Low build electroless baths tend to initiate slowly.  The alkalinity in the
bath may be attacking the Pd catalyst before the Cu gets a chance to plate.
While that may not always happen, you may have stepped out of your process
window in one or more areas that make a net effect.  Like low Pd deposit
combined with a slightly slower initiation rate.  Or some contaminant level
that just puts things over the edge.

This is an OLD bath.  I do believe Shipley has better chemistry now.  

-----Original Message-----
From: TOE /Torben Østeraa [mailto:[log in to unmask]]
Sent: Tuesday, June 29, 1999 3:26 PM
To: [log in to unmask]
Subject: [TN] Voiding, through plating, conventional


Dear collegues, 
 
I have run into a problem of sporadic ' voiding ' in my through plating 
process. the problem has the following characteristics:
- is observed a few times per month
- affects all materials ( fr4, bt/epoxy, kapton, GI, teflon[s] )
- on kapton-containing builds ( i.e. rigid-flex ) , voiding takes place
predominantly
 on the kapton surfaces. Typically, all the kapton is without coverage.
- I have not (yet ) ben able to establish any correlation with time of day, 
batch maintenance or other obvious influences
- Is independent of type of desmear, plasma or permanganate, even
kombinations
have failed. Therefore,  point of my focus at the pth-line itself.
- The obvious cause would be under catalysation, so an alkaline substitute
for my 
present acid conditioner has been tried. Nice results, but the expected poor
hole wall
adhesions is just not good enough to prevent excessive hole-wall-pullaway
during HASL.
- Supplier is no help ( sorry to say.... ).
 
- Line chemistry is Shipley, Cleaner/cond 815 ( acid) , Etch 3330 (
persulfate/acid stabil. ).
Catalyst 3344 + predip 3340, Accel 19H, copper Cuposit 328 ( 'classic' low
build ), Acid strike 4 microns. 
- Water rinses are all 'fresh' tap-water. ( good quality in our part of the
world )
 
Any advice or hint in this matter is highly welcome
 
regards
 
Torben Østeraa
Printca AS
[log in to unmask] <mailto:[log in to unmask]>