#homecmos IRC log for Saturday, 2011-08-06

azonenbergTomorrow night's plan - spin coat a quarter of a <110> wafer in Ta2O5 and bake06:14
azonenbergin preparation for a Cu evaporation run later in the week06:14
azonenbergalso, run some tests of dilute SC2 etch rate characterization06:14
azonenbergsee how dilute an etch is reasonably fast but still controllable06:14
bart416azonenberg, we just tried to launch a little small rocket13:37
bart416But it started13:37
bart416And it shorted the triggers for the stages13:37
bart416so the moment we turned on power all stages fired at once :D13:37
bart416things left: first stage nozzle was still on the ground13:38
bart416nose cone a few meters further13:38
bart416well, at least that means the explosive bolts for the first stage nozzle worked13:39
bart416well yeah with bolts I mean rings, lol13:39
bart416we use two rings, with spacing in between them in the middle13:40
bart416and then clip them together13:40
bart416But the system isn't water tight it seems :(13:41
azonenbergexplosive bolts? What was your separation charge?17:45
azonenbergAnd i wouldnt be sure it's water, could be other interference17:48
azonenbergnext time try doing a power-on test with empty motor casings17:49
azonenbergi'd also strongly suggest a software controlled ignition sequencer of some sort17:49
azonenbergWhich might allow you to isolate from interference better17:50
azonenbergalso, try using detonators that are more resistant to EMI17:51
bart416the controllers are ATtinys18:29
bart416and the separation charge was blackpowder18:30
bart416it's powerful enough for this18:30
bart416cause the energy has nowhere else to go18:30
azonenbergyeah, it'd work18:53
azonenbergAnd is cheaper/safer/easier to get access to than a high explosive lol18:53
azonenbergWhat did you use as the ignition system?18:54
azonenbergOne of the problems with black powder is that especially the finer grades are afaik pretty static sensitive18:54
azonenbergcompared to something like RDX18:54
bart416High voltage spark19:11
bart416we put a small CCFL inverter in the nose cone19:12
bart416and distribute that throughout the rocket19:12
bart416and switch it on when necessary19:12
bart416It's easy, and gets the job done19:12
bart416And its reliable19:12
azonenbergProblem is that sounds very ESD sensitive19:13
bart416The entire setup, yes19:13
bart416If water gets in it, it's sort of fucked19:13
bart416High voltage + water + explosives nearby = bad19:14
azonenbergi wasnt talking water so much as static19:18
bart416static isn't that problematic19:18
azonenbergi think you need a less sensitive detonator and/or an anti-ESD trigger device19:18
bart416all the metal parts are grounded until take off19:18
azonenbergnot necessarily to the extreme of an EBW19:18
bart416actually until about 5 meters altitude it's grounded19:19
bart416Then the launch wire snaps lose19:19
bart416aaaargh, stupid shortcut, lol19:19
bart416anyway, its easy enough to do that19:19
azonenbergSo what was your problem exactly?19:19
azonenbergthe water shorted out something in the trigger mechanism?19:19
bart416water got near the electronics19:20
azonenbergi see19:20
bart416and short circuited the high voltage19:20
azonenbergyeah, that'd do it19:20
azonenberghow were you switching the HV?19:20
bart416cheap relays19:20
azonenbergAnd i bet they werent rated for the voltage?19:20
bart416They are weirdly enough19:21
azonenbergwell in any case you've now learned about a few of the problems involving explosives19:21
azonenbergone of which is that they dont play well with others19:22
bart416I know this is one of the problems :P19:22
bart416I've been doing this for years with a bunch of friends19:22
bart416It's just that it suddenly started raining19:22
bart416so we decided to just go ahead and launch19:22
bart416But water got in the wrong spot it seems19:22
azonenbergyeah - if you arent sealed and rain-rated i'd say rain is enough to scrub the launch19:22
bart416well, it only did a few drops19:22
azonenbergbad luck then?19:23
bart416it wasn't even really raining yet19:23
bart416I guess19:23
azonenbergIts also possible there was a short elsewhere19:23
azonenbergunrelated to the water19:23
bart416Unlikely, we check the wiring before we load fuel19:23
bart416Doesn't matter anyway19:23
azonenbergin that case then i guess it was water19:23
azonenbergNothing left of the rocket? :p19:23
bart416Not really19:23
bart416the nozzle and nose cone19:24
azonenbergany damage beyond the rocket itself? Launch pad, spectators?19:24
bart416nozzle was pushed against the ground and it's made out of graphite with a steel ring, so yeah not much that can happen to that19:24
bart416launch pad is a slab of concrete :P19:24
bart416With a steel rod19:24
azonenbergi meant the rod and related stuff19:24
bart416no real damage19:24
bart416some scorch marks19:25
azonenbergWell, if you can have the rocket literally blow up on the pad and not do any harm19:25
azonenbergit seems you picked a good launch site19:25
bart416open field19:25
bart416wet ground19:25
bart416Nothing in the immediate area19:25
bart416We were all standing at 150m distance19:25
azonenbergWere the motors APCP or black powder?19:26
azonenbergand what was the total magnitude of the charge?19:26
azonenbergMost of the rockets i've done are single stage with one class C black-powder motor19:26
bart416I didn't mix the fuel so not sure what was in it19:26
bart416Probably black powder combined with something else19:26
azonenbergYeah, i dont trust my explosives knowledge far enough to make my own propellant charges19:27
bart416I avoid making my own solid propellants19:27
azonenbergi limit myself to premade motors from a respectable company19:27
bart416blackpowder I'll do19:27
bart416in small quantites19:27
bart416worst that you get from small quantites of blackpowder like I mix would be some minor burns19:27
bart416The only rockets where I do the fuel are liquid or hybrid fuel rockets19:28
bart416Else I only do the electronics19:29
bart416This time all I did was provide the launchpad, lol19:29
bart416It'd be easier if it was destroyed19:30
bart416now I had to drag it home again19:30
bart416The only rocket that has destroyed a launchpad up to now was liquid oxygen + kerosene19:30
bart416Damn that was one hell of an explosion, lol19:30
bart416You could feel the heat from a rather considerable distance19:34
azonenberglox + kerosene? I can imagine19:44
azonenbergWhat was the total propellant mass?19:44
bart416That time, don't really remember19:47
bart416Fairly large19:47
bart416The rocket was 2 meter high19:47
bart416and 40cm diameter19:47
bart416single stage19:48
bart416We launched a few of those19:48
bart416If there's ever a civil war I know what I'm going to use to take out a small armored vehicle...19:48
bart416The issue is, the range of our radio is roughly 600 meter19:49
bart416it went higher than that19:49
bart416so we can't really be sure19:49
bart416But of the 6 we launched in total 2 came back to the ground in one piece19:50
bart416First 2 exploded on the pad19:51
bart416(bad mixing ratios + not enough cooling)19:51
bart416third one acted weird19:51
bart416Huge inferno19:51
bart416We thought it exploded19:52
bart416but it started lifting anyway19:52
bart416But exploded a few meters of the ground19:52
bart416Fourth one broke up on decent19:52
bart416(bad parachute attachment)19:52
bart416fifth launch was a large succes19:52
bart416sixth launch we used the same one as fifth with some replaced parts obviously19:53
bart416But it started showing wear19:53
bart416So we scrapped it19:53
bart416That was our liquid + kerosene adventure19:53
bart416That was totally illegal19:53
bart416Hence why no pictures or videos are posted on the internet19:53
bart416Already had enough complaints due to noise19:54
bart416We then switched to liquid oxygen + ethanol19:54
bart416Far more stable, but a lot less propulsion19:54
bart416but a lot less wear on the engine parts19:55
bart416azonenberg, just heard something rather wow o_O20:40
bart416A friend of mine works at TI20:40
bart416And his wife can't get kids due to cancer in the past20:40
bart416And TI covers legal costs o_O20:41
bart416(They're adopting a kid)20:41
azonenbergThats rather nice of them (TI)21:04
azonenbergAnd yeah, i never looked up the regulations re liquid propellants21:04
bart416Well, strictly speaking launching any rocket we built is sort of borderline illegal21:07
bart416We generally cover ourselves under the rather allowing fireworks laws in Belgium21:07
bart416Here it's perfectly legal to buy a stack of fireworks and store it on your attic21:07
bart416All you need is an identity card stating you're 18 years or older21:08
bart416Also they (TI) seem to have a lot of programs/benefits like that so they can keep their staff for their entire carreer21:10
swkhanhello =)21:30
azonenbergSo whats this about memristors?21:30
azonenbergI'd suggest you start large21:30
swkhani want to pattern a bottom electrode and a top electron quickly, cheaply, and easily21:30
azonenbergMy focus right now is on MEMS so i've been putting a lot of work into bulk micromachining of Si21:31
azonenbergas well as metal layers21:31
swkhani see21:31
azonenberghttp://i.imgur.com/ty4lq.jpg is one of my recent experiments21:31
azonenbergnyanotechnology ;)21:31
azonenbergMight be the smallest nyan cat made to date21:31
swkhanomg so cute21:31
azonenberg200um high and 600 long21:31
azonenbergpixels are 20um square21:31
azonenbergthis is 100x magnification21:32
azonenberghe's patterned in about 2um of SP24 photoresist over 200nm of copper21:32
swkhanyou made that at home? o_O21:32
azonenbergevaporated onto a glass substrate (microscope cover slip)21:32
azonenbergI did the copper deposition in a lab on campus21:32
azonenbergall of the lithography, as well as the photoresist deposition, was at home21:32
swkhani work in a lab on campus too but it doesn't have all the facilities i'd like21:32
azonenbergI overetched a bit (notice the green areas expanded beyond the holes in the mask)21:33
azonenbergturns out even five seconds in full strength SC2 is overkill for going through a couple hundred nanometers of Cu21:33
azonenbergi'm gonna dilute it like 50:1 for the next etch (tonight) so it's slow enough to be controllable21:33
azonenbergI'm also having issues with focuser drift on my microscope objective21:34
azonenbergover the 10-ish minute exposure the stage will move down by a couple of microns21:34
azonenbergand blur really small features21:35
azonenbergi get good results with the 10x objective (so my 200um laser-printed mask becomes 20um per pixel) but the 40x is blurred beyond usability21:35
azonenbergWhich school is your lab at btw? I'm a PhD student at RPI21:36
swkhanwell we're at nasa ames. it's a weird collaboration with a lot of different universities21:36
swkhani know a friend who will be going to RPI next year =)21:36
swkhanit's a weird collaboration because everyone just gives the facade of collaborating but doesn't actually take many steps to actually get work done21:37
azonenbergAnd nice. whats he studying?21:40
azonenbergI did my undergrad work at RPI too (computer science)21:41
azonenberggraduated last spring and i'm now just starting my doctorate21:41
azonenbergBut most of my work, as you can probably tell, is leaning toward the EE side of things21:41
swkhani think he's studying materials science21:42
azonenbergIn that case he'll probably be playing with the same SEM and evaporator i've been using21:43
azonenbergRPI has more than a handful of SEMs21:43
azonenbergthere's a zeiss supra 55 FESEM in the cleanroom (which is expensive to get on but the best on campus)21:43
azonenbergright next to it there's another zeiss, dual beam SEM / FIB21:44
swkhanwow a fib21:44
azonenbergthen outside the cleanroom inthe mat sci department there's an old jeol tungsten-filament SEM, a newer jeol FESEM21:44
azonenberga TEM21:44
azonenbergraman spectrometer21:44
azonenbergthe cleanroom has an AFM21:45
swkhanwow a cleanroom afm21:45
azonenbergLots of toys - great as long as you arent the one paying for it21:45
smedingwish i could play in our university clean room (DIMES at Delft UT)21:46
smedingbut that will have to wait until at least the M. Sc i suppose21:47
azonenbergI actually got in as an undergrad but it was for an outside consulting project21:47
azonenbergand just for SEM imaging of a sample21:47
azonenbergAnyway so re memristors - i'm reading a bit on them now21:49
azonenbergWhat exactly are you trying to build?21:49
azonenbergLol. this is probably the most interesting artwork i've ever put on a PCB22:22
azonenbergA scale drawing of a CMOS ring oscillator22:22
azonenbergbottom copper = diffusion, top copper = metal 1, silkscreen = poly22:22
smedinghmm :p22:23
azonenbergthis is missing vias and body connections, plus all of the other fun stuff you'd need for a real circuit22:32
azonenbergBut as artwork it should wortk nicely22:33
azonenbergoh, the gates on the NAND2 at right should be taller *fixes*22:33
bart416smeding, you're at TU Delft?22:34
smedingnot currently22:34
smedingi'm in between my first and second year of the B. Sc EE22:35
bart416Where are you at?22:35
smedingi live in Noordwijk because living with the22:35
smeding'rents is cheap22:35
azonenbergHere we go http://i.imgur.com/k8BCm.png22:35
azonenbergComplete with two little pics at top right22:36
Action: azonenberg doesnt like empty space in PCBs22:36
smedingme neither! all that copper just dissolving into your etchant, such a waste22:36
smedingsilly PCB artwork is good for the environment!22:36
kristianpaulconsidering crosstalk between layers.. well :)22:37
azonenbergWell i'm not making the board myself - its a 4-layer design22:37
azonenberg3.3V and ground planes are in the middle22:37
bart416smeding, you're not answer my question22:37
azonenbergmarked by + and - vias22:37
bart416Are you enrolled at Delft or not?22:37
smedingbart416: i just said, i am22:38
azonenbergkristianpaul: any overall thoughts on the board?22:38
smedingbart416: however, don't even have my P yet22:38
smedingclose though :p22:38
bart416Just answer yes if people as if you're there >_>22:38
kristianpaulpic32 is to load bitstream?22:38
bart416Also why the hell are you complaining then22:38
bart416At least you guys have the equipment to do something22:38
azonenbergkristianpaul: Lol its overkill for that22:38
bart416We have to beg for new oscilloscopes >_>22:38
smedingi was... complaining?22:39
azonenbergthe plan is for the pic32 to be the main CPU and the FPGA to be a GPU22:39
azonenbergas well as offloading heavy computation to it22:39
bart416"<smeding> wish i could play in our university clean room (DIMES at Delft UT)"22:39
bart416^Sure sounds like it :P22:39
smedingnot so much a complaint as something i would like to do at some point22:39
kristianpaulpci32 is U1 right?22:39
smedingit looks like fun, anyway22:39
azonenbergfpga is U2 - 200k gate spartan-3a22:40
smedingbut yeah, facilities are definitely very, very good22:40
azonenbergin 100vqfp22:40
azonenbergU5 and U6 are 1A linear regulators22:40
kristianpauli dont see to much wires between board fpga and pic, os i asume trougput is low?22:40
azonenbergfor the 3.3 and 1.2v rails22:40
azonenbergand actually its an 8-bit parallel bus22:41
azonenbergwhich could be clocked at up to 40 MHz22:41
kristianpaulhe i jsut countent a nibble ;)22:41
azonenbergIts not a 64-bit bus or something like that22:41
kristianpaul8 bits seems okay then22:41
azonenbergheck, the whole pic is only 64 pins22:41
kristianpaulsure sure22:41
azonenbergThey do have a 100 pin version with a 16 bit parallel port22:41
azonenbergbut realistically i dont need that much bandwidth22:41
azonenbergwhen i outgrow this board in a few months / a year ish i'll do a 6- or 8-layer design with a 256FTBGA fpga and a softcore CPU (no external MCU)22:42
azonenbergand a few hundred megs of SDRAM22:42
kristianpaulthe sram of the fpga is enought for you app i guess too22:42
azonenbergYeah, i considered an external ram but couldnt route it on 4 layers while also having the IO resources i wanted22:42
azonenbergthe main purpose of the fpga is to run the video output22:42
azonenberg800x600 VGA in text mode22:43
azonenbergcurses-type interface22:43
kristianpaulthat GPIO is what for?22:43
azonenbergSeemed a shame to pay $100 for three 4-layer boards and not get access to all the FPGA pins :P22:43
azonenbergso i figured i'd break 'em out to a ribbon connector in case i needed them for something in the future22:43
kristianpaulconsidering is just video out vias cross under it..22:43
azonenbergthis is meant to be a "play around" board, its not for any specific purpose22:43
kristianpaulah ok22:44
azonenbergThey're gonna be relatively low speed signals22:44
azonenbergand there's power and ground planes between themn22:44
azonenbergeverything is single-ended on this board22:44
azonenbergexcept for USB22:44
azonenbergthe v2 design will be using more differential stuff like TMDS video22:44
kristianpaulhe, the cap smile footprint :)22:44
azonenbergthat isnt a cap22:45
azonenbergits just a smiley22:45
azonenbergi like throwing little pictures in that last unused few mm^2 of a board22:45
azonenbergnote the stuff under it? Not hooked up to anything either22:45
azonenbergthats a 7-stage CMOS ring oscillator :P22:45
kristianpaulhum i missed usb, is the CONSOLE?22:45
azonenbergtop copper = metal1, silkscreen = poly, bottom copper = diffusion22:45
kristianpauland yeah, usbon pic32 is host nice22:46
azonenbergand actually, both console and keyboard22:46
kristianpaulwell no more toughts :)22:46
azonenbergconsole is going to an FT232 usb-serial for debugging before i get the video card working22:46
azonenbergthe keyboard connector is for a standard usb keyboard22:46
azonenbergThe ultimate plan is to be able to run PICnix on this board without a PC at all22:46
azonenbergconsole interface with usb keyboard and vga out22:47
kristianpaul(ring osc) yeah :)22:47
azonenbergIts just a little inside joke for anyone who's studied VLSI and looks at the board lol22:47
kristianpauli first tought was for RF or something then read you comments22:48
azonenbergYou cant see in this CAD view but i even cut holes in soldermask where i'd put vias to metal 2 or test points22:48
azonenbergi usually throw artwork on most of my boards if i have space22:48
azonenbergi threw the smiley on two other boards already22:49
azonenbergi stuck a companion cube in the CAD for one but the customer scrubbed the project before prototype fab22:49
azonenbergthe osc is six INVx1 and one NAND2x1 lol22:50
--- Sun Aug 7 201100:00

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