MLB PCB DIMENISIONAL STABILITY AND LAMINATE INTEGRITY

REVISITING MULTILAYER DIMENSIONAL STABILITY AND LAMINATE INTEGRITY It has been said we must learn it all over again and again about every four years. I know Lee Ritchey said that to me about high speed digital circuit design. The same seems true for multilayer printed circuit boards especially concerning dimensional stability and laminate integrity. Phil Hinton and the late Dr. Paul Craven, in the early 1980's (as I recall - ask Phil), performed studies, wrote articles, and presented papers about how MLB shrinkage or growth in the x and y axes could be controlled by using multilayer constructions consisting of specified glass styles and resin contents. We have gotten away from what was proven in the interest lower costs now, but at what price in the finished product. Part of this is due to ignorance and the other is the constant search to reduce price by believing without researching. Recently, using single ply prepreg has been thought to be an answer reducing cost. It is not. Simply stated, using homogeneous, unbalanced multilayer constructions with specified glass styles minimizes x/y growth, shrinkage, and/or warp and twist. This, done using balanced reference plane placement, metal types, weights, and thicknesses (or modified, x-hatched versions), must be done to prevent defect at the design level instead of continually asking what may be done to "fix" problems when it's too late (upon product receipt or during soldering operations). People still offer suggestions saying warp and twist may be "baked out" or relieved somehow without understanding it must be prevented. Constantly, questions arise concerning how to eliminate warp and twist after it is found to be a problem. There is no reason for this as proven answers are available to prevent it. Use appropriate design for manufacturing and concurrent engineering practices, with proven design rules (including material selections, constructions, and relamination processes), to prevent defect at the design level. I will be glad to provide additional, very detailed information to those seeking it. Earl Moon - Proof Of Design (pod@ix.netcom.com)

reply »

We have been told that we should bakeout FR4 boards prior to assesmbly - i gather from your discussion that you disagree - any details will be sincerely appreciated. Regards.

reply »

| We have been told that we should bakeout FR4 boards prior to assesmbly - i gather from your discussion that you disagree - any details will be sincerely appreciated. Regards. Please bake. I agree. Let me know how we have a disagreement.

reply »