You can sometimes tell whether or not a BGA has been reworked by using a
microsection, but not usually, for the following reasons.
You can more reliably determine by visual examination using an
endoscope, but material evidence is required. Evidence of rework flux
residues and full wetting down the sides of non-soldermask defined pads
on the suspect part that are not evident on the rest of the parts are
two tipoffs. Looking at the ball diameter of the part will also
sometimes indicate it has been reworked.
As part of any BGA rework process, the component is usually removed and
replaced, but not always. If it is removed, the old solder needs to be
removed as part of the site preparation for the new part, or at least
"smoothed over" where most of the remaining old solder is removed and
only a very low dome of solder is left behind on the pads. This
effectively pre-tins the pads.
Once the site prep is completed, there are two methods for putting on
the new part to choose from; either apply solder paste using a
mini-stencil, or use a tacky flux and place the part and reflow using
only the solder already present in the form of the solder balls on the
new part.
If the rework operater did remove all of the old solder, applied tacky
flux, and put down the new part, then the diameter of the solder balls
will be slightly smaller on average than the same part on the rest of
the boards. This because no additional solder was added, and this can be
detected in a good x-ray system that has BGA analysis software.
However, if the operator printed solder paste, or left behind the small
domes as part of the site prep process, the final ball diameter could
end up being the same. Typically, during initial assembly, solder paste
is stenciled on for the BGA along with the rest of the SMT parts.
Sometimes, if the failing part is suspected to be electrically OK but
the solder joints are suspected to be bad (confirmed by pushing down
with slight pressure during test), a rework process consisting of a
simple reflow without removal will sometimes be attempted. In this case,
the operator simply refluxes the solder balls, blows away the excess
flux, and then subjects the BGA to the hot-gas rework cycle without
actually removing or replacing the component.
Because different board surface finishes produce variability in wetting,
with a larger or thicker IMF formation on some pads than others, it can
be very difficult to verify the part has been reworked through
microsectioning alone. One cannot tell for sure based on the grain
structure either, as this can vary amongst balls of any given BGA,
caused by variations in the thermal mass (differing rates of ramp to
liquidus and different rates of cooldown. These variations in the
thermal mass are caused by some pads being connected to larger power and
ground planes, and the nozzle turbulence on the rework machine, amongst
other things.
This is especially true of ENIG and ENPIG finishes that have longer
dissolution times, less variable with immersion silver, immersion tin,
or OSP finishes.
If the surface finish is HASL, then the variation can be all over the
place, as the board vendor creates the first IMF during the HASL
plating. HASL is not always a robust process, and oftentimes the pwb
vendor may subject the board to the HASL process several times, forming
an IMF before the assembler even applies paste and reflows the first
time.
So, there are many variables that control or contribute to the IMF and
the solder alloy structure. Time above liquidus, time at peak, and peak
temperature are only 3 of at least 16 variables that I can think of that
could change both the alloy structure characteristics and the IMF
characteristics.
That is why visual detection of evidence is the best method to determine
if rework has been done. If no evidence was left behind, then it is very
difficult to tell.
But absence of evidence is not evidence of absence.
-----Original Message-----
From: Leadfree [mailto:[log in to unmask]] On Behalf Of Mike Taylor
Sent: Monday, April 17, 2006 9:44 AM
To: [log in to unmask]
Subject: Re: [LF] Tin Pest + Aging Performance--from Jennie Hwang
Hello Werner,
I posted a peripheral question about determining if a peak temperature
could be detected by looking at the microstructure of solder joints.
The answer I am getting is that intermetallic growth is both time and
temperature dependent. Thank you to Ryan Grant and others. I had hoped
to find that there was a way to show that a joint saw much higher
temperature than expected in processing, but that is the board bias. We
also practice assembly operations. In that case, show that the boards
did not see much higher temperature than expected in processing.
I fielded the question from an OEM that procures boards and assembly
services from different contractors. From a practical standpoint we
generated several questions.
1) Can you look at a board that has failed in a lead free process and
see if the board saw the same thermal experience as other boards in the
lot? By thermal analysis? By metallurgical examination?
2) Within a board, could you look at a BGA for example, near a failure,
and see if it saw more thermal experience (rework) than the surrounding
board?
The answer to all the questions is probably "yes", but do you know if an
acceptable procedure has been published that we could reference in
support of an analysis? The answers help direct where to go to find a
solution to prevent a future failure.
-----Original Message-----
From: Leadfree [mailto:[log in to unmask]]On Behalf Of Werner Engelmaier
Sent: Thursday, April 13, 2006 7:46 PM
To: [log in to unmask]
Subject: Re: [LF] Tin Pest + Aging Performance--from Jennie Hwang
Hi Ryan,
And what is the point of this exercise?
Werner
------------------------------------------------------------------------
----
---Leadfee Mail List provided as a 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 Leadfree To temporarily
stop/(start) delivery of Leadree for vacation breaks send:
SET Leadfree NOMAIL/(MAIL)
Search previous postings at: http://listserv.ipc.org/archives Please
visit IPC web site http://www.ipc.org/contentpage.asp?Pageid=4.3.16
for additional information, or contact Keach Sasamori at [log in to unmask]
or 847-615-7100 ext.2815
------------------------------------------------------------------------
----
---
------------------------------------------------------------------------
-------Leadfee Mail List provided as a 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 Leadfree To
temporarily stop/(start) delivery of Leadree for vacation breaks send:
SET Leadfree NOMAIL/(MAIL) Search previous postings at:
http://listserv.ipc.org/archives Please visit IPC web site
http://www.ipc.org/contentpage.asp?Pageid=4.3.16 for additional
information, or contact Keach Sasamori at [log in to unmask] or 847-615-7100
ext.2815
------------------------------------------------------------------------
-------
-------------------------------------------------------------------------------Leadfee Mail List provided as a 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 Leadfree
To temporarily stop/(start) delivery of Leadree for vacation breaks send: SET Leadfree NOMAIL/(MAIL)
Search previous postings at: http://listserv.ipc.org/archives
Please visit IPC web site http://www.ipc.org/contentpage.asp?Pageid=4.3.16 for additional information, or contact Keach Sasamori at [log in to unmask] or 847-615-7100 ext.2815
-------------------------------------------------------------------------------
|
