FYI...
Have you brought this problem up with your PCB supplier(s)?
As Michael mentions, PCB design (lyr count, copper wt on externals, board
thickness, etc.) plays an important role. As does the capability of your
PCB supplier(s), notably photo processes.
We spec +/- .002" pattern/feature location to datum on the fab drawing.
However, with your PCB size +/- .003" might be what you're seeing. I
believe PCB shops measure this from the center of their panel, so you might
have better than +/- .003". This board isn't a good fit for a 18" x 24"
panel (widely used) but is a good fit for a 14" x 18" and could result in
better photo registration.
Have considered some other contributing factors? For example:
Stencil stretch - I haven't looked into this recently, but .001" per inch
for chem etched stencils was the best we could get a few
years back.
PCB warpage - contributes to the problem. If using a HASL finish, consider
OSPs or electroless Ni/immersion Au finish. The thermal
gradients seen in the HASL process contribute to both warpage
and stretch.
Screen print - Vision system on the printers? Optical alignment of stencil
pattern (round half-etched fiducials on the bottom side of the
stencil seems to work well) to board pattern is crucial. Of
course foil thickness, aperture design, squeegee material
(we think metal blades work best along with single level foils)
all play a role.
Whoa, better stop now. Sorry for getting off on a tangent.
Tracy Tennant
[log in to unmask]
---------- Start of forwarded message ----------
Message# 783032
23-APR-1996 15:13:29.15
From: [log in to unmask]
To: STHUMMEL,TTENNANT,[log in to unmask]
Subj: RE: dimensional question
Received: from micron.co by BOGEN5 Tue, 23 APR 96 15:11:05 MDT
Received: from is04.micron.com by micron.com (SMI-8.6/SMI-SVR4)
id PAA09590; Tue, 23 Apr 1996 15:10:37 -0600
Received: from simon.ipc.org ([168.113.24.64]) by relay.micron.com with SMTP id
<185376>; Tue, 23 Apr 1996 15:10:59 -0600
Received: from ipc.org by simon.ipc.org via SMTP (940816.SGI.8.6.9/940406.SGI)
id QAA23572; Tue, 23 Apr 1996 16:08:09 -0700
Resent-Date: Tue, 23 Apr 1996 17:08:09 -0600
Received: by ipc.org (Smail3.1.28.1 #2)
id m0uBo7K-0000BqC; Tue, 23 Apr 96 14:51 CDT
Resent-Sender: [log in to unmask]
Old-Return-Path: <[log in to unmask]>
From: "Yuen, Mike" <[log in to unmask]>
To: "[log in to unmask]" <[log in to unmask]>
Resent-To: <[log in to unmask]>
Resent-To: <[log in to unmask]>
Resent-To: <[log in to unmask]>
Subject: RE: dimensional question
Date: Tue, 23 Apr 1996 13:55:00 -0600
Message-ID: <[log in to unmask]>
Encoding: 36 TEXT
X-Mailer: Microsoft Mail V3.0
Resent-Message-ID: <"Id2w31.0.seB.lIJVn"@ipc>
Resent-From: [log in to unmask]
X-Mailing-List: <[log in to unmask]> archive/latest/3601
X-Loop: [log in to unmask]
Precedence: list
Resent-Sender: [log in to unmask]
Q: How much growth (or shrink) should one expect land-to-land over
a 16" distance? This has nothing to do with drilling; just master
pattern to actual board size...
A: I believe that's called stretch effect, the bigger the board gets the
more severe the growth (or shrinkage) it will have. The stretch also depends
on the board house and board thickness. The best way to minimize stretch
effect is to group components which require tight tolorance into the central
location of the PCB.
Thanks
Michael Yuen
----------
From: TechNet-request
To: technet
Subject: dimensional question
Date: Tuesday, April 23, 1996 11:07AM
The proliferation of sub .5mm peripheral leaded SMD's is causing
assembly people to look at the overall size relationship
between the master data and the actual board. When solder paste
is stenciled on larger boards (say, 12-16"), the stencil image
many times will not match the land pattern, causing shorts in the
fine pitch area.
My question:
How much growth (or shrink) should one expect land-to-land over
a 16" distance? This has nothing to do with drilling; just master
pattern to actual board size...
Thanks,
marko
|