John Larkin said:
We sent out a batch of 35 VME module kits to be built by a contract
assembly house. There are lots of high-value resistors on this board
http://www.highlandtechnology.com/DSS/V360DS.html
in the filters and such. Boards started failing in test and it seems
to be caused by ionic contamination trapped under parts. The stuffers
used water-soluble flux (which is contrary to our rules) and obviously
didn't clean the boards enough.
Water soluble is not nice stuff - it is conductive, will tarnish metal, make
solder joints brown if left on for one to several days, and it can be hard
to clean. On the other hand, it is easy to thermo profile as it has a wide
process window, is very active, which means it will wet oxidized parts
better then most other fluxes. It is also easy and cheap to clean, if you
can get to the flux under the parts.
To prevent any problems like you describe, we only consider using water
soluble for thru - hole and mixed boards in the wave, since the parts are
big and chunky (mixed boards would have RMA (to be solvent washed later) or
No Clean on the SMT preventing the entrapped areas from getting water
soluble in them), and wash off easily, and some though hole parts have
lacquer coatings (mainly resistors!) which will wash off with all but the
weakest solvents.
To get the water soluble off, we do a pre wash by soaking the boards with
tap water for several minutes, change the water (the tap water soaking is
only to help make our expensive carbon and DI water filters last much
longer), soak in a water soluble neutralizer for a few minutes, and then
wash in a *closed loop* DI water washing machine, which is basically a
modified stainless steel dish washer. Putting the boards in the washer until
the wash sump resistivity reaches over 2.0 meg ohm (compensated for standard
temp, and is the standard 1 cm cubed measurement), which is mil spec for
ionic contaminants generally ensures the boards are clean. For problem or
high impedance boards, we may set the washer to run until 2.5 meg ohm to
make sure all the hard to clean areas are clean. The wash temp is around 140
deg F and the water entering the sump is generally 16.x meg ohm. Perfectly
clean DI water at standard temp (~20 deg C) is 18.4 meg ohm. DI water very
quickly wants to remineralize it's self and will grab CO2, trace gases,
parts of the container, etc and quickly become under a few meg ohms within
minutes depending on surface area and temp. Over-washing boards with 2+ meg
ohm DI for an hour or two will dissolve some of the metals out of the solder
and leave an interesting spider web like pattern when viewed under a
microscope, so over washing is likely not a good thing, and shows how
aggressive DI water is. Putting 1 clean board in the wash, with a single
touched up lead with a small dab of water soluble flux will drop the sump
resistivity to 0.0xx to 0.1xx meg ohms!
Washing with something that has decent agitation, can monitor the water
resistivity, and keeping it cleaning until you at least meet the mil spec of
2.00 meg ohm should get the boards clean, as DI really wants it's ions back.
Using hot water should reduce the surface tension and increase the
aggressiveness enough to clean under tight SMT parts. You may even be able
to get away with washing the boards in an ordinary kitchen dishwasher
(keeping in mind ESD, as most kitchen dishwashers are not in a ESD protected
area)!
Also don't leave the boards wet for a long time - the solder will corrode,
and any steel / iron parts, leads and wire may rust. Bake at ~ 90 deg C for
at least 1/2 an hour to dry and remove traces of moisture. Note that some
solder masks which may not be fully cured may absorb water and turn murky
looking - heat generally drives off this moisture returning the mask to it's
original appearance.
They claim to use a super
high-pressure conveyerized spray cleaner with super-clean water. I'm
skeptical about the cleanliness of their system, and they just told us
that the cleaning line "just broke" so now they can't rerun the
boards.
If this was working when your boards went through, it should have a DI water
final rinse, however, some companies just use tap water as DI is much more
expensive. You may want to ask them about the resistivity of the water in
the final rinse zone.
So, what's your experience? Can a water-soluble flux be reliably
cleaned off to decent leakage levels? Can they really clean under
surface-mount parts?
Large, tight to the board, dense leaded QFP's seem to be the hardest parts
to clean, but they do come clean.
Last time this happened, some years ago with another assembler, we
nabbed a sample of their wash water, and it was 20x as conductive as
tap water.
As long as it's the wash or pre wash water, that could be OK. If it's the
rinse or especially final rinse water, then that's a BIG problem.
I'm thinking in terms of slowly hand-scanning each board with a
water-pic sort of high-pressure wand, with single-use distilled water,
or something like that. It looks tricky to clean under surfmount
parts, especially with water. Our normal process is RMA flux followed
by solvent wash in a vapor degreaser.
Try a hot water wash in a normal dishwasher as mentioned above.