Removing Glamo from GTA02 (Was: Welcome)
werner at openmoko.org
Mon Apr 20 09:05:28 CEST 2009
Rask Ingemann Lambertsen wrote:
> My view: A NOR-less, but still unbrickable, solution just won't be ready
> in time for a "GTA02.1" release, see you in "GTA02.2". Same thing with
> s/Calypso/MC75i/. You just couldn't have a prototype with that in two weeks
> from now.
Err, even if you had just finished schematics and layout this very
moment, you would at best have the raw PCBs in two weeks ;-)
There are two possible paths:
1) Openmoko releases all the files in Mentor PADS, i.e., not only the
Gerbers (which are required) but also the files in Mentor's
proprietary format. Then someone with access to PADS can make the
2) Everything is redone in some other EDA system. That's schematics,
layout, and parts library (symbols, footprints, maybe packages).
For 1), I'd calculate about one week to make the schematics changes and
to review them, then one more week for the layout.
For 2), 1-2 weeks to copy and review the schematics, then the schedule
from 1). (Assuming that layout work can run in parallel with schematics.
Otherwise, add 1-2 more weeks.)
Then you need to make the PCBs (two weeks), find an SMT factory, book
a slot, get the components, do the SMT (1 day), ideally do some basic
testing on site, then get the boards to developers.
So we're talking about 5-8 weeks until first prototype best-case, and
I don't know how long the setup time for the SMT factory is. (You need
to bring in the components, get them in to their system, etc.)
If taking route 2), the first prototype will almost certainly have some
fatal flaws, so the cycle repeats with the next prototype ready after
Note that this assumes that there are no delays going to the PCB and
the SMT manufacturer. If a lot of people sign up for this, it may be
possible to parallelize some tasks, but I doubt you can shave off
more than a week.
The big task is to take ownership of the design and the process. The
actual changes are almost trivial.
More information about the Gta03