Some additional anecdotes, which might be useful:

* We first started to specify 0.5% in 1991, prior to SMT.  Our PTH
automated pick and place line required it.  Excessively warped boards
would cause the leads to jam in the hole rather than being inserted
through.  Problems were also related to .093 X 15" X 15" boards had to
be .115 maximum to fit through our conveyorized machinery.

* Soon after that, we started SMT.  We worried about planarization of
the SMT pattern and lead off board height.  If there was serious warp
across a large part, there could be a problem.

* So, we instituted 0.5% as our standard.  In the years that followed,
we used a number of vendors.  New vendors would ask if we really needed
it.  After we said yes, they delivered hardware.  I have to assume that
it met the specified requirement.  There was not a significant rejection
rate for warp at either our suppliers or at assembly.  In the last few
years, there have been very few rejections at all.

* We are entirely military aerospace.  As Mahendra pointed out, Class 3A
specifies 0.5%.  (I think that requirement flows from input I provided.)
This discussion will not affect our hardware.

* I do not know what we pay a premium for the tighter warp requirement.

* There are two types of warp; thermoset and thermoplastic.

o The thermoset can be flattened with heat and pressure.  You are using
the pressure to freeze the stress into the laminate.  The next time the
board sees heat, it will stress relieve itself and potato chip.

o Thermoplastic warp can also be flattened with heat and pressure but it
will not appear when heat is re-applied.

o You cannot tell which type of warp is involved by visual inspection of
the board.

* I think that applying bow and twist to the entire board is the
reasonable thing to so.  Bow in acceptable areas could stress relieve
itself in prohibited areas during assembly.

* I would support maintaining the default as it is for some hardware,
but building in a standard variance to be AABUS for boards that need it.

Mike Green
Electronic Packaging Design
Lockheed Martin Space Systems
1111 Lockheed Martin Way
Sunnyvale, CA 94089
office  408-743-1635
Synergy works at LMC.  "Show me the data."

-----Original Message-----
From: IPC-600-6012 [mailto:[log in to unmask]]On Behalf Of John Perry
Sent: Wednesday, May 10, 2006 12:49 PM
To: [log in to unmask]
Subject: [IPC-600-6012] IPC-6012B Bow and Twist Requirements

Colleagues,

IPC continues to receive queries from industry as to whether or not
there will be a tightening of the bow and twist requirements for printed
boards that are given in both IPC-2221A and IPC-6012B.

I would like the members of this forum to provide their feedback on
these requests.

There are currently two arenas where we are getting requests for
alteration to the requirements: BGA devices and screen printing
technology.

Regarding BGA technology, Don Dupriest of Lockheed Martin Missiles and
Fire Control provided a .ppt file at the IPCWorks 2005 meeting for these
groups where the recommendation for bow and twist acceptance for
isolated BGA areas is 0.50%.

This presentation file is available at
http://www.ipc.org/committeedetail.asp?Committee=D-33A

Within "Drafts", under the titled "PCB Bow and Twist - BGA Devices"

Regarding screen printing technology, we recently received the following
request:

We had a board that was slightly warped, make contact with and damage
the camera in one of our screen printers. Of course, when I measured the
bow and twist, it was well within acceptable limits for bow and twist.

My question is this: Are the bow and twist specifications tight enough?
For over 15 years, I have been "jumped on" by the line operators and
maintenance guys to fixed warped boards only to find that the boards are
usually not any where near the limits. Since I have worked for multiple
companies, I see that this is an industry wide dilemma. Board
manufacturers typically can hold bow and twist requirements, but
different SMA machines (especially screen printers) aren't able to run a
board that has bow or twist near but within the specification limits.
Should I be asking why machines built for the surface mount industry
can't run boards that meet industry specifications, or if the
specifications that once fit, are in need of revision?

Larry D. Roberts
Quality Engineer
Andrew Corporation

Thanks for taking the time to review and respond to these industry
requests.

Regards,

John Perry
Technical Project Manager
IPC
3000 Lakeside Drive # 309S
Bannockburn, IL 60015
[log in to unmask]
1-847-597-2818 (P)
1-847-615-7105 (F)
1-847-615-7100 (Main)