Perhaps this is a silly question, but have you looked at any cross sections
in the "as recieved" state? Drilling debris, poor neutralization of desmear
chemistries, and sometimes phenomena related to "pink ring" can sometimes
cause a very small and at times circumferential void. When this plates up
with copper, the copper plates over the top to encapsulate this void
yielding what some would refer to as a "plating fold" (looks like an almost
imperceptible seam and slightly scalloped contour in the otherwise
continuous plating in the hole). A feature such as this will not survive
thermal cycling well and if the only time you've "caught" it is after
thermal stress, it will come apart and  look remarkably like ductility
failure. This thought comes to mind because of your description of the
location. How are you desmearing or etching back?

Hope I didn't muddy the waters too much.

Bob Dube
ES&D

-----Original Message-----
From: Werner Engelmaier <[log in to unmask]>
To: [log in to unmask] <[log in to unmask]>
Date: Friday, August 25, 2000 4:01 PM
Subject: Re: [TN] Copper cracks on polyimide boards


>Hi Tim,
>>Werner,
>>I was a bit confused by your explanation of why copper barrel cracks occur
>>where they do.
>>First, the center of a construction is not necessarily B-stage. If I build
a
>>four layer board with foil construction the center is C-stage and the
outer
>>layers are prepreg.
>The cracks will typically occur at the layer interface region (regions if
the
>center layer is a laminate) between the laminate and prepreg layers closest
>to the PCB center. So in a four layer board with  the center C-stage you
are
>likely to get cracks at both the interfaces of the center laminate layer
with
>the adjacent prepreg layers.
>>Second, why do you say that prepreg always has higher
>>resin content than C-stage? What if I use only 7628 prepreg? I know, on
>>average your statement is probably true, but you said "always". Couldn't
>>that lead someone to a bad conclusion?
>Perhaps always was to much, 'typically' (and in all cases I have seen) may
>have been better since there are ways to make it otherwise. Any bad
>conclusions are not serious since if you got cracks anywhere, they have to
be
>dealt with.
>Werner Engelmaier
>
>##############################################################
>TechNet Mail List 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 web site (http://www.ipc.org/html/forum.htm) for
additional
>information.
>If you need assistance - contact Keach Sasamori at [log in to unmask] or
>847-509-9700 ext.5315
>##############################################################
>

##############################################################
TechNet Mail List 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 web site (http://www.ipc.org/html/forum.htm) for additional
information.
If you need assistance - contact Keach Sasamori at [log in to unmask] or
847-509-9700 ext.5315
##############################################################