't expect great reliability after baking them in a regular
> oven. BGA is fiddly enough before it starts getting stressed from heat and
> flex.
>
>