Once a board warps, there is very little that
can be done to remove the warp.  The board has
memory of stresses that were incurred during 
multilayer lamination.  These stresses will 
be relieved anytime the board experiences a 
thermal cycle after lamination.

Don't point the finger at the fabricator of
the board just yet.  I have rarely seen a 
board warp that was designed with a symmetrical 
stackup.  

First, look at the cross-section stackup of
the board and look for the following:

1.  The dielectric thicknesses should be
    symmetrical about the center of the
    stack.
2.  The copper weights should be symmetrical
    about the center of the stack.
3.  Signal and Ground/Power layers should be 
    a mirror reflection about the center of
    the stack.
4.  If multiple dielectric materials are used
    in the board, the dielectric materials 
    should be centered about the center of
    the stack.
5.  The circuitry should be the same density
    as much as possible in the x-y plane on
    each layer to minimize twist.

The further away from symmetry the board is,
the more it WILL warp and twist.  If the design
is unbalanced, there is very little that the
fabricator or assembler can do to flatten it
out for any extended period of time.  Baking
the board with weights on it will attempt to 
flatten the board out, but once it is heated
back up again in IR reflow, wave solder, etc., 
the warp is back.

If the board was designed with symmetry, then the
fabricator has some evaluation to do.  Is there
any missing prepreg, wrong copper weight, layers
in the wrong order, fast cool down in the press,
etc.????

Darren Hitchcock
[log in to unmask]

[log in to unmask] (Jerry Schwartz) Wrote:
| 
| 
| >
| >Greetings,
| >
| >Does anyone have a good procedure for removing a 
| warp/twist from a pcb?  I 
| >am dealing with a .093 thick assembly that is 7 
| inches by 20 inches.  The 
| >board currenly has smt components placed and 
| soldered (topside only) but, 
| >the through hole components have not yet been 
| mounted/soldered.  I have 
| >tried heating the board and letting it cool while 
| applying weight, that did 
| >not do the job.  Any suggestions ? please respond.
| >
| >Thanks,
| >
| >[log in to unmask]
| >
| >
| >
| If it came to you like that from your supplier 
| SEND IT BACK!
| Also a pre bake of the bare board to reduce 
| moisture should help
| warp later in the process.
| 
| Regards,
| Jerry Schwartz
| [log in to unmask]
| "May The Schwartz Be with You"
|