17 Nov
2006
17 Nov
'06
1:12 p.m.
I'm not sure about this, but it would seem that
using water to cool
the body of an IC while needing to apply more heat to the pin to melt
the solder would make for a steeper temperature gradient through the
packaging. Could this in fact be counter-productive by increasing
stresses on the encapsulation?
I really don't worry too much about overheating a chip by desoldering
it. Consider that what destroys a chip thermally during operation is
localized overheating within the die itself. Using a temperature-
controlled iron would seem to elminate most overheating problems.
Soldering cycles cause latent failures - parts of the chip that still
work, but are much weaker. A typical type of latent failure is a
microscopic crack caused by a thermal shock. As long as the crack is
closed, the electrons will probably not worry too much about jumping
it, much like two wires touching. However, the problems of microscopic
cracks in a semiconductor die make for a weak spot that perhaps can
not handle another solder cycle or current surge.
--
Will