Date: Sun, 13 Aug 2006 10:09:54 -0400
From: Dave McGuire <mcguire at neurotica.com>
On Aug 12, 2006, at 4:37 PM, Jeff Walther wrote:
BGA
components have a specification for the clearance between the chip
package and the circuit board after soldering. I don't think you
could meet that just by putting the thing in the oven.
Very true...but I've looked far & wide for a reason for that
specification to exist, and have found none. The only thing I can
think of is the possible flattening out of the balls causing them to
short against adjacent ones...but is that really possible? Not sure.
I assumed (certainly don't *know*) that the specification is there
because if the chip finishes too high above the PCB, then you didn't
melt the balls enough and you may have cold joints, etc. and if the
chips finishes too low/close to the PCB, then you've squished the
balls too much and they may be in danger of shorting.
At the least, it seems to me that if you meet the clearance
specification, you're more likely to succeed at BGA soldering without
expensive equipment to inspect the hidden solder balls. I could be
wildly wrong though. I never carried through with my experimental
ideas.
Another thought is that if you use the feeler guage idea that I
proposed, then you can heat the chip and board more thoroughly
without the chip sinking any closer to the board. That should help
to ensure good solder joints on the interior balls without allowing
the chip to squish to close to the board.
I think I was looking into it when the PPC750FX and later the
PPC750GX came out and I really wanted to get them onto a ZIF module
of the type that plugs into the Beige G3 model Macintosh. But
Powerlogix came out with a couple of nice cards, so buying commercial
became cheaper and more practicle than trying to do it myself.
(Thank goodness!)
Jeff Walther