kicad-developers team mailing list archive
-
kicad-developers team
-
Mailing list archive
-
Message #13934
Re: build kicad with Interactive Router
-
To:
<kicad-developers@xxxxxxxxxxxxxxxxxxx>
-
From:
Vesa Solonen <vesa.solonen@xxxxxxxx>
-
Date:
Sun, 13 Jul 2014 12:45:05 +0300
-
In-reply-to:
<CAHBYzfQJj_2EHwqZ93k6OA1RfeP+FWaHS7_86p+_ToBViGqY=A@mail.gmail.com>
-
User-agent:
Mozilla/5.0 (X11; Linux x86_64; rv:24.0) Gecko/20100101 Thunderbird/24.6.0
13/07/14 11:28, Javier Serrano kirjoitti:
> There is a group of people doing layout full-time as a service for the
> whole of CERN. Maybe it is because most designs contain many
> high-speed lines with specific layout requirements which autorouters
> struggle with. I'd be interested in the opinion and experience of
> people elsewhere.
Fortunately autorouting can be used for parts of the layout with
critical parts done manually. Some of the slow speed mixed signal
designs work very well by just routing some of the critical signal lines
and power supplies by hand. Freerouting does a pretty decent job
optimizing vias for example. I'm happy doing something else during the
five hours that Freerouting spends on it. On some designs autorouter can
easily handle 90 % of the tracks after just a bit of advice.
> There are things in our plan [2] which we consider
> higher priority, like supporting differential pairs and length
> matching. Regarding freerouting, there are good ideas in it
> independently of the autorouting side of things, which could
> definitely be used in KiCad's push & shove interactive router. Alfons
> deserves credit for his openness and generosity.
I would also consider P&S router an autorouter, just an assisted one and
limited scope ;)
DP and bus routing with length matching is just an easier problem for
the autorouter than the whole board, but it would be very tedious and
error prone for a human. I also consider this kind of limited
autorouting functionality more important for integration than the board
level one.
-Vesa
References