<azonenberg>
then use that as a contact mask in some kind of device with two micrometer stages and an optical microscope or two
<azonenberg>
put the wafer in, apply yellow light, put mask on top, move wafer until it lines up with alignment marks
<azonenberg>
bring into contact, move microscope out of the way, apply UV
<dalias>
another possible process is to commercially fab some sort of gate-array wafers
<dalias>
setup so that you could put them in a modified bluray burner to do the routing
<azonenberg>
That's cheating
<azonenberg>
If you're going to do that might as well just MOSIS everything
<dalias>
well kinda
<azonenberg>
I see the point
<azonenberg>
Realistically, all of the initial work would be on lithography anyway
<azonenberg>
Jeri has the transistor process mostly figured out
<azonenberg>
the big goal is to replace all of her cut-tape litho with photoresist
<rqou>
including consistent Vth?
<azonenberg>
rqou: I dont know how repeatable it is
<azonenberg>
in fact, i don't think *she* does
<azonenberg>
b/c she never made enough on one wafer to do good characterization :p
<azonenberg>
The initial work is going to be on an oxide-coated wafer just making metal patterns
<rqou>
in our microfab class Vth on our transistors would vary from about -0.5V to over a volt
<azonenberg>
Get that down to a micron or two critical dimension
<azonenberg>
heck, i'd be happy with 3-5 um
<azonenberg>
to start
<azonenberg>
Then, try and figure out a good way of getting doped poly down on a surface
<azonenberg>
for gates
<azonenberg>
Alternatively, go with metal gates and do the research on how to make gate material that way
<azonenberg>
At some point during all this build a glove box for critical operations
<rqou>
heh, our class cheated on making the gates
<rqou>
they sent wafers to the "real" cleanroom to do CVD
<azonenberg>
lol
<azonenberg>
I was thinking
<azonenberg>
buy a doped wafer
<azonenberg>
then just sputter it
<rqou>
that's... unorthodox
<azonenberg>
do you think that would give decent uniformity of deposition?
<azonenberg>
It would give a-si
<azonenberg>
i dont know if i could crystallize it with some kind of anneal or something
<azonenberg>
(without destroying the implants)
<rqou>
i don't think that works, but i don't know for sure
<azonenberg>
Another possible strategy is to sputter metals
<azonenberg>
and do metal gate mos
<azonenberg>
Open question
<azonenberg>
I also have to work out details of a process for implantation with spin-on doping
<dalias>
azonenberg, when i said commercially fab gate-array wafers
<azonenberg>
By the time all of this happens, i will probably have access to the SEM, FIB, and RIE at work as well as have Dash etch working in their wet lab
<rqou>
we did implants using spin-on-glass
<azonenberg>
Which will let me cheat a little bit on some process steps, then work out at-home versions of them using simpler equipment
<azonenberg>
i could even do e-beam litho or something
<dalias>
what i meant is doing the gate-array wafers large-scale to sell as a fab-at-home (or in small scale for commercial use) solution
<rqou>
heh, I don't know if UCB even has an e-beam litho setup
<azonenberg>
dalias: i know what you mean
<azonenberg>
its worth thinking about eventually
<rqou>
i think they get masks done outside
<azonenberg>
rqou: We have a SEM and dual-beam SEM-FIB
<azonenberg>
the dual-beam can do patterning in SEM mode too
<azonenberg>
So e-beam litho is totally possible
<rqou>
we probably have an e-beam litho setup somewhere else in the university
<azonenberg>
Some of the guys were tinkering with it a while ago in fact but didnt take it to completion
<azonenberg>
I'm not trained on a lot of that gear yet but it's on the TODO
<rqou>
one thing that UCB has refused to obtain is apparently an ion implanter because supposedly they're too finicky
<rqou>
apparently it's not worth the staff effort
<dalias>
<azonenberg> buy a doped wafer | then just sputter it <-- could you elaborate?
<azonenberg>
you normally buy wafers doped P or N type, then make wells of the opposite type
<azonenberg>
My conjecture was that you might be able to use one of them as a sputter target
<azonenberg>
as a source of Si + doping for producing doped polysilicon gates
<azonenberg>
But i don't know if sputtered silicon will be polycrystalline or amorphous
<azonenberg>
vs PECVD / LPCVD like is commonly used
<azonenberg>
i forget if that typically does just straight Si then separate doping step, or if the precursors are doped during deposition
<rqou>
iirc it's doped during deposition
<azonenberg>
I may save that proble mfor later
<azonenberg>
Once i get litho working the next step will be to get doping working, and test by making some diodes
<azonenberg>
transistors are going to be the very last thing
<rqou>
MEMS? :P
<dalias>
last? how about whole logic gates? :)
<rqou>
mixed logic+MEMS :P
<rqou>
our microfabrication class design was actually supposed to be mixed logic + MEMS
<rqou>
the MEMS doesn't work
<rqou>
:P
<rqou>
apparently one of the professors felt confident that he could fix it to work a few years ago, and he managed to make it less not working :P
the_real_bibor has quit [Ping timeout: 244 seconds]
<rqou>
freenode seems really flaky today...
LeelooMinai_ is now known as LeelooMinai
Lord_Nightmare has quit [Ping timeout: 244 seconds]
lain__ has quit [Ping timeout: 244 seconds]
Lord_Nightmare has joined ##openfpga
dalias has joined ##openfpga
the_real_bibor has joined ##openfpga
JvD_ has joined ##openfpga
azonenberg1 has quit [Quit: Leaving.]
azonenberg has joined ##openfpga
<qu1j0t3>
they're under DDoS
<qu1j0t3>
see @freenodestaff
lain__ has joined ##openfpga
<rqou>
i never understood what the point of ddos'ing an irc network is
<cr1901_modern>
To be assholes.
<qu1j0t3>
I guess that's it
<rqou>
i thought all the irc design flaws that let you take over channels, etc. were duct-taped over a decade ago
<azonenberg>
pretty sure they are
<azonenberg>
The only point i can think of is, if somebody is using freenode as C&C for a botnet somewhere
<cr1901_modern>
they are what? assholes :P?
<azonenberg>
a rival botnet might be trying to take it out
<azonenberg>
Either that or somebody said something somebody else didn't like
<rqou>
both are dumb reasons :P
<azonenberg>
so they decided to DoS the whole network because its easier than flooding the channel :p
<azonenberg>
and it's a sane, though disproportionate and indiscriminate, cyberwarfare strategy in that regard
<rqou>
why does it seem irc servers are so easy to DoS?
<azonenberg>
Because they typically dont have massive pipes
<azonenberg>
because the amount of bandwidth they actually need to use is tiny
<azonenberg>
nobody puts an ircd on a box with a 10gig uncapped pipe
lain__ is now known as lain
<rqou>
what's the reason that irc is a popular botnet C&C mechanism anyways?
<rqou>
doesn't irc usually stand out a lot in IDS logs?
<rqou>
especially on a webserver?
knielsen has joined ##openfpga
tecepe_ has joined ##openfpga
tecepe has quit [Ping timeout: 250 seconds]
<DocScrutinizer05>
((only point i can think of)) occurred to me it might be related to taking down just a particular channel (group), but not botnet C&C but BTC related
<cyrozap>
DocScrutinizer05: BTC as in Bitcoin? Because the original Bitcoin client uses IRC to bootstrap finding other nodes?
<DocScrutinizer05>
yep
<DocScrutinizer05>
not because of clients
<DocScrutinizer05>
it seems to me btc noob as if trading is manually done via IRC chatting now
<cyrozap>
Um, maybe back in 2009, but most people trade on web-based exchanges nowadays.
<DocScrutinizer05>
I heard some announcement as of today that some bitcoin channel needs to stop all activity until freenode recovers
<DocScrutinizer05>
there was actually a lot of noise about all those #*coin channels lately