It appears that we may be looking at two separate problems here.

Problem #1:  Some of the component pads seem to have partial solderability.
This is like the picture I posted in my previous message.  In other words, a
portion of the component pad is still wetted with solder after removal from
the board.  These pads retain the initial texture and grain of the original
pads.  They range from light grey (non-wetting) to shinny solder (dewetting)
appearance.  To me this indicates that the ball was attached to only a
portion of the pad in component assembly.  If this is the case then the ball
would likely fail a ball shear test.  Since only a small sample is tested
for ball shear it would be expected to have some escapees.

Problem #2:  Some of the component pads seem to have no solder attached.
These pads are dark grey and exhibit needle like crystals which we believe
may be nickel-tin intermetalics.  We will perform some Auger analysis over
at Accurel Systems to see if this is the case. This seems to be the more
prevalent of the two problems.  One theory that we are pursuing is that the
balls failed after component assembly and before board assembly.  We are
investigating the test and burn-in processes to see if the strain on the
ball during elevated temperature testing could be causing a solder creep
failure at the ball to component interface.  So far we have seen no
significant amount of gold at the interface so we have eliminated gold
embrittlement as a possible cause.  Does anyone have any data on the shear
strength of (or crack propagation in) nickel-tin intermetalics?

Can anyone supply me with a few BGA components which they received with
missing balls?  It would be very interesting to see they look compared to
the board pulls described above.  If e-test or burn-in is causing this
problem then it is likely that they have some solder balls left behind in
their test sockets.

-Russ Winslow

-----Original Message-----
From: TechNet [mailto:[log in to unmask]] On Behalf Of Russ Winslow
Sent: Tuesday, February 09, 1999 4:14 PM
To: [log in to unmask]
Subject: Re: [TN] Defective BGAs?


Well the preliminary XRF results are in.  The grey BGA pads are definitely
nickel.  The nickel measures 148 to 168 microinches in thickness.  When we
measured the center pads (never soldered) the gold thickness measured 24
microinches.  We were about to do much more testing when the power went out
for a few seconds, wiped out all our data, then came back on and wiped out
the 24KV power supply in the XRF machine.  This was a very expensive XRF
test.

So the nickel appears thick enough and the gold appears thick enough to
protect it through most BGA component assembly processes like wire bond,
mold, and post mold cure.

Any ideas?


Regards, Russ Winslow

Six Sigma

################################################################
TechNet 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 TechNet <your full name>
To unsubscribe:   SIGNOFF TechNet 
################################################################
Please visit IPC's web site (http://www.ipc.org) "On-Line Services" section for additional information.
For technical support contact Hugo Scaramuzza at [log in to unmask] or 847-509-9700 ext.312
################################################################