>Date: 04 Nov 1996 09:08:21 -0600
>From: "Millsap, Pete (AZ77)" <[log in to unmask]>
>To: Technet <[log in to unmask]> (Return requested)
>Subject: Unidentified subject!
>
>The Navy did a surface mount repair study in June 1987.    The conclusion of
>this study is that the "long term reliability of a solder joint will be
>jeopardized with multiple rework cycles. With copper dissolving away during
>thermal processing, the pads will inherently lose their structural
>compliancy.  The copper goes into solution forming brittle compounds such as
>Cu6Sn6.  These brittle intermetallic layers may cause long-term solder joint
>fatiguing problems."  They state that on 2 oz copper  100% of the pad would
>dissolve in 12 cycles.
>
> My question does anybody know of any recent studies that would support or
>counter this study.  Do other companies limit the number of heat cycle on a
>pad?
>
>From the EMPF HelpLine in Indianapolis; prepared by Jack Crawford

I was the project manager for the study you refer to while working for the
Navy base in Keyport, WA. The specific conclusion on the copper solution
and the itnermetallic was based on a dissolution formulas and other
engineering documents, and not by test results.

Subsequent to that study, a more intense evaluation was completed in 1992
that involved more actual testing. That was done by the Navy Surface
Warfare Center, Crane, again coordinated by myself.  We have copies of
those test results in the EMPF Technical Library, and would be glad to
share them.

In summary, the test utilized a matrix of 3 board types (6 layer FR4, 6
layer ceramic copper thick film, and a 4 layer copper-invar-copper board)
with all ceramic components of leadless, J-lead, and gull-lead
configurations; 20 and 68 I/O 50 mil pitch (hey--it WAS 1991-92). The
boards were a SEM "D" size, approx 4.6". Some of the large and small
components were removed and replaced one time, some three times, some not
at all.  We used large and small components both at the edge and center of
the boards. Some components were instrumented with internal thermocouples
to monitor temperatures. We used nitrogen on five of the "state of the art"
hot gas systems available at that time, and one IR. A sample of the boards
were subjected to power cycling and magnified visual inspections.
Additionally, some interconnects were microsectioned.

Relative to the info you are asking about, we controlled the heat cycles
(temp, duration) as tightly as we could. Heat cycles were: (1) HASL (1)
Attach Components (2) single remove/replace and (6) triple remove/replace.
We utilized residual solder to eliminate "partial, uncontrolled heat
cycles" to wick or extract solder from the pads.  We also wanted to see if
the solder could or shouldn't be reused.

The findings were that with controlled cycles at minimal temperatures,
intermetallic growth was minimal, and there was no reason to suspect or
predict solder joint fractures with as many as 8 reflows. Additionally, the
pad thickness didn't seem to decrease at the rate predicted in the first
study--perhaps the intermetallic actually acted as a barrier to further
dissolution.

The study is dated now, because of changes in flux and solder formulations,
new heating control and application methodology, and different kinds of
pad/lead finishes now available, but there is some core info that could
very well be beneficial to process development.  Anyone is free to discuss
this further with me at the phone numbers/e-mail numbers in my signature
file which follows this msg.  Jack

***ALL NEW EMPF PHONE NUMBERS***

                      Jack Crawford
                  HelpLine Manager
    Electronics Mfg. Productivity Facility
****NEW--317.655.3688--NEW****
*****FAX  317-655-3699 NEW ****
          714 N Senate Ave, Suite 100
         Indianapolis IN  46202-3112
          VISIT OUR HOME PAGE AT:
                http://www.empf.org
            [log in to unmask]


***************************************************************************
* TechNet mail list is provided as a service by IPC using SmartList v3.05 *
***************************************************************************
* To unsubscribe from this list at any time, send a message to:           *
* [log in to unmask] with <subject: unsubscribe> and no text.        *
***************************************************************************