Mail*Link(r) SMTP               FWD>warped completed assemblies

Chet,

You asked, "1. Does anyone know a way to straighten a completely stuffed board
without damaging the components?"

In the IPC's-R-700, the basic method to "flatten" out printed boards and their
assemblies is to place the board/assembly in some form of a restraint and heat
the board up to some temperature (ideally above the Tg of the resin), and
allow the board/assembly to cool back down to ambient while still in
restraint.

We've done this for many years with good results and no major end product
reliability problems.  Most "flattenings" have been for printed board
assemblies, though we have flattened some printed boards. Results of
flattening printed boards has been mixed, they're flat for assembly, but due
to design, some will "re-pretzel" in flow soldering or convective IR reflow,
so sometimes we will need to "re-flatten" the assembly.   

1)  We design a 25 mm high, "egg-crate" x-y grid-like structural support
system of 0.7 mm thick CRES (corrosion resistant (Stainless TM) steel) (other
metals could be used, but we designed them for long term general use).  The
centerline spacing of the egg-crate is determined by the arrangements of
components on the board/assembly and should not be greater than about 50 mm
pitch.  The metal spacer/straightners to run the full length and width of the
board/assembly in full contact with the surface of the board/assembly between
components.     

2)  If an assembly, we examine all the components to determine the one(s) with
the lowest (non-use) storage temperature.  Generally aluminum/tantalum
electrolytic capacitors and plastic cased components, such as audible
transducers, speakers, some meters, liquid crystal displays, etc.  This is
used to establish the bake temperature, which is generally in the range of
100-120 deg C. (except for some plastic cased components).  Some times we
removed some of the parts to increase the bake temperature.

3)  We then make a "stack-up" up boards/assemblies to be flattened.  The
stack-up consists of a heavy base plate (13 mm thick aluminum tooling plate
(they're flat)), then one of the egg-crate supports, then one board assembly, 
then repeat the alternating layers of board/assembly and egg-crate support as
required or limited by the size (height) of the oven, lastly another egg-crate
support, then another heavy base plate (13 mm aluminum tooling plate), and
lastly some mass (weights) about 20 kg per 0.1 square meters (40 lbf per
sq.ft.)

4)  Place the stack in a "cold" forced air oven with known good temperature
control or assembly the stack in the oven ( (;-) and cook until done and ready
to be served (;-) ).

5)  We do our "flattening" overnight.  At the end of the work day, we place
the stack in a cold oven, have it slowly ramp up to temperature (about an
hour) hold temperature for 4-hours, turn off the heat, allow to cool with the
forced air blower on until morning (the over has a small make-up air vent
capability).  Remove from the oven and 95+% of the time we have acceptably
flattened boards/assemblies.

Then you asked, "2. How much of warp can be reliably corrected?"

Most boards/assemblies have ranged in size from 150 mm to over 450 mm in
length/width.  Board thicknesses have been up to 6 mm, though most are 1.5-3
mm in thickness.  We've done up to 12-layer multilayer's with not problems.  

As the ol'memory bank recalls, we have routinely "flattened" most assemblies
with about 5% out of flatness.

Results, some to the assemblies were edge-board connector assemblies were the
ol' "quad-wide" DEC M/R-Series computer boards of our design which consisted
of a series of four edge-board connectors that totalled about 300 mm wide.

All "use-environments" were relatively benign indoor/outdoor like operating
enviroments.

Hope this helps,

Ralph Hersey
[log in to unmask]

--------------------------------------
Date: 8/27/96 7:16 AM
From: Chet_Meddles_at_Micro__Industr
 
     1. Does anyone know a way to straighten a completely stuffed board 
     without damaging the components?
     
     
     2. How much of warp can be reliably corrected?

From: [log in to unmask]

------------------ RFC822 Header Follows ------------------
Received: by quickmail.llnl.gov with SMTP;27 Aug 1996 07:14:52 -0700
Received: from ipc.org by simon.ipc.org via SMTP (940816.SGI.8.6.9/940406.SGI)
	 id JAA27761; Tue, 27 Aug 1996 09:01:10 -0700
Resent-Date: Tue, 27 Aug 1996 09:01:10 -0700
Received: by ipc.org (Smail3.1.28.1 #2)
	id m0uvOKc-0000OYC; Tue, 27 Aug 96 08:37 CDT
Resent-Sender: [log in to unmask]
Old-Return-Path: <[log in to unmask]>

Date: Tue, 27 Aug 96 09:36:09 EST
Message-Id: <[log in to unmask]>
To: [log in to unmask]
Subject: warped completed assemblies
Resent-Message-ID: <"E9H4S3.0.wqI.Fel8o"@ipc>
Resent-From: [log in to unmask]
X-Mailing-List: <[log in to unmask]> archive/latest/5970
X-Loop: [log in to unmask]
Precedence: list
Resent-Sender: [log in to unmask]



***************************************************************************
* TechNet mail list is provided as a service by IPC using SmartList v3.05 *
***************************************************************************
* To unsubscribe from this list at any time, send a message to:           *
* [log in to unmask] with <subject: unsubscribe> and no text.        *
***************************************************************************