Searching \ for '[EE:] Government chip foundry' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/index.htm?key=government+chip
Search entire site for: 'Government chip foundry'.

Exact match. Not showing close matches.
PICList Thread
'[EE:] Government chip foundry'
2004\06\26@013411 by Charles Craft

picon face
Too many spy movies on the dish and too many years of watching X Files.   :-)

What's it take to set up a small chip foundry?
I think there have been threads on the PIClist about the government chips that self destruct if
you penetrate the package the wafer is in.

A commercial chip plant is pretty huge but if you didn't care about quantity or speed how small could you make it?
Aren't there universities where students make their own chips?
Can you make chips in a few thousand square foot warehouse?



http://www.utwatch.org/oldnews/aas_sematech_6_15_03.html

Bid to keep Sematech invokes national security

Proposal for government foundry shows extent of Austin effort to head off a move to New York
By Chuck Lindell
AMERICAN-STATESMAN STAFF
Sunday, June 15, 2003

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics

2004\06\26@050026 by Jason Harper

picon face
> Aren't there universities where students make their own chips?

Chips by students in a VLSI design class are generally made via Mosis.org,
not on-site at the university.  Mosis is a clearinghouse that combines many
small chip orders onto single wafers that are manufactured at a commercial
fab.  It looks like they'll take orders from anyone (starting at several
thousand dollars), not just students.
       Jason Harper

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics

2004\06\26@055121 by William Chops Westfield

face picon face
On Jun 25, 2004, at 10:33 PM, Charles Craft wrote:

> What's it take to set up a small chip foundry?

To make what?  And what do you consider part of the "foundry"?  I think
these days the raw wafers are made someplace different than where they
get turned into actual integrated circuits.

If you get to start-of-the-art ASICs or similar, you find that there
are only one or two foundries IN THE WORLD that can make them.  This
does nasty things to your leadtime; everyone else wants THEIR asics
made too.


> Aren't there universities where students make their own chips?

The VLSI chips are usually shipped out to a commercial fab.
I think there are universities doing experiments with odd materials,
odd topologies, and FAR from state-of-the-art geometries (things like
MEMS, biochip stuff, organic circuits; stuff thats so far beyond "art"
that it's still "science."

> Can you make chips in a few thousand square foot warehouse?
>
Chips that do what?  Compete with a P4?  No way.  Solar cells?
Probably.  60s (?) technology RTL logic just for grins?  Maybe.
Don't forget all the OSHA, zoning, insurance, and legal issues you'll
face since you'll have all that hazmat stuff on site...  ("few thousand
feet" is REALLY tiny for a "warehouse."  That's a "small" office
building, or part of one.  After all, a reasonable sized cubical is 100
sq feet or more...)

BillW

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics

2004\06\26@100300 by Dave Tweed

face
flavicon
face
Charles Craft <spam_OUTchuckseaTakeThisOuTspamMINDSPRING.COM> wrote:
> A commercial chip plant is pretty huge but if you didn't care about
> quantity or speed how small could you make it?

I'm not sure, but I think you could go a long way towards making one chip
at a time if you had a supply of virgin wafers and a FIB (focussed ion
beam) machine. It allows you to do things like deposit materials, cut them
away and do ion implantation without the need for resists, masks, etc. You
might still need a separate oven to grow oxide layers, etc. It would be
*very* slow going, though.

-- Dave Tweed

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics

2004\06\26@114055 by M. Adam Davis

flavicon
face
The university of Michigan has its own foundry, but I suspect it's not
nearly as mechanized as a typical industry foundry.  Most of the wafer
movement is done by hand from cell to cell.  It occupies a portion of
the EECS building (you can see maps which include the clean rooms at
http://www.eecs.umich.edu/ .

You can make it fairly small - I suspect one could fit a simple foundry
in a small home (1000 sq feet) but getting the utility hook ups (not to
mention the large quantities of chemicals, water, etc) may be problematic.

But the question I have is why?  The reason UofM has their own VLSI lab
is (aside from teaching) for specialized products that would be
difficult to farm out since they have to be created by hand anyway
(terahertz radio circuits/antennas, neurological interface devices, etc).

You can form most stuff you'd need with an FPGA and then farm that out
for quantity.  If you were planning on doing analog and digital on the
same silicon than a simple foundry simply isn't going to cut it.

-Adam

Charles Craft wrote:

{Quote hidden}

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics

2004\06\26@190212 by T.C. Phelps

picon face
> Aren't there universities where students make
> their own chips?

I took two IC design courses within the last year at
the University of Calgary (Alberta, Canada) and we
were told that all Canadian universities get chips
fabricated through CMC, the Canadian Microelectronics
Corporation (some US universities may have in-house
facilities... I don't know). It's rare for an
undergrad chip to be produced, but one prof had us
design our projects with a 2 micron process because he
thought the University could afford to have one
project fabbed. He himself was doing research on MEMS
applied to biological systems and I think he said he
was doing 2 micron as well because it was the
cheapest... but at any rate, everything up here gets
shipped off.




__________________________________
Do you Yahoo!?
New and Improved Yahoo! Mail - 100MB free storage!
http://promotions.yahoo.com/new_mail

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics

2004\06\26@211201 by hilip Stortz

picon face
the nsa actually operates their own chip foundry, underground.  i saw it
on one of the investigative tv shows (i.e., a semi-real one, not "hard
copy").  if you buy wafers, and do all the design somewhere else, and
only have a small space for packaging (i.e. small volume) i don't think
it would take that much.

for very, very low volume i'd think 10,000 ft would do it, provided that
everything other than the actual chip building was done somewhere else.
for low volume, you wouldn't be dealing with very large quantities of
hazmat stuff, though you would have some (and there are generally limits
on reporting/using hazmats, at low levels it isn't too hard to comply i
don't think, provided you send all of your waste off to an approved
disposal site which even some of the lab supply companies will do,
taking several small bottles in a 5 gallon bucket etc., providing
everything is labeled and there aren't radioactive or other "special"
waste involved).  some of the newer wafer fab stuff is completely sealed
and a human never goes in, wafers and chemicals go in and out.  some of
these fab units are fairly small.  as far as large asics go, there may
only be a few sites in the world that make them, but those sites make
them in high volume and for many customers, and there would be more
sites/larger sites if demand dictated it.

the number of facilities really isn't a good indicator of how hard
something is to do in low volume, for instance half the world supply of
epoxy resins came from one "plant" until it burnt down (very possibly in
an attempt to create shortage and drive prices up, this has been proven
with some other chemical plant "accidents").

a lot of it would depend on knowing what you were getting into and only
providing resources/space for the things a given project would actually
need.  i have heard of people making transistors in their garage...but
for chips a clean room is a must though they can be small and some are
even modular construction now and go up inside another building, like a warehouse.

best bet would be to subscribe to one of the relevant industry journals
and check out the sites of companies that make the equipment.  of course
this equipment is very expensive, both because of quality and
sophistication and because it's meant to earn the owner big money.

for fairly large, far from state of the art geometries, i'm sure you
could do it in a small warehouse - depending on your goals as far as
making money, doing research, etc.   the basic process systems are
fairly simple, other than the need for cleanliness and purity, but all
the raw materials are readily available, most even in very small
quantities.  you definitely need some knowledge of industrial
chemistry/hygeine before starting something like this, many of the
materials are hazardous and tricky to handle and shouldn't be handled by
people who don't know what they are getting into, for waste disposal and
safety issues alone.

chips that self destruct aren't any big deal, simply including some
nitrates and a way to heat them in the package will incinerate the
silicon nicely and heat any remaining material to melting and make it a
blob (which was accidentally discovered by some materials researchers
who i think were using germanium and nitrates to build devices, a little
much current and it went poof).  such chips are rare outside of highly
secure equipment, because they are rather painful to develop and
manufacture with, not to mention maintaining such systems is very hard.
the majority of things that claim or imply damage if the case is opened
are just red-herrings designed to scare the casual cracker away, it's
really not practical to make consumer or mass market goods that are that
sensitive.  any self destruct mechanism built into equipment has a good
chance of accidental or mis activation which is a deterrent to using
such technology.  for instance, if the average users computer wiped the
drive or otherwise self destructed to protect sensitive data, system
crashes would be a real problem as many might trigger software/hardware
designed to detect tampering or probing attempts.  just think what would
happen when a user forgot a password and guessed too many times, a
secure system should assume it's been stolen or otherwise compromised
and destruct before it's too late.

Charles Craft wrote:
{Quote hidden}

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics

2004\06\28@094439 by Paul James E.

picon face
Yes, you probably could, if you were willing to do a lot of the processes
by hand that nowadays are done by machine.   And the machine that do these
processes are not small or cheap.   Another thing is that the warehouse
would have to be virtually vibration free.   This is because unless you
are using large geometries in the chips you are making, just a little
vibration in the floor of the building, and you'll throw off your photo
lith processes.   This will cause the masking processes to be out of
focus, and will ultimately cause components to at bese be out of spec.
Or at worse, not work at all or be shorted to some component it isn't
supposed to be shorted to.

In other words, yes it's possible, but it will be expensive, labor
intensive, and difficult, but it can be done.   The real question
is "would it be worth it"?   Unless you are making some chip that isn't
already available, or is really application specific (ASIC), then the
answer is probably no.

Hope this helps.


                                               Regards,

                                                 Jim





{Quote hidden}

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email .....listservKILLspamspam@spam@mitvma.mit.edu with SET PICList DIGEST in the body

More... (looser matching)
- Last day of these posts
- In 2004 , 2005 only
- Today
- New search...