In a message dated 99-07-02 12:35:06 EDT, you write:
<< Please feel free to contact me with any questions or to receive
additional information about CASTIN.
Thank you and best regards,
David Suraski >>
Hello David,
I am interested in the specifics of your reliability testing of CASTIN SMT
assemblies. Can you e-mail the data or send it by regular mail (see address
in signature below).
FYI: The NCMS lead-free solder project had an alloy "F2" of composition:
Sn96Ag2.6Cu0.8Sb0.5.
is that close enough to the nominal composition of CASTIN?
Solder joints of 1206 discretes assembled with standard Sn-Pb performed
better than with alloy F2 under conditions 0 to 100C or -55 to 125C.
However, the trend was somewhat reversed for solder joints of 44 I/O LCCCs,
with F2 solder joints doing better in the 0 to 100C test (but not in the
-55/125 test).
These experiments were carried out carefully and the NCMS report gives
detailed failure mode analysis (FMA), failure cycles and statistics (to order
their report, see their web site which I believe is ncms.org). FMA showed
that, under all conditions, the recorded failures were fatigue / wear-out
failures in the solder itself.
I realize that the NCMS results constitute a small dataset (about 2 component
types, 2 conditions for which the reliability information is rather
thorough), so we cannot draw general conclusions, or develop empirical
acceleration factors, until a wide range of data is available.
I bring this up since we went thru a similar experience for standard Sn-Pb,
with reliability rank-ordering being sometimes inconsistent for different
components tested by different organizations. The reason for it is that
fatigue of solders (or other metals) can only be estimated, at best, by a
factor of 2 (in test as well as in modeling).
So, to get true reliability trends, we need lots of data, covering a wide
range of SM components (BGAs, discretes, QFPs, TSOPs, LCCC, flip-chip, CSP
etc...), different test conditions, and at least three orders of magnitude in
fatigue life. My experience has been that this takes at least 12 to 24 data
points / experiments (50-100 experiments is better but that takes time), with
fatigue lives covering the range 10 to 10,000+ cycles.
Several companies have done it for Sn-Pb, enabling the development of life
prediction models and acceleration factors for extrapolation to field
conditions.
The same should be feasible for SMT assemblies using promising lead-free
alloys, but this requires a large number of data points and experiments.
With best regards,
Jean-Paul
____________________________________________________
Jean-Paul Clech
EPSI Inc., P. O. Box 1522, Montclair, NJ 07042, USA
SMT / BGA / Flip-Chip / CSP Assembly Reliability Specialists
tel: +1 (973)746-3796, fax: +1 (973)655-0815
home page: <A HREF="http://members.aol.com/Epsiinc1/index.html">epsi</A>
URL: http://members.aol.com/Epsiinc1/index.html
(last updated: March '99)
################################################################
Leadfree E-Mail Forum provided as a free service by IPC using LISTSERV 1.8c
################################################################
To subscribe/unsubscribe, send a message to [log in to unmask]
with following text in the body:
To subscribe: SUBSCRIBE Leadfree <your full name>
To unsubscribe: SIGNOFF Leadfree
################################################################
IPCWorks -October 25-28 featuring an International Summit on Lead-Free Electronic
Assemblies.
Please visit IPC's Center for Lead-Free Electronics Assembly
(http://www.ipc.org/html/leadfree.htm ) for additional information.
For technical support contact Gayatri Sardeshpande [log in to unmask] or 847-790-5365.
################################################################
|
