azonenberg changed the topic of #homecmos to: Homebrew CMOS and MEMS foundry design | Wiki: http://homecmos.drawersteak.com/wiki/Main_Page | Repository: http://code.google.com/p/homecmos/ | Logs: http://en.qi-hardware.com/homecmos-logs/
GeorgeHahn has joined #homecmos
kristianpaul has joined #homecmos
kristianpaul has joined #homecmos
nmz787_i has quit [Read error: Connection reset by peer]
XgF is now known as LDREX
LDREX is now known as XgF
nmz787_i has joined #homecmos
nmz787_i has quit [Client Quit]
X-Scale has joined #homecmos
GeorgeHahn has quit [Read error: Connection reset by peer]
mrdata has quit [Read error: Connection reset by peer]
mrdata has joined #homecmos
nmz787_i has joined #homecmos
nmz787_i has quit [Client Quit]
nmz787 has joined #homecmos
chris_99 has joined #homecmos
<berndj> SpeedEvil, i've often wondered if, in say 1960, if you had a "money is no object" budget if you could make semis that are decades ahead of their time
<berndj> that avogadro project made me think of it again, in a sense that involved making a boule that would be useable for 2030s' chips
<SpeedEvil> It gets asymptotically harder..
<SpeedEvil> For example - to make in 1960 a intel I3 haswell, you need to first invent really powerful computers to handle the design.
<SpeedEvil> At some point simply throwing labour or money at stuff doesn't help - you're limited on the cognitive advance.
<SpeedEvil> Or tools to make tools to make tools.
<SpeedEvil> As technology regresses with time, this results in stuff asymptotically going insane.
<SpeedEvil> I can see in principle getting a generation or several on with infinite funds, but much beyond that, you're going to have to advance much of global technology first.
<SpeedEvil> http://www.baenebooks.com/chapters/0671578499/0671578499.htm?blurb - I found amusing. Small town in america with only its own resources gets dropped in 1630s germany in the middle of the hundred years war.
<SpeedEvil> Issues around how to start an industrial revolution arise.
<SpeedEvil> For example - how on earth do you make large stainless steel vessels for a chemical industry.
<_Sync_> SpeedEvil: to make current gen semis in the 60s you'd need a whole lot more materials science
<_Sync_> not so much the machines
<SpeedEvil> That too.
<SpeedEvil> I was assuming that you magically knew how stuff should work partially
<SpeedEvil> If you don't, it gets even harder, as you're going to try stuff that we know know won't work
<berndj> SpeedEvil, yeah, i didn't even mean making a fancy cpu; just something cognitively accessible back then, maybe a gigasample ADC or something
<SpeedEvil> gigasample ADCs were available.
<SpeedEvil> Sort-of.
<berndj> and your point about needing computers to design computers is what makes me an anti-singularitarian
<berndj> single-chip ADCs?
<SpeedEvil> No.
<SpeedEvil> sampling oscilloscopes
<_Sync_> that's something completely different than an adc
<berndj> i meant more like the scenario where earth had jupiter's gravity well and we were doing the apollo project
<berndj> where you'd want to reduce the mass of all the electronics even more than you already want to
<SpeedEvil> There is nothing in principle impossible about nuclear propulsion - it's just harder.
<berndj> i think i read somewhere that there was literally tons of electronics (not even counting the wiring) in the rocket?
<SpeedEvil> You absolutely can't do apollo without nuclear propulsion.
<SpeedEvil> Even with a zero mass electronics and computer budget
<SpeedEvil> Ballparking numbers, the saturn 5 launched a mass of some 5% of its mass to the moon. A delta v of about 10km/s. Jupiter you need 50 km/s or so
<SpeedEvil> This is a required mass ratio not of 20, but of 20^5. For a 5 ton payload, you'd need 3200000 tons of rocket.
<SpeedEvil> err 16000000
<berndj> haha. /me handwaves some more to focus the problem better
<berndj> i'm not fixated on reaching 2010 tech in 1960, since there's some stuff in the last 20 years or so that i'd consider close enough to "magic" that the problem becomes uninterestingly hard
<berndj> so maybe just to the apex of "normal" photolithography where it's just lasers and masks, not x rays and electron beams
<berndj> that said, they had electron beams way before lasers
<SpeedEvil> I note we've not yet got to x-rays yet.
<SpeedEvil> They're doing sex patterning immersed in oil with extreme UV or something.
<_Sync_> I wonder when we will get nano embossing
<SpeedEvil> I don't know, but when we do, we will do it like a boss.
nmz787_i has joined #homecmos
nmz787_i has quit [Ping timeout: 246 seconds]
nmz787_i has joined #homecmos
RoBo_V has joined #homecmos
<RoBo_V> what you guys create silicon cmos wafer at home ?
<SpeedEvil> Creating wafers is generally insane - it's possible to buy cheaply
<RoBo_V> so what homecmos stands for
<_Sync_> the plan to do so
<SpeedEvil> MAnufacturing chips on wafers
<RoBo_V> have anyone done it so far at home ?
<nmz787_i> someone in here got some electrical characteristics embedded in a wafer, and I think they did it all at home
<nmz787_i> personally I'm here mostly for micro/nano fab related chat
<RoBo_V> nmz787_i: u in fab lab, or just learnign about it ?
<nmz787_i> i am a programmer, but I have a biotech degree and want to do lab-on-a-chip stuff
<nmz787_i> I just bought a used/needs-work SEM
<nmz787_i> lots of electronics and physics knowledge./intuition needed in my opinion
<RoBo_V> biotech degreee and programmer both are very diff indeed
<SpeedEvil> Not so much anymore.
<SpeedEvil> Computational ology is getting to be a major thing.
<_Sync_> almost all tech degrees are now coders
<_Sync_> if you look at what people actually do in the end
<davidc__> just domain-specific programming
<nmz787_i> heh, I work at a semiconductor company though, doing PCB design :P
<nmz787_i> and most of my biotech was heavily lab course focused
<nmz787_i> i took a few programming classes as part of my bioinformatics minor/option thing
<RoBo_V> I seee
<nmz787_i> instrumentation has always been my interest from the first summer internship I had where the manual labor was just insane and I knew a webcam and some code could do my job faster and better
<RoBo_V> nmz787_i: thats cool
RoBo_V has quit [Ping timeout: 256 seconds]
nmz787_i has quit [Quit: Leaving.]
nmz787_i has joined #homecmos