|
|
|
|
|
Frequently Asked Questions: Assembly
|
Process Sequences
Q. What is the source of my problem
and how do I solve it?
A. This is, of course, a
very complex question. There are more ways to screw up a process than
you can shake
a stick at, and often several solutions for fixing any problem.
In general, when we
backtrack through a manufacturing process, we look for several
things: Has something changed in the process recently which transforms
good product
to bad? What is the failure mechanism? Electrical leakage, corrosion,
a visible residue, metal migration? In a vast majority of cases,
you can isolate the problem by looking at the materials of construction
and the
materials and processes used in manufacturing.
>top
Cleaning
Q. How do I determine if my cleaning process is adequate?
A. Ion chromatography is a good first step. By examining the
baseline process, we can give a pretty good indication of the
cleanliness of the process. We can then examine either variations
in the cleaning
process to see if the harmful ions are decreased. A good explanation
of the different residue tests can be found in section 8.0 of
IPC J-HDBK-001.
>top
Q. What kinds of cleaning processes are available and which are
dominant?
A. Excluding the niche applications (such as CO2 super-critical
fluid cleaning), cleaning processes can be broken into three
areas in electronics cleaning: solvent, semi-aqueous, and aqueous.
Solvent cleaning
is used primarily where components are water intolerant or have
unsealed components. Solvent cleaning used to be the norm for
military and high-reliability
applications, but is now greatly reduced with the elimination
of Freon and trike. Co-solvent blends, hydrofluoroethers (HFEs)
and hydrofluorocarbons (HFCs) are taking the place of Freon.
Semi-aqueous cleaning is
done primarily
on older military contracts as the semi-aqueous cleaning was
a good transition material to other technologies. Axarel and
Bioact from Petroferm are the
primary materials here. Aqueous cleaning is the more dominant
cleaning today and has many advantages. The greatest cleaning
is saponified-aqueous.
>top
Components
Q. How clean do components have to be?
A. That depends on the components and the applications. A high-impedance
device, which assumes an infinite resistance between leads, is
more sensitive to contaminants and plating residues than will
a low-impedance component.
High voltage and high frequency components are also very sensitive
to contamination. As to what cleanliness levels you need, you
would have to do some correlation testing to determine the needed
cleanliness levels
for the application.
>top
Equipment
Q. What equipment should I be using in assembly?
A. That depends on what you are making and whether you are a
high volume or low-volume/high-mix manufacturer. Different manufacturers
have products with different strengths and weaknesses. Our sales
manager, Mr. Terry Munson, has extensive equipment experience and
can answer your questions.
Give him a call.
>top
Conformal Coating
Q. What IPC specifications should I use to make sure the coating
sticks to my assemblies?
A. The short answer is, there are no such specs. The IPC specifications,
over the past few years, have gone to a system of materials qualification,
using standard substrates and standard tests to give a level
playing field between coating manufacturers. The work of determining
if any particular
conformal coating will stick to any particular assembly or flux
residue is left to you, the assembler.
>top
Q. What resources exist to help me understand conformal coating
issues better?
A. The IPC Conformal Coating subcommittee has a task group under
its domain that is working on putting together a conformal coating
handbook, compiling the collective wisdom of IPC members on this
topic. This handbook
should be available in late 2000 or early 2001, but drafts of
the document can be obtained from the IPC.
>top
Acceptance Specs
Q. What IPC specifications are there that tells me if I have
good hardware?
A. A complex question. Ideally, IPC-A-600F and IPC-A-6012A look
at bare board quality issues. IPC-A-610C and J-STD-001C look
at assembly issues. All are good guidelines for assembly quality,
but no spec is perfect.
Only you the assembler can determine how "bulletproof" your
hardware is to your end-use environment. Unfortunately, IPC has
no specifications at this time on how to do accelerated testing
to determine how reliable
your hardware is for various end-use environments.
>top
Q. Which specification should I be using, A-610 or J-STD-001?
A. This depends a great deal on what your contract administrators
and your company attorneys say. A-610 is a visual guide and is
not a specification per se. It does not have any testing or criteria
with it. The recourses
available to you if your subcontractor does not comply are limited.
J-STD-001, on the other hand is a standard/specification, with
all that implies for
legal purposes. In our experience, J-STD-001 is used by the Class
3 "big
boys" (e.g. defense contractors and other high-rel shops). A-610
is used by the Class 2 and Class 1 assemblers. J-STD-001 is a good specification,
but does not guarantee that the hardware is "good".
>top
Q. I have
several IPC manuals such as IPC 610B, IPC 700B, IPC 6011, IPC
6012, etc. Unfortunately, I have not been able to trace any information
regarding
ionic contamination for completed printed circuit assemblies
with regards to any pass/fail criteria and/or quality standards.
Can you refer me to
any IPC manuals or publications that contains information regarding
these issues?
A. When you talk of pass-fail levels
for ionic contamination, the
ONLY IPC specification is for the standard ROSE test (e.g. Omegameter,
Zero Ion, etc.). It is the historical military limit of 10 micrograms
NaCl equivalent per square inch, adjusted for the instrument
used. Most assembly level committees, recognizing that this level
is valid ONLY for
high solids rosin fluxes, allow for an alternative pass-fail
as agreed upon between vendor and customer. As an example, if
you don't have an
agreed upon value with your customer, the default is 10. If your
customer wants to see 7, then the pass-fail is 7. It is largely
a contractual issue.
The bare board specifications, such as 6011, have a pass-fail
for bare boards of 5 or so, which was a carryover from MIL-P-55110
and MIL-P-28809.
In my view, the pass-fail levels there are antiquated and unreliable
as a protective measure for assemblers.
We would suggest that if you are looking to become better educated
in the area of ionic cleanliness and how it is determined, that
you read Chapter 8 of the J-STD-001 Handbook. You should also
get a copy of EMPF
report RR0013 (An In-Depth Look at Ionic Cleanliness Testing).
You should be able to contact them on line at www.empf.org
The test methods of interest here are found in the IPC test methods
manual IPC-TM-650. The methods are 2.3.25, 2.3.26, 2.3.26.1.
We are working on a new test method for bare boards, 2.3.25.1,
but it is still quite
a ways from acceptance. A more advanced measure of ionic cleanliness
is ion chromatography, which is test method 2.3.28, which we
wrote. Comparing ROSE to IC is like comparing a broad blade ax
to a surgeon scalpel.
All
of our failure analysis and process troubleshooting is done with
ion chromatography.
For assembly level documents, you can look in section 5 of A610B,
or section 7 of A610C (proposal now), or J-STD-001B (section
8) or J-STD-001C (proposal now) to find pass-fail requirements
for ionic cleanliness. Basically,
they take the old military limit of 10 and perpetuate it, unless
you have other arrangements made, as I outlined above.
At Foresite, we don't believe there are any "golden" numbers, or
an ionic cleanliness criteria that can be blindly applied to
all technologies and all flux types. From our own failure analysis
work, we have recommended
limits based on the metallization type (e.g. HASL vs. gold), based
on flux (water soluble vs. no-clean) or application (hybrid vs.
PWB).
Foresite does not place much confidence in ROSE tests as a guarantor
of quality. We have had dozens of customers who had adequate
numbers in their test instruments, only to have corrosion and
metal migration in
the field. The RR0013 report can help you understand why this
can be. Our Process Rx column in Circuits Assembly magazine also
covers this.
>top
Q. Why does IPC-TM-650, method 2.6.3.3A
specify FR-4 and bare copper as the substrate? Wouldn’t HASL
metallization
be better?
A.
This test method is controlled by the J-STD-004 specification
on fluxes. This specification uses a "level playing field" approach, and
so specifies FR-4 as the base laminate and bare copper as the
metallization.
HASL, or some other coating, might be more representative of the bare
boards you get, but then you have the complicating factors of HASL and
post-HASL cleanliness complicating the materials characterization testing
of J-STD-004.
>top
Q. What’s the difference between
Bellcore NWT-TR-000078, issue 3, and Bellcore GR-78, from the standpoint
of materials
testing?
A. Not a great deal. The SIR test remains essentially
the same with the same
sample sizes and test patterns. The SIR test humidity level went
from 90% minimum to 85% minimum to allow greater control. The
minimum levels
required to be considered as compliant remained the same.
The Resistance to Electromigration test is essentially the same,
but the test environment went from 596 hours at 85oC / 85% relative
humidity to 596 hours at 65oC / 85% minimum relative humidity.
The requirement of no more than a one-decade drop and no corrosion
or electromigration
remained the same.
>top
Rework / Repair
Q. Are there any rework or repair operations that I should stay
away from?
A. First, determine what solder defects are cosmetic and what
are true defects. More damage is done touching up solder defects
that don’t compromise reliability than any other action. Second, never
use the addition of liquid flux of any kind as a touch-up soldering aid.
This causes far more problems than it helps, especially for low solids
fluxes. Third, localized cleaning with isopropanol is not recommended.
In essence, this only spreads the contaminant out on the board, "infecting" a
larger area. Fourth, do not mix cored wire solders in the work
area. Keep workstations as water soluble cores or no-clean cores
only. Mixing them
will assure that at some point you will put the water soluble
on the no-clean assemblies.
>top
Q. How should I strip conformal coating for field repairs?
A. That depends on the coatings. Acrylics are fairly easy to
remove locally with a variety of materials. Most coating manufacturers
have a coating stripper formulation for acrylic. The next easy
coating to strip is urethane, followed by silicone, paralene,
and epoxy. If you
only have to repair a solder joint or two, burn through the coating
with the solder mask and do a spot recoat afterwards. If you
have to remove a single component, then strip the coating in
that area. Placing
an RTV
Silicone dam around the area allows you to selectively strip.
The IPC Conformal Coating Handbook and IPC-7011 and IPC-7021
should be consulted
for more information.
>top
|
|
|
