X400-Content-Type: |
P2-1984 (2) |
Old-Return-Path: |
|
Date: |
Tue, 15 Oct 1996 11:01:14 +0100 |
MIME-version: |
1.0 |
Precedence: |
list |
Resent-From: |
|
Resent-Sender: |
|
X-Status: |
|
Status: |
O |
X-Mailing-List: |
|
X-Loop: |
|
Resent-Message-ID: |
<"40FE71.0.R58.vMsOo"@ipc> |
Message-ID: |
<"672*/I=J/S=Drake/OU=kid0112/O=icl/PRMD=icl/ADMD=gold 400/C=GB/"@MHS> |
X400-MTS-Identifier: |
[/PRMD=icl/ADMD=gold 400/C=GB/;kid0112 0000051200000672] |
TO: |
|
Return-Path: |
<TechNet-request> |
X400-Originator: |
|
In-Reply-To: |
|
X400-Recipients: |
non-disclosure:; |
Received: |
by ipc.org (Smail3.1.28.1 #2)
id m0vD6d8-0000QMC; Tue, 15 Oct 96 05:21 CDT |
From [log in to unmask] Tue Oct 15 09: |
41:42 1996 |
Content-type: |
multipart/mixed; boundary="6moRsMTYaceDZaGYDbuNAblNQ70o7vTU" |
Subject: |
|
From: |
|
Alternate-Recipient: |
Allowed |
X400-Received: |
by mta bath.mail.pipex.net in /PRMD=pipex/ADMD=pipex/C=gb/; Relayed; Tue, 15 Oct 1996 11:18:36 +0100
by mta feldr1 in /PRMD=pipex/ADMD=pipex/C=gb/; Relayed; Tue, 15 Oct 1996 10:59:27 +0100
by mta kid01c1 in /PRMD=pipex/ADMD=pipex/C=gb/; Relayed; Tue, 15 Oct 1996 10:59:40 +0100
by /PRMD=icl/ADMD=gold 400/C=GB/; converted (ia5-text); Relayed; Tue, 15 Oct 1996 11:01:14 +0100 |
Original-Encoded-Information-Types: |
ia5-text |
Content-Identifier: |
672 |
Parts/Attachments: |
|
|
One answer I recently heard on this from an SMT assembly machine
supplier was that two local fiducials can actually make theta
location of the device worse. The reasoning was that the
distance between the two local fiducials is very small. The
error in theta location due to the camera resolution was higher
than the error in theta you get going off the widely spaced
global fiducials. Their take was that you use the board global
fiducials for theta location of fine pitch devices and a single
local fiducial for x,y correction of the device. Localised x,y
drift of the board location was much more likely than localised
theta skew. In addition the extra time needed to locate a second
local fiducial would make a serious impact on high speed placement
throughputs.
I have no direct experience of this myself and I have not
done the calculations on theta error, but it makes some sense and
might be worth considering.
Jeremy Drake
Design to Distribution. Stoke on Trent. England
|
|
|