Al Slagle ([log in to unmask]) asked:

>As a follow on question about solder paste;
>
>If the SMT parts are glued onto the bottom side of the board, and thus are held
>into place with the glue and not the paste itself, How can a proper solder
>joint exist?

Most of the time, parts which are "glued" in place are later wave soldered.
There is no paste. The wave wets the pads and the leads, and solder joints
will form nicely if the geometry is OK. The height of the glue dot needs to
be sufficient to ensure the belly of the part will contact the dot when the
pick and place machine puts the part down. The volume of the dot must not
be so great that the adhesive will smear over the pads at device placement.

The adhesive we use is thermal epoxy, and after placement we put the
assembly thru a furnace which cures the adhesive - then we wave solder.
This process can accomodate most discretes and 0.050" pitch gull wings if
you can tolerate some shorts, and if the device orientation and pad shape
is optimized (which unhappily ain't normally the case).

For two-side reflow, Interphase designs are not glued, they are paste
printed, pick & placed, IR reflowed, cleaned (optionally), inverted, paste
printed on 2nd side, pick & placed, then reflowed again. The parts on side
one will stay in place (even during the second reflow) due to the surface
tension of the molten solder.

We do not put large parts (or parts whose mass is high relative to the
soldered leads) on both sides. Such parts might drop off during the second
reflow.

>SMT parts on top of the board are seated down to the surface of the board by
>gravity/convection.

Most SMT placement machines are controllable for Z-axis force, too.

>If glue holds SMT parts, won't these back-side parts be permanetly held off
>the surface of the board? and thus require a different solder-paste approach?

I have heard of a process where 2 pass reflow is used with _both_ adhesive
and printed paste. If this were the case, then the adhsive dispense would
be a little more critical, but I'm sure it could be done. The adhesive we
use is much less viscous (prior to cure) than paste. I would think that one
problem would be how to b-stage the adhesive without thermally degrading
the printed (but not yet reflowed) paste.

Some other more clever process type will probably tell us how.


cheers,


Jerry Cupples
Interphase Corporation
Dallas, TX (how 'BOUT them COWBOYS?!) USA
http://www.iphase.com