> ======== Original Message ========
> I work in a material and failure analysis lab for Lockheed Martin.
>         We have a set of PCBs that are failing after hand soldering
> operations.
> Similar hand soldering operations on other boards are successful. Cross
> sections of the failed areas shows pad lifting and a seperation of the
> barrel and innerconnection on layer 5. The board is a 6 layer glass
> epoxy GFG board. Suspect problem with board material between layers 5 &
> 6.
>         Looking for information on types of tests to run, printed material
> relating to testing PCBs, or other sources for information.
>         Thank You.
>
>         Russ Salvo
> ======== Fwd by: Richard MacCu ========
> Is there only one operator doing the soldering or many?
>
> If there are many, check all iron temps and specifically check the operator
> who incurred the failure.  I have experienced a case where one employee was
> not following correct procedure even after seven years of service.
>
> The other item to check is if there were mixed lots of laminate involved.
> If so have the vendor of the laminate retest a sample for cure of the resin.
>
Additional comments:

In studying the root causes of ICD's one of the most
effective test methods to create innerconnect separation
was to hand solder.   Because of the high temperature on the
surface and the quick cool down the Z expansion was
substantial.  Failure analysis should be done to see if the
separation is at the plated copper to the innerlayer or
in the innerlayer foil cracking.  I have always felt the inner-
connect should be the strongest bond in the holewall.  If
it is separating you can reduce this by using higher Tg
materials, or baking before thermal shock, or by managing
the thermal cycle (a very good idea) to lower the chance
of ICD's.   But the bottom line is you have a weak
innerconnect.
Joe Dickson
Dir. of Engineering
Sigma Circuits