Bob,

                My direction and findings in DS are in complete agreement
        with what you've said. I've established a fairly sizable extensive
        empirical database that I used to establish scale factors for front
        end plotting. Just to add more twist, vendor A vs. vendor B adds
        further variation within the stack. I believe this to be related to
        resin source (?). 

        [log in to unmask]
______________________________ Reply Separator _________________________________
Subject: dimensional stability
Author:  [log in to unmask] at SMTPLINK-HADCO
Date:    12/13/95 12:15 PM


To: [log in to unmask]
In-Reply-To: your message  of Tue Dec 05 08:32:58 EST 95 
Content-Type: Text
     
I definitely agree that the current IPC test has no correlation with 
re-lamination results.
     
In my opinion there are two issues: amount of shrinkage and consistency of 
shrinkage.  Many factors effect both: lamination pressure, thermal expansion 
(TCE) mismatches, residual strain, Cu pattern effects, etc.  Consequently 
there is no simple test that predicts shrinkage!  I gave 2 papers (published 
in the proceedings) at the last IPC meeting on the effect of C-stage 
parameters on re-lamination shrinkage.  In our experiments we looked at total 
shrinkage and shrinkage variation in actual MLB boards. Independent variables 
included C-stage resin content, C-stage lamination conditions, C-stage 
stress relaxation, glass style, and B-stage cure level.  The bottom line is 
that resin content is the major factor.  I believe this is because resin 
content has a first order effect on TCE, and lamination shrinkage is mostly 
driven by the TCE mis-match between the innerlayer and the finished board.  
     
This means that looking at the C-stage properties is only half the story.  We 
also have to look at the B-stage, the pattern, and the other C-stage layers 
used in the MLB.  In other words, there is no single measurement that can be 
made on a laminate to predict re-lamination shrinkage.  The only possibilities 
for predicting re-lamination behavior is to have a physical model that 
includes all important parameters in the MLB or to use empirical results based 
on actual measurements on similar designs.  We are now using the latter 
approach, but hope to move to the former as our understanding improves.  I 
plan to give a paper at the Spring IPC meeting on this topic.
     
Bob Holmes   AT&T  Bell Labs
[log in to unmask]

------------- Begin Original Message -------------
     
Subject: Re: LAMINATE DIMENSIONAL TESTING 
Resent-Date: Tue, 05 Dec 95 08:32:58 EST
     
     
In my experience the current IPC test used by laminators to check dimensional 
stability has no correlation to dimensional stability results after multilayer 
lamination at a PCB shop.  Is IPC investigating a different test method or