I agree with John's responses to Roger shown below and would like to 
add the following comments.

It is worth mentioning that the following thoughts only apply to 
thermoset resin systems, of which epoxy is one among others. Once a 
thermoset resin cures, it does not flow anymore, it will soften and 
expand rapidly if heated above it glass transition temperature, but 
it will not flow and is locked in place.

In summary, it's been my experience that baking prior to assembly is 
done to remove moisture from around the areas where soldering, 
and thus its related thermal expansion, is going to take place. 

Bow and twist or warpage is caused by uneven stresses built 
into the circuit board during the lamination process.

The uneven stresses are caused by various factors, some of which are: 

1) How well balanced is the construction of the base material (on 
   double sided) and laminated fabrication panels (resin, glass 
   and/or copper circuitry symmetry from the center plane).
   
2) For multilayers: the rate of heatrise and cooldown at lamination   
   and how well are the lamination plates and presses are maintained.
   
There is a possibility that boards that are received flat, and that 
warp during reflow may have been warped and reworked by heating them 
between heavy metal plates. This kind of rework temporarily hides the 
warped and unevenly stressed nature of the PCBs, but as soon as the 
printed circuits are heated up they go back to their less stressed 
state.

Luis Rivera
CompuRoute, Inc.
Dallas, TX
[log in to unmask]
(214) 340-0543



> From: "John Gully" <[log in to unmask]>
> Organization: CompuRoute, Inc.
> To: [log in to unmask] (ROGER HELD), [log in to unmask]
> Date: Mon, 24 Jun 1996 15:25:34 +600 CDT
> 
> > Date:          Mon, 24 Jun 1996 13:55:33 -0500
> > From:          [log in to unmask] (ROGER HELD)
> > Subject:       FAB: PCB Delamination and Warp
> > To:            [log in to unmask]
> 
> >      
> >      To everyone,
> >      
> >      We have, in the past, had problems with bare board warp and 
> >      delamination in our assembly process (reflow).  We have solved this 
> >      problem by instituting a baking process at our vendor or in house 
> >      prior to assembly.  My questions are as follows:
> >      
> 1. Is baking a standard PCB fab process (before shipping)?
> 
>    Yes and No - If the lamination recipe (asymmetrical) is prepared 
>    correctly and cleaning methods are followed you should have no
>    problems with warp/twist or delamination.  However, during storage 
>    and shipping the boards will absorb moisture.  This is due to 
>    laminate and prepreg materials being hydroscopic.
>           
>    Recommmended baking, say at 230 F for 1 to 2 hours, as shown
>    in the HICAM fabrication notes, removes retained moisture
>    from the board.  Especially from around the vias and/or
>    PTH where moisture can easily work its way through the 
>    plated side wall, causing blowholes, which is a unwanted
>    defect.   
>  
>    From an assembly standpoint, large components and tight fitting
>    terminal pins on a connector may induce delamination failures.
>       
>    In your case the pre-bake prior to ship was pushed off to the
>    supplier versus in-house.
> 
>    Remember also that a preheated board is less likely to fracture
>    under stress prior to subsequent heat excursions than unbaked 
>    boards.
> 
> >      2. If so, what is the standard profile?
> 
> see above for baking profile.
> 
>       3. If not, why do some vendors' boards need to be baked to prevent     
>          warp and delamination while others don't?
> 
>    Its sometimes good practice to bake boards prior to shipping and
>    bareboard assembly.  Remember the material is hydroscopic, the
>    materials characteristics are to absorb moisture.  This absorbtion
>    increases as the humidity increases.  
> 
>    NOTE: Do not bake boards near the Tg temperature without handling
>    procedures in place.  Additional warp and twist are prevalent 
>    if not handled properly. 
>       
> >      4. Does baking cure both delamination and warp or does it just cure    
> >         warp?
> 
> Baking does not fix and/or cure delamination.  Delamination is a 
> Class A defect and is more likely to proprogate further when  
> thermal excursions are known to occur.  Again the warp and twist
> go back to how the lamination recipe is done.  If you have warped
> boards, baking prior to assembly (using weights in the oven helps)
> will flatten boards or in spec for SMT placement.  It may or may not
> help if the board is two side SMT or multiple passes are required.
> Remember also, the additional thermal cycling may cause unwanted 
> material fratures that can lessen the boards longevity.
> 
> >      
> >      5. My understanding is that moisture is getting trapped in the board  
> >         during laminate press (or contamination of some sort) which is      
> >         leading to the problem during assembly.  Can the moisture           
> >         entrapment be avoided?  Should we claim to our vendor that he       
> >         better control contamination in the process?
> 
> Volatiles from the prior solvents and moisture or resin starvation 
> can attribute to delamination.  Its these pockets volatiles that 
> attribute to delmaination.
> 
>  
> >      5. It seems to me that if boards need to be baked before delivering to 
> >         a customer that there is some poor process control going on.
> Not always.  Read above.
> 
> 
> Please call should you need additional information.  Thank you
> Roger Dodger.
> 
> John Gulley - PE
> Compuroute Inc.
> 214-340-0543 wk
> 214-847-0343 pgr
> >      
> >      Someone, please educate me.  Maybe, my perspective is in error.
> >      
> >      Thanks,
> >      
> >      Roger Held
> >      Hitachi Computer Products (America), Inc.
> >      [log in to unmask]
> > 
> > 
> 
>