#homecmos IRC log for Wednesday, 2011-12-07

« Monday, 2011-12-05 Index Thursday, 2011-12-08 »
--- Wed Dec 7 201100:00
bart416azonenberg, http://www.keil.com/forum/11867/11:31
bart416^you should totally try to gravedig that thread :')11:31
azonenberglol11:32
bart416Also, any suggestions for a transistor level flank triggered D latch?11:46
bart416*rising edge11:46
smedinghahaha that thread11:58
smedingbart416: do you need a schematic? i can rip one from a book of mine11:59
bart416smeding, anything is welcome :P11:59
bart416My design is not that uhm, great :p11:59
bart416Uses way too much transistors in my opinion11:59
smedingeh, latches use a fair bunch of them usually11:59
bart416Too much heh :P12:00
smedingthe one here uses 22 transistors12:01
smedinglet me see what the actual SoG implementation for our uni project does12:02
bart416well, if you're going VLSI or better it's not much of an issue if it works fast12:02
bart416The thing is, I don't want to be using 160 transistors per 8 bit word12:03
smedingdo you want one with a reset12:03
smedingand if so, sync or async?12:03
bart416that doesn't matter that much actually12:03
smedingi'll take a picture12:03
smedingsec12:03
smedinghttp://imgur.com/VbG3812:05
smedingthat's an implementation that's actually used12:06
smedingit's for a 1.6u SoG process though, hehe12:06
smedingbut i don't think it should matter all that much12:06
bart416Urgh, I'm using BJTs :P12:07
bart416I have enough books on CMOS design to build a desk out of them by now I think12:07
smedingwell you should have mentioned that :p12:08
smedingi don't have any books on TTL or anything12:08
bart416It's RTL as well :P12:08
bart416TTL is tricky in discrete components for reasons I'm so not getting into lol12:08
smedingah12:09
bart416I'll leave it at you're quite often largely dependent on somewhat similar hFEs12:09
bart416I can't count on that12:09
smedingah yeah12:09
bart416I'm using 2N3904 bjts in my design12:10
bart416so min hFE is about a hundred12:10
bart416It should work as long as hFE stays above 50 or so12:10
azonenbergWell you're also dependent on multi-emitter transistors if you build them the classic way12:13
azonenbergI like CMOS12:13
smedingme too12:13
azonenbergAnd I want to build it :p12:13
bart416BJT is cheaper and actually faster in this case12:13
azonenbergSomething as complex as a FF might take a while to be able to make12:13
bart416It's easier to design a "high speed" BJT circuit than a MOS one12:13
smedingECL!12:13
azonenbergI do wonder if Jeri's process would work with BJTs12:14
azonenbergBecause I have a pretty good idea of how to do a home MOS process12:14
azonenbergI actually want to try prototyping a CD4000 series chip on a PCB with discrete FETs12:15
azonenbergmake sure i have the circuit right12:15
azonenbergUsing a standard-cell based design12:15
azonenbergBut actual fab is still a ways out12:15
azonenbergI want to do a comb drive soon12:16
azonenbergJust have to find the time to actually work on it ;p12:16
azonenbergAnd make a new spin coater controller12:16
bart416Best I've got right now is actually the very old trustworthy digital design, replace gates with transistor equivalents12:16
bart416that'll do it with 18 transistors12:16
bart416I fear I'll have to switch my design to 4 bit12:17
bart416or 6 bit12:17
azonenbergbart416: I would not attempt anything more complex than a single CD4000 chip12:17
azonenberguntil you have VLSI down well12:17
bart416VLSI?12:18
bart416I'm building discrete...12:18
azonenberg... oh12:18
bart416That's why I'm worried about transistor count12:18
bart416I don't give a damn about speed at this point actually12:18
bart416I'm happy with 1kHz in fact12:18
azonenberglol12:19
bart416My current method should alllow me to go past 100 kHz though!12:20
azonenbergVery nice12:20
azonenbergBtw, I was just thinking12:20
azonenbergI can hit 12um on a film mask12:20
azonenbergdesign rule12:20
bart416yes...12:20
azonenbergif i feed that into 10x reduction thats 1.2um design rules12:21
azonenbergOn an area about 1.6mm across12:21
azonenbergWhich should be enough to actually build something of useful complexity12:22
azonenbergI think some time in the next week-ish i am going to try making a flip chip BGA12:22
azonenbergSomething of trivial complexity, not even projection litho12:22
bart416If you do that you might be able to get some donations (to buy new equipment)...12:22
azonenbergjust contact at few-hundred-micron feature sizes12:23
azonenbergIOW an inert packaging dummy12:23
azonenbergbut if i can conduct a signal from balls A1 to A2, B1 to B2, C1 to C212:23
azonenbergand not short between the pairs12:23
azonenbergThat will tell me the packaging technique is sound12:23
bart416and you won't need to be nearly as careful at those feature sizes...12:28
azonenbergYeah12:28
azonenbergThe goal is to build a FCBGA in an amateur process12:28
bart416and you could make 7400 series ICs already12:28
azonenbergEven if the device is inert12:28
azonenbergAnd i cant do 7400 yet12:28
azonenbergi dont have dopants12:28
azonenbergor a furnace12:28
bart416oh :(12:28
azonenbergI mean, my litho is good enough12:28
azonenbergI just have to get the other tools12:28
azonenbergits probably a few more months of work12:28
azonenbergi'm right on the edge12:29
bart416I guess I'll just ask a professor about it tomorrow12:39
bart416btw, I pissed somebody off at an auction yesterday evening azonenberg12:49
bart416I did a Google12:49
azonenberglol12:49
azonenbergexplain?12:50
bart416Was placing bids based on physical and mathematical constants :P12:50
azonenbergLol12:50
azonenbergI approve strongly12:50
bart416So did the fellow engineering students that were there, sadly we weren't in charge of the auction :P12:50
azonenberglol12:50
bart416So the polsoc major that was doing it got pissed cause he couldn't remember the numbers we were throwing at him12:51
azonenberglol12:51
bart416ok, in his defence he might not know that the square root of pi is 1.77245412:52
bart416But that's no excuse to suck that much :P12:53
azonenberglol12:53
azonenbergsqrt(pi)?12:53
azonenbergWhy not sqrt(2)12:53
azonenberg1.414 ish12:53
azonenbergor were you trying to beat 1.5012:53
bart416Well, multiplied by 10 that is :P12:53
bart416So 17.72 euro was our bid lol12:53
bart416ofc we were shouting "TEN TIMES THE SQUARE ROOT OF PI"12:54
bart416Actually, that might have been why they kicked us out halfway12:54
bart416Well, two reasons12:54
bart4161) We were buying everything12:54
bart4162) We were being assholes12:54
bart4163) Constant math jokes12:55
bart416Number 3 not being a real reason to get kicked out12:55
bart416(Number 1 is problematic cause it's for some good goal and they were auctioning dinner seats :P )12:56
bart416But we bought the entire table of 20 people with 7 xD12:56
bart416well yeah, 20 pairs of people I meant sorry12:56
bart416Considering most of us don't even have a relationship we'll take up 40 places with 7 people12:57
bart416And that all for the price of 173.8 euro in total...12:58
bart416Considering the restaurant it's a bargain lol12:58
azonenberglol12:59
bart416The polsoc idiot also managed to sell his own bike without realising it :')13:00
azonenberglol13:00
bart416One of his "friends" put it in between the items13:00
azonenberglol13:00
bart416And it was described on the list as "bicycle, used"13:00
azonenbergDummkopf :p13:00
azonenbergWhat'd you buy it for13:00
bart416And he was too lazy to bring it up to the stage so he never noticed it13:00
bart416Nah, we didn't buy it13:00
bart416We didn't know it was his bicycle until the "friend" in question told us13:01
bart416I'll just leave it at, it went cheap13:01
bart416There were some nice items though13:05
bart416Several rare books13:05
--- Thu Dec 8 201100:00
« Monday, 2011-12-05 Index Thursday, 2011-12-08 »

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!