X-Authentication-Warning: delorie.com: mail set sender to geda-user-bounces using -f
X-Recipient: geda-user AT delorie DOT com
Date: Thu, 30 Jul 2015 03:30:46 GMT
From: falcon AT ivan DOT Harhan DOT ORG (Spacefalcon the Outlaw)
Message-Id: <1507300330.AA07472@ivan.Harhan.ORG>
To: geda-user AT delorie DOT com
Subject: Re: [geda-user] Converting footprints from PADS to PCB
Reply-To: geda-user AT delorie DOT com

Erich Heinzle wrote:

> Even better than using a background image is snapping to features.
>
> The gerber -> footprint method allows new tracks to be snapped to the
> locations of pad locations derived from the gerber.

For the quadband version the components will need to be arranged in a
totally different way than how they are arranged in Openmoko's triband
design, so the old locations of footprints relative to each other
won't be of any use, only the individual footprints in themselves.

OTOH, if we are going to build the first prototype as triband, copying
Openmoko, then I say that copying the original element pad locations
and then having to redraw the traces by hand isn't enough automation -
why not copy/translate the connecting line/polygon/via geometry from
the historical gerbers as well?

Speaking of vias, they are blind and buried in the original.  The
drill structure of Openmoko's 8-layer board is L1-L2, L2-L3, L3-L6,
L6-L7, L7-L8.  From what I remember, someone made a patch for PCB to
support blind & buried vias, but it was never integrated - probably
because it wasn't the general solution which the purists demand.
Translating Om's modem layout into PCB would require resurrecting that
patch and using it in a private fork.

As an additional observation, I don't know of any other Calypso GSM
cellphone or modem design that was done without blind/buried vias -
all of the ones I've examined use them.

SF