It is really application and chip package specific, and even makes a big
difference on where the chip is mounted on the board. It also makes a big
difference on the various lead finishes, process steps and how they are
done, and which of the common 23 lead-free mixes you use. Would you believe
ENIG finishes can even cause issues with lead-free?

Personally, if you look there is a consortium lead by Boeing, of areospace
industries, who have been doing a lot of testing on the subject, and ended
up settling on just 1 or 2 alloys that perform the best in the early phases,
to pick out which ones to run thru the gauntlet- with control subject of
regular leaded solder.

If you do searches on key documents/terms listed within these, you will find
more than you ever wanted to know:
http://www.aciusa.org/leadfree/LFS_S...P_WG_Brief.pdf
http://www.calce.umd.edu/lead-free/SMTAExemptMay8.pdf
http://www.jgpp.com/projects/lead_fr...entations.html - a
whole list of documents
http://www.jgpp.com/projects/lead_fr...M-0409-991.pdf
http://www.jgpp.com/projects/lead_fr...ll%20Voids.pdf
http://www.jgpp.com/projects/lead_fr...404Woodrow.pdf
http://www.jgpp.com/projects/lead_fr...up_3_16_05.pdf

Another key term to search for is Thomas Woodrow, who has dedicated years
attempting to unravel this puzzle for best practices.

When looking at these presentations, it is important to look hard into the
nitty gritty of the source references, or you will easily end up with the
wrong conclusions. Things like a good number of the tests were done with
immersion silver, and ENIG (gold) and other board finishes can have a big
impact on the results of the test. ENIG has it's own issues, such as a big
increase in what is becomming known as "black pad" failures when lead-free
solders are used with it. As I recall, it had something to do with the ENIG
process steps, which vary from board house to board house.

In the second one, several working groups have found the failure shown in
page 33 of the document( pdf page 33), where you can see cracks running all
through the BGA ball, and a rather clear separation off the pad. Caused by
temperature cycling, and failures in less than 150 thermal cycles!

Tin Whiskers Theory and Mitigation Practices Guideline:
http://www.jedec.org/DOWNLOAD/search/JP002.pdf

"The amount of damage required for
a BGA to fail on an assembly was used to predict lifetimes for other BGAs,
at other locations, on the same board. They tested: Sn3.9Ag0.6 for reflow
soldering, Sn3.4Ag1.0Cu3.3Bi for reflow, Sn0.7Cu0.05Ni for wave, and
63Sn37Pb for reflow and wave. The surprising part is that BGAs using
tin/lead will outlast SAC BGAs by a factor of 20x. Thus, SAC BGAs in
high-reliability electronics could be problematic in high-vibration
environments.
....
There has been no other transition to affect all aspects of our industry
as fundamentally as lead-free. Predicting reliability will prevent future
disasters, and that's the best reason to glean information from all
approaches. There's still so much we need to know."
http://listserv.ipc.org/scripts/wa.e...eadfree&P=1444


----- Original Message -----
From: "Eeyore"
Newsgroups: sci.electronics.design,sci.electronics.repair
Sent: Tuesday, July 24, 2007 3:42 PM
Subject: So what's the truth about lead-free solder ?


The debate about lead free solders seem to be nearly as politically
charged as
that about anthropogenic global warming and a casualty seems to be useful
data.

I've read plenty of comments to the effect that lead-free is less reliable
in
the long term (vibration seems to be a key weakness AIUI - maybe also
thermal
cycling) which presumably explains the exemptions for certain categories,
yet
I've also seen some studies that claim it can out-perform lead containing
solders.

Is there any real hard and fast information out there that one can rely on
?

Graham


"Eeyore" wrote in message
...
The debate about lead free solders seem to be nearly as politically
charged as
that about anthropogenic global warming and a casualty seems to be useful
data.

I've read plenty of comments to the effect that lead-free is less reliable
in
the long term (vibration seems to be a key weakness AIUI - maybe also
thermal
cycling) which presumably explains the exemptions for certain categories,
yet
I've also seen some studies that claim it can out-perform lead containing
solders.

Is there any real hard and fast information out there that one can rely on
?

Graham