[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: gEDA-user: FYI [Fwd: [Balloon] Balloon 4]

Hi, all.
  I am that person... Balloon4's design and layout is (at least
partly) my baby.
I'd like to do it in open tools, and for that, gEDA seems to be the
only game in town.

(Balloon4 is roughly credit card sized, and has a TI OMAP(3or4) SOC,
Xilinx FPGA, power supplies, USB, Ethernet, camera, display interfaces
- imagine a Beagle Board XM, with rather more stuff, in about half the
space. There a few design challenges, other than the density - fast
wide buses (DDR2/3 and others), balanced and impedance controlled USB
and Ethernet lines, >Gbps MIPI buses, many power domains and planes.
We need to be able to manufacture it reliably, in volume, and have it
pass EMC standards. We don't have funds to do many aborted runs due to
DRC oopses. We don't have time to burn. The project sponsor needs
working boards.)

Things in gEDA's favour: It exists, it works (in that people are
generating boards), it's live and under active development and use.
It's free. The autorouter is alleged to work.
If I can usefully get other people to collaborate on the design, they
don't need expensive tools. It would be great to have Balloon designed
in open tools. (I'm a happy user of plenty of other open tools, and
electronic design would be fine too. I'd be happy to chip in my Altium
support money...)

Things in Altium's favour:
Momentum. I use it day in, day out. I can get stuff done in it,
quickly and correctly, including boards of this complexity.
I can import TI's Beagle or Panda reference designs and work from
there. No need to generate schematic parts. PCB footprint generation
is speedy, and has integrated 3D models. I also have large libraries
of tried & tested components.
(I do have a full Specctra autorouting license, but Altium's
interactive router is my preferred layout scheme, on any but the most
autorouter-friendly boards).

From a perspective of 'just getting it done and shipped', it's (to me)
clearly going to be faster in Altium. I know the tools, I know the
tools can do the job, and I can import the reference design.

I've seen some reasonably complex boards done in gEDA - but have no
idea how long they've taken. It's possible to do PCBs in MS-Paint or
by cutting tape, given sufficient time, the tools just make things
faster and less error-prone.

So: Given the board spec, and the constraints, would _you_ say it's
sane to use gEDA for this? I'd expect it to take 5-7 weeks end to end,
in Altium, from spec to gerbers (with the usual level of spec
revisions, datasheet reading, component unavailability, coffee
drinking - or 3-4 weeks of uninterrupted work).

Or should I fire up gEDA on something smaller and less visible first?
Anyone want to hold my hand while I do it?


geda-user mailing list