MIME-Version: 1.0
In-Reply-To: <525AC3B0.5080703@penguindevelopment.org>
References: <525AADB0 DOT 8030508 AT penguindevelopment DOT org>
	<525AB9A5 DOT 2060109 AT estechnical DOT co DOT uk>
	<525AC3B0 DOT 5080703 AT penguindevelopment DOT org>
Date: Sun, 13 Oct 2013 17:10:58 -0400
Message-ID: <CAM2RGhQSY2EOj+Kz+C=jV_QvVOpSs2tXJ17RrMW8p0DK2HA0cQ@mail.gmail.com>
Subject: Re: [geda-user] Dealing with internal connections
From: Evan Foss <evanfoss AT gmail DOT com>
To: geda-user AT delorie DOT com
Content-Type: text/plain; charset=ISO-8859-1
Reply-To: geda-user AT delorie DOT com
Errors-To: nobody AT delorie DOT com
Precedence: bulk

For things like FPGA I/O lines the kind of functionality you are
looking for would be very helpful. Shame it does not currently exist.

I have the same situation and I use the 4 pin symbol if only because
the symbol should mirror the part.

On Sun, Oct 13, 2013 at 12:00 PM, Peter <link AT penguindevelopment DOT org> wrote:
> On 13/10/13 17:17, Ed Simmons wrote:
>>
>> On 13/10/13 15:26, Peter wrote:
>
>
>> Perhaps you could instead use a symbol with four pins like you have
>> 'drawn' above... then you could also use the footprints with pads
>> numbered 1-4. You could show the connection in the symbol easily, with
>> the design intent clearly set out in the schematic.
>>
>> When it comes to PCB, it would still complain about the pads not being
>> joined, but ignoring the rats between switch pads doesn't seem too awful.
>>
>> When I complete a PCB layout, it's quite common for me to short ground
>> planes to connector shells... these don't always have the corresponding
>> pin in the schematic. Usually by the second revision of the design I've
>> tidied these up and added the appropriately numbered pins to symbols to
>> show these connections and prevent PCB from complaining. ;-)
>>
>> Hope that helps...
>>
>> Ed
>>
> Using a 4-pin symbol is what I'd like to avoid having to do, because when
> routing the PCB, I'll sometimes decide it's useful, for example, to connect
> only pins 1a and 2b -- but then, after some reconsideration, I'll suddenly
> see a more efficient way that instead requires pins 1b and 2a to be
> connected, or requires "hijacking" the internal connection between pins 1a
> and 1b as a bridge across tracks routed underneath the switch. All of those
> things would require changing the schematic during editing.
>
> Your idea of going with the "wrong" version first and cleaning it up in a
> later revision does sound good, though; I could just use two
> nearly-identical footprints, one with multiple identically-named pins for
> development, and one with different pins for "release"; then I'd just have
> to change the schematic to match the PCB once, and using "import schematics"
> would hopefully magically "fix" the routing.
>
> Cheers.
>
>
> Peter
> --
> http://www.penguindevelopment.org/ -- Free and open source software and
> hardware


-- 
Home
http://evanfoss.googlepages.com/
Work
http://forge.abcd.harvard.edu/gf/project/epl_engineering/wiki/