TECHNET Archives

January 2014

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:
Julie Silk <[log in to unmask]>
Reply To:
TechNet E-Mail Forum <[log in to unmask]>, Julie Silk <[log in to unmask]>
Date:
Mon, 6 Jan 2014 17:53:21 -0600
Content-Type:
text/plain
Parts/Attachments:
text/plain (11 lines)
Weighing in with some thoughts and our experience:

The only QFN parts we've seen that didn't wet or that opened up over time have been parts with thick-gold plating (resulting in my various gold in SAC studies) or egregiously bad process (rare).  Usually, QFN parts solder just fine as long as we follow our current design rules.  However, we still see too many voids for some situations, so need better design rules or processes, so there is work to do here.
The process and the design have to be worked on.  Simple process changes are unlikely to give the needed result.

The main issue with a set and strict void percentage on QFN central pads is the cost, as is mentioned in other posts.  Not just of the process development or rework costs.  Reworking a QFN that has a ground pad needed for good heatsinking is a real PAIN to rework because it's connected to ground planes, sometimes several planes and has several vias.  The heat runs out into the board, and the part has to get very hot for a long time to melt the joint, and damage to the board is likely.  Also, in the case where there are several QFNs near each other, the QFN _next to_ the reworked QFN will have larger voids than before the rework and will fail earlier.  Each rework heats the board up to soldering temps and reduces the life of the PCB.  The scrap rate or field failure rate will be higher with reworked boards.  In the situation Rueven describes, with several QFNs, he may need to require some of the new techniques described in recent papers, especially by Indium -- to use preforms to add extra solder without adding extra flux.  

While an IPC requirement is very desirable in many ways, it would end up very lax in order to prevent unnecessary rework on less sensitive parts.  And that would not work for those with high heat transfer.  In many cases, part-specific requirements have to be developed for each application.

When we have a standard and inexpensive  process that can produce <25% max voiding in QFNs, the IPC spec will be useful.

ATOM RSS1 RSS2