[libre-riscv-dev] [Libre-silicon-devel] [OT only slightly] STEAM Camp OpenSourceEcology 22jan2020 for 9 days

[David Lanzendörfer]

Hi Marcin

> This is great - https://libresilicon.com/
Thanks :-)

> What are the current equipment needs that Open Source Ecology can help
> with?We design and build industrial machines, and on our roadmap for 2022
> is production of photovoltaics. How much overlap is there for equipment
> needs?
In general, the operation of a clean room, even when minimal sized, is pretty
expensive because of all the chemicals and the power consumption of the
machines.
What would be cool would be simplistic plasma etcher machines with batch load
capability (from wafer cassette), as well as coaters and developers for the
application of the resist.
Then of course we need some furnaces, CMP engines and so on.
Maybe you go through the documentation I wrote to get a better idea.

> Great diagrams at
> https://download.libresilicon.com/process/v1/process_hightech_steps.pdf -
> thanks for producing that.
Thanks for the compliments and thanks! Took me quite a while to make those
with tikz (It's all in LaTeX).

> For some of the machines - Ion implanter • Plasma etcher • Sputter engine
> (Metal deposition) • Diffusion furnace • CVD/LPCVD machine • Exposure unit
> - are these the high cost items that constitute the majority of the $1B
> cost of  a fab? Or are there other major costs? My starting assumption is
> that we can reduce the cost of a $1B fab 100x with open source design and
> collaboration.
Yeah. Those machines cost quite a lot even when bought second hand,
and also the rent to already established labs, like the one at HKUST or
the one in Lisbon, I'm probably going to use instead, (Portugal is a fantastic 
country, totally looking forward to go there), is pretty high.
In order to set up a lab you'll have immense initial costs, just because
manufacturing the required equipment will be very expensive, even if
it's open source, you will still require the man  hours and materials to
build them.

> Are you aware of any viable open source versions of the above equipment?
Nope. That's why we have to design them from scratch.

> Regarding smart-contract based rewarding of IP contributors in your
> Lightning Talk <https://download.libresilicon.com/propaganda/talk.pdf> -
> does that mechanism have traction yet?
Nope. That was more of a strategic move in the hope to get some crypto 
hippsters involved into the financing. Didn't work out so well.
On the other hand, having the checksums for the verified GDS2 files in a
decentralized database makes it way harder for an attacker to modify the
layout somewhere on GitHub after said layout has been silicon verified
and checked for security issues.

-lev