Well, as Brother Brian copied me on his response about no-cleans and
conformal coat, I took it as a subtle hint to throw my two
cents/shillings/micro-Euros into the mix.

I am of the general opinion that you can make almost any material set do
what you want with sufficient time and attention to the selection of the
materials and the associated process parameters.  I know of companies out
there who successfully use no-clean assembly processes and conformally coat
over the residues.  Delco Electronics is one such company.  They make more
assemblies in a month than Rockwell makes in a year.  Many of their product
lines are no-clean and they coat over them.  And the under-the-hood
environment of a car is a pretty nasty environment with lots of thermal
cycles, humidity, and salt environment (if you are from the north).  I know
that they put a tremendous amount of time into the selection of the flux
and the selection of the process parameters, such that conformal coat would
adhere to whatever they left on the board.  There are other companies as
well that manufacturer to Class 3 high-reliability standards that use
no-clean and conformally coat.  And all of them spent a lot of money tuning
the process.  As Brian said, the person who tries to combine these two
without diligent process control and process qualification is
irresponsibile.  I would agree.

Keep in mind what a low solids flux residue represents.  The amount of
residue depends on how much you apply and the thermal reflow profile you
use.  If you throw huge amounts of 5% flux on an assembly with lots of
entrapment sites, then expect to have lots of residue after soldering.  If
you lightly spray on a 1% flux, using no more than you absolutely must
have, and you tune your reflow accordingly, then most or all may flash off
in reflow.  Secondly, the flux residues after soldering will have different
characteristics, depending on how you processed the flux.  A low solids
flux (LSF) must reach some minimum activation temperature before the
residues fully react and reach the "benign" stage.  If the flux is not
fully reacted, then portions of the high molecular weight high boiling
point solvents remain, making the residues mildly conductive.  This may be
why Lorraine Reid saw arcing in a high voltage application.  The flux
residues may not have been fully reacted, and so did not have the same
dielectric strength as a fully reacted residue.  I have seen this failure
mechanism numerous times where the assembler did not fully realize that
throwing more flux at the assembly was not a wise thing to do.  One
critical mistake that many assemblers, new to the field of no clean make,
is that they do not understand how a LSF works.  With a water soluble flux,
you have a chemical etching that goes on from the point of application to
the point of reflow.  For a LSF, you do not get fluxing action until you
drive off the carrier solvent and the crystals of the weak organic acids
(usually one or more dicarboxylic acids) melt, which then provide the oxide
stripping mechanism.  So OA flux experience tells an operator that if you
have soldering problems, throw more flux at the assembly, but with a LSF,
if the board reaches the wave wet, you get no oxide removal, so the
solution is to throw less flux at the problem.  For the same reasons, you
don't want to squirt liquid LSF on a board as part of hand soldering.  The
flux goes a lot farther than the heat from the soldering iron reaches.
That isopropanol carrier takes that flux places you would not believe.

Keep in mind that flux manufacturers can work to tailor the properties of
the residual flux.  One example that I know of is Indium SMQ-92 and
SMQ-92J.  The residues in the 92J are harder and glossier than those of the
92.  Indium did that to make the 92 easier to penetrate in probe testing.
I have found it useful to look at the flux residues as if it were a
material that I design into the board, rather than a process effect.  If I
have a great big honkin' electrolytic capacitor, I know that I have
laminate between the through-holes, and solder mask between the through
holes.  I would evaluate both the laminate and the solder mask to determine
if they had sufficient dielectric strength for the application.  Now, add
to that a layer of flux residue remaining after processing.  Does that flux
residue have adequate dielectric strength for the application.  Then, if it
does in the fully reacted state, what kinds of processing controls do I
need to have in place in order to make sure that the flux residue reaches
that dielectric state during processing?  This is where I see many no-clean
assemblers failing.  They fail to take into account the residues as a
dielectric material, affecting dielectic strength, cross talk, impedance,
rise time and propogation in RF, etc., or they fail to make their assembly
process robust enough against these factors.

As Brian pointed out, many of the LSF residues are hydroscopic in nature,
which has to be remembered when considering conformal coating.  If you have
a case where you have unreacted flux residues, or a board that has an
unacceptable level of ionic contamination on it, conformal coating will NOT
help you.

I had a number of clients when I was at CSL that had problems identified on
their boards.  Maybe it was the unreacted flux, maybe dirty bare boards,
maybe dirty assemblies.  Some did not want to spend the time or money to
attack the true cause of the problem, which were the ionic residues.  Their
solution was to slap a layer of conformal coating over the problem and ship
product.  Of course, choosing the cheapest solution, the acrylic coatings,
which have the higher water permeability of them all.  If you remember the
old commercial with the auto mechanic "You can pay me now or pay me later",
that was what happened.  Nine months to a year later, they came back saying
"Uhhhhhhh, it didn't work".    Conformal coating only buys you time since
moisture WILL go through the coating.

You can't really fault the coating manufacturer either.  Assemblers often
expect the coating manufacturers to make a coating that sticks to any
surface, no matter what the assemblers have done to the board, and for low
cost.  Most components have plastic release agents build in, most solder
masks have smooth surfaces that may be difficult to adhere to, and now you
throw in residues with a highly variable surface tension, and you have an
adhesion nightmare.  You have this nighmare in fully cleaned assemblies
too, so don't be too quick to blame the coating for not sticking to your
residues.

If you have a choice in conformal coating, I would recommend a urethane, I
would recommend it over an acrylic.  The acrylic is simply a dried film
over the surface, with limited chemical interaction to the substrate in
terms of bonding.  In addition, you might also have the solvents in the
coating soften and attack the LSF residues, changing the properties (keep
that in mind).  The urethanes or epoxies tend to chemically react better
with the substrates, forming a better bond, but they are also more
difficult to rework.  If your manufacturing process is mature and
controlled, the tech staff of the coating manufacturer should be able to
help you choose a coating for good adhesion.  On the other hand, if your
process is not controlled, you will probably get the generic "Well, the
assembly has to be clean before you coat" response.

So, if you are working with a no-clean assembly processes, you should be
doing the following:
1.  Working to understand the dielectric properties of the residues
2.  Determine what your processing windows need to be in order to make sure
the flux residues are fully reacted
3.  Minimize the amount of flux you are applying
4.  Work to use a lower percent solids in the process to leave fewer
residues on the board
5.  Select your flux for desirable properties for conformal coat adhesion.
6.  Tune your process to get the residue properties you want (in
conjunction with #2), such as a matte finish rather than a hard glossy
7.  Select your conformal coating so that the solvent system is compatible
with the flux residues and so that the coating sticks to the residues.

As this message is nearing "the Ellis Zone", I will wrap up by saying that
you CAN manufacture hi-rel equipment with no-clean, you CAN make it
reliable for long life, you CAN make coating stick to it for the design
life, IF IF IF you have done your homework and know what you are doing.

Doug Pauls
Rockwell Collins

(And all done without Mountain Dew, which will be immediately remedied)

---------------------------------------------------------------------------------
Technet Mail List provided as a free service by IPC using LISTSERV 1.8d
To unsubscribe, send a message to [log in to unmask] with following text in
the BODY (NOT the subject field): SIGNOFF Technet
To temporarily halt delivery of Technet send e-mail to [log in to unmask]: SET Technet NOMAIL
To receive ONE mailing per day of all the posts: send e-mail to [log in to unmask]: SET Technet Digest
Search previous postings at: www.ipc.org > On-Line Resources & Databases > E-mail Archives
Please visit IPC web site http://www.ipc.org/html/forum.htm for additional
information, or contact Keach Sasamori at [log in to unmask] or 847-509-9700 ext.5315
---------------------------------------------------------------------------------