Open Fabs

[grarpamp]

On Tue, Jul 28, 2015 at 8:15 PM, jim bell <jdb10987 at yahoo.com> wrote:
> From: Seth <list at sysfu.com>
> On Tue, 28 Jul 2015 15:40:55 -0700, oshwm <oshwm at openmailbox.org> wrote:
>>
>>> So is anyone working on building an 'openfab' or is it such a big task
>>> that everyone just backs away in horror? :D

Of course you're going to fail if you keep saying no and FUDding yourself.
Is anyone building "cars" or is it such a big task that...
http://www.teslamotors.com/

>>My understanding is that the capital costs involved with building and
>>operating a chip fabrication plant are astronomical, although the
>>situation may be getting better.
>
>>http://spectrum.ieee.org/semiconductors/design/the-new-economics-of-semiconductor-manufacturing

Competition, so what? Leech what's been done before you.

> Even 30 years ago, there were custom fabs that were designed to allow small
> organizations to get chips fabbed.  This may be one of the modern version of
> them:  http://www.globalfoundries.com/manufacturing/manufacturing-overview

You don't need a $50B 1Msqft setup to start making chips, you need
a floor in a warehouse and some people who believe.

> In the mid 80's, they typically purchased older fabs (not state of the art)
> and allowed small companies to prototype their semiconductor designs.
> Today, some of them apparently do near-state-of-the-art production.

Universities have always had fabs too.
The cost to get a basic line going at some tech level is not prohibitive,
it's an adventure.
Open is your differentiator and your ROI, even private runs will come
paying to you because they know your ground up construction and
monitored production process can't inject warez in their silicon.
Can you as a reasonably learned hobbyist go observe an Intel run
through GloFab from start to finish? What about for the chip inside
your phone? Your makerbot? Your RPi? Your USRP? Your own mask?
No? Well... that's a problem.