kicad-developers team mailing list archive
-
kicad-developers team
-
Mailing list archive
-
Message #02031
Re: pcbnew - thermal connections patch proposal
-
To:
kicad-devel@xxxxxxxxxxxxxxx
-
From:
Vesa Solonen <vsolonen@...>
-
Date:
Thu, 25 Dec 2008 21:09:03 +0200
-
In-reply-to:
<49539934.70205@...>
to, 2008-12-25 kello 15:31 +0100, Rok Markovic kirjoitti:
> Hi
>
> I have coded a new thermal connection algorithm which uses tracks
> (SEGZONE) instead of the polygons. I have started doing this because
> the results of polygon thermal conection is sometimes ugly (see
> atachment so8 footprint).
Nice patch indeed. It seems to handle "dead end" directions in zones
too. There is a bug (supposedly) in KBool that has been keeping from
fixing the thermal stub issue among other things.
> Now you get more freedom with setting minimal zone thicknes (it
> doesn't
> need to be less than copper track width) and you can choose to have
> different thermal patterns (not yet implemented).
I suggested using track stubs in thermals to Jean-Pierre too, but it was
uncertain which would be the most efficient way of doing thermals.
Anyway there are small details still missing like what would happen when
fill strip next to the pad is thinner that thermal track. You see, there
is currently no way for straight edge track ends and round one would
leave gap or go over. Jean-Pierre is working on the problem for
microwave use though.
First hand it seems track stubs would be easier to implement with
flexibility in mind. At least the fill code would be simplified
dramatically. Also rounding all inside corners should be considered as
reasoned in my previous posts about polygons. I was going to add a small
miters in all thermal cut outs for that.
> I have a problem with generating connection to the zone (ratsnet).
> What do I have to do for ratsnet calculation to work correctly.
DRC issues was a major reason for polygons too. Otherwise negative
Gerbers would be easiest.
Happy holidays,
-Vesa
References