just for your information, there are 2 phenomenons mentioned in the literature:
head in pillow and ball in socket.
For head in pillow you can see photos in the following articles, and the shape of the defectis similar to what Ioan has sent.

http://www.aecouncil.com/Workshop/7A.2Varnau-DelphiES.pdf
http://www.dage.de/download/semi/papers/xray_inspection_for_electronic_balled_devices_kbryant_2005.pdf
http://www.alphametals.com/techservices/pdfs/Case1.pdf

For ball in socket the story is not the same and you can see the picture in the following article:
http://smt.pennnet.com/Articles/Article_Display.cfm?Section=ARTCL&ARTICLE_ID=260263&VERSION_NUM=2&p=35
By the I have seen both  By X-ray and subsequent cross section. Head in pillow is much easier to see by X-ray because you can see the convex shape against another convex shape and the ball is smaller in diameter.
Just my 2 p.
Gaby
  ----- Original Message ----- 
  From: [log in to unmask] 
  To: [log in to unmask] ; [log in to unmask] 
  Sent: Thursday, November 16, 2006 7:29 PM
  Subject: Re: [TN] Separation in BGA joints


  Hi Gaby

    http://alphametals.com/products/solder_solutions/pdf/pillowhead.pd


  This report is interesting froma number of perspectives.
  The report states as the "Failure Mechanism:"
  It appears most reasonable that the opens have actually been created during
  the second soldering operation. This can be easily confirmed by comparing electrical
  test results on some BGA’s before and after the second soldering operation. During this
  second soldering operation, warpage has been experienced by the BGA package due
  to initial package conditions (amount and direction of ‘set’ and moisture content), and
  the thermal conditions of the soldering process (preheat and reflow temperatures and
  times).

  The extensive thermal environment during the second soldering operation has
  resulted in the liquification of the previously reflowed connections, and along with the
  extreme warpage in certain areas, particularly around the BGA package periphery, has
  extended the now molten solder connections beyond their cohesive limits. The
  separations thereby created, along with the lack of active flux at these locations, have
  resulted in a skin of oxide on the two surfaces resulting from the separation. Upon
  cooling, reduction of the warpage condition, and depending upon the exact timing of
  solidification, these two separate solder globules have made contact and have resulted
  in the various conditions observed during the microscopic investigation of their
  microstructures.

  The observed separations have occurred where the generated oxide films on
  the two solder surfaces have not fractured upon reinitiation of contact and allowing
  them to rejoin into one continuous globule. Instead, the significant oxide film has
  created both a physical and electrical barrier as detected during ICT.

  The observed deformation of the larger globule on the component side, is the
  result of the slightly faster solidification of the small solder globule left on the
  motherboard attachment site. Upon subsequent contact of the ‘oxide enveloped’, and
  ‘still molten’ larger globule, deformation and deflection of this larger globule takes
  place as it is pressed against the now solid mass on motherboard attachment site.

  Where the oxide layers have broken, the two globules have merged to varying
  degrees again, forming a number of peculiar shapes as observed during microscopic
  examination of the microstructures and as illustrated also in the attached
  photomicrographs.

  Unfortunately, this is pure technical nonsense and physically impossible.
  There is no way that a second soldering operation  can cause the observed problem—separating a properly formed solder joint into 2 globules, with the same volumes as the initial BGA balls and the initial solder paste volume, no less.
  If solder joints are liquid, component warpage can not separate them, because they will 'stretch' because of surface tension to more than double their thickness without separation.
  This shows that even very reputable companies, like Alpha Metals, cannot control what some of their employees publish in their name, anonymous, no less.

  Werner

---------------------------------------------------
Technet Mail List provided as a service by IPC using LISTSERV 1.8e
To unsubscribe, send a message to [log in to unmask] with following text in
the BODY (NOT the subject field): SIGNOFF Technet
To temporarily halt or (re-start) delivery of Technet send e-mail to [log in to unmask]: SET Technet NOMAIL or (MAIL)
To receive ONE mailing per day of all the posts: send e-mail to [log in to unmask]: SET Technet Digest
Search the archives of previous posts 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
-----------------------------------------------------