Thanks Ian,

I will try your suggestion next time.

IW> If you specify "Pour over net" then it should not matter if the router has
IW> routed the nets - you can still pour the poly over the top (just make sure 
IW> you specify Pour Over and the correct net name.
IW> As Brad says, sound like you might like to investigate the pour options you 
IW> used.

I did have pour over net checked.

When the VCC and GND nets are routed,  some VCC tracks will be placed on top,
others on the bottom.  Same for the GND's,  some on top, some on bottom.
Now when you do the pour,  lets say on the bottom GND side,  all the GND traces
will be "absorbed" into the pour,  but any VCC tracks which were routed on the
bottom layer will still be present.  The same thing happens with any GND traces
with the (topside) VCC pour.  These were the unnecessary traces I was
looking to do away with.

(hmm...thinking now maybe I could have restricted routing GND to bottom,  VCC to top,
and tried that... for so simple a board this might have worked too)

The other problem is that if you first route the board and then do the pour,
you do not get great thermal reliefs.  The routed net seems to "step on" the
thermal reliefs from the pour.  (as viewed from the gerbers)

>>Then I tried making the trace width impossibly large.  It's a 2" X 2.7"
>>board,  so I made the (min/max/pref) trace width all 5".  That
>>way it can't _possibly_ route those nets.... then I'll just do the poly
>>pour over it...  The router happily placed 3-4 VCC/GND tracks down
>>on the board.  At maybe 50-100 mil width.  This has to be a bug?
>>The router should not put down tracks that fail to meet the design

IW> I agree.  Sounds like a bug.  I have had problems with the rules at times, 
IW> mostly order of priority issues.  What is specified as the width rule for 
IW> the Whole Board?  Maybe the router is incorrectly picking up the whole 
IW> board scope and using it.  Does sound like a bug.

The whole board width constraint at the time was set to 8 mils min.
(I'll have to look and see what the max was) The traces it put down were
wider than 8 mils.  I had previously set the power/GND nets to 50/100/100.
Perhaps it was somehow using the previous 50/100/100,  even though 5000/5000/5000
was currently set as the rule to follow?

However if so,  it does not explain why only 3-4 of the segments were routed.
All but one small GND segment was easily routeable at 50/100/100.  Why route only the 
and not all (but 1 of) the rest?  Either way I feel something "weird" was going on 

IW> If you can investigate further whether the router was picking up another 
IW> width constraint rule and let me know I will get on and add the report to 
IW> the bug list.  I have about 10 bugs and possible bugs to add I think at 
IW> this stage, all reported over the past couple of months on this list, but 
IW> not yet added to the bug list.
Maybe I'll spend some time creating a minimum test case of it.

IW> Ian Wilson

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
* To post a message: mailto:[EMAIL PROTECTED]
* To leave this list visit:
* Contact the list manager:
* Forum Guidelines Rules:
* Browse or Search previous postings:
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

Reply via email to