Mark,

Our immediate response is:  You have a problem.
1)  There would always be a concern in allowing vias
to "plate shut".
   a.  Were any plating chemicals trapped within the
hole as it was plated shut?  Probably
These could react with the hole wall since they are 
acidic and acids dissolve plated copper.

   b.  Is the hole completely plated shut or is there a
small orifice which will allow other chemicals and 
cleaning products to be trapped within and react
similarly as plating chemistries?

   c.  Why in the world would  a board mfg need to 
plate 13 mil vias in a 93 mil or 62 mil thick board closed.
It would take 7.5 mils of plating at the knees to close these
holes.  That is not satisfactory plating.  Plating 1 mil of copper
in 13 mil holes in these thicknesses/aspect ratio boards is 
not uncommon.  Plating them shut leads to all sorts of questions
regarding the plating on the rest of the board.
    a.  If the holes are plated shut, it is most likely that there
is very little plating in the centers of the holes.  How thick is it?

    b.  There could be voids (not circumferential) and thin copper 
in the barrels which would pass continuity testing, however could 
crack during assembly. 

   c.  If this is the case - all holes are suspect.  

3.  Opens in any plated through hole can be caused by
    1.  Absence of copper plating  due to voids or cracks.  These 
are both directly related to plating.  A break down of the etch
resist can also lead to voids.

4.  Electrical tests would detect complete circumferential break
in hole wall plating, however it will not detect plating conditions
which degrade as a result of assembly operation. 

5.  You should be concerned about the via and plated through 
component hole integrity of all board supplied since it appears
that you have a degrading product.

How do you proceed?
1.  You must determine the exact failure mechanism.  This is 
going to have to be done with microsections to determine
why the via has failed.  In addition component holes must be 
microsectioned to determine if they have the same condition 
which could degrade during the life of the board.

2.  Did you buy these boards to an IPC spec, or an inhouse spec?
     Did you require any testing of the boards by the board mfg 
     prior to their being shipped to you?

     If so you need to review their microsection data.


Susan Mansilla
ROBISAN LABORATORY
317-353-6249
317-357-1270  fax