At 09:43 AM 1/25/2002 -0800, Afshin Salehi wrote:

>I have a question that has had me thinking every so often.  When the
>autorouter connects traces to small smt pads of IC's it never goes straight
>into the pad.  Instead it will go along the side of the pad and then turn
>left or right and into the pad.  This looks like a potential problem to me,
>and on top of that it just looks cheesy!  :)  I am not sure what the most
>efficient method would be to solve this problem (hint, hint).

It is a quality issue from two points of view. One is that it often causes 
pockets to be formed, i.e., included angles of less than 90 degrees, which 
traditional wisdom says can trap etchant, causing possible trace failure 
even a long time later. The other is cosmetic, which is important when you 
are making your living designing boards.

I remember, many years ago, when I was prone to make very curvy traces 
instead of sticking to 45 degree angles for bends, an engineering manager 
nixed my work because it was "messy." (This was manual tape, and "messy" 
was a value judgement: a board had been designed with very high density 
made more possible because the traces packed in a curved manner; but he was 
used to everything being nice and orthagonal.)

On the other hand, changes on that board would have been a bit of a 
nuisance, one would simply have ripped up all the routing. It *might* have 
been easier to make changes if the track had run on-grid in an orthagonal 
manner. In any case, I went to cut-and-bend with tape shortly thereafter 
and films of my layouts looked like they had been done with CAD unless you 
looked *very* closely (the 45 degree corners were very slightly rounded).

Anyway, to the present point. The autorouter cleanup routines are almost 
brain-dead, there is no excuse, really, for the state in which it leaves 
layouts. It is a *lot* more difficult to do what they have already done -- 
autoroute a board with moderately good completion -- than to clean up 
meandering segments, since the latter process can be done very locally, one 
layer at a time, with one trace at a time. Simply eliminating segments and 
determining if the resulting new trace causes clearance violations would 
vastly improve routing appearance. (There is no requirement that the last 
segment of a trace be orthagonal or semi-orthagonal, it may be any-angle 
unless it is long.)

To clean up traces manually, be sure to have Loop Removal and snap on. 
(tools/Preferences/Options/Automatically_Remove_Loops) You can then simply 
draw over the traces, very quickly, and the original segments will usually 
disappear. (With a little practice, you will learn how to drive this loop 
removal feature so that it almost always does what you want.) Loop Removal 
is the manual router's friend!


[EMAIL PROTECTED]
Abdulrahman Lomax
Easthampton, Massachusetts USA

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
* To post a message: mailto:[EMAIL PROTECTED]
*
* To leave this list visit:
* http://www.techservinc.com/protelusers/leave.html
*
* Contact the list manager:
* mailto:[EMAIL PROTECTED]
*
* Forum Guidelines Rules:
* http://www.techservinc.com/protelusers/forumrules.html
*
* Browse or Search previous postings:
* http://www.mail-archive.com/proteledaforum@techservinc.com
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

Reply via email to