Many of the boards we manufacture are 11" X 17".  Our customer requires that we
hold the dimensional tolerance on their corner tooling holes to +/- 0.005".
When we  started producing these boards several years ago, we had a problem
holding this tolerance.  We found that baking the panels prior to drilling
allowed us to achieve this tolerance.  We found that if shrinkage occured, it
was during the final cure of the LPI solder mask (310 deg. F for 45 minutes).
Baking panels at 310 deg. F for two hours at temperature solved the shrinkage
during LPI cure.  

Several weeks ago we started experiencing shrinkage of 0.015" - 0.020" over a
16" Y dimension.  The X dimension shrinkage is about 50% of the Y direction
shrinkage or less.  The annular ring mis-registration is 0.001" or less.  We
have confirmed that the baking oven is properly calibrate and working correctly.
This shrinkage occures on more than one lot of laminate but does not happen on
all lots.  Furthermore, the shrinkage is not consistant within the lot.  The
boards are not warped or twisted.  The problem is not related to a specific
board design.  It occurs with and without extensive ground planes.  The laminate
is produced by one of the larger U.S. manufactures.  

This is a first for us.  Has anyone seen anything like this?  I would appreciate
any inormation (fact or theory) on the possible cause.

Hans Rohr
Printed Circuits of America