Z
zwsdotcom
tech said:huh? are you saying the GPU ASIC and the PCB are made of the same
material (FR4)? i don't believe so, but even they are, what does that
have to do with the different thermal coefficients of the copper and
solder? that's the real reason the ASIC's BGA detached itself from the
board, whilst being subjected to heat/mechanical stresses.
Sigh. No it isn't. You don't understand what you're talking about, or
what I'm saying. I'm saying the chip's ball plane and the PCB are made
of the same material, so that as it heats and cools, the pads on the
chip stay vertically aligned with the pads on the PCB, and there is no
lateral stress on the joints.
There are all sorts of possible failure points due to overtemperature,
starting with the die itself, then the bond wires, then the
encapsulation, then the ball plane, then (...) but you have no evidence
that it is in fact the case here. It is equally possible that it's a
combination of the PCB layout and the process. Without isolating which
component has actually failed - it could be the GPU, the RAM, the PCB,
or the solder joints interconnecting them, or it could be something
totally unrelated like a decoupling capacitor going open circuit - you
are just guessing.
suggested re-melting the solder with a blowtorch? that's perhaps the
most harebrained idea i've ever heard of... even if it does work, i
It was me. We use this process - more or less - for hand-prototyping
BGAs. As does everybody who needs to lay down just a few now and then.