TECHNET Archives

1996

TechNet@IPC.ORG

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
"Thad McMillan" <[log in to unmask]>
Date:
Thu, 11 Jan 96 08:51:02 CST
Content-Type:
text/plain
Parts/Attachments:
text/plain (84 lines)
     Doug,   
     
        I'll address the issues of BGA's associated with bare boards.
     
        In general boards with BGA's are overall not much different than a 
     standard SMT board without.  The differences can however give a lot of 
     headaches.  You and your process guys need to be familiar with them.  
     Some differences are:
     
        BGAs create a spot of very high density on a PCB.  Typically a BGA 
     involves a grid of vias on a 50 mil grid our less.  This may sometimes 
     make making test fixtures more difficult.
     
        BGAs tend to drive up layer count because of the high I/O density 
     in the area of the BGA.
     
        BGAs tend to require fine lines (at least on the inners) of 5/5 mil 
     line/space.  This is due to the high density.
     
        The vias associated with BGAs almost always are 13.5 mil drill or 
     less.  Again due to density.  
     
        The geometry associated with a BGA footprint typically involve a 
     BGA bond pad connected to an adjacent via.
     
        Many OEMs require that the adjacent via to a BGA bond pad be filled 
     with soldermask.  This is for several reasons.  One is to prevent 
     solder wicking up the via and bridging to the BGA bond pad.  The other 
     is to prevent heat transfer during post SMT wave operations causing 
     the BGA joint to reflow.  Thus a PCB shop needs to know how to plug 
     vias, consistantly.
     
        When OEM's do not fill the via with soldermask there is generally a 
     soldermask web between the via and BGA bond pad.  This web needs to 
     stick.  Other OEMs use a soldermask defined pad, in which the 
     soldermask opening over the BGA copper bond pad is smaller that the 
     BGA bond pad.  I call this a soldermask defined pad.  I've seen 
     solderability problems with soldermask defined pads occurring for 
     different process reasons. 
     
        One key to BGA designs is that different customers specify 
     different BGA footprint geometries.  Please have you process engineers 
     look at the particular geometry of a given design.  They need to 
     review hole/pad locations, how BGA vias are connected to the BGA pad, 
     the general geometry of the footprint, and where the soldermask is.
     
        Soldermask issues are generally the major issues with BGA boards.  
     Please note that I am not talking about boards used in a BGA package, 
     but the boards that have BGA packages attached to them using a SMT 
     assembly process.
     
     Hope this helps.
     
     [log in to unmask]    



______________________________ Reply Separator _________________________________
Subject: BGA
Author:  [log in to unmask] at Dell_UNIX
Date:    1/10/96 7:33 PM


-- [ From: Doug Jeffery * EMC.Ver #2.10P ] --
     
     
Question?
     
What does BGA technology mean to board fabricators?  Do we have to 
become solder paste experts and reflow solder again.. I thought we got 
rid of that problem.....:)....
     
  No Seriously, what are the responsibilities at fabrication for
creating boards to accept BGA's?
     
Thanks?
     
Doug Jeffery
Electrotek..
     
     



ATOM RSS1 RSS2